Ultra Precision Reference LTZ1000

[TiN]

I don't use special transformers, but the battery power for critical measurements.
If anyone have problems with my design, I'd like to see feedback from those people, otherwise it's just a guesswork.
Using DMM as the measure of reference (is it reference? to me even those people you linked are not so sure ) spikes is somewhat far-fetched.

Same would be saying ref have a problem when observing shift few ppm by connecting noisy DMM like F8846A. 
Film SMT capacitors are not some cents, but that is (together with 10$ for shielding) IMO irrelevant, as it is not commercial product, but a learning tool.
Also in your design it is not just two capacitors, but other modifications required as well, to make it all working properly.

[MisterDiodes]

Andreas:

It's OK if you don't understand the structure of a crystal lattice and aging effects, and some of the factors that might make that worse - or better.  That's well beyond the scope of what I can show you here.  I'll give you a hint below though.

From my experience after building hundreds of LTZ's, we've never needed to alter the basic LTZ circuit.  Ever.  Yes we've added shielding and Yes on the test bench we take pains to keep everything as quiet as possible.  For units going into the field they include more ruggedized shielding techniques and that works in every situation we've encountered.  Like how it works on 3458a. The reason we can't always add CPU's to the circuit for DAC controls is we're in a situation where WE can't be the ones making noise, and that's why sometimes pure analog is the way to go.  Sometimes an assist with a CPU / FPGA is a good way.  It just depends on application.

If you had to alter the circuit for your use, that's one thing - but I know from experience that doesn't apply everywhere.  It doesn't apply to 3458a, it doesn't seem to apply to TiN's circuits, it doesn't apply to our circuits, LT/AD doesn't need or recommend they be used, etc.  Now of course there might be -some- effect on YOUR board or test setup where the altered circuit got better results - but that doesn't mean it applies in -every- case, does it?  If someone is having trouble, and adding the caps helps - then Yay!.  But I suspect the problem really might lie elsewhere.

I would invite you to look at a new Re-Spin of an LTZ application, now that AD owns the Company - Look at the LTZ circuit here, and in this case you want a low noise Vref to show off the 20 bit AD5791 DAC they're selling - see page 20:

http://www.analog.com/media/en/technical-documentation/user-guides/eval-ad5791sdz-ug-1152.pdf

After a chat with AD, I'm told there are some typos that will be corrected on the next doc release...But the note on C44 refers to C13 C12, The note about R30 should be R48, and the note re: U3, U4 should refer to U1, U2.

The deviation from the LTZ1000 datasheet is they use SMT parts (board picks up vibration and stress changes), the '1013 is swapped for an AD4077-2 (Basically newer gen precision OP07) because AD wanted to showcase some of their own brand.  The AD4077 current driver likes an extra cap compensation, and the usual 22nF cap has been upped to 100nf - otherwise the circuit topology is just really a basic LTZ.

Notice the RC between the LTZ and the 10V boost.  That a good place to remove a little noise and isolate the LTZ from any downstream activity / current noise.

Notice the 15k over 1k heater resistor ratio.  They went with this "Hotter" ratio because on an bare naked board they liked the warmer run temp of the LTZ - it would resist casual air drafts better.  This may cause slightly lower long term drift but that is not the goal here.

Something else to notice is the standard resistors used for 10V boost:  Because this circuit is feeding a DAC you'd like to have a little OVER +-10V cardinal output on this Vref, which means as designed you want the LTZ with a minimum 7.14VDC using the 25k/10k boost ratio.  Since AD owns LTZ part, they can pick and choose the LTZ demo chips to make sure the LTZ demo board  reference output has enough headroom to let the output of the '5791 output run to +-10VDC full scale (The AD guy was pretty sure on this when I pointed that out, but he is going to check).

Andreas (BIG CLUE):  After 36 years of LT / AD manufacturing LTZ chips and getting customer feedback, please take a look at that schematic topology and let me know how many low impedance caps you see straddling any single PN junction on the LTZ?.  How many chopper amps are used?  What could be the reason for that "Fairy Tale" ??

Now back to a real test:  I haven't tested this board on my bench, but my colleague tried one of these new LTZ ref boards on his '5791 demo, and wanted to double check these test results - and he does have access to a .1 to 10Hz LNA (Jim Williams Wet Tant style):

http://www.analog.com/en/analog-dialogue/articles/high-precision-voltage-source.html

So I emailed him last night to see how the results went.  In general his comments were positive.  Testing the noise at the output of the LTZ was typical <<1.2uV p-p, and the '5791 output was fairly close to the published test report IF you pay attention:

A) Gotta add the goofy little afterthought foam LTZ draft cover in the kit. Looks like you're on your own for covering the bottom side (AD says they are changing that on next rev).

