ADR1399 reference

[branadic]

I think it's worth having a separate thread for the ADR1399 to not mess up the threads about LM399.
About 2000 h ago I've started monitoring one of two ADR1399 with a Prema 5017. After playing with the integration time in the first days I kept the setup as is afterwards and recorded the output voltage. Unfortunately, the Prema has quite some t.c. and noise contributing to the measurement, so it's not the meter of choice to observe this kind of references, but all I had available at that time. Will change that setup soon and add the ambient conditions as well.

-branadic-

[jonpaul]

Brandiac, have connexions at ADI, hoping to get the 1399s and setting up tests here, have latest Keysight 34465A in CAL.

Any tips on your set-up? How does 34465a compare to your meter?

Bon Soirée


Jon

[macaba]

Attached is output voltage deviation vs. heater voltage. 0.37ppm/V.

[Kleinstein]

Brandiac, have connexions at ADI, hoping to get the 1399s and setting up tests here, have latest Keysight 34465A in CAL.

Any tips on your set-up? How does 34465a compare to your meter?

Bon Soirée

Jon

Directly measuring a reference with a DMM will give the combined drift and noise of the DMM and the DUT.  The Prema 5017 by now likely has a well aged LM399 reference and could thus be low dirft.  For the KS34465 it depends on how much that meter was used before. It may help if one has a calibration history.

The noise in both case would be to a large part the meters internal noise form the LM399.

[Andreas]

Attached is output voltage deviation vs. heater voltage. 0.37ppm/V.

Hello,

How is this measured? with or without thermal isolation of the ADR1399? (on both sides of the PCB)

Additional question:
Does the output voltage change also (like on LM399) when changeing the orientation. (laying on left / right side or legs above plastic cap ?)

See also for a LM399 based instrument here:
https://www.eevblog.com/forum/metrology/project-pimp-a-keithley-2000/msg1106829/#msg1106829

with best regards

Andreas

[bastl_r]

hello guys
I found on ebay very cheap resistors with only 2ppm for 1,5€ each.
With 2 sets of these and a Trimmer you can create a voltage divider for the ADR1399 to 10V.
2pcs 31k623 in series and 1pcs 26k1 plus a 200Ohm trimmer.
Here are the links to the resistors:
https://www.ebay.de/itm/144374605346
https://www.ebay.de/itm/144475379838



regards

[branadic]

Why bothering with expensive single resistors, if you can use a resistor network for less the price, that can be configured in a ratio required?

https://xdevs.com/guide/rnet_ratio/

They offer good t.c. matching and the network can be trimmed for the final output voltage by paralleling resistors to a single element within the network, that way the t.c. of the parallel resistor is dampened by the network.
Final t.c. trimming can be accomplished by adding copper in either the upper or lower branch of the network, depending on the sign of the remaining t.c. (see attachement).

-branadic-

[bastl_r]

Hi
Thank you for the tip.
But i don't understand how i can adjust a calculatet ratio of 1.41520412358094.

Can you explain what is to dofor this ratio?
One of my ADR has 7,06624V and the other 7,07143V...

Regards

[branadic]

Simply start with a 2:5 ratio = 0.4 and trim the "5" part of it close to 2:4.817 = 0.4152 with parallel resistors.

-branadic-

[dietert1]

The proposed design with 26100 and 26300 Ohms isn't wrong, except one should not use a trimmer. A trimmer isn't stable enough after some years. Also it is easy to lose calibration without notice.

Better calculate the value needed and get a good ( < 15 ppm/K) resistor of that value. The tolerance of the 26100 resistor will be about 2.6 Ohms. Better choose a series trim resistor on the high side, then one can do ppm calibration by some parallel resistor to that trim resistor, without touching the foil resistors anymore.

Regards, Dieter

[branadic]

New datasheet available https://www.analog.com/en/products/adr1399.html showing ADR1399 in LCC package. Good thing about that package is: "...The LCC version adds four pins, two of which are not internally connected (NIC) and the other two split the active reference into force and sense action on the top and bottom of the shunt..." but that comes at the cost of "...The heater power for the LCC version is about 3× to 4× greater than the TO-46 version...".

-branadic-

[alm]

A higher heating power means stronger thermal gradients, which I'd expect to cause higher temperature coefficients.

[Kleinstein]

The actual heat needed for the LCC package version will depend a lot on the layout / PCB. With some isolation and thin traces it can be quite viable to create quite some thermal impedance on the PCB level. With a suitable PCB and cap the heater power may be comparable.

The higher hyteresis would be a weak point for the LCC version - though this may also depend on the PCB / holder.

[miro123]

I'm looking forward on LCC version. Leaving kovar and adding sense pins are welcomed changes. I wonder what can be made by hobbyist in insulated "dead bug" configuration.
Professional world will most like appreciate LCC voltage reference @5/4.096/2.5V

[branadic]

Meanwhile, 3000 h have passed and measurement is still ongoing. As announced ambient sensors were added, hopefully giving a better clue what's going on. Maybe I should switch from Prema 5017SC to Prema 6048 to get better resolution and stability plots. But that's a task for the near future.

-branadic-

[branadic]

I was asked to perform a 1/f noise (0.1 - 10 Hz) measurement on the ADR1399, so I did exactly that, result attached. A nice match for post 2000.
By the way, this reference was used to verify, whether the Prema 5017 is really limited by its ADC in terms of noise as Andreas assumed or by the low frequency noise of the LM399 used as a reference in there and Andreas was totally right. There is absolutely no improvement when replacing the reference. A short article about it can be found here.

-branadic-

[WillTurner]

The ADR1399 datasheet revision 0 (not the new datasheet listing the LCC package, which I am yet to read) has the following statement listed under "Thermal Resistance" on page 4 of 11 which may be missing in the new revision :-

  "To reduce heater power, avoid bringing power planes close to the device except on the bottom of the PCB, and use a mesh ground instead of a solid plane. In addition, mounting the device at the full height of its leads, approximately 1cm above the PCB surface reduces the heater power."

Does that put an end to the long/short leads debate[5]?

The LCC package is an attractive option. I guess mechanical relief slots, and edge of board location[3, 4] come into play, or dead-bug style[1, 2], which looks promising :-).

References
[1] 

ADC20 is built with a hermetic LS8 package mounted dead bug style in a PCB cutout to avoid PCB stress.
:
:
The LS8 hermetically package shows +/-14 = 28uV or 8 ppm hysteresis.
:
:
With best regards

Andreas

[2]

I have a small pcb with a voltage vref LTC6655B. Unfortunately, the long-term drift specified only for LS8 is 20ppm/sqr(1000) ... I have MSOP package though.

:
:
it has a reason why the (typical) drift is only specified for the hermetically package.
:
The LS8 package is better, but only if you mount it dead bug style on the PCB so that there is no influence from the PCB.

With best regards

Andreas

[3] 

https://www.eevblog.com/forum/metrology/slot-holes-around-ltz1000(a)/msg1531931/#msg1531931

[4] That App Note that discusses mechanical relief which was quoted long in the past ...

[5]

I once investigated LM399 with short and long leads with and without slots.

-branadic-

Edits 1. : Add some references, dead-bug bit.

[Kleinstein]

For the LM399 there was never much debate over long or short leads. Long leads are preferred, just need a bit extra effort for soldering (e.g. a temporary spacer) and the extra hight.
The longer leads also give some extra isolation from mechanical stress.

The long / short leads discussion is with the LTZ reference with higher current (and thus more effect of the lead resistance) and a usually lower temperature setting and thus less heat flow through the leads.

The question is more how much thermal/stress relief is needed / sensible on the PCB. A point to keep in mind here is that it is not only about the stress to the reference, but also stress cause by the heated reference effecting other parts (e.g. resistors) on the PCB.  The heated reference is not so much a sensitive part, but in the thermal sense also a source of disturbance. While the reference itself stabilizes fast, it can take quite some time before the parts around the heated reference get a stable temperature (and even worse a stable humidity level).

[KT88]

"To reduce heater power, avoid bringing power planes close to the device except on the bottom of the PCB, and use a mesh ground instead of a solid plane. In addition, mounting the device at the full height of its leads, approximately 1cm above the PCB surface reduces the heater power."

This statement is about power consumption. The ADR1399 in LCC package has a junction to ambient thermal resistance of 125°C/W with a JEDEC board (I didn't find that layout though).
The layout recommendations are meant to further increase the thermal resitance to save power. The LTZ1000A has 400°C/W in comparison.
Other than the TO-46 package the LCC package doesn't use Kovar thus having a lower influence of the Seebeck effect. This means that thermal gradients would have a lower impact on errors than with a TO-46 case.
If highest performance is the goal I would still aim for lowest thermal gradients on the board and sacrifice a bit more power...

[branadic]

For the LM399 there was never much debate over long or short leads. Long leads are preferred, just need a bit extra effort for soldering (e.g. a temporary spacer) and the extra hight.
The longer leads also give some extra isolation from mechanical stress.

There was a discussion about short vs. long leads before, you must have missed that. Why was there a discussion? Well, you find short leaded LM399 in HPAK meters sitting in sockets, so Keysight must have missed your arguments for long leads. There are other examples were short leads were used, but there are also also examples for long leads, same goes for mounting the reference upside down or straight into the air.
I also made a few thermal imaging videos about short vs. long leads, slots vs. no slot that you can find here:

https://www.youtube.com/user/TheHendi78/videos

Everyone can draw his conclusion based on that results.

The heated reference is not so much a sensitive part, but in the thermal sense also a source of disturbance.

I wouldn't sign that either, the reference is quite sensitive to tilting, as the heater resistor is located only at one edge of the reference, so it's to be expected.

-branadic-

[Kleinstein]

I was asked to perform a 1/f noise (0.1 - 10 Hz) measurement on the ADR1399, so I did exactly that, result attached. A nice match for post 2000.
By the way, this reference was used to verify, whether the Prema 5017 is really limited by its ADC in terms of noise as Andreas assumed or by the low frequency noise of the LM399 used as a reference in there and Andreas was totally right. There is absolutely no improvement when replacing the reference. A short article about it can be found here.

-branadic-

I had a quick look at the 5017 circuit. One point is that this meter, like the HP3457 has a prime range of some 3 V. The 30 V range already goes through the input divider and correspondingly has 10 M input resistance. So the range to test would be the 3 V range. The 30 V range may be limited by the divider / amplifer. The ADC part is a bit confusing and may also have quiite some noise.

[MK]

For the LM399 there was never much debate over long or short leads. Long leads are preferred, just need a bit extra effort for soldering (e.g. a temporary spacer) and the extra hight.
The longer leads also give some extra isolation from mechanical stress.

There was a discussion about short vs. long leads before, you must have missed that. Why was there a discussion? Well, you find short leaded LM399 in HPAK meters sitting in sockets, so Keysight must have missed your arguments for long leads. There are other examples were short leads were used, but there are also also examples for long leads, same goes for mounting the reference upside down or straight into the air.
I also made a few thermal imaging videos about short vs. long leads, slots vs. no slot that you can find here:

https://www.youtube.com/user/TheHendi78/videos

Everyone can draw his conclusion based on that results.

The heated reference is not so much a sensitive part, but in the thermal sense also a source of disturbance.

I wouldn't sign that either, the reference is quite sensitive to tilting, as the heater resistor is located only at one edge of the reference, so it's to be expected.

-branadic-

Long leads means the heat loss is less, but it is easier to get the kovar/copper thermocouples at different temps, short leads I suspect may make it easier to keep the thermocouples under control, but give you other thermal problems.

[Andreas]

Hello,

i do not agree if I look at the pictures branadic did with some LM399 with short/long legs and slotted/unslotted PCB.

https://www.eevblog.com/forum/metrology/ultra-precision-reference-ltz1000/msg278247/#msg278247

The Pins on the PCB are much more uniform on the unslotted PCB with long leads.

With best regards

Andreas

[MK]

Hi Andreas,

I do not think the camera has the resolution to prove it for either of us, 40uV per degree.
We need the pads to have less than 20 millidegree difference, the photos cannot provide that.

[Kleinstein]

Hi Andreas,

I do not think the camera has the resolution to prove it for either of us, 40uV per degree.
We need the pads to have less than 20 millidegree difference, the photos cannot provide that.

The requirements for the reference are not that strict: the LM399 reference is not that accurate to need better than 1 µV stability. The popcorn type jumps are usually in the 4 µV range and 1 ppm drift corresponds to 7 µV.  Another point is that a stable temperature difference only adds to the reference and is thus not a problem. Only the change in the temperature difference is that causes problems.

The camera is still no sufficient to directly resolve the temperature differences. However the camera can still help and show the temperature gradients a bit away from the pins. Due to thermal coupling the differences at the pins is expected to be quite a bit smaller than the differences some 5 mm away. The larger differences around the reference is that can be resolved and compare different configurations.
One has to balance between temperature gradients and voltage drop with long thin copper traces.

[Andreas]

Hello,

for me the resolution is sufficient to see that on the bottom side of the PCB the temperature difference is ~4 deg C from center (KOVAR) to edge (Copper) for the long leads and > 10 deg C for the short leads on the unslotted side of the PCB.

with best regards

Andreas

[MK]

Hi Andreas,

My thinking is that it is necessary to keep the pads at the same temperature, a gradient in the copper is not as important.

[oz2cpu]

THANKS for a seperate ADR1399 thread :-)
see the picture mine just showed up

[iMo]

Hello,

for me the resolution is sufficient to see that on the bottom side of the PCB the temperature difference is ~4 deg C from center (KOVAR) to edge (Copper) for the long leads and > 10 deg C for the short leads..

Fun with 1399 

[iMo]

@1399 owners: could you somehow compare the 1399 vs. 399 popcorning?

[branadic]

I couldn't observe any popcorn on my ADR1399 sample by now.

-branadic-

[iMo]

LCC version in LTSpice..

[Andreas]

Hello,

my first LS8-samples from DigiKey showed up yesterday.
From datecode the LS8 version seems to be much older than the TO-46 devices here in this thread. 

Unfortunately I still do not have a "ageing board" for those.

with best regards

Andreas

[iMo]

Dead bug style?

[Andreas]

Hello,

dead bug is not planned as I want to see the influence of the PCB (humidity) on the output voltage.

With best regards

Andreas

[branadic]

Hi Andreas,

I would appreciate if you would split that group into halve, one group mounted to a rigid PCB, the other one with the reference on a flexboard that is attached to the rigid PCB. This way you could proof if a semirigid solution is better, since it is a standard industrial solution, reproducable and without magic.
If you need the flex board made by someone else for this investigation, I could provide you with some.

-branadic-

[Conrad Hoffman]

Just got an LCC8 package today. Hope is to dead bug it and mount it inside an old Eppley (or Weston) standard cell enclosure as a cool looking . Not sure what circuit or output voltage I'll use yet but probably a traditional opamp with the output feeding back around to the zener, though I'm not sure the 1399 needs that. Maybe just a buffer?

[iMo]

For example in my replay #31 above there is the simplest version with the buffer transistor only.
The ADI's eval kit 1399 has got the transistor too..
PS: the R4 there in my sim is not part of the schematics..

https://wiki.analog.com/resources/eval/adr1399e-ebz

PPS:
Below a Sim of ADI's Eval schematics - the TC at 1mA out.
The max +REFS current is 2.19mA in this sim.
The max +REFF current is 5.47mA in this sim.

Interestingly the ADI Eval schematics has got the RC filter at the +REFF pin (my sim above has it at +REFS pin), not sure ADI's schematics RC wiring is the correct version (ADI's typo?)..

[Conrad Hoffman]

Would have been nice if they had revealed their board layout and done a buffered application. Hey, more to experiment with!

[iMo]

The I_load char with the by transistor buffered 1399.
-1ppm at aprox 0.4mA..

[Kleinstein]

Driving the reference with just 1 external transistor is an interesting option.
The Eval board has the 1µF + 5 ohms RC element at the ref force side, not at the sense side. This does make a difference for the stabilty against oscillation.
The capacitor at the force side should be closer to the case with the 4 pin TO46 version and likely the better placement.

It is somewhat the question if one really needs the extra external transistor. The main advantage of the external transistor is unloading the reference chip from possibly variable currents.  For most cases the load to the reference is essentially constant and this essentially no gain from the extra transistor. The extra transistor as emitter follower may make sense when there is a variable load to the reference. So the extra transistor does not replace the bootstrapping ciruit to provide a very stable ref. drive current, but is more like an optional extra buffer for the 7 V reference, that may follow the reference.
The separate pins allow for a lower ref. impedance and thus less need for the bootstrapping part. To a large part this already applies to the simple circuit without the transistor. The extra transistor allows for a larger resistor at the force output (so the 1.3 K are rather low) and thus a little less effect of supply variation. So it helps to some degree (e.g. a factor of 5), but would still need a stable supply.

It is a bit unclear how the AC stabiltity of the reference is effected - the DC load current from the sense input is essentially constant, but it is a bit unclear how this part behaves AC wise.

