Ultra Precision Reference LTZ1000

[Dr. Frank]

Thanks Frank,

Glad I haven't missed any too obvious discussions then. 

Yes, agreed, Drift and TC need to be considered separately but I don't think moving from active heating to passive TC compensation should adversely affect drift. With well aged external resistors and low overall power dissipation (eg, no op-amp and transistor drive for the heater) hopefully at room temperature drift would be very low indeed.

For the TC, I'm not sure if your 50ppm/°C includes the Vbe of the compensating transistor or not, I'm assuming that it does (I can't find a figure in the datasheet). That particular application circuit seems to be tweaking the zener current to optimise the tracking of the zener and compensation transistor TCs (hopefully by shunting a 200R long term stable resistor with a high value less critical shunt one). It's a question of how close that tracking / compensation can get over a limited (room) temperature range, which it sounds as if nobody's really looked into.

It also occurs to me that if the heater is not in active use then there is also the sense transistor available which could maybe be employed in a second order compensation circuit?

At least having the on-chip heater would make the thermal cycling nice and easy for trimming... and maybe for 'relaxing' any hysteresis too.

Of course I don't actually have an LTZ1000 but the possibility is making me edge towards about a Digikey order!

EDIT: Though I suppose if adverse effect on drift is already so low when heated, then there may not be too much milage in this approach. I just like the elegance of a room temperature solution that doesn't have as many issues with thermocouple effects with a heated package.

Well, afair, it is nearly impossible for the LTZ reference, to trim its T.C. (unheated, of course superposition of zener and BE diode) to near zero.

Therefore, a room temperature solution w/o oven is absolutely wishful thinking.

The other Fluke topology, i.e. BE diode on top of the zener, that is SZA263 and LTFLU, they may be really trimmed to near zero T.C.

Fluke used that feature in the 731B, transfer standard, and in the 5200A, AC standard.

But these achieve 1 ppm/°C at best, and maybe the initial T.C. trimming may not last forever.

So you may try that, but such a solution has nothing to do with an Ultra Precision voltage standard, as we discussed here.
W/o an oven, you simply cannot achieve ultra high stability with Reference Amplifiers.

It is also worthless to make things complicated.

With the intended mechanism (oven), stable resistors, and a LTZ1000CH at 45°C, with the simplest circuit possible, you will definitely achieve  stabilities in time and temperature, which you will have big difficulties to verify with amateur based tools.

Thermo voltages are also not a problem at all with proper layout and thermo-mechanical setup.

The existence of all these ovenized references and standards are a simple proof for that!

The LTZ oven is not a good choice for temperature cycling, if you have a relatively low oven temperature.
This cycling, intended to remove hysteresis, only works correctly, if you have a symmetrical temperature swing below and above the stabilization temperature.

With 45°C, you may cycle between 25°C (r.t.) and 65°C. For these low temperatures, you might see no hysteresis at all, it's simply not working.

So you might select 65°C, to have a +/- 40°C swing, but this setting increases the timely drift, as a big disadvantage.

At 45°C, you may try to avoid any excess temperature of the LTZ, or do the temperature cycling of the whole assembly from 0°C to +90°C, in a climate chamber.


Frank

[Gyro]

Well I've long been known for my wishful thinking (for want of a better term)

My interest really was piqued by the LT application circuit - they must (still) think it's possible for them to have kept the circuit in the current datasheet for what is, after all, a very high end device. There are lots of cheaper solutions for average TC performance (or better - My SVR-T is under 1ppm/C at room temperature for what WAS similar money to the bare chip, which thankfully all I need). Maybe it's an app circuit intended to be used in an external oven (part substitution?), after all, an on-chip oven is always going to be a 'poor mans' alternative to a full circuit oven. Who knows, curious thing to include without a relevant comment though.

I agree, there's no point in arguing against the more or less universal use of ovens for high precision references, still feels a bit more brute force than finesse though. I think everyone would like to get away from warm-up, operating current, Flukes with batteries and transport time issues given half a chance.

Thanks for the additional thoughts and comments.

[TiN]

As little offtopic, reading this discussion reminds me myself a year ago when I started my reference design. I was thinking too about improving existing appnote design, but after experimentation on first prototype and eventually killing first LTZ chip, I roll back to proven circuit used by industry with just lower temppoint. We talking about 300$ BOM in parts just for one module (most of cost are VPG resistors), so it's rather limiting for testing many various options without proper multiK$ gear around... Sorry for interrupt

[Gyro]

Glad it brings back some memories - blue sky thinking is great until reality drags you down to earth again  You'll always end up being dragged down the route trodden by those with a larger R&D budget than you do!

Sooner or later I'm going to run out of my little pile of  '70s wirewound resistors (as in my Hamon divider) and have to start buying instead like you and Dr Frank - that comes hard to an inveterate hoarder!