B) He couldn't do it with the USB cable from a laptop to the demo controller card - he used an isolated cable from a desktop PC box  located on the floor a little farther away.  The output of the DAC got cleaner when you unplug the digital controller cable, as expected.

C) The Keithley switcher power supply added a little more noise to the output (not a total disaster), but switching to batts or Protek 303 linear (transformer) supply worked fine.

D) This circuit didn't like the soldering iron cycling on and off on the same outlet strip, nor did it like the LED desk lamp.

E) BUT: Putting the circuit and LNA amp in a leftover Xmas Cookie tin got very good results, and after that the desk lamp was fine.

F) If he set the board on a soft foam bleacher pad, the board quit reacting to every mechanical noise on the test bench (Noise pickup from SMT caps).  Without the pad, if he flipped a switch or tapped a pencil on the bench, that would show up on the output.

G) He saw more noise on the '5791 output at full scale - he was getting more like 4 or 5uV (or 7uV on occasion) peak.  The datasheet DAC spec's 1.1uV added noise but that's at mid-scale, at 90%+ full scale you'll see more noise.

EDIT:  He noticed that if you averaged the peaks from several 10 sec periods the peak noise could be called ~4uV peak.  So he was able to get in the same ballpark as published results.

I realize that not everyone has access to LT / AD apps engineering.  I realize people like to play.

Again:  My only suggestion is - If you're finding that adding caps helps your circuit, you might sniff around a bit to find out what the true cause of the problem might be.

The standard circuit topology -should- work fine, and I suggest if you're having trouble with a modified circuit:  Try stripping it back to the standard circuit and that might help you troubleshoot where the real problem is.  You might find that correcting a thermal / noise issue will work as well or better as adding extra caps - and that in turn might give your LTZ the very best chance at aging gracefully.

[MisterDiodes]

One other suggestion while you play with Vrefs, and a head's up if you're just starting out:

Sometimes there's an "uncertainty principle" involved in measuring the LTZ (or any precision Vref) circuit nodes proper while the circuit is running:  If you probe the circuit nodes with a meter or scope, that can change the thermal balance of the LTZ enough that you're suddenly looking at a different LTZ circuit operating point when the probe is in the circuit.  Especially probing the traces anywhere near the LTZ can suck enough heat out - or add enough capacitance - that suddenly your LTZ will show a shifted Vref output, or sometimes add a bit of oscillation.  The LTZ gets cooler, then adds more heat, then your probe heats up a little and suddenly less heater power needed, etc.  That can show up on the output as a wobble.  Sometimes it's fine; you'll see the Vref output glitch and then re-stabilize a short time after the probe goes in or out of circuit.

Even if you attach a measuring access wire or trace to a circuit node (and even a very fine wire) or try to get a measure from a component lead -  that becomes part of the thermal balance, and that may or may not have an effect.  Lower mass probes can work better here, or sometimes a longer stainless steel probe needle will help (less heat flow), and try to keep your warm paws away from the circuit as much as possible!

So it's a good idea to keep an eye on your Vref absolute value if you're trying to read for example the Vbe of the heater sense transistor.   Try to grab your "hot" measures in a way where the Vref output doesn't change (much) during the measure, with your probe in or out of circuit.

Otherwise you might be chasing a "phantom" signal.

[Insatman]

Hello Voltnuts,
                      I have attached schematics of the EV-LTZ1000-REFZ carrier board as well HiRes photo of the actual board and R48 can be found here: http://www.vishaypg.com/docs/63109/vfcd1505.pdf

Best regards
Fred Flinstone

Hmmmm....why the 1N270 Diode is being used in this board rather than the 1n4148 used before.   Data sheet attached.

[zhtoor]

Hello Voltnuts,
                      I have attached schematics of the EV-LTZ1000-REFZ carrier board as well HiRes photo of the actual board and R48 can be found here: http://www.vishaypg.com/docs/63109/vfcd1505.pdf

Best regards
Fred Flinstone

Hmmmm....why the 1N270 Diode is being used in this board rather than the 1n4148 used before.   Data sheet attached.

interesting... maybe MisterDiodes can explain this. (lower Vf maybe Ge vs Si)

-zia

[martinr33]

The small SMT 0.1 uf capacitors on the board look like X7R material, because they appear brown. These devices are microphonic - x7r has piezoelectric properties.

Replacing them with C0G parts would likely eliminate the mechanical sensitivity.

