Ultra Precision Reference LTZ1000

[eurofox]

Hello Volt Nuts,

I would like the add that I did some test with this unit and the Geller Lab that is mounted in the same box with the same power supply.

I use Pomona special cables (Shielded and I think they are twisted as well).

No influence of change on the latest digit (10nV) by touching the cable or moving the reference box, it is immune to noise from outside (of course in the area of my lab but there is a RF phone and a computer near with no influence)

This mean that if you have some instable measurement you have to look to your cabling, grounding, connector but basically it is not that critical.

eurofox

[Dr. Frank]

Hello Volt Nuts,

..
No influence of change on the latest digit (10nV) by touching the cable or moving the reference box, it is immune to noise from outside (of course in the area of my lab but there is a RF phone and a computer near with no influence)

..
eurofox

Hey, eurofox.

Your LTZ circuit is really extremely insensitive, extremely accurately trimmed (10^-9), and your measurement is extremely high resolution.

Could you please tell me the model and manufacturer of your 9 1/2 digits DMM? 

And at which NPLC setting the DMM measures stable on the last digit?
Thanks!

And the slotted PCB is from China?
At which temperature did you set the LTZ?
(Or which resistor over the 1k did you use?)

Frank

[eurofox]

Hello Volt Nuts,

..
No influence of change on the latest digit (10nV) by touching the cable or moving the reference box, it is immune to noise from outside (of course in the area of my lab but there is a RF phone and a computer near with no influence)

..
eurofox

Hey, eurofox.

Your LTZ circuit is really extremely insensitive, extremely accurately trimmed (10^-9), and your measurement is extremely high resolution.

Could you please tell me the model and manufacturer of your 9 1/2 digits DMM? 

And at which NPLC setting the DMM measures stable on the last digit?
Thanks!

And the slotted PCB is from China?
At which temperature did you set the LTZ?
(Or which resistor over the 1k did you use?)

Frank

Hello Frank,

Maybe I did not express myself corectly, I mean that moving the reference box (box with inside the power supply closed), touching the cables, there is no influence on the reading with my LTZ1000A or the Geller lab, and yes of course after stabilisation periode and call of the HP3458A the last digit (10nV) change value but stay in the range of the last digit (usually less than 5 with NPLC 100).

I will post the schematic after cleaning up, it is basically the application note wiith a few changes.

The PCB is comming from Germany (at least I bought it there).

eurofox

[Mickle T.]

HP3458A the last digit is much more than 10 nV 

[Dr. Frank]

Hello Frank,

Maybe I did not express myself corectly, I mean that moving the reference box (box with inside the power supply closed), touching the cables, there is no influence on the reading with my LTZ1000A or the Geller lab, and yes of course after stabilisation periode and call of the HP3458A the last digit (10nV) change value but stay in the range of the last digit (usually less than 5 with NPLC 100).

I will post the schematic after cleaning up, it is basically the application note wiith a few changes.

The PCB is comming from Germany (at least I bought it there).

eurofox

Hello eurofox

(the heck, what's your real 1st name?)

Sorry, I just wanted to tease you a little bit....

The 3458A in the 10V range resolves 8 digits only, that's 100nV, not 10nV!

Well, 5 digits of fluctuation (500nV) fits to what I have measured @ 100 NPLC on LTZ references from babysitter and my two ones.

See 10min stability diagram here: https://www.eevblog.com/forum/projects/ultra-precision-reference-ltz1000/270/

Who designed the PCB? And what temperature did you chose?

Well,

each of those LTZ references is very susceptible to external AC sources.
So I banned a nice new LED lamp from my analogue lab, because its switch mode PSU disturbed the reference up to 1 ppm, inside its shielded box.
So I think, your even less shielded case will show even worse behavior, if treated with the "right" HF noise source .
Frank

[eurofox]

Hello Frank,

Yes you are right on the 10V it is 100nV, I'm confused because when use my Time Electronics calibrator I really go to the 10nV in the 10mV range of the calibrator.
I'm quite new with this instrument and start now to be familliar with it.

eurofox

[CaptnYellowShirt]

Hello Everyone!

I've taken the last month (or there about) to read though this thread, the posted pdf's, associated forums on the LTZ (volt-nuts, etc), and several scientific papers (IEEE, Metrologia) on modern voltage transfer standards.

Let me just say... Wow!

I'm overwhelmed by the shear amount of high quality "amateur" knowledge that is available on this subject. Thank you all for your contributions!


After reading all this information, I've had a few ideas, but I wanted to check with the experts here before I let my mind go too far...

There seems to be a body of information on the environmental conditions that effect the LTZ's long term stability -- atmospheric pressure, humidity, mechanical stress*, etc etc.

There have been a few suggestions in this forum on how to 'seal' a LTZ and its associated hardware away from most ambient conditions (O2, H2O), but has anyone seen a design that has placed the reference board in a vacuum? If so, can you post a link?

Secondly, has anyone seen a study on the effects of ionizing radiation on the LTZ (or semiconductor references in general)? If so, a link would be helpful.