The simpler case is to just connect the force and sense pins externally to the chip and use the extra pins just for compensation of the exffect of the bond wires. So the circuit is than essentially the same as for the 4 pin version.

[iMo]

No idea how the ADR1399 model in LTSpice has been designed, but I've tried a simple test with pulsing 1mA load and the peak-peak transients with RC on the force side are 4x higher than when RC is on sense side (with the transistor buffer).. So the ADI's eval kit schematics with RC on the force pin (with the transistor buffer) is not correct, imho..

[Kleinstein]

With 2.2 mA the transistor as emitter follower has an output impedance of some 12 Ohms. So the somewhat worse reaction to fast peaks is not such a surprise. The extra transitor changes the circuit and can change the suitable compensation quite a bit. 1 µF at the base looks a bit high - chances are a smaller cap there and than some RC (could be different from 1 µF and 5 ohms) at the ouput would be OK.
The spice model could be helping, but it really depends on how good the model actually is. A first test on the model would be how well it reporduces the output impedance in the LM399 like mode without the extra transistor. Even than it is not clear how much of the internal compensation is from which pin.

[branadic]

Just got an LCC8 package today. Hope is to dead bug it and mount it inside an old Eppley (or Weston) standard cell enclosure as a cool looking . Not sure what circuit or output voltage I'll use yet but probably a traditional opamp with the output feeding back around to the zener, though I'm not sure the 1399 needs that. Maybe just a buffer?

You could consider building a circuit similar to the LM399-PWM-DAC, which works pretty decent and this reference was used for the EU CalClub test run within Germany.

-branadic-

[Conrad Hoffman]

After playing with the LTSpice model for a bit, I'm sure I don't understand how the F & S lines work and where/why current flows where it does. On a board where everything is close together, it seems safest to just hard wire them together.

[KT88]

With 2.2 mA the transistor as emitter follower has an output impedance of some 12 Ohms.

That's not what is happening...
The emitter follower is the external part of a closed loop amplifier - it is just only a uni-polar output stage. No need to sink current...
Running the sim in LTSpice gives an output impedance of only 1mOhm.
Some significant advantages come with using an emitter follower:
-The current through the shunt reference remains basically constant - the classic circuit shares the bias current with the load.
-The bias resistor won't need to provide the load current - it can be larger.
-A wide range of load currents are challenging for a shunt circuit - in paticular if larger load currents than the max bias current are required.
-The power dissipation within the reference IC remains constant with changing load currents.

Cheers

Andreas

[MiDi]

You could consider building a circuit similar to the LM399-PWM-DAC, which works pretty decent and this reference was used for the EU CalClub test run within Germany.

Documentation for PWM-DAC is in thread LM399 based 10 V reference:
Schematics & Gerber
Firmware

[Conrad Hoffman]

Had some time for a crude measurement using manganin wire to isolate the package thermally. 0.011" diameter and about 0.2 ohms per inch. Just wired the heater and ran it on 15 VDC. Current stabilized at about 10 mA in free air, so I think this might be a useful trick. Probably better with some sort of insulation. I used to do the same thing with cooled CCD chips to lower the heat loss through all the wires. I got an IR reading off the edge of the chip of about 150F (65C) which is probably too low. Don't know the emissivity- usually I put a dab of flat black paint on things. Pretty sure the heater is working correctly. If you blow on the chip the current doubles. Still thinking about how to handle the reference wiring.

[iMo]

You could consider building a circuit similar to the LM399-PWM-DAC, which works pretty decent and this reference was used for the EU CalClub test run within Germany.

Documentation for PWM-DAC is in thread LM399 based 10 V reference:
Schematics & Gerber
Firmware

Is that the latest schematics and firmware? I think there were some changes after the evaluation here by Andreas and Kleinstein..

[iMo]

The manganin leads shift the voltage a little bit (the refsense is rather fake here), but otherwise it should work.
In the ADI's schematics (with the transistor) the higher the bias resistor R1 the worse regulation under variable load.
@Conrad: here are the 3 classic examples - the worse load char is with the ADI's transistor, those with an opamp are the best (the graphs below are "ppm vs. load current" from 1uA to 5mA/10mA/20mA).

[branadic]

Is that the latest schematics and firmware? I think there were some changes after the evaluation here by Andreas and Kleinstein..

I have a newer, slightly modified board version that also covers the ADR1399, so with the additional required filter.

-branadic-

[Conrad Hoffman]

Continuing with the measurements, I painted the top of the package black so I could get a good temperature reading. Seems to sit about 150-155F. With a small plastic cup to shield it, the heater current drops to 9.4 mA. I wonder if the die has some thermal insulation between it and the package.

@branadic is that board available to purchase, or Gerbers?

[branadic]

Attached is the latest version of that board.

-branadic-

[miro123]

R11 can be ten.. twenty times higher. 10K is good starting point.
On such way you are avoiding the the cross couple with the noisy +15V line. For more info look at HP34401 schematics

[DavidKo]

I have tried to design the PCB for LCC version, but I'm not sure which way to go. The schematic is based on Analog design, I have only tried to add the NTC and have as much variability as possible (1 or 2 separate power supplies, with or without protection diode, added measurement/power supply point). But where I'm not sure is PCB design. I have prepared "neck" design, which is based on Analog PCB and later I have completely redrawn it to spider design, but I'm not sure which way to go.

Additionally if you have the idea to make it even more universal you are welcome. At the end I want to have 2 PCB on 100x100mm plate with power supply in between and maybe also the amplifier to have 10V output.

[branadic]

My recommendation for LCC package still is to put it on a flex board or use a flex-rigid board with the LCC package located on the flex part.

-branadic-

[Conrad Hoffman]

No expert by any means but the neck design looks more straightforward. Not sure the spider buys much. I'm convinced that lowering the conduction out of the LCC package is important. Hollow an area under it- have to run traces outside only if you do that. I don't see any reason for the NIC pads to be there at all. I really want a 10V output on my board and it will almost certainly use thru-hole resistors in some pad-able/trimmable arrangement. On a small board I don't know if the sense connections buy anything if the traces have decent width. Caveat- I haven't sat down with PCB software and tried to do this yet!

[DavidKo]

@branadic I have checked the flex board pricing and for the first experiments it does not make too much sense. At least I have found prizes > $100. I will rather use 0.6mm pcb (prize is the same as for standard one $2 + shipping). But maybe it make sense to create on thin PCB only the "adapter" with pin header  and create separate power and amplifier PCB from standard 1.6mm FR4. The idea with header have also advantage, that later the reference can be used in different design without need to desolder it.

@Conrad Good point, I will remove the NIC pads


The enclosure around LCC package should be 3D printed or foam to hold the reference package in "stable" temperature generated by the inner oven.

[cellularmitosis]

I have tried to design the PCB

Hi David,

Are you familiar with the "bootstrapped zener" circuit?

[Kleinstein]

IN principle the kelvin connections with seprate sense and drive make some sense. However with the ADR1399 there is quite some current (2.2  mA) through the sense pins and this makes the separate sense pins essentially useless to compensate for trace and bond wire resistance, as most of the current still flows through the sense side. The separate force and sense pins still make sense to provider a lower output resistance, so to replace an extra buffer. However this is a relatively rare use case.

Another point may be a few more options for the compensation, though the externa 1 µF + 5 ohms does not look so bad. For the shown impedance curves it looks like 1 µF is more on the low side and if a MLCC is used one needs to take into account the reduction in capacitance with bias. So I would more like plan for 2 x 1 µF or 2.2 µF or even more if this is a X7R or similar capacitor.
 

To get a 10 V reference, the classical bootstrapped supply to the zener should be the best solution. There is than no real need to serparate the drive and sense side and one could join them right at the chip.
One could still seprate the current and voltage path after connecting sense and drive.

Even if mainly used as a 10 V reference it would make sense to also have the option to measure the raw 7 V. This would allow to see if drift is from the reference itself or the resistors to set the gain.

[DavidKo]

@shodan@micron Kelvin point should be the two "crosses" close to the output. But I have messed that out during my rework of the PCB (the crosses were right to the reference without cut out). The output should be directly connected the REFS not from current crosses. It will be easy modification.

@cellularmitosis Such an approach can be used since the the heat power and reference power are separated. I think that the bootstrapped circuit is used in LM399 portable calibrator mentioned in datasheet. The bootstrapped circuit can be part of the free space behind the reference.

I'm aware of quite high current to the sense part, that is why I have implemented transistor between power and REFS as close as possible to decrease the current as in Analog proto board. The board itself is intended to be experimental to try everything what came to your mind. This is also the reason for jumpers. For example I want to try to drive the reference sense with 2.2 mA from external current source (bootstrapped circuit will be good for it). Right now I have 3rd iteration and still I'm finding things to improve and every additional idea counts. Firstly I had planned to be able to use also TO version for ADR/LM 399 reference, but I had left that idea since there are better designs already available.

Final design should be on 100x100mm board with 2 references, stabilized power supply and amplifier to 10V (ideal will be 2, but with option to use 1 amplifier and paralleling the references to experiment with noise). Thank you for your remarks, I want to remove big flaws before I publish the design.

[Conrad Hoffman]

I've used the bootstrap circuit with good success when using 1N829 type diodes. It's what I had in mind for this, as I really want something just a few PPM under 10 VDC. Classic opamps might be OP-07, but what makes sense in 2022?

[Kleinstein]

For the amplifer on could consider an auto zero type like LTC2057,  MCP6V51 or OPA189.  There are a few slightly more modern alternatives to the OP07, like OPA207, OPA202 or ADA4077. The drift and noise from an OP07 or similar would still be less than the reference. So there is no absolute need for an auto zero OP-amp, but it would not really hurt either. Most of the modern ones come in SMD only.

[iMo]

..I will rather use 0.6mm pcb..

You may create "microslots around the pads" (no idea what is the thinnest slot they can mill), say 0.3mm thick slots - that gives some flexibility too..

[Conrad Hoffman]

For the amplifer on could consider an auto zero type like LTC2057,  MCP6V51 or OPA189.  There are a few slightly more modern alternatives to the OP07, like OPA207, OPA202 or ADA4077. The drift and noise from an OP07 or similar would still be less than the reference. So there is no absolute need for an auto zero OP-amp, but it would not really hurt either. Most of the modern ones come in SMD only.

I used an auto-zero in the Mini-Metrology reference and it worked OK. It didn't have much drive current, which messed me up the first time I tried to use it with low value feedback resistors. It was also easy to damage, which is OK if in a socket, but not so much for SMT parts. I'm surprised the OP-07 is still competitive! I'll look at the other choices.
 

[Kleinstein]

To get a little more drive power, one can add an NPN transistor as emitter follower in the loop. The gain setting divider already provides in minimum load.

The OP07 is not really that competative - it is just one that is available in DIP and can be relatively cheap (some versions). Especially the power consumption is a bit on the high side compared to modern alternatives.

[Extrasolar]

My Take to the 8LCC package, I opted for symmetric copper polygons or traces.

[DavidKo]

@Extrasolar: My solution was symmetrical till I have added the transistor between REFF and REFS.

@Kleinstein: I want to have DIP and SOIC option on one PCB, idea is to have "integrated" SOIC adapter.

@imo: I can try that if there will be space on PCB. Additional transistor and NTC added additional connections.

[DavidKo]

I have simulated the bootstrap circuit.  I have added TDP1603 in second iteration (LT5400-1 is used since it is in LTspice database). 200k start-up resistor is used since it is in original LM399 datasheet. I had played with "tuning" resistor placement and the 4R7 at the top was the winner. When I had placed the resistor in the bottom, the resistor was quite big in comparison with 1k - 1k resistor had tuned the resistor placed there.

[iMo]

The problem with your LT5400 design is you are using 2 packages.
The beauty of the LT5400 is its "ratio tempco" lower than 1ppm/C (ADI told me). That is achieved by placing all 4 resistors on the same chip. When you combine two packages you break that principle.
I think there is no way to achieve <1ppm/C other than with a single LT5400 package.
I did with single package w/ 10k resistors and my several measurements resulted in 0.5-0.6ppm/C (LM399), but you would get 10.57V or so. Any other discrete resistor added to the 5400 will kill that results, sure.

PS: LT5400's resistor tempco is typ 8ppm/C and max 25ppm/C. Only the "ratio tempco" between the resistors (like tempco of  R1/R2, or (R1+R2)/R4, etc.) on its chip is <1ppm/C.

[DavidKo]

The problem with your LT5400 design is you are using 2 packages.

Probably it was not clear from my statement. LTspice does not have TDP1603, which I want to use (8 resistors in one package). Thus I have used 2x LT5400-1. I can have 8 standalone resistors in schematic instead, but this was simpler.

[iMo]

Ok, when you want to simulate tempco you may simply add to the resistor value the tempco, like  "10k tc=10e-6" for +10ppm/C and thus simulate N resistors with the same/similar tempco when in the same package.
".step TEMP 20 50 1" will then step your simulation from 20 to 50C step 1C.

[Kleinstein]

The TC specs are usually for a relatively large range, considerably larger than relevant to a precision measurement. Much of the TC extremes usually happen at the temperature extremes. This especially applies to the better grade resistors, where it is not just the linear TV but also contributions from higher order terms. For the reduced temperature interval alone one can expect resistors to behave quite a bit better than the limits for the larger range.
I have done some tests with NOMCA and ORN resistor arrays. In the tests involving different combinations the relative TCR was usually below 1 ppm/K, at least near room temperature. This is despite having much more tolerance in the specs.
I would anyway less worry about the TC. The critical point is more the hard to predict long term dirft.

From the excess noise the NOMCA type did not perform that great (barely meeting the specs), so I would not really recommend them. The noise of the LM1399 should still be a bit higher, though not much.

In some circuits one can still combine the resistors from 2 chips, if they are mixed equally between the chips. Even if not perfect some interleaving can help quite a bit.
A gain of 1.5  (e.g. use 3 equal resistors) combined with a -5% drop would be a real option, with the resistors for the 5% being about 20 times less important.

[iMo]

I burned X megawatts of electricity while simulating all possible variants of the divider, but at the end of the day I always finished with a single LT5400 and 10.57V as the simplest and cheapest solution with lowest achievable tempco..    Of course the long term stability of the divider is even more important (as you can measure the temperature pretty well)..

[DavidKo]

@imo: Mouser pricing TDP16031001BUF 6.62EUR (500ppm@2000h) vs. LT5400BCMS8E-4#PBF 10.03EUR (not in stock, <2ppm@2000h, cheapest version with worse matching that is not important in our case). For the testing purposes the Vishay is at least 3 times cheaper.

[iMo]

For you the most important params are the long term stability of ratio matching and ratio matching tempco. It looks like the Vishay array is much more worse than the ADI in those parameters. As the price both are comparable. There is a difference in the technology afaik, the Vishay is NiCr, the ADI is BiCMOS.

[Kleinstein]

There is a difference in the technology: the LT5400 is NiCr on silicon.  AFAIK the TDP1603 is NiCr on ceramics and in a larger size case.  The case effects on how much mechanical stress from the PCB transfers to the actual resistor.

The specs for those higher end resistor arrays are tricky, as some just represent the test-limits and not the actual performance.  Things like the long term drift are very hard to test - at least not fast, so they can include them in the data-sheet when the part comes out.
The TC specs are also hard to compare, as they are for different temperature ranges. The actual relevant range is likely much smaller again.
For a single unit build one may not care that much about the spec limits, more about the typical performace. One would do some tests anyway and find out if one gets a poor sample.

The big plus for the TDP1603 is that it has 8 resistors and not just 4. This allow to get a nomianl 7 to 10 ratio and thus no extra -5% step. With more resistors one also gets additional statistic averaging which can help quite a bit with the TCR matching.

Despite the not that great spec limits ( < 5 ppm/K rel. TC for -55 to +125 C), I would consider the TDP1603 a reasonable choice. I would expect the TC matching more in the < 1 ppm/K range for the near room temperature range and thus good enough for the ADR1399.

[branadic]

It looks like the Vishay array is much more worse than the ADI in those parameters.

Don't be fooled by the datasheet specifications, measure yourself. They behave way better than what the datasheet suggests. We already learned that for the noise figure and same is true for the matching. It only reflects that measurement limits of Vishay.
If they were as bad as per datasheet, references such as W/F7000 wouldn't work as good as they do.

-branadic-

[iMo]

FYI - I've found this PhD paper on the ADI's SiCr resistors in their 0.35um CMOS process, with an interesting measurement setup (most of you do at home, afaik)..
Here is the paper..

PS: it could easily be ADI will use them in the upcoming ADR1001 - that will solve all our problems, hopefully 

[cellularmitosis]

200k start-up resistor is used since it is in original LM399 datasheet.

Just a note: I don't believe the LT1001 inputs nor output extend to the negative rail (0V), so the startup resistor probably isn't needed.  However, not a bad idea to include the footprint on the PCB anyway, in case you want to try other op amps.