[Cerebus]

Hmmmm....why the 1N270 Diode is being used in this board rather than the 1n4148 used before.   Data sheet attached.

One immediate thought is that these evaluation boards tend to be used lying around in the open on the bench. The 1N4148 has a clear glass case, the 1N270 is opaque. One hopes the implication of this is obvious in present company.

Beyond that I'm not sure that there's anything in it - the two places a diode is used are both the kind of place where almost any old signal diode will do. Ge V_f tempco is about -2.5mV/°K versus Si's of -2.1mV/°K, so nothing helpful going on there. V_f is lower, but it's not as if this circuit is scrabbling to find the odd 0.3V. Junction capacitance is much the same.

In fact, unless I've missed something, the 1N4148 and 1N270 are about as similar as it's possible to get while using a different semiconductor material except one's black and one's clear.

[Cerebus]

The small SMT 0.1 uf capacitors on the board look like X7R material, because they appear brown. These devices are microphonic - x7r has piezoelectric properties.

Replacing them with C0G parts would likely eliminate the mechanical sensitivity.

In point of fact they have X8R dielectrics, the TDK part number (C1608X8R1E104K) is called out in the eval. board documentation.

[zhtoor]

Hmmmm....why the 1N270 Diode is being used in this board rather than the 1n4148 used before.   Data sheet attached.

One immediate thought is that these evaluation boards tend to be used lying around in the open on the bench. The 1N4148 has a clear glass case, the 1N270 is opaque. One hopes the implication of this is obvious in present company.

Beyond that I'm not sure that there's anything in it - the two places a diode is used are both the kind of place where almost any old signal diode will do. Ge V_f tempco is about -2.5mV/°K versus Si's of -2.1mV/°K, so nothing helpful going on there. V_f is lower, but it's not as if this circuit is scrabbling to find the odd 0.3V. Junction capacitance is much the same.

In fact, unless I've missed something, the 1N4148 and 1N270 are about as similar as it's possible to get while using a different semiconductor material except one's black and one's clear.

no 1n4148 opaque equivalent in silicon? why germanium? gold-bonded at that (supposedly lower noise)

-zia

[blackdog]

Hi,


I cant see why there is a 1N270 and not a 1N4148, the diode is in the feedbackloop, so normaly you can forget about the properties of this diode.
But.. i'am never to old to learn 

Kind regards,
Bram

[martinr33]

Here's a link to an article on piezoelectric effects in MLCC devices.

http://www.kemet.com/Lists/TechnicalArticles/Attachments/88/2006%2007%20ArrowAsiaTimes%20-%20MLC%20Noise.pdf

x8R is a nice dielectric, but  does make the design vibration and shock sensitive. Replacing those caps with C0G devices will help with the mechanical sensitivity.

[MisterDiodes]

I posted that schematic of the AD5791 just as an example of how the standard LTZ1000 topology works for noise - the point being is that even a fairly casual, less than perfect design should easily get you in the ballpark of <1.2uV noise, even on a bare board.  This is not the best design if you're going for low long term drift on higher precision system, and that was not the goal of the demo board.  This is designed just as a quiet, stable-enough-for-a-demo voltage source that can demonstrate low noise in the few uV range even after boosting to ~10V.

By the way that demo board LTZ Vref was a cut & paste from another piece of in-house gear at LT/AD where they needed a low noise source, my rep says, and that is Rev 0.  He says that they have another rev drawing coming up from the docs department.  I suspect some of the parts on the BOM are there because there aren't critical and are what they had in stock (see below about diodes).

If you were to buy a Vref from me, all critical parts in the analog section would be Thru hole, because vibration noise is a P.I.T.A. and where our modules go they have to be vibration-resistant.  If you need very low TC we might talk about a chopper amp for the VBoost section (after the RC to isolate the LTZ from chopper noise), and we'd use PWW resistors when required.  If the customer has other needs then we go from there.

We would also consider air draft shields very important around the LTZ and we also especially look at the traces and component leads around Pin 6 on the LTZ.  Since whatever you connect to Pin 6 is also part of the very sensitive thermal balance on the Temp sense transistor.  All the pins on the LTZ are critical, but Pin 6 can cause a fair amount of trouble.  If you suck heat out of that base junction, you're un-intentionally changing Vbe, and well...you kiss heat stability goodbye.   The other pins are sensitive too.

When the LTZ tells you to shield the LTZ from air drafts, by extension that also means the traces and leads directly connected TO the LTZ might be causing trouble if they see air drafts. 

