1V Reference @ 1ppm

[Rafael]

Hello,

Does anyone have any suggestions on how to find / build a 1v @ 1ppm reference?

Thanks a lot!

[Wimberleytech]

Hello,

Does anyone have any suggestions on how to find / build a 1v @ 1ppm reference?

Thanks a lot!

I assume you mean: 1ppm/°C.  Over what temperature range???

[Rafael]

Hello,

Does anyone have any suggestions on how to find / build a 1v @ 1ppm reference?

Thanks a lot!

I assume you mean: 1ppm/°C.  Over what temperature range???

You're right, sorry to forget this detail.

Range, around 23°C... I can imagine a hard task!

Thanks!

[Wimberleytech]

Hello,

Does anyone have any suggestions on how to find / build a 1v @ 1ppm reference?

Thanks a lot!

https://www.apexanalog.com/resources/products/vre3050ds.pdf

[Wimberleytech]

Hello,

Does anyone have any suggestions on how to find / build a 1v @ 1ppm reference?

Thanks a lot!

I assume you mean: 1ppm/°C.  Over what temperature range???

You're right, sorry to forget this detail.

Range, around 23°C... I can imagine a hard task!

Thanks!

Yes, it is challenging.  You must curvature compensate a bandgap reference...and pay attention to EVERYTHING.

Solution: Buy it.  I posted an example product.

If you limit the range of temperatures ("around 23°C"), you do not have to do curvature compensation. 

[beanflying]

I have used quite a few of the MAX6350 & 6341 (5 & 4.096V 1PPM/C)  on different jobs very good initial accuracy and stable over time but can still be tweaked if needed. One was a Mercury Button Cell (1.35V ) replacement for an old HP anchor I own. Buy some good quality Vishay or similar resistors (low temp co) and divide it down.

https://www.eevblog.com/forum/testgear/hewlett-packard-740b-dc-standard-digital-voltmeter-(and-740a)/msg1438536/#msg1438536

[spec]

Hello,

Does anyone have any suggestions on how to find / build a 1v @ 1ppm reference?

Thanks a lot!

https://www.apexanalog.com/resources/products/vre3050ds.pdf

Nice device, but the price: £138UK   Ouch

[Wimberleytech]

Hello,

Does anyone have any suggestions on how to find / build a 1v @ 1ppm reference?

Thanks a lot!

https://www.apexanalog.com/resources/products/vre3050ds.pdf

Nice device, but the price: £138UK   Ouch

 ...you gotta pay for precision!  The Maxim parts are much cheaper.

[mycroft]

Many, many years ago, i read an article by NBS (now NIST) on using a mercury cell as a voltage transfer standard. The major problem was temperature sensibility. I would like to find this article again. I searched a lot without success. Anyone remembers this article and has a reference?

[beanflying]

Many, many years ago, i read an article by NBS (now NIST) on using a mercury cell as a voltage transfer standard. The major problem was temperature sensibility. I would like to find this article again. I searched a lot without success. Anyone remembers this article and has a reference?

Not the article you wanted but some of us still have some Standard Cells and play with them from time to time.  https://www.eevblog.com/forum/metrology/1000-my-(hi)story-of-the-weston-cell-of-the-volt-and-of-being-a-volt-nuts/msg1039421/#msg1039421

I had mine out of the boxes a couple of days ago in 35 degrees ambient. It was 0.06mV up on last time I recorded it at about 20C and some of that drift would have been the 34401A. The un-saturated ones had a better tempco than the saturated ones. Generally run in banks and averaged when used as references too.

EDIT: NIST is currently offline thanks to the USA shutdown BS but try this search on Google 'nist weston standard cells' will get you some pdf hits when it gets back up.

[spec]

Does anyone have any suggestions on how to find / build a 1v @ 1ppm reference?

Hi Rafael,

I couldn't find a 1V voltage ref to meet your spec, so the only option is a higher voltage voltage ref, say 1V25, two resistors, and a precision opamp.

You could make a miniature oven quite easily and pop the whole circuit in the oven to improve the stability with ambient temperature changes.

As you no doubt know, the power supply would need to be pretty fancy, but all within practical bounds.

How much current would you want to draw from the 1V reference?

By the way, I am no expert on extreme high-spec voltage references, so these are just my thoughts.

If you are interested in this approach, just say and we can discuss it some more.

[tggzzz]

Hello,

Does anyone have any suggestions on how to find / build a 1v @ 1ppm reference?

Thanks a lot!

There is a lot of information in other threads on this forum, in the metrology sub forum.

Voltnuttery is expensive, very expensive.

But your question is easy to answer - but is it the right question? Hint: state your problem, not your presumed solution. For reasons, see https://entertaininghacks.wordpress.com/library-2/good-questions-pique-our-interest-and-dont-waste-our-time-2/

[tggzzz]

Many, many years ago, i read an article by NBS (now NIST) on using a mercury cell as a voltage transfer standard. The major problem was temperature sensibility. I would like to find this article again. I searched a lot without success. Anyone remembers this article and has a reference?