[DavidKo]

I have updated the schematic and added the power supply. I have prepared it as symmetrical, but with half wave rectifier to be able to run from single output transformer.

[Andreas]

Hmm,

do you really think 10uF with one way rectification are sufficient for up to 200mA current on heat up?

with best regards

Andreas

[DavidKo]

It is only placeholder. I have not calculated/simulated the correct value. The values can be different. Now I'm trying to fit 2 references + PS + amplifier to one PCB and be able to split it to 2 separate refences (without amplifier) if needed - I do not want to de-solder the refence from PCB.

I'm still not sure if it make sense to use such a simple split power supply or to create one power supply with full rectifier and power both heating and reference from it. Reference part have this possibility, but the power part on PCB was intended for mainly for burn in. I have thought to use LM317+LM336 for reduced temperature coefficient as one supply for both parts.

[Andreas]

Hello,

first ADR1399 populated on my ageing board.
So after some initial checks/measurements the automated ageing measurement can start soon.

with best regards

Andreas

[Andreas]

Hello,

first 1/f noise measurement with the LS8-package here:

its ~2.1uVpp in average with up to ~1uV popcorn noise on sample #01.
So much more than the 1.4 uVpp from the datasheet.

So following questions arise:

- simply a bad sample?

- is the 1.4 uVpp in the data sheet only for the metal can package. (branadic showed a measurement at beginning of this thread)
  https://www.eevblog.com/forum/metrology/adr1399-reference/msg4213732/#msg4213732

- has the heater voltage a influence (data sheet condition 30V heater voltage against my 15V)?

someone else with 1/f noise measurement for different packages/heater voltages?

with best regards

Andreas

[Andreas]

Hello shodan,

guessing that your 2 devices are both with the LS8-package.

I now have measured my #01 LS8-Sample with 15V and 30 V heater supply from a linear power supply.
(Zener supplied with 2K4 from 15V regulated battery voltage).

Result with this setup is in both cases 2.0 uVpp average 1/f noise. -> not heater supply voltage correlated.

with best regards

Andreas

[Kleinstein]

I don't see the RC damping element for the ADR1399 on Andrea's PCB (it could still be on the back side). This may in theory add some noise from added rinding.
The 0.1-10 Hz noise usually means a 10 second window for the observation - a longer window will add some noise from lower frequencies, from an amplifier. Over 10 s windows the noise looks a little smaller.

Some of the rather low frequency part could be from mechanical stress / acoustic coupling. Without a cover there could also be thermal effects, even though there are no covar leads and thus hopefully less thermal EMF. 

[Andreas]

Hello,

after populating the other ADR1399 in LS8-package I checked the Zener voltage with heater against the voltage with Heater off.

on the x-Axis the zener voltage (in mV) with heater active.
On the y-Axis the temperature drift (Heater active - Heater inactive) / Temperature difference.

As temperature difference I assumed 60 K (95 degC - 35 deg C assumed temperature for self heating of the 3 mA zener current).
Result is ppm/K between heater on and heater off.

For me it looks like the the lower zener voltage has a lower unheated tempco.
Maybe there is some zero T.C. near a zener voltage of ~ 7010 mV.

with best regards

Andreas

[Kleinstein]

It is normal to have a correlation between the zener TC and voltage. Normally a higher voltage means a more positive TC.

The temperature range from some 90 C heated to some 30 C unheated is quite large and the 2nd order TC part can be relevant. So the parts with the least change from cold to hot would not have the lowest TC around 90 C.

To get an idea on the TC at different temperatures one may have to look on the zener voltage chamge just after turning the heater off. I would not be surprized to see something like like voltage first going up and than down again.

[Andreas]

Hello,

overview over ADR1399#01-#04 with noise and (unheated/heated) T.C.
All measured 1/f noise (0.1-10Hz) is above 1.9 uVpp/100 sec as average over 19 measurements.
I found more or less popcorn noise on every sample with #4 being the worst with 1.6 uV.

I want to operate the devices for at least 1-2 weeks before doing next measurements after initial ageing.

with best regards

Andreas

[Kleinstein]

The cut-out around the references are nice for the thermals, but it would be better to have them in slightly different way.  One could have them more in the 90 deg. angle scheme, so that if the center part expands from heat the center part could rotate a little in one direction to relief the stress.

[Kleinstein]

The center island with the reference is still quite large. I think one could reduce it a bit by removing parts from the edges and corners. Less area mean less heat loss.

The outside part with the OP-amps is quite dense packed and may get a bit warm.  Some of the OP amps should get away with just a +15 V and GND for the supply to save on the power consumption and reduce the heat.

[Andreas]

Hmm,

what is 0 deg and 90 deg (with regard to the pin position in room)

the largest difference should be when the heater area is below the temperature sensor versus above the themperature sensor.
https://www.richis-lab.de/images/REF01/21x06.jpg

So best guess: Pin 4 on top side versus Pin 8 on top side (with the LS8-package) would give maximum difference.

Additional thermal isolation around the PCB should reduce the effect.

with best regards

Andreas

[Kleinstein]

The orientation effect can have two components: a chance in the thermals and a change in mechanical stress. Both would likely depend a lot on the PCB design and case. So interesting to note an order of magnitude, but a different design may behave quite different.

With the TO46 metal case part of the effect can be local from the plastic cover that comes with it. Even there the air flow can be different.

[opa627bm]

Stupid question here, any obvious improvement if I replace my LM399 in my Power Design 2010 to this ADR1399?

[MegaVolt]

Stupid question here, any obvious improvement if I replace my LM399 in my Power Design 2010 to this ADR1399?

It is possible to reduce the output noise. This will work if the output noise is dominated by the noise of the reference source.

[xerbo]

Mouser have recently got some evaluation boards in stock: https://www.mouser.co.uk/ProductDetail/?qs=4ASt3YYao0UVSWYehgmbvQ%3D%3D, users guide: https://wiki.analog.com/resources/eval/adr1399e-ebz.

Bought one to test, for the price (£54 including VAT) it seems like a pretty good deal for a fully tested and assembled board.

[Andreas]

Hello,

some days ago I started with initial drift measurements on ADR1399#1 - #4.
Initially I had placed the board above my LM399 ageing box.
To see better the self heating of the ADR reference board I put it on a cooler place (a unused analog power supply).
This gave up to 3 ppm jumps for 2-3 deg C lower board temperature.

To exclude a rather large T.C. or a influence of the on board multiplexers (MAX4052A) I made T.C. measurements.
So except for ADR1399#4 the box tempco near room temperature is below 0.1 ppm/K.
All 4 samples had a negative T.C. near 25 deg C. (whereas on my LM399 I have more or less only a open hysteresis depending on rising/falling temperature).
On ADR#1 I checked direct measurement of ADR output versus multiplexer output which gave no significant difference.

So I have to look for different influence on "position dependant drift".

with best regards

Andreas

[dietert1]

Mouser have recently got some evaluation boards in stock: https://www.mouser.co.uk/ProductDetail/?qs=4ASt3YYao0UVSWYehgmbvQ%3D%3D, users guide: https://wiki.analog.com/resources/eval/adr1399e-ebz.

Bought one to test, for the price (£54 including VAT) it seems like a pretty good deal for a fully tested and assembled board.

One of those arrived here yesterday. It works, yet i don't know whether those AD people understand what they are doing. Not only the switchmode converter and the missing low thermal EMF output.
In their wiki they explain they wanted to make something "portable calibrator style". But where is the calibration value? I got "7.05 V" that's it. Probably they can do a precision measurement, but they didn't do so?
Guess their 573 € LTZ1000 kit is similar and the chinese offering an LTZ1000 board on ebay for a fraction of the price understands better what a reference is good for.
Sorry for the rant.

Regards, Dieter

[xerbo]

The utilised switchmode converter is isolated and has a built in low noise linear regulator so should be fine. As for the terminals I think they are going for the "give it as much thermal mass as possible to minimise thermal gradients" approach, weather or not that is effective I don't know (probably not very).

But as a portable calibrator it is totally and utterly useless, the fact you already need a already calibrated voltmeter is probably a killer for most people.

[iMo]

You have to burn-in it for 1k5 hours at least.. ADI will not do it for us, imho. But a good opportunity for the voltnuts here to offer the "calibration" of the module after the burn-in period. The switching regulator - best to look at the input/output voltage with an o'scope or similar.. And it is an evaluation kit, so we have to evaluate 

[Kleinstein]

Many of the commercial, isolated DCDC converters produce quite quite some common moder interference. So there is some capacitive coupled current between the 2 ideally isolated side. This may not effect the reference circuit itself, but it may effect some DMMs or amplifiers used for the test.
The problem with the common mode interference is, that it is hard to filter.

[dietert1]

You have to burn-in it for 1k5 hours at least.. ADI will not do it for us, imho. But a good opportunity for the voltnuts here to offer the "calibration" of the module after the burn-in period. The switching regulator - best to look at the input/output voltage with an o'scope or similar.. And it is an evaluation kit, so we have to evaluate 

Started evaluation by hooking it up to an unused 6.5 digit meter and it shows 7.04264 or 7.04265 from the start. Most people would do this with a pocket calibrator. I mean, the reference is a reference at the 6 digit level, yet of unknown voltage except 7.0xx V. When i ordered references from Mr. Geller 13 years ago, they came with calibration info. This know-how has been around for many many years.

When there is some spare time, i will probably add a linear supply and a 10 V output stage/buffer with a NOMC array divider in order to make it 10 V and integrate it into my scanner setup.

Regards, Dieter

PS: First step already finished. On a LM723 based linear supply with 15.002 V (switchmode supply inactive) the reference yields 7.04277 V into the same meter. An up shift of 17 ppm.
After undo and running it on USB-C again it comes back to previous value, so it isn't damaged. Supply is 14.4 V now and a bit noisy.

[iMo]

Many companies go Agile these times.. The calibration of the module will be done by ADI in the next Sprint, after the Sprint Review, Retrospective Meeting, Backlog Refinement Session, new Sprint Planning, provided the Product Owner agreed that with the Scrum Team and the Scrum Master and the Calibration will be put into the Product Backlog as the new Story in the 1399 Epic, with enough Story Points to be Burned-down during the Calibration Sprint.

[dietert1]

My previous supply test detected another design problem of the evaluation board: The ADR1399 won't be low noise and stable to 1 uV without a well made bootstrap gain stage. Calculate: 120 uV shift for a 0.6 V supply change results in a +/- 5 mV requirement for the supply. Also i am wondering about the resistor they used to feed current into the ADR1399.
Already replaced the two zero ohm resistors next to the reference by pieces of copper wire.

Regards, Dieter

[miro123]

My previous supply test detected another design problem of the evaluation board: The ADR1399 won't be low noise and stable to 1 uV without a well made bootstrap gain stage. Calculate: 120 uV shift for a 0.6 V supply change results in a +/- 5 mV requirement for the supply. Also i am wondering about the resistor they used to feed current into the ADR1399.

How do you calculate this numbers? Here are mine.
R4=1.33k R=0.04Ohms for TO46 R=0.011Ohm for LCC
To46 sensitivity is =0.04/1330 =1/ 33250  - 33,25m shift make 1uV
LCC  = 0.011/1330 = 120909   - 121mV makes 1uV

[dietert1]

My numbers are an observation. When i connected a 15.002 V linear power supply to the test points, reference output shifted up about 120 uV (17 ppm of 7 V). To check whether i damaged the reference, i reverted the change and it returned to the previous output voltage. Then i measured the supply from the switchmode converter to be 14.4 V. So a 0.6 V supply change caused 120 uV output voltage change.
There could be other side effects, like heat or RF emitted by the switchmode converter.
Did you notice that everywhere in the datasheet they recommend to feed 3 mA into the reference, but the 1K33 in their evaluation kit feeds about 8 V/1K33 = 6 mA.
Meanwhile i made the bootstrap/gain stage and i need some idea how to feed 10 V into the nice thermal protection without damaging things.

Regards, Dieter

[MiDi]

Made the rejection ratio calculations for bootstrapped LM399 & ADR1399 some time ago, Excel attached:

[opa627bm]

Apology if I didnt read through the entire thread.
Anyone is having the EVM drifting due to the 1Mohm isolation resistor between the AGND and PDNG?
For the ADR1001 it is using the same design and I have 2 EVMs that is showing the same (kind like EMI ) issue, and I ended up bridging this resistor.

[dietert1]

By "bridging the resistor" you mean a short? This shouldn't be necessary. You can keep isolation by using a small capacitor to bridge the resistor, like the 4.7 nF found in many switchers. Then you keep the reference voltage isolated.
My impression was the switchmode converter isn't stable enough, so i am using an alternative linear supply. My ADR1399 EVM also got the bootstrap/gain stage. Not the one calculated by midi, but the "real thing". It has a chopper amplifier and a NOMC divider and outputs 10 V to some ppm. Put the whole circuit into a metal case. Running in my scanner setup, hourly averages exhibit noise of about 2 uVpp plus strong initial drift of -2 uV/day. As this is a new build, both noise and drift should calm down soon.

Regards, Dieter

[opa627bm]

By "bridging the resistor" you mean a short? This shouldn't be necessary. You can keep isolation by using a small capacitor to bridge the resistor, like the 4.7 nF found in many switchers. Then you keep the reference voltage isolated.
My impression was the switchmode converter isn't stable enough, so i am using an alternative linear supply. My ADR1399 EVM also got the bootstrap/gain stage. Not the one calculated by midi, but the "real thing". It has a chopper amplifier and a NOMC divider and outputs 10 V to some ppm. Put the whole circuit into a metal case. Running in my scanner setup, hourly averages exhibit noise of about 2 uVpp plus strong initial drift of -2 uV/day. As this is a new build, both noise and drift should calm down soon.

Regards, Dieter

Yes, I can try the capacitor too. The ADI engineer told me the PSRR from the part should be good enough for it ...

[Andreas]

Hello,

some further measurements regarding PSRR of the heater voltage.
compare also the measurement of macaba:
https://www.eevblog.com/forum/metrology/adr1399-reference/msg4114465/#msg4114465
In my case I use a positive heater voltage from 9-30V (supplied by a external analog power supply).

Below 15V the PSRR gets quickly worse from ~0.5ppm/V up to 1.2 ppm/V.
2 samples of LM399 (in TO-46) seem to be slightly better especially at low heater voltages.

unfortunately I still have not found the "position dependant drift"
https://www.eevblog.com/forum/metrology/adr1399-reference/msg4468918/#msg4468918

maybe there are several different reasons:
Partly a explanation may be that on the LM399 ageing box the temperature distribution at the output connector is more equal.
Measured 2-3 degrees difference near left and right pin of the connector due to the heaters of the ADR1399 on one side of the pcb.
But also direct soldering of the measurement lines near the "far" multiplexer did not solve the differences between hot and cold place.

I additionally found out that the temperature of the multiplexers varies by up to 3 deg C depending on the level of the Adress/chip select lines.
So obviously the level shifters in the MAX4052A need considerably high power when not being at the Gnd/VDD level. So I might need external level shifters to keep heating away from the multiplexers.

with best regards

Andreas

[Kleinstein]

The quite variable power level for the max4052 is interesting. This could indeed be a problem for a precision measurement. There can be 2 parts: the level shifters and inputs that get a relatively weak signal. 74HCT... also show surprisingly high supply current with 3 V at an input, though a perfectly OK input level.

[Andreas]

Hello,

yes current consumption is unexpected high.
With 15V supply voltage and 5V from processor on the adress and CS lines I get up to 85 mW power consumption for one MAX4052A (with the same date code).
So I am waiting for the level shifter boards. (BSS138 based).

with best regards

Andreas

[dietert1]

The ADR1399 evaluation board i mentioned above was put into a hermetic enclosure. It got an OPA189 buffer that is a 7 to 10 V gain stage with a NOMC divider and a Sallen-Key low pass filter with two Wima MKP4 10 uF/250V. After observing a TC of about +0.2 ppm/K i made the enclosure a simple thermostat with an analog PI controller. Image shows heater resistor, transistor and NTC. Controller and power supplies are still external.
After the most recent temperature controller mod about 32 hours ago (arrow) reference noise is down to about 0.8 uVpp in comparison to the average of our two LTFLU references. ADR1399 datasheet allows about 2 uV per 10 V. Drift estimate after one month of operation would be 180 nV/day = 6 ppm/year (probably overestimated).
There are still events decreasing voltage, so it needs more "education".

Regards, Dieter

[branadic]

Why don't you add a bit of copper in the 7V --> 10V gain stage to compensate for the remaining t.c. overall?

