3458A / SN18 in a non-temperature controlled environment

[alanambrose]

Hi all,

Anyone have a good procedure for SN18 tests in a non-temperature controlled environment?

From old notes I seem to have calculated a temperature co-efficient in ppm 10V/C of ~0.3 using the internal temp. 

Is the drift for drifty instruments so obvious that a great deal of science isn't needed?

Alan

[Kleinstein]

 The internal temperature should be OK to compensate for a possible temperature effect. With additional temperature effects it may just take longer to get a decisive result. Besides temperature there is also some possible hysteresis / recovery if the meter was off before for a longer time. 

It is rare to have a boarderline meter. The good ones are usually considerably better than the limit set by HP and the bad ones are often quite a bit (e.g. a factor 10) worse. If the drift is borderline it would just take longer to see if the meter is good. The closer to the limit the longer it would naturally take to see of the ADC still looks good.

[TiN]

Long time no see, Alan. Sounds like you got another 3458A ? 

You can follow usual SN18 procedure, it does not really matter if you have stable or not temperature if you collect TEMP? data together with ACAL values and if your temperature changes are random ups and downs. 0.3ppm/K sounds about typical for older 3458A (newer RoHS ones have higher TC on DCV). Since you'd only look for trending over few weeks on ACAL values, tempco usually would be minor contributor overall.

Good A3 stability should get numbers better than 0.05ppm/day, while bad ones are everything over 0.1ppm/day of drift. I have few A3 boards that have drift +0.5-1 ppm per hour. 

[alanambrose]

Thanks for your thoughts both.

Yeah TiN, I've been a bit busy

OK, well I guess the answer is to 'suck it and see' if you'll forgive that English idiom.

https://www.merriam-webster.com/dictionary/suck%20it%20and%20see#:~:text=Definition%20of%20suck%20it%20and,just%20suck%20it%20and%20see.

Alan

[maat]

In case this helps. This was, what my dying 3458A looked like, before it died.

Drift: (-2.857e-02 ± 2.892e-05) ppm/d
Tempco: (-8.501e-07 ± 4.026e-10) 1/K

The temperature was measured using a 10k thermistor stuffed into the side of the device. Then I called CAL? 72,1 in a loop.

[branadic]

I used multiple ACAL measurement performed at one day with the room at different temperatures and used CAL?175 data to compensate the CAL?72 values over temperature. This way I was able to create a chart of temperature compensated CAL?72 data.

-branadic-

[alanambrose]

OK well, reporting back to add a couple more meter drift data points...

Below graphs of:

(1) straight PPM drift and internal temperature for two meters (I think the room temperature data on the graph is bogus btw).

 - which give these regression results (units are CAL 72 PPM/day and /C):

    Meter 1                       Coefficients  Standard Error
    Date/Time                       -0.086         0.000
    Internal Temperature        -0.230         0.003

    Meter 2
    Date/Time                       -0.188          0.001
    Internal Temperature         0.520          0.005

Curiously one meter is +ve temp coefficient and one -ve (these are both properly old meters). Also, one reports internal temperatures 40-44C and one 35-39C (I cleaned both filters). Worth noting I guess is that the temperature coefficient is substantially bigger than the time one - so we might expect a bit of effort needed to get the time one out from under the temperature-based 'noise'.

(2) PPM/day over time calculated in the SN18 sense (i.e. ignoring temperature which moved about -1.5C over the 20-day period). This gives average 'drifts' of about +0.03 and -0.2.

(3) 'Temperature corrected' drifts using the regression results in (1) to correct the CAL 72 gain numbers. Fitting a line to the 'adjusted CAL 72' numbers gives drifts of -0.08 and -0.19. That is, roughly the same as the regression numbers - as it should. Curiously, there still seems to be a temperature variation observable and I believe I've done my calculations right. Maybe some delay or non-linear effects?

Alan

[alanambrose]

>>>
This was, what my dying 3458A looked like, before it died.

Drift: (-2.857e-02 ± 2.892e-05) ppm/d
Tempco: (-8.501e-07 ± 4.026e-10) 1/K
<<<

Thanks Matt. That is: -0.03 ppm/day and -0.85 ppm/C? They don't look crazy to me except the tempco a bit high.

Alan

[maat]

That is: -0.03 ppm/day and -0.85 ppm/C? They don't look crazy to me except the tempco a bit high.

Well, 0.03 ppm/day of statistical uncertainty is not bad, but this is a linear drift in one direction. so after a year this will be roughly 10 ppm, which is god-awful for a precision resistor network.

After Keysight replaced all of its innards, this drift is at least an order of magnitude less. In a temperature stable ennvironment ACAL is typically not needed. I use it if the humidity has changed quite a bit, but that it.

[alanambrose]

Duh, I did my back-out of the temperature effects wrong - of course, you need to deduct the fluctuation due to the temp coefficient to remove its effect.

See below - that looks a lot better and it also lets me plot the (cumulative) drift/day number to see the trend - both meters seem to be slightly improving after a couple of weeks.

Lastly, I fished out TiN's early coefficient #s from this page https://xdevs.com/fix/hp3458a/#a3drift for comparison. I trust he wont mind me quoting those here.

I guess a tabulation of the time and temperature coefficients for many meters would give interesting population stats - if we could get the methods approximately comparable.

Alan

[srb1954]

I have just been evaluating my 3458A (2005 build) after having a new A3 board fitted.

The drift appears to be dominated by temperature effects. After compensating for the effects of temperature I get:

CAL72 drift:   -0.006ppm/day
                    +0.45ppm/K

This is after after only 3 days monitoring with the new A3 board so it may still be too early to determine the true time-related drift. With the old A3 board it was very easy to pick out the time-related drift as it drifted a total of 50ppm over a 31 day test period! 
At times the drift was up to 0.55ppm/hour!                                 

[srb1954]

Update after a total of 9 days testing:

CAL72 drift:  +0.003ppm/day
                   +0.43ppm/K

The new A3 PCB seems to have virtually elminated the time related drift seen with the old A3 PCB.

Pity about the temp co though.   Looks like I will have to install some air con in my lab to tame down the temperature variations.

[Dr. Frank]

Hello,

I think it's convenient to publish here in this old thread my long-term measurement of the drift measurement of my 3458As U180, i.e. the drift of the CAL? 72 parameter.
I determined the T.C. of U180 to be about -0.24 ppm/K, so I corrected all my monitored data relative to its internal temperature (CAL? 72 versus CAL? 175 - CAL DCV Temp, or TEMP?) .

The new DCV calibration in June 2019, following our volt-nuts metrology meeting, shifted the absolute value of CAL? 72 by the same amount.
Anyhow, both time frames, 2016 - 2019  and 2020 - 2023 show a similar drift of roughly + 1ppm / 3 years, i.e. less than +1ppb/day.

Frank