[Galaxyrise]

After damaging the heater on my LTZ, I did play around with compensating it, and I could get its TC very close to zero.  I experimented with two different approaches, both of which were pretty successful.  I did the datasheet approach, and I also did one where I used an inverting opamp to amplify vbe. But getting it to near zero and keeping it there are very different levels of challenge.

[Gyro]

Thanks, interesting that the app circuit does 'sort of' work then. I wonder what drifted off over time then (don't worry, I don't expect anyone to answer that one!). I suppose it could be anything from die damage due to the heater accident to drift in the opamp or gain setting components, even Zener TC shifting that would normally be masked.

[bitslice]

(Apparently) some pictures of LTZ1000.
Only posting because they look different to the ones posted earlier

[plesa]

Assembled 4 boards with LTZ1000. For testing I used 25ppm resistors. but soon will be replaced by PWW.
LTZ1000 has been aged in oven for three weeks at 125°C. After PWW will be placed I will cycle it according to John Pickering.
All 4 units I would like to put into one box (similar to which use e.g wekomm use for their resistance standards).
I'm impressed how stable it is on table with 25ppm resistors, temperature of reference is 12k/1k so about 40-45°C.

[TiN]

Cool. Perhaps you can make bunch of pics and post in my kx thread? That would revive it a bit.

Happy to see it in use.

I would suggest to put resistors parallel to PCB, horizontally. Having them standing like this can cause thermal gradient over resistor body (due different length of pins and thermal mass), producing higher thermal EMF.  Just form leads in C-shape and solder horizontally.

Why all four in one box? Perhaps better keep them separate, so you can compare directly?

[ltz2000]

A (low resolution) photo from the US Air Force Primary Standards Lab. I was able to count 56 Fluke 732B Voltage standards...

[Dr. Frank]

Nice visit,
Do you also have a high resolution photo?

Frank

[plesa]

A (low resolution) photo from the US Air Force Primary Standards Lab. I was able to count 56 Fluke 732B Voltage standards...

Multiplied by quoted price (8k USD) it is fortune. I'm looking for some decent surplus ones.

[ltz2000]

Nice visit, Do you also have a high resolution photo?

Unfortunately I wasn't there. Just an internet find that I had saved in the funny pictures folder. Well, probably absurd rather than funny...

[Dr. Frank]

Assembled 4 boards with LTZ1000. For testing I used 25ppm resistors. but soon will be replaced by PWW.
LTZ1000 has been aged in oven for three weeks at 125°C. After PWW will be placed I will cycle it according to John Pickering.
All 4 units I would like to put into one box (similar to which use e.g wekomm use for their resistance standards).
I'm impressed how stable it is on table with 25ppm resistors, temperature of reference is 12k/1k so about 40-45°C.

I don't know, which oven temperature you'll use, maybe 45°C?
Then I wish you good luck for your try to remove the huge hysteresis, you've induced.
You were aware, that you have to apply symmetrical and very precise temperature changes around oven temperature, if Pickerings patent should work at all..
At 45°C oven temperature, vs. 125°C max., that would be -35°.. +125°C with decreasing amplitude. That needs a well controlled temperature chamber.

These parts are good for  drifts well below 1ppm/yr, if handled with care, and no excessive heating upfront.

Don't be disappointed, if you (very probably) encounter drifts of several ppm in the first years, until they finally / hopefully stabilize again.

Frank

[plesa]

Assembled 4 boards with LTZ1000. For testing I used 25ppm resistors. but soon will be replaced by PWW.
LTZ1000 has been aged in oven for three weeks at 125°C. After PWW will be placed I will cycle it according to John Pickering.
All 4 units I would like to put into one box (similar to which use e.g wekomm use for their resistance standards).
I'm impressed how stable it is on table with 25ppm resistors, temperature of reference is 12k/1k so about 40-45°C.

I don't know, which oven temperature you'll use, maybe 45°C?
Then I wish you good luck for your try to remove the huge hysteresis, you've induced.

These parts are good for  drifts well below 1ppm/yr, if handled with care, and no excessive heating upfront.

Don't be disappointed, if you (very probably) encounter drifts of several ppm in the first years, until they finally / hopefully stabilize again.

Frank

I'm going to cycle them according to John Pickering. Or in case the drift will be huge I will buy new ones LTZ1000 references.
If Mr.3458A (TiN) will have more PCBs I can assemble another 4 pieces for direct comparison.
At least after that we can make the final conclusion what is better

[plesa]

You were aware, that you have to apply symmetrical and very precise temperature changes around oven temperature, if Pickerings patent should work at all..
At 45°C oven temperature, vs. 125°C max., that would be -35°.. +125°C with decreasing amplitude. That needs a well controlled temperature chamber.