-branadic-

[dietert1]

I didn't have the time and right now i don't know whether
- the NOMC divider suffered a minor TC while fine tuning for 10 V or
- whether it is the ADR1399 or
- whether it is the Pomonas' fault (thermal EMF).
It's an experimental setup. Next time i'd probably use soldered connections for the reference voltage output or move the hot reference further away from the Pomonas.
The temperature controller and the precision parts of the power supply should go into the box. One of the tests to do yet is running the ADR1399 with its heater off. The thermostat is good for +/- 20 mK.
My conclusion is that the ADR1399 with its increased zener current and its current price is
very interesting. Once the TO-46 version arrives at the distributors one should try an array.

Regards, Dieter

[Andreas]

Hello,

I want to give some update to the "position dependant drift".

The main factor is obviously thermal drift (ppm/K) of the references.
If I do not use the environment temperature but the PCB temperature the value increases up to 0.3 ppm/K.
(see overview).
 
I tried to compensate the readings with the PCB temperature.
But the improvement is much lower than I would expect it.

There is a remaining "low frequency noise / instability" of the values of about 2 ppm.
This is more than that what I usually get with my measurement system.

So the questions for me are:
- Is this related to my setup or do others also have larger instabilities with ADR1399 than with LTZ1000 or LM399?
- Is it only the LS8 package or is it also characteristic for the TO package?

At least it seems that the initial ageing drift is rather low.

with best regards

Andreas

[Andreas]

Hello,

I also did some tilting tests in 6 orientations of the PCB.

0 = top = component side of PCB on top
5 = bot = solder side of PCB on top
1 = north = Pin8 ADR1399 on top
2 = east  = Pin6 ADR1399 on top
3 = south = Pin4 ADR1399 on top
4 = west  = Pin2 ADR1399 on top

I had guessed that between north and south (PIN 4 and 8 ) there would be the largest differences.
https://www.eevblog.com/forum/metrology/adr1399-reference/msg4457818/#msg4457818

But there it seems that I was not completely correct.

Pin 4 on Top = south = 3 gives mostly (not always) the highest output voltage.
Pin 6 on Top = east   = 2 is likely to give the lowest output voltage.

The tilting drift amplitude with some thermal isolation is comparable to that what I also usually have on LM399 devices.
(see overview in previous post)

with best regards

Andreas

[branadic]

Some of you already know that I've revised the PWM-DAC board, now called xx99-PWM-DAC, that allows to populate the ADR1399 and its snubber network too. One of the boards was assembled and tested. I populated the same components such as 20 MHz crystal that I've used on the LM399-PWM-DAC and used the same firmware (8.16), but measurements weren't satisfying and I observed huge spikes.

I finally ended up sacrifying my LM399-PWM-DAC board which was used for the EU-CalClub test run and frankensteined the ADR1399 to it. I also changed the 2W NTC connection to a 4W one and took, after some adjustment of the gain, first measurements. For some yet unknown reason no such spikes are visible here, but with the filtering as it is by now the noise at the output is larger than it should be, but still an improvement with respect to the former populated LM399.

-branadic-

[Andreas]

Hello branadic,

is it the TO-46 package?
how much thermal isolation has been used?

with best regards

Andreas

[branadic]

Yes, TO-46 package as I don't have trust in that ceramic SMD packages without proper low strain assembly solution such as FlexPCB.
There is a bit of foam block around it, similar to what was used for LM399-PWM-DAC.

-branadic-

[Andreas]

as I don't have trust in that ceramic SMD packages without proper low strain assembly solution such as FlexPCB.

Hmm,

I measured 55 mA Heater current @15V and room temperature for 4 populated ADR1399 in LS8 package on my PCB with some thermal isolation:
one cotton pad around the edge of the PCB + some thin (2 mm) foam foil.
So only 14 mA per LS8-package (+3mA Zener current)

I fear with a FlexPCB the environment temperature limit to no thermal stability is reached soon.

with best regards

Andreas

[branadic]

I don't see why the FlexPCB should be worse in performance. What I have in mind is a small breakout board LS8 to through hole connections on one side only, so that the FlexPCB is floating underneath the LS8 footprint to prevent mechanical stress to the package. Does that sound reasonable to you?

-branadic-

[iMo]

I mentioned this issue in the adr1001 thread too - a need for a flex board as well..
@Andreas - what is the env temp limit to no thermal stability? Is that defined somehow in the DS?

[Andreas]

Hello,

no its not specified explicitly.
But you should stay away from the 90 deg C setpoint with the chip temperature.
And that with the 3 mA zener current.

I don't see why the FlexPCB should be worse in performance.
Does that sound reasonable to you?

It is worth a try. (in comparison to the TO-46 package).

By the way did you test already the T.C. of the devices in TO-46 package?
My LS8-values are higher (>=0.08 ppm/K) than the values that I have from the LM399 measurements (<= 0.07 ppm/K) near room temperature.

with best regards

Andreas

[branadic]

By the way did you test already the T.C. of the devices in TO-46 package?

No I didn't and probably won't for the individual samples, but for the overall circuit including the gain stage. In the end I'm not that much interested in what every component contributes, but in compensating the final circuit for low (almost no) t.c.

-branadic-

[Andreas]

Hello,

to come back to the "position dependant drift"
https://www.eevblog.com/forum/metrology/adr1399-reference/msg4597733/#msg4597733

after having added the level shifters between processor and multiplexers I added 2 LM399 to the PCB which I had in the drawer to exclude some drift by the multiplexers.

Unfortunately both LM399 (from LT) had a large initial drift.
Additionally #2 of the both LM399 had initially popcorn noise which I recognized on a first T.C. test

After some checks with constant temperature I recognized that the temperature compensation with respect to the PCB temperature (0.3 ppm/K on #4) showed some "over compensation".

So I finally measured the T.C. of ADR#01 - #04 again which was surprising to me.
The PCB temperature range with the same environment sweep has been reduced (and shifted to higher temperature) by the adittional "heaters" on the PCB.
The ADR in LS8 package do not correlate well to the PCB temperature.
But they obviously correlate good to the "environment" temperature (which is sensor glued to a heatspreader aluminium plate below the ADR1399 test PCB).
Attached a "overlay" between a old and new measurement of PCB related TC and environment related TC of ADR1399#03 which obviously driftet nearly zero since last T.C. measurement which was end of november.

So now I can explain the position dependant drift:
I relied on the PCB temperature sensor near the multiplexers to compensate for environment temperature.
Initially I had 2 degrees self heating of the multiplexers due to missing level shifters creating PCB temperature rising which does not correlate to the behaviour of the ADR1399 which react more to the environment temperature.
Also placing the ADR1399 PCB above another heat source gave a large temperature gradient between the both PCB temperature sensors of +7/-2 deg C difference.

Additional point:
We had some rather cold days where I had down to 17 deg C in my lab.
Normally the T.C. of my measurement devices (24 Bit ADCs) is temperature compensated so this should be no issue.
But with my other daily measurements I recognized that the standard deviation of my ADCs has nearly doubled from 0.25 ppm to nearly up to 0.5 ppm.
And partly some devices developed some kind of popcorn noise.
In the beginning I only evaluated one ADC (ADC#13). Now I am evaluating all 3 stable ADCs (#13, #15, #16).

comparison ADR1399 vs. LM399
- the 2 lm399 show a much larger initial ageing drift as the 4 ADR1399LS8 samples.
- The T.C. of both LM399 (TO-46) samples is lower than the 4 ADR1399LS8 samples.

lessons learned:
- the PCB temperature is not usable for temperature compensation of ADR1399LS8.
  instead I should use a Metal housing and measure the temperature below the PCB.
- Never rely on one single measurement device for ageing measurements.
- Build a ageing PCB with some kind of injection channel for a proven stable reference (Like my LTZ#4 and/or LTZ#9) as sanity check.

with best regards

Andreas

[r6502]

Hello All,

I will get some ADR1399 in LCC version  and I lokked at the data sheet if there are informations given, where to place the damping network, sould it be placed cloce to the pins, like you do with caps for decoupling, or can it be a litttle more far away? Should they be close by pins 1/7 or better close by pins 2/6 or where both connections are joined when usig kelvin connection. Or connect pins 1/2 and 6/7 direct at the chip and place the damping network there?

Guido

[Andreas]

Hello,

if you look at the user guide (evaluation board) schematics:
https://wiki.analog.com/_detail/resources/eval/adr1399e_schem.jpg?id=resources%3Aeval%3Aadr1399e-ebz
its more on pin 2 + 6.
But on the other side they have used the bridges instead of the transistor.

So I have simply connected both pins at the chip and put the network directly under the chip on the soldering side (together with the cap for the heater).

with best regards

Andreas

[Kleinstein]

The damping part is mainly relevant for frequencies of some 50 kHz to some 1 MHz. So the exact position should not really mater. More distance from the chip (e.g. away from the heat) and also more to the sense or force should not matter that much. A position more to the force side makes sense, especially if there is an additional transistor driver in between as in the evaluation board.

From the shown imedance curve it looks like the capacitance is more in the low side. Especially with MLCC capacitors I would plan with space for 2x1 µF to at least compensate of the biasing effect and not get much below 1 µF.

[alex-sh]

Hello all,

Did anyone compare the LS8 package vs TO?
Particularly, I am interested in the long term drift, but suppose nobody has this info at this stage?

[Andreas]

Hello,

I still have no TO-46 samples.

The only thing that I can say:
I have difficulties with the LS8 package to monitor any useful ageing data.
The tempco is rather high on my samples.
And the LS8 package is rather picky about the location of the temperature sensor.
I hope that I will have a better setup for ageing monitoring at end of this week.

with best regards

Andreas

[alex-sh]

The tempco is rather high on my samples.

Andreas
Do you think this is just your samples or this is a general rule of thumb for LS8?

[Andreas]

Hello,

We have still not enough measurements (4 on my side) to make any statistics.

with best regards

Andreas

[branadic]

I have measured my ADR1399-PWM-DAC 10 V voltage reference over temperature (18 ... 28 °C) against an internal temperature sensor with the result of -0.223 µV/V/K.

-branadic-

[JodelJonny]

Hello there
I am new here. After sucessfully repairing a HP3456A I solwly but surely waddle alogng the path towards voltnuttery. And after years of silent observing I now have decided to try to build a 10V reference based on the ADR1399. The LTZ1000/ADR1000 is just unreasonably expensive and overkill for now.

I designed a simple PCB based on the circuit from the datasheet.
I think that should work out. Do you argee? Are there any errors or things I should improve before ordering the parts?
The resistor divider to amplify the aprox. 7V to close to 10V is planned to be build from Vishay TNPU e3 0805 smd resistors with 0,1% and 5ppm. There are a multiple footprints to be able to tune the output close to 10V and they are all encloed with copper planes on both sides of the pcb to keep the temperature differences between them small.

I plan to supply it with stable 15V DC from a simple linear transformer power supply. Or would it be better to include a 7815 on the pcb?
The reverse polarity protection is handled by a DMP3099L-7 P-channle mosfet. I think it can do that job without any additional components, or did I overlook something?


Thanks for the feedback! :-) 

[iMo]

I would not place the 7815 on the board (it changes its temperature based on the mains voltage and adds a heat source).
The 7815 should be there but after the protection mosfet.
I would use the 1399_zener_bootstrapping from the opamps output. That would require to add a transistor in order to de-load the opamp's output.

[dietert1]

Maybe produce a reference modul that fits into the HP 3456A? The ADR1399 can be used as an upgrade for the existing LM399. The 10 V to 7 V divider and the buffer transistor that are external to the HP reference module can be optional parts, maybe hidden on the bottom of the board.

Regards, Dieter

[JodelJonny]

@imo Good point about the bootstrapping. Didnt really understand that before but now I think I do. I added Q2 and reduced the value of R1 to 1k to maintain 3mA zener current.
What transistor should I use? I dont have much expereince with all this. I guess the specs are not that important.

@dieter1 The 3456A ist right now the only thing in my posession of wich I can at least assume that it is reasonably stable and likley even still kinda accurate.
Long story short. I wont touch it until I am sure to be able to monitior the cosequences of my actions. But this might be an interesting projet for the future.

[dietert1]

I did not propose to touch the HP 3456A, but to contribute something useful. It's a plug-in module with connectors, easy to exchange. I would never unsolder a LM399 from a HP reference module, but make a replacement module.

Sooner or later compatible reference modules with ADR1399K will be requested. They also fit the HP3457A. We have in total 7 of those meters.

Regards, Dieter

[iMo]

@imo Good point about the bootstrapping. Didnt really understand that before but now I think I do. I added Q2 and reduced the value of R1 to 1k to maintain 3mA zener current.
What transistor should I use? I dont have much expereince with all this. I guess the specs are not that important..

Any modern transistor will do. It also depends whether you want an output current limitation for safety reasons, that would require 2 transistors, and also sometimes people shift the opamp's output voltage to a lower value by a zener diode, like below. But is depends what you want..
PS: the R7 sets the current limit roughly..

[Kleinstein]

There is usually no need for the extra diode D1 in series with the zener.
In many cases one would also need additional capacitive feedback directly from the OP-amp for stability.

For the OP-amp there is no need for the relative expensive LT1012. A relatively cheap OPA202 may be a better choice.

[JodelJonny]

Thanks for the great advice!

So in the circuit shown by imo the opamp is sinking current or pulling Q1 down towards the Out_level against R4, correct?
May I ask why one would like to shift the opamp's output voltage lower? And what is the idea behind D1?

I copied the LT1012 from the ADR's datasheet but I'll happily save the money if OPA202 is good enough.
Overcurrent protection would be nice to have. Little accidents do happen.

What I want is a simple and cost effective voltage standart to give me a stable reference to compare my 3456A to and use is as a transfer standart so I could bring home a known voltage if I come across an oficially calibarted precision multimerter.  That combined with learning about analog desing, pcb layout, metrology and so on. This part is already working.

@dietert1 Sure but right now I have no way of evaluating the perormance of such a desing. Thats one of the reasons I am building a Voltage refernece
 

[Kleinstein]

For a 10 V reference there is no real need to use the zener diode to shift the level. One sometimes does this to get an output voltage closer to the supplies (e.g. 14 V out from 15 V supply).

The diode D1 may be wanted, so that the base current to the transisor is only from the resistor and not from the OP-amp. Usually the op-amp outputs are current limited and thus no real problem without D1, but the diode would not hurt either.
The OPA202 is nearly rail to rail out, and one may still need some startup mechanism, if the zener current is taken from the 10 V. The zener diode part would provide the start-up.

Ideally one would want to have both an output for the 10 V level and the raw 7 V from the reference, so that the gain drift does not matter and can be checked separately. It could be enough to just have a series resistor (e.g. 1 K range) as a crude protection for the raw reference.

[Andreas]

I copied the LT1012 from the ADR's datasheet but I'll happily save the money if OPA202 is good enough.
Overcurrent protection would be nice to have. Little accidents do happen.

What I want is a simple and cost effective voltage standart to give me a stable reference to compare my 3456A to and use is as a transfer standart so I could bring home a known voltage if I come across an oficially calibarted precision multimerter.  That combined with learning about analog desing, pcb layout, metrology and so on. This part is already working.

Hello,

I think: for a single device the price of the OP-Amp should not play any role.

I think you should take more care about star grounding. (A power plane easyly can carry some 10 uV from heater current).

Thermal management is also to consider. (Thermally isolating and packaging into a metal housing to avoid large temperature gradients).
Depending on how stable your lab temperatures are you should take also care to select the right reference and divider resistors.
A LM399 might give better results than a ADR1399.
At least some temperature sensor would be needed if you want to transfer the "volt".

I would at least put some 100 nF capacitors directly at the Heater pins and the zener output to avoid EMI drifts. (for the ADR the 5R 100nF should be placed very close to the zener).
And the buffer amp output also should be also shorted with a 100 nF (of course with a isolation cirquit against capacitive loads).
see example here:
https://www.eevblog.com/forum/metrology/emi-measurements-of-a-volt-nut/msg2684100/#msg2684100

I would build at least 3 devices to observe ageing drifts or drifts related to temperature shocks during transport of one device.
Otherwise you will never know wether the volt that you measured on calibration is the same that you imported into your lab.
I had a 4 ppm drift on a LM399 before and after calibration. I finally recognized that the LM399 output is dependant on orientation.
So I now mark the housing of the LM399 reference where the "Top side" has to be.

with best regards

Andreas

[JodelJonny]

Thanks again!

I tried to understand and implement your suggestions. To be honest, the amount of information in all those threads is quite overwhelming at times, especially since I have no formal education in electronics.

I plan to mount the pcb in a metal housing with temperature sensing but I am not sure were I want to measure the temperature.

@Andreas I have now added two 100nF caps (C6 &C7) directly at the pins of the ADR1339. One of which is now in parallel with the 5.1Ohm 1uF (R3 & C3) RC circuit from the datasheet which I also moved on the backside and very close to the pins. Is that correct or are those circuits redundant?

I am not sure about what you meant with the 100nF to short the buffer amp output and the isolation circuit against capacitive loads. I added C5 in the schematic.