Like I keep saying: LTZ is a Power In / Power Out device.  Get the power flow stable and a stable Vref is the by-product.  That means all the thermal circuits have to be stable.  Generally we won't ever run an LTZ on a bare board except for burn in or initial testing...when it's going to work it is always in a shielded enclosure.  At least that's how we -always- use them.

The ADA4077-2 amp is just a newer gen OP-07 (and closest thing AD had to an LT1013, for the sales dept.).  Even that is not the most perfect amp to use since it is not spec'd for low current noise at 10Hz.  The LT1013 is spec'd for low current AND low voltage noise @ <10Hz, AND comes in a DIP.  The LT1013 IS -the- amp to use with the LTZ because it was designed with that exact LTZ application in mind.  There has never been any true upgrade for it.

The diode is not super critical.  It is only there to ensure that the LT1013 (or other amp) would start the current loop.  On the newer process LT1013's often you don't even need the diode because they can make sure the Voffset is slightly shifted.  Sometimes you can run an LTZ without the diode at all and there's enough Voffset when the amp powers on that the current loop starts and settles in a stable condition.  Without the diode you run the risk of the current loop starting in its second stable condition, where current = 0 for this circuit - which is a perfectly valid operating point also. 

In fact if you go back to very early prototype 3458a's, I think there are might be a few of those Vrefs around without the diode - but LT learned pretty early on what can happen if a 3458a powers up with a Vref sitting at 0 volts.  Oops.

SO:  You need that diode to guarantee startup across all temperatures, and that's about it.  I don't think diodes optics plays a huge role but that would be something to look at.  Once the LTZ Vref goes into an enclosure that should NOT be an issue at all.

The diode on the heater line is just a safety prevent reverse bias on the substrate.  It is not critical in operation, but it should always be there during power supply power up / down.

Remember the LTZ circuit needs a quiet CURRENT driver for its Zener, and the Voffset of the loop is not too important.  The gain of the transistor and Zener is so large and overwhelms any small Voffset drift in the current loop amp - it doesn't really have a huge effect on the final Vref output.  It's more important to keep current noise very low.

That's why LTZ's never need or want a chopper amp in the current loop, and that's why LT/AD doesn't recommend using chopper amps there, even though they would be more than happy to sell you a million or two choppers.

[kj7e]

I found the LTZ1000A was many times more sensitize to air currents around the pins, top and bottom than air over or around the can.  My best results where with the chip just above the board and with the leads shielded from air currents, but with the can exposed.







As an unintended benefit, the PCB traces are also protected from air currents near the chip.

[ArthurDent]

There is a reason that the 1N270 (and other Ge) diodes are painted black. Years ago I followed the instructions of an article that described removing the black paint from a common Ge diode and using it for a photo sensor in a flash meter for photography. The diode with the paint removed was placed under a frosted plastic dome removed from an old neon panel lamp. When the flash went off the diode (which formed half of a voltage divider) conducted proportional to the intensity of the flash and then a regular diode connected the voltage to a capacitor used as a sample-and-hold circuit with an IGFET to display the voltage on an analog meter with the scale calibrated in aperture values.

As far as using Ge vs Si diodes within a feedback loop, the Ge would have lower voltage drop but I’m not sure if that is the reason or how it would matter.

[MisterDiodes]

The voltage drop of the diode in the current loop really makes no difference. The current driver amp will always drive it's output to bring its input pins into balance - so a silicon or germanium diode doesn't really matter.  You just need to create a voltage drop in the loop to guarantee the current amp will start the loop flowing in the correct direction at power up.

The need for LTZ air draft protection is clearly shown on the datasheet, in fact there are couple sample output noise plots showing the difference between no air draft shield vs air draft shield.  This has always been a key requirement of the LTZ since Day One.  Typically you cap the top and bottom but protecting the pins is most important.  If you OVER insulate the LTZ system that can create problems as well, since the heater is depending on heat flow -out- of the chip to maintain regulation.  The heater can only add heat - the system depends on at least a little heat flow out of of the device pins to maintain regulation if it gets too warm.  Normally you want that to not happen, and normally it stays in good control.  That's why you always pick your heater ratio resistors to run the chip at a comfortable temperature margin over ambient.

[MisterDiodes]

..On some of the 3458a's there was also a high value resistor from V+ going to the Vref line - and that was an additional startup aid to trickle a little current into the current loop before the current amp really started.  According to LT that shouldn't be needed on newer process LT1013's - but often nobody wants to change a working LTZ circuit so that resistor is still included. The regular datasheet circuit doesn't show it these days, the older sheets did have that.

I remembered something else:  With that ADA4077 there may be a "sweet spot" where if if its operating output is raised a certain offset from the rails that made for a bit lower output noise.  The ADA4077 also needs an extra comp cap on the output.