Not the article you wanted but some of us still have some Standard Cells and play with them from time to time.  https://www.eevblog.com/forum/metrology/1000-my-(hi)story-of-the-weston-cell-of-the-volt-and-of-being-a-volt-nuts/msg1039421/#msg1039421

I had mine out of the boxes a couple of days ago in 35 degrees ambient. It was 0.06mV up on last time I recorded it at about 20C and some of that drift would have been the 34401A. The un-saturated ones had a better tempco than the saturated ones. Generally run in banks and averaged when used as references too.

EDIT: NIST is currently offline thanks to the USA shutdown BS but try this search on Google 'nist weston standard cells' will get you some pdf hits when it gets back up.

I can use my working 1949 Weston standard cells as a thermometer, 40uV/°C

[beanflying]

How much current would you want to draw from the 1V reference?

By the way, I am no expert on extreme high-spec voltage references, so these are just my thoughts.

If you are interested in this approach, just say and we can discuss it some more.

If it is a 'reference' the answer would be as little current as possible to maintain voltage stability. Op amp buffers can be added but generally they add to errors and tempco problems not help.

There is for example a 0.01mV variation on my HP735A Transfer Standard between 10M and the GigOhm impedance settings on my 34401a so if it is being used as a transfer it has to be set to allow for this depending on the intended use. So load matters even if tiny.

[spec]

How much current would you want to draw from the 1V reference?

By the way, I am no expert on extreme high-spec voltage references, so these are just my thoughts.

If you are interested in this approach, just say and we can discuss it some more.

If it is a 'reference' the answer would be as little current as possible to maintain voltage stability. Op amp buffers can be added but generally they add to errors and tempco problems not help.

There is for example a 0.01mV variation on my HP735A Transfer Standard between 10M and the GigOhm impedance settings on my 34401a so if it is being used as a transfer it has to be set to allow for this depending on the intended use. So load matters even if tiny.

Yes, I do appreciate that that is a very specialist area and that 1ppm is only 1uV in 1V, so what I was going to do was to do the best I can and knock out a circuit and some notes for discussion. To  my mind, the opamp is the biggest problem because choppers would be too intrusive.

[beanflying]

The Maxim Ref I pointed to earlier in the thread https://datasheets.maximintegrated.com/en/ds/MAX6325-MAX6350.pdf is about 1ppm/mA typical (seems to be correct from my testing) so a 10k total divider would keep it fairly close to that or better and if you fit up the extra pot shown on the datasheet it can be tweaked up if needed without needing an op amp.

[spec]

The Maxim Ref I pointed to earlier in the thread https://datasheets.maximintegrated.com/en/ds/MAX6325-MAX6350.pdf is about 1ppm/mA typical (seems to be correct from my testing) so a 10k total divider would keep it fairly close to that or better and if you fit up the extra pot shown on the datasheet it can be tweaked up if needed without needing an op amp.

I was thinking on similar lines after you said that the loading would be very low. Will have a look at your suggestion.

If you had a 1V25 precision Vref (the lowest extreme precision voltage you can get), and connected a precision 250R and 1k resistor across it, as a potential divider, that would produce 1V with a source resistance of 200R (taking  the source resistance of the Vref as 0R), which would mean that a 200M load would only result in a delta V of 1uV or 1ppm, all other things being equal that is.

[spec]

The Maxim Ref I pointed to earlier in the thread https://datasheets.maximintegrated.com/en/ds/MAX6325-MAX6350.pdf is about 1ppm/mA typical (seems to be correct from my testing) so a 10k total divider would keep it fairly close to that or better and if you fit up the extra pot shown on the datasheet it can be tweaked up if needed without needing an op amp.

OK had a look at this. To have a 10K potential divider to produce 1V from a MAX6325 (2V5), would require 6k north  and 4K south. This would give a source resistance of 2k4, which would mean (ignoring the MAX6325 for the moment), that the minimum load for a 1ppm voltage delta would be 2G4.

I have also been thinking about the OP's requirement of 1ppm/deg C. But the specification we need is what voltage delta is acceptable over what temperature range and with what delta loading resistance, if there is a delta load- perhaps the load is fixed, which would simplify the design a bit.

[tggzzz]

Hello,

Does anyone have any suggestions on how to find / build a 1v @ 1ppm reference?

Thanks a lot!

I assume you mean: 1ppm/°C.  Over what temperature range???

You're right, sorry to forget this detail.

Range, around 23°C... I can imagine a hard task!

Are you concerned about drift over time? If so, you may need lengthy burn-in times.

[beanflying]

Without wanting to go way down the rabbit hole of voltnuttery a full and complete spec of the OP's requirements would be needed or we are just speculating on the requirements outside the 1V @ 1ppm of ? (voltage, temperature, drift over time etc) and what is the references intended use? So some more info is really needed.