I switched to a ground plane on the back side and added some slots to break temperature gradients and thermally isolate the ADR a bit.

I hope it is now better and would appreciate further feedback :-)

[Kleinstein]

The temperature sensor should be more at the resistors to set the gain - not directly at the ADR1399, where the temperature is regulated in parts..
For the temperature sensor I would more prefer an anlog one, like PT1000 , 10 K NTC or maybe just a diode.

The ground plane is not a valid replacement for a star ground. I would prefer less ground around the LM1399 to reduce the heat loss even more.

The configuration for the gain setting resistors looks a bit odd. The required gain is a bit less then  3/2. So the fist approximation could be 3 identical high grade resistors (e.g. 10 K) and than some additoinal fine trim pallel to the one on the high side.  To get fine trim without very high values the fine trim would ideally be 1 resistor (~ 65 K ?) in series with some 3-4 in parallel (e.g. 5 K + .... ).

[iMo]

Here is an interesting setup with 8 resistors (an array), I've tried to sim it and a resistor in series with the low side 10k one, around 450ohm, could be used for the fine trimming (do simulate the stuff with your values first).

https://www.eevblog.com/forum/metrology/statistical-arrays/msg4703825/#msg4703825

Also, as I wrote above, the 15V voltage regulator should be placed "after" the mosfet, otherwise you get -2mV/C at the 15V supply.

The 22k and 15ohm in the buffer were examples only in my schematics, you have to adjust that values based on your requirements.

To the 100nF feedback capacitor at the opamp - do we really need it? Where the instability comes from? From the output load capacitance? Could we somehow calculate the value really required? The output buffer is an emitter follower with no gain, imho.. Moreover I did some simulation in past and I saw decreasing PSRR of the opmap with increasing the value of that capacitor. But perhaps that is just an issue with the opamp model used..

[Kleinstein]

Some capacitance for C5 is likely needed due to the capacitive loading. How much depends.  With the output capacitance one shoud aim for C5 * R5  > 2*C6*R12 . So less than 100 nF should be OK,  more like the 1 nF range.

The MOSFET used as a kind of ideal diode may give less than -2mV/K, but still not ideal. I would not mind a LDO on the PCB that much, at least as an option to populate.

[DavidAzulay5000]

Next week I receive several units of the ADR1399
I started doing simulations in LTSPICE
to volt reference 10 volts
In your opinion, there is something to change in the circuit
I have a lot of vishay precision resistors 0.01%
of
10K
5K
2.5K
1K
100 ohm
Etc. etc
So I will use them
I have dozens of units

[iMo]

Here is an example re PSRR - the injected 200mVpp 50Hz hum into 15V Vcc translates into 3.7uVpp with the 1nF, and into 310uVpp with the 100nF feedback capacitor..

[Andreas]

@Andreas I have now added two 100nF caps (C6 &C7) directly at the pins of the ADR1339. One of which is now in parallel with the 5.1Ohm 1uF (R3 & C3) RC circuit from the datasheet which I also moved on the backside and very close to the pins. Is that correct or are those circuits redundant?

Hello,

C7 can only be used with a LM399
C3+R3 is needed by the ADR1399 for stability. C7 will lead to oscillations here.

With the output capacitance one shoud aim for C5 * R5  > 2*C6*R12 . So less than 100 nF should be OK,

its more like C5 * (R5||R4) > (C6 + CLoad) * (R12 + semiconductor resistances)
100nF is plenty but you do not need to optimize for speed.

with best regards

Andreas

[Andreas]

Here is an example re PSRR - the injected 200mVpp 50Hz hum into 15V Vcc

200 mV is insane.
if you have more than 5 mVpp you should work on your star grounding concept.

with best regards

Andreas

[iMo]

Here is an example re PSRR - the injected 200mVpp 50Hz hum into 15V Vcc

200 mV is insane..

That is just an example to visualize/demonstrate the issue. The larger the feedback capacitor the lower attenuation of a potential noise coming from the Vcc (provided the simulation works ok).. What is important is to see the ratio 1nF/100nF == 3.7uVpp/310uVpp.. Thus with your 100nF FB capacitor the Vcc noise attenuation (related to the 10V output) is 100x lower than with the 1nF FB capacitor..

[dietert1]

Yes, the simulation shows that for all practical purposes there is no problem. Use a power supply with less ripple and try various capacitors to find a small capacitance that gives good stability. I'm not shure the simulation can resolve this. One possible method to improve stability is a small capacitor from the opamp output to the output transistor base, once more like 1 nF or below.

One can also replace R4 10K by a JFET constant current source. With the zener between opamp and transistor one can run the opamp from the 10 V output voltage in order to eliminate supply drift/ripple going into the opamp. Then the only entry point left is the collector of the output transistor.

Regards, Dieter

[iMo]

@dietert1: Something like this..

[dietert1]

Yes, appears to work. As far as i understand ripple with 100nF is down from 310 to 9 uV. With a 1 nF capacitor you should be down from 3.7 uV to about 120 nV. At 200 mV supply ripple.

Regards, Dieter

[DavidAzulay5000]

 

[Kleinstein]

The use if the LT5400 for the gain is an interesting solution. Combining a separate resistor with the LT5400 array is not ideal, though may still be acceptable for only some 2% of the final voltage.

The suggested LT1113 in the plan is however not a good choice, as it has too much 1/f noise.  I would more look at LT1013, OP07, OPA202, OPA207 or ADA4077 or maybe an auto zero type like LTC2057 or MCP6V51.

It would make sense to plan for at least some filtering and an output stage that is capacitive load tolerant like discussed before.

[DavidAzulay5000]

The use if the LT5400 for the gain is an interesting solution. Combining a separate resistor with the LT5400 array is not ideal, though may still be acceptable for only some 2% of the final voltage.

The suggested LT1113 in the plan is however not a good choice, as it has too much 1/f noise.  I would more look at LT1013, OP07, OPA202, OPA207 or ADA4077 or maybe an auto zero type like LTC2057 or MCP6V51.

It would make sense to plan for at least some filtering and an output stage that is capacitive load tolerant like discussed before.

i am going to use the ltc2057 i order them the 925 ohm resistor is a 1k 20 turn trim pot

[iMo]

Do not use trimpots (10t/20t), those are source of instabilities. When you have to use a trimpot then with as small as possible value compared to the rest of the divider (1k is rather large value, imho).
The LT5400 ratio TC is <1ppm/C, thus the 925 ohm resistor contributes heavily to the overall TC. The trimmers are +/-50ppm/C typically.
Do a simulation over TEMP, for example:

.op
.step TEMP 20 40 1

with "R3=925 tc=30e-6"

as an example..

[DavidAzulay5000]

Do not use trimpots (10t/20t), those a source of instabilities. When you have to use a trimpot then with as small as possible value compared to the rest of the divider (1k is rather large value, imho).

My method is always to put TRIM POT resistors
At first
Check what the exact resistance value is after everything stabilizes
And after that replace the resistors with the exact value

[iMo]

Yes, appears to work. As far as i understand ripple with 100nF is down from 310 to 9 uV. With a 1 nF capacitor you should be down from 3.7 uV to about 120 nV. At 200 mV supply ripple.

Regards, Dieter

Output ripple so low that is not measurable in the LTSpice with 1nF and 200mV Vcc hum..

The situation changes of course with squared pulses (ie a switcher), with 200mVpp 2us/10us pulses the output ripple is around 600uVpp with both 1nF and 100nF and the OP07.

[DavidAzulay5000]

I have a problem with an error in the simulation in LTSPICE
Once I try to use the LTC2057

[iMo]

Attach your .asc file..

[DavidAzulay5000]

Attach your .asc file..

[iMo]

For example with 30ppm/C trimmer..

PS:
and the Vrefout change in ppm: with combination of a 25ohm 30ppm/C trimmer and a 900ohm -4ppm/C resistor (as an example).

[DavidAzulay5000]

For example with 30ppm/C trimmer..

PS:
and the Vrefout change in ppm: with combination of a 25ohm 30ppm/C trimmer and a 900ohm -4ppm/C resistor (as an example).

thanks i fix the error problem i update the ltspice and the error no more

[DavidAzulay5000]

For example with 30ppm/C trimmer..

PS:
and the Vrefout change in ppm: with combination of a 25ohm 30ppm/C trimmer and a 900ohm -4ppm/C resistor (as an example).

How do you make the graph of temperature in relation to voltage in LTSPICE

[iMo]

See the commands in my above simulation..
.op
.step TEMP 20 50 1
and after the simulation stops do select a voltages or currents (click on the node or wire in the schematics), etc..

[DavidAzulay5000]

See the commands in my above simulation..
.op
.step TEMP 20 50 1
and after the simulation stops do select a voltages or currents (click on the node or wire in the schematics), etc..

Attach your .asc file.. thanks

[DavidAzulay5000]

I managed to create the graph
But it's strange that for this I have to delete the entire part of the LT3083

[DavidAzulay5000]

I managed to arrange the simulation to include everything
And I also did a TRIM POT simulation
From a temperature of 20 degrees to 80 degrees
Strange the stability at 55 degrees

[mawyatt]

Maybe a dumb question but will throw it out anyway.

Do we know IF the models utilized in these simulations reflect the actual devices being simulated?

Are second/third order effects such as PSRR, chip heater effects, thermal gradients and so on included is such device IC models? Not to mention effects related to actual PCB implementation and the environment. Also, noted the discrete resistor TC seem to include the linear temp coefficients and not the quadratic effects, and what about the LT resistor array TC effects?

Best,

[iMo]

..
..Strange the stability at 55 degrees

That is because there is a TC of the 1399 zener combined with the TC of the trimmer - see below an example, trimmer's TC from -50ppm/C to +50ppm/C. Also your LT3083's TC may contribute as well.
Vrefout in ppm.

[Kleinstein]

I doubt the models for the parts are that accurate to worry about the 2nd order TC.  The main point would be checking the coarse part (e.g. how much is the TC of the trim resistor attenuated) and for the simulation one needs some x-axis (e.g. time or temperature) to get the graphs and more conventional interface.

Beside the 100K/10K LT5400 array, one could also use an array with 8 equal resistors (e.g. TDP or NOMC )  for a 1.5 to 3.5 R ratio. This starts at the nominal 3:7 ratio and would need trim options for up and down.  The ADR1399 is nominal 7.05 V and thus in between the 2 approximation. The shown trim resistor in the LT5400 version only works up to some 7.2 V. A rare but still in spec higher voltage unit would need a different position trim resistors to reach 10 V.

[Conrad Hoffman]

Has anybody come up with a PCB for the 1399 ceramic SMT version that's generally considered to be good?

[Andreas]

Hello,

I think I have now solved my "position dependant drift". see also:
https://www.eevblog.com/forum/metrology/adr1399-reference/msg4683005/#msg4683005

I have put the PCB into a die cast housing (Hammond 1590U)
https://www.reichelt.de/alugehaeuse-1590-u-120-x-120-x-58-6-mm-aluminium-1590u-p221457.html?&nbc=1

The temperature sensor from CPU is now cabled + taped on the bottom of the housing.
The PCB with the 4*ADR1399LS8 + 2*LM399 sits with some cloth (thermal isolation) within the metal housing.
Of course the effective T.C. with respect to the "case" sensor also is different from "environment" and "MUX" temperature sensor.
So I had to determine a new set of T.C. coefficients.

Now the measurements are stable (as I am used from my ADCs) and the T.C. of the references can be compensated.
The temperature drift of the ADR1399 is still much higher than the LM399 samples.
The LM399 shows more stable uncompensated values (except for some popcorn noise on day 4 since Hammond packaging).

The T.C. compensated values are additionally averaged between 3 ADCs to reduce further measurement uncertainities.

The overview shows the new effective linear T.C. values which are minimum a factor 3 higher for the ADR1399LS8 samples than for the LM399s.

with best regards

Andreas

[DavidAzulay5000]

LM399 + OP AMP LT1001 COMBO
PERFECT 10V Reference 10.000.00
v

[gamalot]

The simplest ADR1399 comes, only two 1uF capacitors and a 1K ohm resistor, measured with 34465A, it seems not bad! 
 

[Kleinstein]

The 34465 compares the ADR1399 to it's internal LM399.

[gamalot]

The 34465 compares the ADR1399 to it's internal LM399.

I also did a test with my 6581T (1000 samples, 20 NPLC, 14uV Span, 2uV Std dev), these results were almost identical to the 34465A.

[iMo]

Evaluating an ADR1399 (or an LM399 and better refs) with a 6.5digit voltmeter is difficult. You have to know your meter well at the 10V range - you have to know its TC to compensate for its internal temperature and/or ambient, you have to log an 8-10 hours long measurement (at 100NPLC) and evaluate several thousands of samples at minimum (analyze the graphs, compensate for internal/ambient temperature, remove the outliers, provide filtering/averaging/smoothing). And you get a single number (I do it this way). That all with an assumption (not valid one, btw.) the long therm drift of your meter is zero.

The LM399 inside those 6.5meters does a random walk typically within 5-10uV at 10V range (therefore the minimal resolution with those meters is 10uV at 10V) and there are sporadic random jumps 4-5uV large (the popcorn) on top of that walk as well. And your DUT does the same too, basically..

[dietert1]

As long as the ADR1399KHZ isn't available, that minimum setup with the SMD part as "dead bug" is a good solution. It reduces stress on the package and the heating power. With that minimum setup one can make two or more of them. Once you observe more than one, it gets a little more conclusive.
And the differences can be measured to 100 nV with a good 6.5 meter. Except one needs to avoid systematic errors affecting all of them in a similar way.

Regards, Dieter

[iMo]

Btw. - the DCV Ratio (34401a) always returns 7 significant digits after the decimal point over the serial, could be somehow useful when having a 10V ref and the DUT is an LM399 (7V). I have not used it this way yet. Do not forget to wire the LO and the LO together 

[Andreas]

Hmm,

I have here 7.5 life digits with 34401A ratio measurement.
(Measuring the Ratio of a ADR1000A between 10V and 6.6V output)
Ratio is in mV/V VOut/VRef.
The pitfall is only  1 measurement all 12 seconds with 100 NPLC.

21:57:08;  1499.99030;               1
21:57:20;  1499.99020;               2
21:57:32;  1499.99040;               3
21:57:44;  1499.99030;               4
21:57:56;  1499.99010;               5
21:58:08;  1499.99030;               6
21:58:20;  1499.99060;               7
21:58:32;  1499.99060;               8
21:58:44;  1499.99040;               9
21:58:56;  1499.99020;              10
21:59:08;  1499.99040;              11
21:59:20;  1499.99020;              12
21:59:32;  1499.99040;              13
21:59:44;  1499.99020;              14
21:59:56;  1499.99030;              15
22:00:08;  1499.99030;              16
22:00:20;  1499.99040;              17
22:00:32;  1499.99040;              18
22:00:44;  1499.99050;              19
22:00:56;  1499.99030;              20
22:01:08;  1499.99020;              21
22:01:20;  1499.99040;              22
22:01:33;  1499.99050;              23
22:01:45;  1499.99030;              24
22:01:57;  1499.99040;              25
22:02:09;  1499.99060;              26
22:02:21;  1499.99000;              27
22:02:33;  1499.99030;              28
22:02:45;  1499.99020;              29

The Ratio measurement is rather stable (shown by Allan Deviation mV/V of 1 minute averages)

with best regards

Andreas

[iMo]

Yep, as I wrote above there is always 7 sig digits after the decimal point. There is 9.12345670e-01 or 1.12345670e0 I get.
It works such it makes 100NPLC with V1 and then 100NPLC with V2 and calculates the ratio. Thus the internal LM399 is still involved, afaik..

[Kleinstein]

With many modern DMMs the ratio measurement is just measuring the the 2 external voltages relative to the internal reference and than calculating the ratio.
This may not be that bad with a reasonable fast and low noise ADC. Many of the old Keithley 19x meters used a 3 reading cycle with a reading of the reference voltage on there normal AZ mode. So using the ratio mode with a reading of the external reference instead would at least not slow down things much or add extra noise.

Requiring 12 seconds for a 100 PLC ratio measurement on the 34401 looks a bit slow, as 4 conversions (high and low side for the input and reference) should be enough and thus 4 x 2 seconds.
I don't expect the ratio mode to be high priority for the 34401.

Only very few older meter (e.g. Solartron 7071/7081) do an a true direct ratio by using one of the voltages for the ADC reference.

The often much more practical solution is to looks at the difference of 2 external references, if they are comparable in voltage (e.g. less than 20 % difference).

[JodelJonny]

Thanks for the feedback. I tried to implement your suggestions.

-The temp sensor is moved to the resistor network. I'd like to stick to the digital one because it makes it easier to display and save the measurements with minimal external stuff needed. If it interferes with the reference it could only be powered on for brief periods between measurements.

-I switched to the NOMC resistor network with pads for external trim resistors.