At one time that was another reason for the 1N4148 on the LT1013 - it raised the normal amp output drive level  to a "sweet spot" where you see a bit lower noise on the amp output.  LT says you probably don't need it but they suggest using a regular Si diode in that location to cover all bases for best performance. OK!

[TiN]

Perhaps attached chart might be interesting to some in this thread.
After some greedy bidding on recent Wavetek 7000 bits and pieces I got myself one 7000 cell (based on LTZ1000CH and TaN statistical resistor network to provide 10V) and 7000S switch modules.

My friend Todd hooked it up to K2400 to power with +12V, and we get next log, showing conditioning procedure over the 10V output:



Now it settled to about -8 ppm.

.

Peak-to-peak wobble is about 1ppm, which is a LOT for LTZ1000-based ref. Maybe this was the reason why Fluke killed 7000 series ref? 

Cover unit with the blanket...



Aha Another reminder to air drafts not allowed when talking <ppm noise resolution. 

Article and reverse engineering of the module is planned for future projects.

[doktor pyta]

Article and reverse engineering of the module is planned for future projects.

Wow! This will be a hit!

[Echo88]

You never fail to amaze TiN.  But im more interested in the used DCDC-converter which boasts with very high isolation, than the hysteresis-elimination technique used.

Wavetek 7000 DC reference testing from eBay at MM lab:

[Cerebus]

Perhaps attached chart might be interesting to some in this thread.
After some greedy bidding on recent Wavetek 7000 bits and pieces I got myself one 7000 cell (based on LTZ1000CH and TaN statistical resistor network to provide 10V) and 7000S switch modules.

Oh, goody. I'll be very interested to see how these settle out across several cycles of the conditioning process. That is, I'm interested in how close the end values of several conditioning procedures are to each other.

Can I please suggest that this gets its own thread/topic rather than being tacked on as a subsidiary to this, already overlong, thread.

[TiN]

Is it just me, or Analog devices does not allow to buy LTZ1000 from website anymore? Once I click "Purchase" it redirects me to ecommerce cart site, but does not add anything into cart.
I tried both ACH and CH, still same. If add other parts, like opamp - no problem. 

[Cerebus]

Is it just me, or Analog devices does not allow to buy LTZ1000 from website anymore? Once I click "Purchase" it redirects me to ecommerce cart site, but does not add anything into cart.
I tried both ACH and CH, still same. If add other parts, like opamp - no problem. 

Possibly a glitch in the transition to the Analog Devices website? If I were you I'd drop them a line and ask.

[Macbeth]

Is it just me, or Analog devices does not allow to buy LTZ1000 from website anymore? Once I click "Purchase" it redirects me to ecommerce cart site, but does not add anything into cart.
I tried both ACH and CH, still same. If add other parts, like opamp - no problem. 

I tried and get the same, actually no I get worse - if I choose any sample, like LM399A which I gladly sample for free, I am given the red carpet treatment... https://shoppingcart.analog.com/ and my old Linear email and password works fine.

The same LM399 and I click Purchase (or indeed the LTZ1000 which has never been available for sample so I can only click Purchase)... it takes me to https://checkout.na3.netsuite.com/app/center/nlvisitor.nl/... and requires a login. I've tried both my old AD and Linear logins with no joy. The email password reset isn't working either. 

So my advice is to load up with all the AD and LT samples for free while they sort it out... 

[eurofox]

Is it just me, or Analog devices does not allow to buy LTZ1000 from website anymore? Once I click "Purchase" it redirects me to ecommerce cart site, but does not add anything into cart.
I tried both ACH and CH, still same. If add other parts, like opamp - no problem. 

I tried and get the same, actually no I get worse - if I choose any sample, like LM399A which I gladly sample for free, I am given the red carpet treatment... https://shoppingcart.analog.com/ and my old Linear email and password works fine.

The same LM399 and I click Purchase (or indeed the LTZ1000 which has never been available for sample so I can only click Purchase)... it takes me to https://checkout.na3.netsuite.com/app/center/nlvisitor.nl/... and requires a login. I've tried both my old AD and Linear logins with no joy. The email password reset isn't working either. 

So my advice is to load up with all the AD and LT samples for free while they sort it out... 

You guy's must be on a black list because I can buy without problem, I just try it!

 

[SvanGool]

None of my linear or AD accounts worked either. But if you register new and just add the item to your purchase order, you can purchase. The only thing I had is that minimum UPS shipping cost for two LZ1000ACH to The Netherlands (Europe) is USD 54, which diverted me back to Digikey