Just run over the rough numbers based on the 10K guess I used this morning. -

To get the divider of 6 and 4k (assumed dead accurate) you would actually get 0.9996 volts read assuming the nominal 2.5V 10Meg meter so it would need the compensation pot fitting to the circuit to tweak it accordingly if it's purpose is for calibration and if not the input impedance of the driven circuit needs to be taken into account regardless.

My couple of tweaked MAX 6350's I built have remained within spec and have been powered on now for over a year so they seem to hold up well after the first week or so of settling. There was some questions raised about the epoxy version available (ceramic is off the market) could be susceptible to moisture over tim but that is well into VoltNuttery @ Andreas level territory (nothing wrong with that at all btw) 

[ArthurDent]

I appreciate the irony that someone asking about precision voltage has a tag line of "LTZ1000 for voltages"

[Rafael]

Hello,

Does anyone have any suggestions on how to find / build a 1v @ 1ppm reference?

Thanks a lot!

There is a lot of information in other threads on this forum, in the metrology sub forum.

Voltnuttery is expensive, very expensive.

But your question is easy to answer - but is it the right question? Hint: state your problem, not your presumed solution. For reasons, see https://entertaininghacks.wordpress.com/library-2/good-questions-pique-our-interest-and-dont-waste-our-time-2/

Thanks a lot for all answers... As a beginner, I need to learn a lot before try some adventures in Metrology section.

I have an old russian voltmeter https://www.eevblog.com/forum/beginners/31-voltmeter-sch-31-nixie-russian-voltmeter/ and I need to calibrate it with 1v reference.

@ArthurDent, I really really admire the precision that the component can generate, but for 1V, it gets a little further.  edit: Or maybe the answer? Please, be gentle with me...

[beanflying]

Nice Nixies I hadn't seen the thread hiding away in the Beginners section. I have seen similar ones on evilbay from time to time and occasionally tempted.

Given the instrument you really need a better 1V source than even 1PPM assuming it's own internal reference is at least that stable. Those cheap Chinese references are reasonable to 3.5 and 4.5 digit meters or as a sanity check on whatever you may build is close to the actual 1V but no more. The initial accuracy of the Maxim references and the easy trimming I would have confidence to better than 4.5 digits against another good meter but not to 6.5 digits.

Nothing I have for example is up to 'recalibrating' my Agilent 34401A other than I have a pair of 34970A dataloggers with 6.5 digit multimeter cards fitted that are near the same spec. All of my references held at a static temperature give me a reasonable level of certainty as to how the 34401A is behaving over time and temperature but as it is currently out of Cal it really needs to take a trip to a lab or I need to scrape the $ together for a 3458A and keep that in Cal to keep the rest in line

What should be important to you is that YOU are happy you have the required level of precision, accuracy and certainty YOU need unless you go to the cost of getting the meter calibrated at a known Lab. Pick a day or conditions at a known temperature and record it preferably 20-25C. Then raise or lower just the meter if possible and see how the drift is with temperature while the reference hopefully remains at a static temperature. Also watch your connections to the meter you will get several PPM of error with a bad connection. The Russians use an oddball size under 4mm

Should you build a reference YES but bear in mind what that reference will give you. The Max will get you about as good as you can get for initial accuracy without breaking the bank. To go to an LM399 or LTZ gives you more ultimate stability but a wider range of initial accuracies so if you don't trust the meter a worse result 

[tggzzz]

Hello,

Does anyone have any suggestions on how to find / build a 1v @ 1ppm reference?

Thanks a lot!

There is a lot of information in other threads on this forum, in the metrology sub forum.

Voltnuttery is expensive, very expensive.

But your question is easy to answer - but is it the right question? Hint: state your problem, not your presumed solution. For reasons, see https://entertaininghacks.wordpress.com/library-2/good-questions-pique-our-interest-and-dont-waste-our-time-2/

Thanks a lot for all answers... As a beginner, I need to learn a lot before try some adventures in Metrology section.

If you follow the points in that reference, you will find that most people on most forums will be friendly. A small proportion might not, but that's people for you.

I have an old russian voltmeter https://www.eevblog.com/forum/beginners/31-voltmeter-sch-31-nixie-russian-voltmeter/ and I need to calibrate it with 1v reference.

You probably want to calibrate each range separately; a stable known voltage and a 7 decade Kelvin-Varley Divider is a good starting point. You will realise there is no end point in voltnuttery, only an acceptable level of dissatisfaction

You might like to do some initial experiments.

For example find a battery which is reasonably stable, keeps its temperature reasonably constant, connect it to your meter, turn on the meter, and record the measured value until it becomes stable. Some meters (particularly old ones) take quite a few hours to stabilise. Keep an eye out for popcorn noise in the meter's reference, and variations with temperature.

When at the uV levels, seebeck/thermocouple effects become noticeable.

[iMo]

The answer is simple: there is not such an 1V/1ppm source available in form you would expect (not talking volnuttery capable calibrators). As I can see on the pictures from your link you are already volnut-positive, so you are asking in a wrong section