-I tried my best at star grounding. Just for the heater it felt wrong to wire it all the way to the star point and back. I don't see any problems wiring it directly to the input. 

-I will try the 7815 on the board. If it causes thermal problems it is easy to short it out and use an external regulator. I want the reverse polarity protection on the pcb in any case because between the power supply and the pcb is the most likely place to mix up the polarity.
Is there a better regulator than the 7815 for this application?

If no one spots any big problems anymore I'm going to order the parts and the pcb and give it a try. This is just supposed to be the first step to get a feel for things.

[Conrad Hoffman]

I'd probably use an LM317. I think it's a better regulator and it gives you the option of trying different voltages. Reg on board worries me, but give it a try.

[Andreas]

-I tried my best at star grounding.
-I will try the 7815 on the board.

You obviously have not understood the principle: remove high courrents (heater, zener supply) from critical wiring (zener voltage sensing).
Doubling the lines (seen on the zener ground) has no effect (the voltage drop from the 3 mA Zener supply will be seen on the output connector).

There is nothing wrong with a 7815/LM317 if the cirquit is mains powered.
With battery supply I usually prefer a LT1763.

with best regards

Andreas

[iMo]

I would also fill the circle on the bottom pcb side beneath the 399 with copper.
Also - what is the pitch of the 399 pins (between the adjacent pins)?

[Kleinstein]

The optional trim resistors R4 and R5 would have no effect as they are shorted. The trim part with 4 resistors are 4 different ends still looks a bit odd. I would more prefer 2 positions to get a trim up or down. Suitable are the resistors that are in series with another and thus have a lower weight to start with and than 2 or 3 optional resistors there.

The star ground idea is still way off. 
The heater, regulator ground and supply to the OP-amp can be treated separate and no real need to use a star ground for these.
One should use the star ground part mainly for the critical parts: the neg output terminal, the zener neg side, the low end of the resistor array and one link to the normal ground.
All 3 points have or can have some current flow. The use of a start ground thus does not fully compensate for voltage drop on the ground lines. This would need exrta current compensation from a negative supply, which may be a bit overkill. Just using relatively short and not too thin traces should be good enough. The main point is to keep the heater current isolated from the critical part.

[JodelJonny]

The doubled up ground trace was left after thinking about using the SMD package with separate sense pins. I forgot to delete it. It does nothing in this case but as long as the enclosed are is near zero it shouldn't hurt either.
 
The pitch between the 399 Pins is 1,796 mm.

I'll think again about the ground configuration but I honestly don't see the problem right now.

R4 and R5 are in series with the circuit so they have jumper links to short them out between the pins of the resistor network that could be easily cut if I'd like to use those resistors instead. I should have made that more obvious in the circuit diagram.

[dietert1]

C9 could be a 1nF film capacitor.
Also i would try to avoid vias in the reference ground connection and at the negative opamp input. I'm not shure how much thermal EMF vias can contribute and whether it is possible to "untwist" the design.
The thermal ring can be ground and under the sensor signal lines there can be ground, too. Same applies to the pad under the resistor array.

Regards, Dieter

[iMo]

An example of a star grounding for a typical LM399 circuit (an example only).. @Kleinstein - plz do advice us directly in the schematics..

[Andreas]

I honestly don't see the problem right now.

Just one example:

Your connection between output pin and LM1399 has about 25 mOhms (assuming 1 mm width and 35um Copper with around 50 mm length).
It carries the reference current of 3 mA.
This gives a increase of 75 uV of the zener voltage (or 10-11 ppm increase).
If all would be constant that would not  be a problem.
Now the T.C. of the copper trace with nearly 0.4%/K increases the zener voltage T.C. by 40 ppb/K.

An example of a star grounding for a typical LM399 circuit

The connection between output and LM399 has to be as short as possible in this case (ok for LM399 its a factor 3 less critical).

with best regards

Andreas

[Kleinstein]

The ground is indeed not that bad as it looks at first sight. Going to both sides makes it a bit hard to follow though.
The neg output terminal is used as the star ground point. This is a coice for a very low output impedance, but external wires woud add anyway.

The link from the neg output to the resistors should be thicker - that resistance adds to the divider and may add a tiny bit to the TC. If really needed one could still add parallel copper to correct.
The link from the neg side of the reference could be made a bit shorter and maybe use 2 or 3 vias. I am not that afaid of thermal EMF of the vias, more the resistance of small vias.
The 2 parallel lines are confusing, but could be used for some TC tweek if really needed, as the restances adds a little to the ref. voltage. So they are not that bad.

There is no need to also have the neg supply of the OP-amp from the star ground, but as the supply current is pretty stable it is also not that bad. Ideally the ground here would not go through the star ground for the critical levels.

One should avoid isolated copper islands and better connect those areas made for thermal reasons to ground or a simlar potential.

Cutting a trace directly at the SO16 chip can be a bit tricky. I would prefer trim by just adding resistance in parallel and not opening a link. There are 2 suitable position for a trim up and down.

[JodelJonny]

Thanks. Good to know, that my thoughts were not totally off after all.

I'll work on improving the "readability" of schematic and layout.

One of the thoughts was to be able to support up to 10 mA output current without major problems. That's way more than needed, I know. Therefore using the output terminal as the star point was chosen.

[Andreas]

One of the thoughts was to be able to support up to 10 mA output current without major problems.

with 10 mA I would think about 4 Wire output.
with best regards
Andreas

[iMo]

One of the thoughts was to be able to support up to 10 mA output current without major problems.

with 10 mA I would think about 4 Wire output.
with best regards
Andreas

Also doublecheck the output buffer's capabilities - whether the buffer is able to provide those 10mA without a significant drop at the +10V output star.. (ie. the 10k base resistor and the 20ohm shunt sizing)..


@Andreas: 4w is needed, sure.. 10mA is a huge current. I had 16uV drop with the same buffer (but 25ohm shunt) with aprox 12mA load directly at the pcb's posts, with a ~15cm (EDIT: actually longer as I have there a ferrite bead with 2 or 3 turns) long 0.8mm dia (!) copper wiring from the pcb posts to the posts on the box I get ~160uV with 10mA. Not counting additional leads.. 

[DavidKo]

I'm not sure how much it is critical. For the start-up there is 200k resistor from the supply in "portable calibrator" on page 6.

[Kleinstein]

With the OPA202 the startup could indeed be an issue, as it has rail to rail outputs. Chances are it would still work even with just 1 diode vor the shift at the output, as the diode forward voltage for a small diode tends to be higher than transistor BE voltage. To be on the safe side, maybe have 2 diodes in series, so that the output would never go below some 0.6 V.

[DavidAzulay5000]

perfect 10
LM399AH
LT3042 OUTPUT 15V
LT1001
aluminum PCB

[branadic]

@ Andreas

Did you observe jumps on all your LCC packaged ADR1399? Is it already confirmed popcorn noise or could it also be what is known for some LTZ1000ACH and its glass bead filled die attach?
Can someone else report similar jumps on his devices?

-branadic-

[Andreas]

Hello Branadic,

as stated here:
https://www.eevblog.com/forum/metrology/adr1399-reference/msg4453792/#msg4453792

I have found popcorn noise up to 1.6 uV on every device during 1/f noise measurements.
Unfortunately this is so low that I cannot confirm this by absolute voltage measurements with my equipment.

The largest popcorn noise (> 1ppm at certain temperatures/times) on the same PCB is one LM399 from LT (LT#02).
This is well above my standard deviation of ~0.25 ppm for 7V measurements.

So now after proper packaging the only remaining issue is the rather large T.C. of the ADR1399LS8
(compared with a typical LM399 near room temperature).

with best regards

Andreas

[JodelJonny]

Hello,

the parts for my reference arrived and I successfully build it without major problems. Just that I stupidly ordered the SMD package ADR1399 :/
I made it fit with some copper wire. Not Ideal but the best I could do for now.
I'll likely build a second one in the future.

It has run for a few days now and seems stable at 10.02271V with very slight variation of the last digit.

I am neither equipped nor able to actually characterise its performance right now. That is the next project to figure out.

No worries, the ADR got some insulation after I took the picture.

Thanks for your help. I am happy with the result and this gives me a good starting point to dive deeper in the topic.

[iMo]

Is the D2 the zener? Btw., I've been always putting a diode in series with the zener (see my schematics around). Even often commented here the diode is redundant, you will see it in action when your output posts get shorted..

[JodelJonny]

D2 is a standard diode. There is no Zener.

[dietert1]

Looks great, already better than the AD evaluation board, that does not have a 10V gain stage.
The metal enclosure is connected to Gnd? One can use a small 50 or 100 Ohm resistor to do that. Or give it a "Guard" terminal.

Regards, Dieter

[Andreas]

Hello,

update of ADR1399LS8 ageing 1kHr after putting into Hammond enclosure:
see also
https://www.eevblog.com/forum/metrology/adr1399-reference/msg4737989/#msg4737989

The ADR1399LS8#01 - #04 are now stable (except for temperature drift which is large compared to the 2 LM399 which are running in parallel).
The LM399LT#01 - 02# in comparison are much more temperature stable.
But LM399LT#02 shows a rather large popcorn noise > 1 ppm

The ADR1399 seem to drift upwards by around 0.5 ppm/khr after 4.3 kHrs in total now.
The LM399 after 2.5 kHrs in total show nearly no ageing drift (except for popcorn).

with best regards

Andreas

[aronake]

Would anyone have kicad (or other circuti design program) or Gerber files to share for a ADR1399KEZ based circuit? Best if it have 10V boost stage, but if not, that would still be helpfull.

[JodelJonny]

I could post the KiCad, Gerber and BOM file for the circuit I build if that fits your need. But they are pretty messy since that was my first design in KiCad.
I would need at least a few days to clean up the most important errors.

[aronake]

I could post the KiCad, Gerber and BOM file for the circuit I build if that fits your need. But they are pretty messy since that was my first design in KiCad.
I would need at least a few days to clean up the most important errors.

That would be awesome! And I am sure many other here would appreciated that too!

[gamalot]

Got an interesting one, maybe it could be a good candidate for a 10V reference based on the ADR1399. 

https://www.vishay.com/docs/60002/osop.pdf

[Andreas]

Hello,

update of ADR1399LS8 ageing 2kHr after putting into Hammond enclosure and with temperature compensation:
see also
https://www.eevblog.com/forum/metrology/adr1399-reference/msg4809413/#msg4809413

The ADR have now 5.3 kHrs in total operating time.
The ADR1399LS8#01 - #03 are drifting slightly upwards ~0.5 ppm / kHr in average over the last 2 kHrs.
The ADR1399LS8#04 has nearly no ageing drift.

The LM399 after 3.5 kHrs in total show still nearly no ageing drift (except for popcorn noise).

The measurement gap around day 70 is due to T.C. measurements of all devices.
The repeated T.C. measurement (in ppm/K) shows nearly no change against previous measurement.

with best regards

Andreas

[Andreas]

Hello,

in the mean time I have built a 2nd PCB in a Hammond housing for ageing of the ADR1399 references.
And while waiting for the metal can packages which are delayed from month to month
I could not resist to do some T.C. testing with the 4 "fresh" LS8 devices on the pcb.

The PCB has been operated in total 8 hours within 4 weeks before this tests.
Since there was a relatively large initial ageing drift I had to wait some temperature cycles
until the T.C. could be measured with a reasonable ageing drift.
So total measuring time was 100 hrs.

here the results of the 4 devices on PCB #2
Again the T.C. of the LS8-packages is relatively high compared to typical LM399 devices.

with best regards

Andreas

[Conrad Hoffman]

I guess I need to ask, having bought one of the SMT versions a while back, is this reference really a good choice if you have room for a regular 399 or other device? I'm not seeing it as a "serious" reference to use as a lab standard or in a high precision instrument.

[Andreas]

Hello,

as always it depends on your needs:
- low noise (better than LM399)
- low T.C.  (my LM399 samples are better)
- low ageing drift (I am still evaluating but the ADR1399 seem to be better than the ADR1000A)

For me with 18-33 deg C in my "lab" the T.C. is a enemy.

with best regards

Andreas

[dietert1]

After buying the SMD ADR1399 evaluation i ended up building a 10 V gain stage - also as bootstrap for the reference. This includes a low pass filter and outer oven to keep everything at 38 °C. It performs very well. I have seen 0.2 ppm RMS over one month.
Our LTFLU references are better, as they are 4x zener arrays. I am wondering what happened to the 4x ADR1399 array designs shown before in this thread.

Regards, DIeter

Edit: After fixing a workstation i looked at the logs once more and the difference voltage between the references ADR1399 and LTFLU1 was 0.059 ppm RMS of hourly averages between 2023-04-02 and 2023-05-08.

[Andreas]

Hello,

update of ADR1399LS8 ageing 3kHr after putting into Hammond enclosure and with temperature compensation:
see also

https://www.eevblog.com/forum/metrology/adr1399-reference/msg4875467/#msg4875467

The ADR have now 6.3 kHrs in total operating time. Tempco is calculated out for the ageing curves.
The ADR1399LS8#01 - #03 are drifting slightly upwards ~2 ppm / 3 kHr in average over the last 3 kHrs.
The ADR1399LS8#04 has nearly no ageing drift.

with best regards

Andreas

[branadic]

Unfortunately, ADR1399KHZ delivery date is shifting each month by another month. Ordered in Feb 2023 this is the history of updated delivery dates so far:

MAR 28, 2023
MAI 02, 2023
MAI 09, 2023
MAI 16, 2023
JUN 20, 2023
JUL 14, 2023
JUN 26, 2023
JUL 03, 2023
JUL 07, 2023
JUL 14, 2023
JUL 21, 2023
AUG 01, 2023

Seems like ADI has some serious problems in getting them fabricated.
Luckily, some forum member was kind enough to send me a few samples and they will arrive within the next days, so I can proceed with my experiments soon.

-branadic-

[DavidKo]

During my order at 26th July 2022, the delivery date was unknown. At 28th November, they have sent update that ADR1399KHZ will be available 15th August 2023 (and ADR1399KEZ on stock). Ordered through Mouser.

[Andreas]

you are not alone
my order from Jan 22 is still waiting on delivery ...
with best regards
Andreas

[branadic]

Attached are multiple noise measurement of my ADR1399-PWM-DAC. Please replace µV by nV and you end up with about 2 µVpp.

-branadic-

[Kleinstein]

The noise looks good, with quite some of the peaks that are realatively short, possibly on a time scale where a little more filtering can help.
The real question is how good the long term drift will be for the TO46 version.
Let's hope it is driften less than the expected date for availabilty.

[deepfryed]

I have one ADR1399H-EBZ on order from Mouser, they say ETA end of September. Let's see.

[maxwell3e10]

It is in stock on Digikey in TO46 can right now

https://www.digikey.com/en/products/detail/analog-devices-inc/ADR1399KHZ/15848880

[gamalot]

Four ADR1399KHZ from Digi-Key have just been delivered. My money is running out and I can't afford more. 

I don't know what printing technology in 2023 can make the silk screen on top look so blurry and dirty!

[iMo]

Does not look like ADI's quality.. A fake?

[gamalot]

Does not look like ADI's quality.. A fake?

No, that just proves that they must be genuine, perhaps using eco-friendly printing techniques. 

[branadic]

Looks really weird, as it is ADR1399 with LT logo instead of ADI logo, beside the poor printing quality.

-branadic-

[gamalot]

Even the thermal shield itself is not the same as before, what happened? 

[iMo]

..ask your supplier..

[NNNI]

@branadic
Have you already posted about the ADR1399 PWM DAC somewhere? I would love to learn more about it. I've recently been playing around simulating something similar using a passive 3 pole RC filter (3 stages of 10K/100nF C0G) which does indeed reduce ripple, but assuming I want 1uV settability, I would need a 23-bit resolution PWM signal. While that is possible, it pushes the PWM frequency very low, and I'm not sure how to make a practical implementation. Perhaps reduce settable resolution to 10uV?
@gamalot: Someone I know also just got an ADR1399, the printing was similarly crusty but the package looked to be in better shape.

[Kleinstein]

From the picture the "printing" look like laser writen, with small dot matirx depressions.  Getting the LT logo is indeed odd.

[Andreas]

While that is possible, it pushes the PWM frequency very low, and I'm not sure how to make a practical implementation. Perhaps reduce settable resolution to 10uV?

Hello,

it is all in the LM399 thread.
We are using Sigma Delta modulation for the PWM.
So it is possible to do several kHz @8 Bit PWM with 24 bit resolution.

https://www.eevblog.com/forum/metrology/lm399-based-10-v-reference/msg2189460/#msg2189460
https://www.eevblog.com/forum/metrology/lm399-based-10-v-reference/msg1955149/#msg1955149
https://www.eevblog.com/forum/metrology/lm399-based-10-v-reference/msg2200098/#msg2200098
https://www.eevblog.com/forum/metrology/lm399-based-10-v-reference/msg2082544/#msg2082544

with best regards

Andreas

[iMo]

@Andreas - perhaps you may open a new thread, like "x399 based PWM voltage standard" or something like that, and start with an update on the effort you, branadic and others did (like the latest schematics, firmware, etc). Otherwise it remains hidden in the rather long 399 thread..