* metal packaging with long leads aside (Dr. Frank, et al)

[Andreas]

There have been a few suggestions in this forum on how to 'seal' a LTZ and its associated hardware away from most ambient conditions (O2, H2O), but has anyone seen a design that has placed the reference board in a vacuum? If so, can you post a link?

Secondly, has anyone seen a study on the effects of ionizing radiation on the LTZ (or semiconductor references in general)? If so, a link would be helpful.

Hello,

I think placing a reference board in vacuum is usually beyond the capabilities of a amateur.
And further it does not solve all problems since you would have to use kovar leads to get through the glas sealing.

I would never place a metrology grade reference into ionizing radiation.

I can remember that there is a application note from Intersil on the ISL21009 on the output voltage shift of their FGA references. But since the voltage is stored into a EEPROM-like cell the shift is more likely than on other references.

And there are some annotations on the Linear Technology LT1021 in the /883 version.
http://cds.linear.com/docs/en/datasheet/RH10215fe.pdf

With best regards

Andreas

[branadic]

And not to forget to mention the problem of gas emission of the pcb itself. I would prefer a solder process of the case under shielding gas conditions with feedthrough capacitor for electrical connections trough the case. IMO the most promissing approach for homebrew application.

[CaptnYellowShirt]

I think placing a reference board in vacuum is usually beyond the capabilities of a amateur.
And further it does not solve all problems since you would have to use kovar leads to get through the glas sealing.

I would never place a metrology grade reference into ionizing radiation.

I can remember that there is a application note from Intersil on the ISL21009 on the output voltage shift of their FGA references. But since the voltage is stored into a EEPROM-like cell the shift is more likely than on other references.

I've read that note from Intersil. As I remember the gist there is they have a MOS-capacitor that stores a reference charge level. That charge level is set before it leaves the factory -- relying on the near-zero leakage rate of the capacitor to store the same charge level for 10+ years. The worry is that an ionizing particle will whiz though the capacitor, altering the charge level.

The effect that I am thinking of in the case of a zener reference is a bit different (changes to the semiconductor structure).  However, I have to admit its simply that -- my imagination. So I was wondering if anyone had seen an expert talking on that subject.

While I don't know much about the subject, but common sense would tell me not to (intentionally) expose a reference to high energy...  anything. But we're awash in radiation day-to-day.  A cosmic ray probably interacts with the zener cross section at some rate of 1's of strikes per day. Of course, you can't really shield against cosmic rays, but other lower energy (terrestrial) sources can be shielded or avoided -- concrete, ceramics, etc etc.

As far as the vacuum idea is concerned, it really depends on the level of vacuum. Ultra high vacuum? Yeah, forget it you might as well build yourself a JJ array. A vacuum on the order of 1's or 10's of Pascals.... definitely achievable.  Sustainable in the long term? Maybe.  It'd have the advantages of isolating the reference from atmospheric conditions and removing convection as a heat transfer mode. But as you've pointed out, it would add construction and design difficulty. I am considering building one, but before I sit down to design it, I wanted to ask for some feedback from the group here.

[CaptnYellowShirt]

And not to forget to mention the problem of gas emission of the pcb itself. I would prefer a solder process of the case under shielding gas conditions with feedthrough capacitor for electrical connections trough the case. IMO the most promissing approach for homebrew application.

I have a few of these feed-though "gizmos" ( http://www.omega.com/pptst/MFT.html ) laying around from another project. Three to be exact.

I considered the de-gassing problem and I have to admit, I have no experience with that. Slow degassing could be handled by several pump-down sessions. But I would be worried about placing a large stress on the components during the (relatively) quick pump-down. Has anyone ever seen an LTZ explode? Sounds fun, but I can think of better ways to spend my money.

I considered back-filling with argon. Which would take care of the oxygen, water vapor, and (most) pressure changes, but it would.... I forget a lot of my heat transfer... increase convection(?).

[TiN]

Almost ready for thermal sim

Board - FR4 4-layer 40x80mm

+15VDC input
LTZ1000A schematics from datasheet
LT1112A Opamp
PZT3904 NPN in SOT223
Three MAX6610 temperature sensors.
LT1761ES5 5V LDO for temp sensors
Planed resistors: Vishay Z202

Inner layers right now pretty much solid power ground (opamp GND and heater- GND).

Temperature sensor locations:

Sensor 1: Near Opamp with copper plane under it.
Sensor 2: On wires, glued to top of LTZ1000A in dead-bug position.
Sensor 3: On bottom near LTZ1000A

Will run temperature simulation to try get approximate temperature gradients, and tune accordingly before gerber out.

Any ideas? 

[sync]

A general question about the LTZ1000 circuits from the data sheet. The output is directly connected to the LTZ1000 (ZENER SENSE). Can it used for a reference with external output (long leads, current injection from the DMM, ...) or is a buffer needed?

[CaptnYellowShirt]

Almost ready for thermal sim

What program do you use for the thermal simulation?

[Andreas]

A general question about the LTZ1000 circuits from the data sheet. The output is directly connected to the LTZ1000 (ZENER SENSE). Can it used for a reference with external output (long leads, current injection from the DMM, ...) or is a buffer needed?