I have controlled environment, so it is not big deal. Yes, I going to little bit further ( safety margin ) -40°C  .. 130°C. I only needs to set the time on each step. I suppose that 5°C step with 20 minutes stabilization will be enough. Maybe I should check the 7001 manual how long this whole procedure takes.

[TiN]

I have some boards. If you don't mind bodging older rev (red one, have power on zener opamp reversed and heater transistor incorrect collector-emitter) I can send you 10 of those

I was able to count 56 Fluke 732B Voltage standards...

Who cares 732B, if you have two JVS rigs

[HighVoltage]

Nice visit,
Do you also have a high resolution photo?

Frank

I found a higher resolution picture and a pdf that explains a little about this military lab.
http://www.wpafb.af.mil/shared/media/document/AFD-150707-006.pdf
Page 9 of the PDF
It seems like money is no limitation factor for this lab.

[TiN]

Well, they still use LHe cooled PJVS (~300k$), not standalone cryocooled version (~400K$), so money are a little factor 

[Dr. Frank]

Well, they still use LHe cooled PJVS (~300k$), not standalone cryocooled version (~400K$), so money are a little factor 

In any case you'll need liquid He(4) @ 4.2K for these Josephson arrays.

In this picture, a commercial transportable dewar vessel (50 or 100l of liq. He4) is on display.
This lasts for a few weeks, until the He4 has evaporated.
Then you need to order a fresh vessel, and then transfer the PJVS over room temperature from the old vessel to the new one.

That's always a temperature shock, but modern array chips are assumed to be more ruggedized.

There also exist He4 circulation refrigerators, which only require an initial reservoir of liq-He4, and then may operate without further service.. that's probably what you meant.

You are not depending on the delivery of new vessels, and theoretically never have to remove the PJVS from the 4.2K environment.

Frank

[TiN]

No, i meant cryocoolers, some of which NIST used to go for 4K. JJA die does not generate lot of heat, so multistage cryopumps working on Stirling effect are suitable to do the job, I'd expect.

[Dr. Frank]

No, i meant cryocoolers, some of which NIST used to go for 4K. JJA die does not generate lot of heat, so multistage cryopumps working on Stirling effect are suitable to do the job, I'd expect.

Yeah, that's clear. But these cryo-coolers also operate with liquid He4, what else?

[TiN]

They don't need liquid He supply, only gas, which supplied by He compressor, forming closed loop.

[Kleinstein]

Such big labs usally have a supply of liquid He anyway - some even have there own recycling system. Still it is expensive to operate. Usually on can refill the cryostat with new Helium. So there is no need to warm up the setup so often. Transfering liquid He from one container to another is not simple, but it works with well insulated tubes. 

With liquid He one usually can get very stable temperature and no problem with a power outage.  The cryocooler also needs quite some electricity and maybe service every few years. The cooler may also produce extra noise.

An interesting question would be if one could get away with liquid nitrogen and high temperature superconductors.

[Dr. Frank]

Such big labs usally have a supply of liquid He anyway - some even have there own recycling system. Still it is expensive to operate. Usually on can refill the cryostat with new Helium. So there is no need to warm up the setup so often. Transfering liquid He from one container to another is not simple, but it works with well insulated tubes. 

With liquid He one usually can get very stable temperature and no problem with a power outage.  The cryocooler also needs quite some electricity and maybe service every few years. The cooler may also produce extra noise.

An interesting question would be if one could get away with liquid nitrogen and high temperature superconductors.

Well, on the picture from this AF lab, and also at the PTB, you'll see the standard transport He4 dewars, like from Air Liquide, Linde or in the past from Messer-Griessheim.
These have one standard flange for He4 transfer only, but this usually is used to stick the probe holder inside.

If they would use a stationary Dewar, like from Oxford-Instruments, these would have separate probe and He transfer flanges, and a refill would be  possible.

But it seems easier and cheaper to transfer the probe holder, instead.

At KfA Jülich they already demonstrated a Josephson standard with High Tc superconductors, an array with YBCO @ 77K, as far as I remember.
This was ok for about 1E-7.. 1E-8 uncertainty.
At these high temperatures, and with these amorphous superconductor structures, the Shapiro steps were sort of washed out, or the noise was too high, I don't remember that quite well.

So currently no option, also SQUIDs on HTcSC is still mediocre.

Frank

PS.: Here's the title of this paper:
"DC Voltage Calibrator Based on an Array of High-Temperature Superconductor Josephson Junctions
Alexander M. Klushin, Alexander V. Komkov, Valentina D. Gelikonova, Solomon I. Borovitskii, and M. Siegel"

Biggest problem with these materials still is, that monocrystalline films and structuring is not possible like Nb based SC.