[Kleinstein]

The PWM DAC part is not directly related to the reference, only indirectly through the voltage and noise level / stability expected.
AFAIR the LM399 10 V ref. part is not a classic PWM DAC, but in the 7 to 10 V amplfier feedback. This adds some limitations but also gets up from 7 V to 10 V without critical resistors setting the gain.
A more universal voltage source would more use 2x7 V to start with 14 V and than use a more conventional PWM DAC to get the 14 to 10 V (or another voltage) step.
Filtering would likely use film capacitors of some 1-10 µF, maybe even PP types to get fast settling / low DA.

[dietert1]

That's what i started building long time ago. There is an array of 10 LM399 to generate 14.1 V and to control the mains power supply that is built as a bootstrap amplifier to 19 V. There is an aluminum enclosure.
The 19 V will be used to implement the main PWM switch with a dual n-channel FET. I got the control board already with a 16 MHz MSP430 with isolated RS232, the PWM filter and the ripple compensation switch. So little missing. Just need to finish the main PWM switch and it will be a voltage calibrator.
The LM399 array has been powered for one or two years now and was already used for many different precision measurements.

Regards, Dieter

[gamalot]

From the picture the "printing" look like laser writen, with small dot matirx depressions.  Getting the LT logo is indeed odd.

Yes you can even feel the rough surface with your fingertip. At least on voltage reference components, I would prefer LT's logo to be there instead of ADI's.

[JodelJonny]

I just recieved my ADR1399 from DigiKey as well and it looks the same.
Unfortunately they were out of stock of the NOMC 8x10k Resistor arrays. If I can find one of those and a little bit of time I am going to build a second ADR1399 10V Ref.
Number 1 has been running for approx. 2800h.

I am very sorry that I havent uploaded any PCB/Gerber files yet. As this was my first real KiCAD project they are a mess and need signifacant work to be really usable.

[KT88]

There is a PCN related to the top marking:
https://www.analog.com/media/en/PCN/ADI_PCN_22_0025_Rev_-_Form.pdf

[Grandchuck]

How about counterfeiting?  Laser marking might discourage it.

[gamalot]

There is a PCN related to the top marking:
https://www.analog.com/media/en/PCN/ADI_PCN_22_0025_Rev_-_Form.pdf

It seems that there is no mention of the change of the logo.

[NNNI]

In case someone is interested, a slightly more high resolution picture of the laser marking. The carbonization, I assume, is because of the laser process itself, but is it a useful side effect or intentional? One can only guess.

[Echo88]

20€ on Mouser and the marking looks like piss in snow.

[iMo]

ADI hired a special tattoo saloon in Parañaque City..
PS: it could be it is a novel anti-counterfeiting technology..
Or, simply they reuse their existing laser engraver - all other epoxy and metal chip's markings are laser engraved.

[Kleinstein]

The relatively exotic plastics for the case could be tricky with conventional printing. With a relatively low volume product the laser labeling makes absolute sense, as there should be little fixed costs.

[iMo]

But why LT logo? Perhaps some legal stuff? Like patents, trade marks, IP, etc.? The Design remained mostly the same, afaik.

[DavidKo]

Mouser is shipping mine today. They still have 783 for dispatch today ;-)

[maat]

*knock* *knock* Oh, look what arrived today.

On the left and and right are samples I got earlier from ADI. These have a higher quality look than the production units (and also carry the ADI logo). It seems ADI tries to undercut the Ebay scammers selling refurbished diodes in terms of look and feel 

Off they go into the test setup. Lets see how they perform for the next 1 kh.

[analityk]

I've bought ADR1399KHZ and it arrived in weird ceramics package with standard that legs.
It is a bit curious because in datasheet was specified only ceramics smd package and metal can hermetically sealed to-42, but mine one is hybrid? I do not know, maybe someone here have any experience.
This ceramics shields act as heat insulation and inside is metal can?

[iMo]

Yep, that is the plastic heat insulating cap, inside is the metal package.

[Andreas]

Hello after removing the cap it should look like this:

and you get a thermal anemometer instead of a reference.
see patend of Joe Geller.

https://www.freepatentsonline.com/7380471.html
https://www.nutsvolts.com/?/magazine/article/micro_gust_thermal_anemometer_mgta

with best regards

Andreas

[iMo]

and you get a thermal anemometer instead of a reference.
see patend of Joe Geller.

Frankly, I doubt he could get the patent for that - as it is a known general principle, nothing new..
You may apply for such a patent with any similar heated device, like my old OCXO - the crystal frequency and heater current behave the same way.. 
PS: ..but perhaps no Einsteins sitting in the Patent Office anymore..
PPS: on the other hand we were laughing years back on a patent where the guy claimed uploading the binary into the mcu's flash via a bootloader, fortunately he indicated a single TI's mcu partnumber only.. 

[gamalot]

Hello after removing the cap it should look like this:

and you get a thermal anemometer instead of a reference.
see patend of Joe Geller.

https://www.freepatentsonline.com/7380471.html
https://www.nutsvolts.com/?/magazine/article/micro_gust_thermal_anemometer_mgta

with best regards

Andreas

Looks like they didn't change the structure of the thermal shield.

[BU508A]

Does not look like ADI's quality.. A fake?

On Feb. 2022 I've ordered 18 ADR1399KHZ directly from Analog Devices.
They arrived on 1st of August 2023.
https://www.eevblog.com/forum/chat/what-did-you-buy-today-post-your-latest-purchase!/msg4991827/#msg4991827



I have no idea, why they are using the LT logo and why the printing or engraving quality is so poor. I was surprised by myself when I saw it at first.

[iMo]

Details from my uscope..
It looks like laser engraving, but the intensity of the individual pulses differs.
Perhaps a code with a secret message 
PS: The intensity of the pulses could be constant, but the reflectivity of the surface might differ from place to place, therefore the mess.

[Andreas]

Hello,

update of ADR1399LS8 ageing 4kHr after putting into Hammond enclosure and with temperature compensation:
see also

https://www.eevblog.com/forum/metrology/adr1399-reference/msg4939813/#msg4939813


The ADR have now 7.3 kHrs in total operating time. Tempco is calculated out for the ageing curves.
The ADR1399LS8#01 - #03 are drifting slightly upwards ~2.5 ppm / 4 kHr in average over the last 4 kHrs.
The ADR1399LS8#04 has nearly no (<1 ppm) ageing drift.

with best regards

Andreas

[JodelJonny]

Hello,

this Saturday I used the offer of the PTB at the Maker Faire Hannover to get inofficial calibrations. I got some quick and dirty measurements of my 10V ADR1399 ref with a calibrated 3458A.  The measurements start quickly after I got the reference, that had already been powered on for a few hours with a battery, out of my backpack.

I just took pictures with my phone and used the timestamps to get a graph out of it. This is why there is inconsistent timing between measurements.
This was after approx. 3000h burn in time. 

I guess that the reference got relatively hot in the backpack and is cooling of during the measurement so that a negative tempco causes the rise in voltage.

 Thanks to the PTB team! I was nice to meet NNNI there as well.

[Andreas]

Hello,

my 2nd PCB is now populated additionally with 4 metal can devices. ADR1399KHZ.
Ageing drift for the first 2 weeks (333 hrs) starting after 1-2 hrs total operation time.
Unfortunately I have high changes in temperature in my lab between morning and evening.
So two measurements per day are not optimal.

Next I will check tempco of the 4 samples.

with best regards

Andreas

[Andreas]

Hello,

Tempco measurement on the 4 Metal can devices (KHZ) of 2nd PCB after 333 hrs run in phase.
I had hoped that the T.C. is lower than that of the LS8 packages.

But in comparison this is not the case. T.C. is very individual.

See also the (non pre-aged so with large drift) measurement of the LS8 packages..
https://www.eevblog.com/forum/metrology/adr1399-reference/msg4879958/#msg4879958

With best regards

Andreas

[iMo]

Interestingly the "red" noise decreases with the sample N..

[Andreas]

Not shure what you see.

But it could be that it is due to large changeing ramp speed (actual temperature differing from setpoint due to environment) during test.

with best regards

Andreas

[iMo]

Ok, your ppm scale on the right side differs..

[miro123]

Hallo,
I want to build few boards with ADR1399 and nomc/ot lt5400 resistor networks - are there any kicad/ altium reference design from other eebvlog members.
i need the design files since i want to do some modifications.  I have non-standard requirements
- ADR1399 & LM399 TO-46
- High stability 7V, 5V and 2v5,
- Something in 10....12V is nice to have.
- Small size is nice to have feature.
BR
Miro

[dietert1]

Not shure what you see.

But it could be that it is due to large changeing ramp speed (actual temperature differing from setpoint due to environment) during test.

with best regards

Andreas

One should be aware that a TC of 3 uV/°C is always in the range of thermal EMF. Those references run hot.
Maybe the residual TCs exhibit a pattern. If the numbering of references reflects their geometrical position on the test board, the apparent TC of the references follows a linear dependency - with one exception: ADR1399#2. Maybe it had some extra solder blob..

Regards, Dieter

[Andreas]

Hello,

news from the 2nd ADR1399 PCB:

See also
https://www.eevblog.com/forum/metrology/adr1399-reference/msg5024692/#msg5024692

now I am measuring only 1 time per day (in the morning) so the temperature differences between measurements are smaller.

the 4 metal can devices have now 1kHrs. The values are now temperature compensated after having determined the T.C.

the 4 LS8 (SMD) devices have 100 hrs for initial measurements including T.C. measurement (not shown on the diagram) and after that 1 kHr temperature compensated ageing drift.

with best regards

Andreas

[Andreas]

Hello,

update of ADR1399LS8 on PCB#1 ageing 5kHr after putting into Hammond enclosure and with temperature compensation:
see also
https://www.eevblog.com/forum/metrology/adr1399-reference/msg5011936/#msg5011936



The ADR have now 8.3 kHrs in total operating time. Tempco is calculated out for the ageing curves.
The ADR1399LS8#01 - #03 are drifting slightly upwards ~2-3.5 ppm / 5 kHr in average over the last 5 kHrs.
(now drifting in different directions).

The ADR1399LS8#04 has nearly no (~1.2 ppm) ageing drift.

with best regards

Andreas

[Andreas]

Hello,

update from the 2nd ADR1399 PCB now at 2 kHrs for the metal can devices (KHZ) and 2.1 kHrs for the LS8 devices (first 100 hours not shown):

See also
https://www.eevblog.com/forum/metrology/adr1399-reference/msg5068999/#msg5068999

The values are all temperature compensated (temperature sensor at the bottom of the hammond die cast housing)

with best regards

Andreas

[branadic]

That is some interesting result Andreas. While the ADR1399KHZ seems to drift upwards (similar to what I observe for the Valhalla 2720GS/HSR-ADR and my ADR1399-PWM-DAC) all LS8 devices seem to follow the picture we can see in the datasheet.
This seems to indicate, that Figure 27 was achieved using LS8 devices, not TO-46.

-branadic-

[Andreas]

Hello branadic,

I have 2 PCBs with 4 LS8 devices each.
On PCB#1 all ageing drift is rising (datecode #2043)
here the drift is rising.
https://www.eevblog.com/forum/metrology/adr1399-reference/msg5077309/#msg5077309
https://www.eevblog.com/forum/metrology/adr1399-reference/msg4308397/#msg4308397


On PCB#2 (see post above) I bought the devices much later
There I have falling drift for the LS8 devices. (unfortunately I did not write up the date code).

The metal can devices have date code #2316

I fear the direction of the drift is more related to the date code.

with best regards

Andreas

[Andreas]

Hello,

update of ADR1399LS8 on PCB#1 ageing 6kHr after putting into Hammond die cast enclosure and with temperature compensation (sensor on bottom of the hammond case)
see also
https://www.eevblog.com/forum/metrology/adr1399-reference/msg5077309/#msg5077309


The ADR1399LS8 have now 9.3 kHrs in total operating time. Tempco is calculated out for the ageing curves.
These LS8-devices are drifting upwards.

The voltage shift at the 2 LM399 (only for comparison) occurred after repair on the output connector. (while operating the cirquit).
Most probably I have damaged the Multiplexer for the "upper half" of the PCB during soldering.
(will have to check that when I have some time).

with best regards

Andreas

[iMo]

Is there any recommendation what type of the 1uF (ser 5R1) output capacitor should be used?
Foil or MLCC?

[Kleinstein]

The capacitor should not be hat critical. Even leakage is not a problem. When using MLCC I would plan for 2 x 1 µF to conpensate for the drop in capacitance under voltage and the 1 µF value is already on the low side to start with.

[Andreas]

Hello,

I have no issues with 1uF/50V X7R in 0805

with best regards

Andreas

[iMo]

@Andreas: inspired by your low noise 399 in your 34401A I put the HP399 off my DMM and been just doing some measurements..  Thinking to replace it with something better, currently running the DMM with a pops free LM399H, but thinking about one of my 1399 (<2uVpp).. Not sure I've been doing the right thing with my DMM, anyhow.. 

[Kleinstein]

If one has a rather rare LM399 without the typical popcorn noise, this would be a relatively low noise ref. Not sure that the ADR1399 is better than this. Some the LM399 still have rather rare jumps from the popcorn noise (I have one that is at about 1 jump per hour) - this can be great for some short time tests, but bad for other (no averaging for many tests, just shifting the noise to even lower frequencies).

The HP34401 also has a relatively noisy ADC. Though not directly comparable, the ADC noise can be limiting on the shorter time scale. To a large part there is not much one can do about it, as it's the somethwat limited resolution / quantization noise. Significant lower noise would need more like a redesign of the ADC.
With a well aged reference there is also low dirft - a new ADR1399 would need quite some time to get the drift rate of a LM399 with some 10000 hous of use.

[iMo]

I know, but - see below (the same setup), I will run the pops free LM399H#2 (1500h BI) over night.. Both 1399 (0h BI) I have are half of the LM399H#2 noise..

[Kleinstein]

Looking at the standard deviation of a 7 V ref reading is bit unusual. It still reflects the ref. noise, but only for a limited times scale of some 4 - 400 seconds. So more normal way would be to look at the alan deviation.
Anyway the difference is not that large. It is mainly having the jumps in the external LM399 ref and less or smaller jumps for the ADR. One would always see a combination of both references: The good lm399 in the 34401 and than either a more normal LM399 or the slightly lower noise ADR1399.  The fact that there is quite some difference shows that the reference in the meter is way better than normal. With a more normal ref. in the meter one would not see much difference (70% noise even with a noise free external ref., though the way the test is done may be more sensitive to peak then RMS noise).

Even for the relatively long time scale use in the test, the ADC noise is still a good part of the noise.

[iMo]

That is 10V external Vref. Both HP399 and LM399 are the 34401A's refs (inside the meter). The peaks in the STD show where a reference shifted its voltage abruptly (or it is EMI). Those pictures are not about any stability over time..

I have the data thus tomorrow, after the Starship's orbital flight, I will show you ADEV/MADEV as well

But sure, I do understand the replacement of the old HP399 with the fresh ADR1399 is tricky.. 

PS: CONT in the ADR1001 thread..

[diodak]

The capacitor should not be hat critical. Even leakage is not a problem. When using MLCC I would plan for 2 x 1 µF to conpensate for the drop in capacitance under voltage and the 1 µF value is already on the low side to start with.

What about the capacitor itself in parallel (instead of the RC circuit) according to the datasheet:
"Another single element passive that works directly with the ADR1399 is a 10 μF tantalum capacitor, even though the series resistance can
measure less than 5 Ω on an impedance analyzer."
Has anyone checked this?

[dietert1]

I did not try that as i think the better solution is a low leakage cap with a known resistor instead of unknown capacitor ESR. Trying to save a 1 cent part on a 15 $ reference?

Here i have a three month log of the ADR1399 SMD evaluation kit i got one year ago. It got an outer oven at 38.2 °C. These are 4 second measurements with residual TC corrected. The log shows a drift of about 1.4 ppm and the internal temperature of the P6048 in red. One can assume the drift of about 1.3 ppm is caused by the Vishay NOMC divider used for the 10 V gain stage. Maybe a confirmation of results Andreas showed above.

Regards, Dieter

[ivo]

Is there any recommendation what type of the 1uF (ser 5R1) output capacitor should be used?
Foil or MLCC?

It's a 1uF good MLCC + 5.1R resistor, AFAIK essentially to emulate a tantalum cap, same with older regulators wanting tants/ECs to settle on their output pin and you can generally pull the same trick with them

In the LM399 datasheets they advise you can do a similar thing with the heater: you can add 10~500R above it to limit the current draw on initial heatup, but to do so you must add a 2.2uF *tantulum* cap across it.