If you short the output (a uncharged capacitor is sufficient) then the temperature setpoint will rise to infinity. -> The heater goes to 100% -> The chip will get a (more or less permantent) voltage shift on the output.
So I do not recommend to use a unbuffered LTZ1000 (like mine) as working standard.

With best regards

Andreas

[iTist]

Secondly, has anyone seen a study on the effects of ionizing radiation on the LTZ (or semiconductor references in general)? If so, a link would be helpful.
* metal packaging with long leads aside (Dr. Frank, et al)

Hi,

here is some information from the  JPL (jet Propulsion Lab)
http://nepp.nasa.gov/docuploads/6A935FBA-65D1-42CE-827076A3DE0C7384/PrecRef-97.pdf


Greetz

Oliver

[CaptnYellowShirt]

Hi,

here is some information from the  JPL (jet Propulsion Lab)
http://nepp.nasa.gov/docuploads/6A935FBA-65D1-42CE-827076A3DE0C7384/PrecRef-97.pdf


Greetz

Oliver

I didn't even think of space electronics. Good find!

It would seem the crux of article would suggest its probably not an issue for anything that's kept on the Earth's surface. If I'm doing my maths right, an equivalent exposure for an Earth-based LTZ1000 would occur over the period of 10,000's of years (for natural radiation ~2000 strikes/m^2/s @ 1GeV). And since the LTZ in the article only shifted by 10's of ppm for that level exposure....

It makes me a little sad to think my DNA is drifting more than the LTZ is.

[Galaxyrise]

If you short the output (a uncharged capacitor is sufficient) then the temperature setpoint will rise to infinity. -> The heater goes to 100% -> The chip will get a (more or less permantent) voltage shift on the output.

The datasheet mentions multiple times about using PWM to drive the heater, and you're implying that's bad?  I've been wondering about that myself... PWM would seem to have a lot of undesirable effects on the rest of the circuit.

[TiN]

Almost ready for thermal sim

What program do you use for the thermal simulation?

I'll try to use Solidworks Flow Simulation, using still air box primitive. Problem is to transfer traces from PCB CAD to Solidworks.
Plan to use LTZ1000 datasheet temperature data to reach as a target first, so will have that as a reference for simulation jig

[CaptnYellowShirt]

And not to forget to mention the problem of gas emission of the pcb itself. I would prefer a solder process of the case under shielding gas conditions with feedthrough capacitor for electrical connections trough the case. IMO the most promissing approach for homebrew application.

I've been considering both the vacuum and inert gas ideas in the past few days. However I'm stuck on one problem -- trimming the circuit.

My common sense tells me that if there's any merit to either idea, one would want to place the entire circuit (op amps, current source, resistors, divider, etc) in the shielded environment. But if everything is nicely setup inside of a sealed container, how does one access the circuit to adjust it for the first time?

One method might be just to build it, insert it into the sealed container, start it up, and accept whatever value it starts with (7.xxxxxx V) as being the reference point -- valuing the (potential) higher stability above having a nice round number (7.000000 V).

But if there was a way to adjust a resistor (or a set of dip switches, etc?) from outside the container, I'd sure like to use that method.

Any ideas?

[iTist]

Hi,

you can use a DAC for controlling the see:
Maybe at the FLUKE website or her:

http://www.elcal.ch/files/11749-eng-01-a.pdf

Page3 Figure 5


Greetz

Oliver

[SimonSatCom]

This is how far I am with my LTZ1000A project. It is not an easy project...but I have learned a lot in here.

http://www.simonthenerd.com/LTZ1000A.htm

[Andreas]

Hello,

Nice project, from where do you get the housings for your references?

The temperature setpoint of your LTZ is rather low for a LTZ1000A.
In my lab (up to 32 degrees) I would not go below 12K5 for the temperature setpoint resistor.

Do you have a temperature stabilized lab?

With best regards

Andreas

[Dr. Frank]

This is how far I am with my LTZ1000A project. It is not an easy project...but I have learned a lot in here.

http://www.simonthenerd.com/LTZ1000A.htm

...but you also made the mistake to buy the A version and 2nd to set it to a too low stabilization temperature.
A temperature difference of 15-20°C above maximum room temperature is necessary for proper operation.

The LTZ1000A is intended for battery operation and for elevated temperatures.
It is badly abused @ 95°C in the HP3458A and in the Keithley 2002, with mediocre 8 ..10ppm/yr.

Burn-in (storage at high temperatures) is also a fault, due to the pronounced hysteresis, which will in turn create a longterm creeping drift, which would not be present by avoiding any elevated temperatures.

The LTZ1000 (non A) can be set to 45°C with <1ppm/yr, which is easily achievable.

Fluke 7000 and Datron 4910 are the most stable references (<1ppm/yr.) and are based on the LTZ1000 (non-A) with 45/55°C!
 
Frank

[TiN]

But what would be the best operation point for LTZ1000A then?
Asking in practical reason, as have A-based reference in design as well.