Trying to save a 1 cent part on a 15 $ reference?

My guess is because people don't want to have to deal with production scale soldering polypropylene (can't stand up to high soak temps) if they don't have to, and prefer to save cost on a tant. And yes two 1 cent parts vs a 10-50c part is a big difference!

[Andreas]

Hello,

update from ADR1399 PCB#2 now at 3 kHrs for the metal can devices (KHZ) and 3.1 kHrs for the LS8 devices (first 100 hours not shown):

See also
https://www.eevblog.com/forum/metrology/adr1399-reference/msg5145858/#msg5145858

The values are all temperature compensated (temperature sensor at the bottom of the hammond die cast housing)
Now the LS8 devices, which started with negative slope for ageing, trend to a slow upwards ageing slope like the TO-46 metal can devices.

with best regards

Andreas

[Andreas]

Hello,

update of ADR1399LS8 on PCB#1 ageing 7kHr after putting into Hammond die cast enclosure and with temperature compensation (sensor on bottom of the hammond case)
see also
https://www.eevblog.com/forum/metrology/adr1399-reference/msg5152335/#msg5152335


The ADR1399LS8 have now 10.3 kHrs in total operating time. Tempco is calculated out for the ageing curves.
These LS8-devices are drifting upwards.

The voltage shift at the 2 LM399 (only for comparison) occurred after repair on the output connector. (while operating the cirquit).
The investigation of the PCB did not give a obvious damage.
So I decided to replace the multiplexer for the LM399. This was obviously the reason for the defect as the readings reverted to the old level.
For the future I will drop the defective readings.

with best regards

Andreas

[Andreas]

Hello,

update of ADR1399LS8 on PCB#1 ageing 8kHr after putting into Hammond die cast enclosure and with temperature compensation (sensor on bottom of the hammond case)
see also
https://www.eevblog.com/forum/metrology/adr1399-reference/msg5226999/#msg5226999


The ADR1399LS8 have now 11.3 kHrs in total operating time. Tempco is calculated out for the ageing curves.
These LS8-devices are drifting upwards.

I had again a 5-6 ppm voltage shift on the 2 LM399 (only for comparison) on the 2nd Multiplexer (MAX4052A).
This time I had no obvious reason for the shift. So it was very unreasonable that the Multiplexer should have died again.
I reflected what I did last time:
- Switching power off and on (on another desk) to do some measurements.
- Measurements did not show any obvious error.
- Exchange the MAX4052A. (with power off).
- reassembling and putting all back to the measurement desk.

So I decided to do power cycling only and actually this "cured" the 5-6 ppm voltage shift.
Obviously there is some "latchup effect" which affects the upper channels.

But what can the reason for this latchup effect be?
I had introduced some level shifters to reduce the self heating of the multiplexers when having 5V logic levels.
But the level shifters are relatively high ohmic (BSS138 with 330K pull-up resistors on the 15V side) to reduce self heating.
Simulation with a similar FET (BSS123) showed around 20 us rise time with 20 pF for wiring and input capacitances of the multiplexers.
Should this be the reason for the latchup-effects?

with best regards

Andreas

[Andreas]

Hello,

update from ADR1399 PCB#2 now at 4 kHrs for the metal can devices (KHZ) and 4.1 kHrs for the LS8 devices (first 100 hours not shown):

See also
https://www.eevblog.com/forum/metrology/adr1399-reference/msg5226966/#msg5226966

The values are all temperature compensated (temperature sensor at the bottom of the hammond die cast housing)
The LS8 devices continue to trend to a slow upwards ageing slope like the TO-46 metal can devices.

If I look closer there are also 0.5-1 ppm jumps on the measurements.
One jump in down direction is on the same day as the power cycling together with PCB#1 (same transformer).

so I started to investigate (see zoomed diagram).
And actually: every time when I did a power cycling (half a minute followed by >15 minutes warm up time) the measurements started at lower levels.
After one measurement cycle mostly the higher level was again active.
So also here I have a less pronounced "latchup effect" of 0.5-1 ppm.

So the question arises: Are there any multiplexers like MAX4052A which can be used as replacement (hoping that there will be no latch up effect)
- low leakage current
- operating from 12-15V (single supply)
- in SO-16 package with same pinout as 4052 devices
- at best with integrated low power level shifters from 5V to 15 V

with best regards

Andreas



with best regards

Andreas

[Kleinstein]

A possible choice for the MUX chips could be the relatively new TMUX4052. The specs are rather loose because it is a cheap part (so not much testing done), but the typical performance is still expected to be relative good. They accept a low control level (e.g. 2 V) at least.  The DG4052 should be similar with tighter specs.

[Andreas]

A possible choice for the MUX chips could be the relatively new TMUX4052.

Hello,

Electrically it would fit (I think). But I did not find a SO16 package. Or did I overlook something?
(I do not want to resolder the references, only the multiplexers).

The DG4052 might be a replacement.

with best regards

Andreas

[Mickle T.]

A simple, general-purpose, dual-voltage (10V and 1V) reference based on the ADR1399 in a TO-46 package. Temperature coefficient <0.2 ppm/°C (10 V), noise amplitude 0.25 ppm p-p (0.1-10 Hz). Resistors are selected with TCR 0.2...2 ppm/°C.

https://youtu.be/guZnwp_DdUM

[ivo]

I swear most people's definitions of "simple" differ wildly from mine on this board 😅

[dietert1]

Yes, it seems to be more like a calibrator for voltage, current and resistance. Very nice.

Regards, Dieter

[Andreas]

Hello,

update from ADR1399 PCB#2 now at 5.4 kHrs for the metal can devices (KHZ) and 5.5 kHrs for the LS8 devices (first 100 hours not shown):

See also
https://www.eevblog.com/forum/metrology/adr1399-reference/msg5303665/#msg5303665

I have exchanged the MAX4052A multiplexers against standard HEF4052 (NXP) CMOS devices.
As result ther is a relative large jump on all references.
But also the "ageing" slope has changed significantly on all channels (mostly near zero).

So I fear that up to now I have measured the "ageing" of the MAX4052A instead of the references.
That also would explain the difficulties with my "position dependant drift".
I think that the MAX4052A are not usable for higher supply voltages than ~10-12V since I never recognized something unusual with 5V.

Should have used a relay multiplexer instead?
In the mean time I also have some DG4052 to give them a try.

with best regards

Andreas

[CalibrationGuy]

A simple, general-purpose, dual-voltage (10V and 1V) reference based on the ADR1399 in a TO-46 package. Temperature coefficient <0.2 ppm/°C (10 V), noise amplitude 0.25 ppm p-p (0.1-10 Hz). Resistors are selected with TCR 0.2...2 ppm/°C.

https://youtu.be/guZnwp_DdUM

This is a beautifully executed project. How long did this simple project take? How much did it cost you when all is said and done? Anyway, thanks for sharing.

PS - It costs my lab over $2700 per 3458A to backup our numbers. Voltage tests here are a bit subjective because even a Keysight ISO17025 calibration has an uncertainty of 1.1ppm.

[ivo]

I wonder if anyone with capable equipment, would be interested in testing if there's any appreciable difference in noise, and tempco ppm sweeps, over different reference currents? Say 3mA/6mA/10mA?

[CalibrationGuy]

I wonder if anyone with capable equipment, would be interested in testing if there's any appreciable difference in noise, and tempco ppm sweeps, over different reference currents? Say 3mA/6mA/10mA?

If such a request were made, you would have to clearly define what tests you would want including tolerance, number of samples, test conditions, etc. People who have labs are usually busy people. *cough* *cough*

TomG.

[iMo]

I wonder if anyone with capable equipment, would be interested in testing if there's any appreciable difference in noise, and tempco ppm sweeps, over different reference currents? Say 3mA/6mA/10mA?

The "zener current" inside the 1399 is set to a specific "constant value" by ADI, the same as with 399 (the chip is almost the same). You cannot change that current from outside.

[CalibrationGuy]

I thought they were referring to changing the heater current maybe?

TomG.

[iMo]

In 399 and 1399 the heater current is internally set such the temperature of the chip is kept "constant" at 95degC, afaik.
Thus you cannot mess with zener nor heater current with 399 and 1399.

For such experiments with zener/heater currents you want LTZ1000/ADR1000/ADR1001 for example..

I thought they were referring to changing the heater current maybe?

TomG.

[CalibrationGuy]

In 399 and 1399 the heater current is internally set such the temperature of the chip is kept "constant" at 95degC, afaik.
Thus you cannot mess with zener nor heater current with 399 and 1399.

For such experiments with zener/heater currents you want LTZ1000/ADR1000/ADR1001 for example..

I thought they were referring to changing the heater current maybe?

TomG.

I was just perusing the LTZ threads, thanks for pointing that out.

TomG.

[Andreas]

Hello,

on the other side the idea is not uninteresting.

Depending on location of heater, zener and temperature sensor on the chip
the temperature distribution could be affected when the cirquit around the zener gets more current
(and the heater less to keep the temperature).

with best regards

Andreas

[iMo]

Hello,

on the other side the idea is not uninteresting.

Depending on location of heater, zener and temperature sensor on the chip
the temperature distribution could be affected when the cirquit around the zener gets more current
(and the heater less to keep the temperature).

with best regards

Andreas

With a higher "excess" current you start to influence the heater's equilibrium, sure.
1mA more "zener excess current" will create 7mW of heat somewhere on the chip, with Rthja 220C/W (DS value for the TO-46 version) it increases the chip temperature by 1.54C, that will force the oven to decrease the heater current to go back to 95C. The excess current through the "zener" will increase the drop at the zener's bond wire(s) as well.
So you may try to increase the "zener current" till the oven gets off the regulation (my guess would be aprox 10-20mA) and then the "excess zener current" starts to dominate..
Another version is to switch off the heater, and heat up the zener with the "excess zener current" only. You would need aprox 25mA of current for say 70C (23C Tamb), imho.
The question is how the "zener voltage" with the excess current dependence is in reality.
I would not do it, however.. 

[dietert1]

In their evaluation board they use about 3 mA. If you use a 0.1 % resistor to generate that current, you expect at most 3 uA variation. Times the 0.04 Ohm dynamic impedance spec gives 120 nV. Good enough!
In the datasheet they explicitly show that the same dynamic impedance applies for a 3 to 13 mA change (delta I = 10 mA gives 0.4 mV).

Regards, Dieter

[Andreas]

Hello,

before changeing the mulltiplexer to DG4052 on PCB#2 I wanted to see wether the multiplexers have any influence on tempco of the references.
So I started with T.C. measurements.

Unfortunately I recognized that I did not repeat the T.C. measurements of the LS8- parts after I had populated the metal can parts.
Due to the additional heaters on the PCB the measured temperature range on the bottom of the die cast housing has changed even with same environmental temperature profile.

On the LS8 parts the temperature range was 12-45 deg C without population of the metal can parts.
Now it is 18-46 deg C with the metal can parts populated.
And it looks like the effective T.C. relative to the temperature sensor has increased.
So most probably my temperature correction for the LS8 parts was wrong after population of the metal can parts.

Further I recognized that there is still some latch up effect (2 times) after having changed from MAX4052A multiplexers to HEF4052.
I looked at the power supply values at the beginning of the latchup.
And surprise: on entry of the latchup the 15V supply of the LDO (LM2940) increased by 50-60 mV. Whereas the raw value 17V was about constant.
At the same time the voltage appeared more noisy. (Oscillations ?)

I double checked the data sheet and found that I was missing the 22uF capacitor at the output of the LM2940.
There are only 12 pcs 100nF capacitors on the 15 V supply (one for each ADR1399 heater) 2 for the Multiplexers and 2 directly at Output of LM2940 / input of LM2950A.
I tried to measure the oscillations with the scope. But obviously this changed something at the cirquit so no oscillations visible.

I removed the Jumper for the heaters since the "latchup" occured more at higher temperatures = lower heater power.
So in this condition (lower 15V load and 8 capacitors removed) I saw the oscillations 2Vpp with 45 kHz.
A 100uF capacitor which I had on hand fixed the oscillations.
So now I can repeat the T.C. measurements hopefully without any further "latchup" effect.

But I fear the ageing measurements up to now where useless:
Wrong T.C. and one never knows how the unstable 15V supply influenced the measurements.

I am feeling like a bloody beginner.
On the other side: the oscillations are only visible under rare conditions otherwise I would have seen them on cirquit start up.

with best regards

Andreas

[Alex Nikitin]

Guys, just for clarification - the ADR1399 can not work properly below about 2.5mA reference current, with 3mA is not just a datasheet figure but the correct minimal current. For that reason if you replace LM399 with ADR1399 you must take care of the current, in many old units LM399 is run at just above 1mA reference current and without modifications to the circuit the ADR1399 as a replacement will work poorly.

Cheers

Alex

P.S. - did you really think you can get lower noise for free, without an increase in current? 

P.P.S. - I've scanned through the circuits uploaded in this thread, it looks like I am overreacting  .

[Andreas]

Hello,

if it calms down your nerves: I am using 2K4 to 15V in my ageing box.

with best regards

Andreas

[Alex Nikitin]

Hello,

if it calms down your nerves: I am using 2K4 to 15V in my ageing box.

with best regards

Andreas

It certainly does  ! I would be heartbroken if all your work on this reference could be wasted!

Cheers

Alex

[Alex Nikitin]

Now I can explain a bit, my earlier strong reaction. A few days ago I bought HP3245A unit which was modified by replacing the original LM399 with ADR1399 (on top of other service work, a new mains filter, new Dallas chips etc). Unfortunately, the minimum current issue was missed and the unit (recently calibrated using HP3458A) was not very stable. I've added a 2K7 50ppm/C resistor in parallel to R401 and had to recalibrate manually, as the reference output (and hence the unit output) increased by 2.35mV or over 300ppm! Here is the stability test results overnight in my lab with the ambient temperature variation during nighttime about 4 degrees C, before the mod and after the mod and calibration. The voltage was measured by one of my Keithley 2015 units (NPLC10 +10 averages, a rough equivalent of NPLC100), so essentially against a very well aged LM399.

Cheers

Alex

[iMo]

I reported that 1399 instability in my "noise indicator" thread - the 1399 really needs at least those 3mA and the RC "snubber" (!!!). So for standard 7->10v you need 1k cathode resistor or even smaller..

[IanJ]

My 3245A with a newly fitted ADR1399 (24hrs ago).
Recently re-capped.
3245A Vref circuit op-amps changed out for Chopper (yes I know....but I am experimenting!).
Red trace is workshop temperature (regulated).
PS. My workshop is quite noisy, LED lights, PC running, Server cabinet.......all within fairly close proximity.

No snubber fitted to ADR1399:
10NPLC, 5sec interval, 78mins total.
Max-Min across graph = 2.224uV

Snubber fitted to ADR1399:
10NPLC, 5sec interval, 83mins total.
Max-Min across graph = 2.127uV





Ian.

[iMo]

Also do not solder the snubber directly on the 1399 pins (ie from bottom side of the pcb). I did it in my 34401A and the noise (stddev) rose by 20%. After removing it it went back. Caused by the thermal flows as the capacitor and resistor worked as a heatsink. I did it with the stock 399 there as the preparation for 1399 replacement, but stopped the experiment as that would require an adjustment of the meter (adjustment of everything).. 

[Alex Nikitin]

Ian, the reference noise in not a problem in HP3245A. Here are some interesting measurements I've just made using a fully passive 0.1-10Hz filter and Keithley 2015M at 100mV range, 1 NPLC, changing voltage output from HP between 1.2 and 1,3V.

Cheers

Alex

[IanJ]

Ian, the reference noise in not a problem in HP3245A. Here are some interesting measurements I've just made using a fully passive 0.1-10Hz filter and Keithley 2015M at 100mV range, 1 NPLC, changing voltage output from HP between 1.2 and 1,3V.

Cheers

Alex

I have seen this before when I was playing with WinGPIB. I guess the PWM isn't properly controlled when changing output voltage. Am guessing in Lo Res mode the 3245 doesn't do it?

Ian.

[Alex Nikitin]

It is not PWM control, it is a range change, 1.25V is the maximum output for the lower range. Moreover the level of noise changed with the range but doesn’t change with the voltage, it is the same voltage p-p for 3V and for 10V.

Cheers

Alex

P.S. - Low Res behaves in the same way just the ranges are different, for example if you go from 1.250V to 1.251V in Low Res, the noise doubles from about 12-13uV p-p to 25uV p-p.

[branadic]

An update of my ADR1399-PWM-DAC.
Recap: I'm observing the bootstrapped reference, that uses PWM in the 7 V --> 10 V gain stage, with a S7081 and a Prema 6048 at the same time.

I repeated t.c. measurement, result attached.
Long-term drift observation unfortunately were corrupted by a data crash, resulting in a total loss of data 
Since then 1000 h of new drift data were recorded, which appear rather stable.

-branadic-