Keysight 34401A

[cemelec]

Just acquired a new old stock Keysight 34401A (6.5 digit) at a good price, never used, original packaging, complete with all accessories and calibration certificate dated January 2016.

Its accurate in all functions according to the best standards I have - 0.01% dc low voltage, 0.1% resistance, drift after warm up is negligible. Full self test all pass.

A general question - what is the main cause of drift in such an instrument? Is it usage or age?  It has some age but no usage.

I'm contemplating getting it to a calibration service, but it won't be cheap. I know I don't need that level of precision, just want it. Also a cal certificate might improve its value?

Thanks
Charles

[Andreas]

A general question - what is the main cause of drift in such an instrument? Is it usage or age?

Can be both. (usually effect from usage is larger since the voltage reference is heated).

Another effect may be humidity during storage.

So I would let the device run at least for a week before comparing it to a standard.

With best regards

Andreas

[HighVoltage]

Also a cal certificate might improve its value?

These instruments are very stable over time.
If you get it calibrated, most likely it will be within specification and no adjustments are needed.

The value of the instrument is greatly increased with a new cal certificate, especially if you want to sell it.
In addition you have a good reference point with the certificate because the cal certificate should tell you how far off the instrument is in all ranges.

Just be careful falling in love with such instrument, soon you have two and you do not know which one to believe!


 

[cemelec]

Just be careful falling in love with such instrument, soon you have two and you do not know which one to believe!

You are so right - I now have 5 oscilloscopes, at least 7 AF/RF sig gens, 9 or 10 multimeters, etc.etc. plus a mountain of components enough to stock a small shop.

Still, its mostly hobby so I don't need to justify it. The 33401A is so nice I'll be keeping it, I'm now searching for a decent voltage calibrator I can afford. Saw an Analogic AN 3100 DC Voltage Standard recently, supposedly 0.005%, maybe worth getting....

Charles

[Andreas]

Hello,

I'd rather go for Ians PVDS2 instead of the AN 3100

https://www.ianjohnston.com/index.php/onlineshop/handheld-precision-digital-voltage-source-2-mini-detail

with best regards

Andreas

[Keith956]

I wouldn't bother getting it calibrated. They don't drift much - I have 3 of them, of varying ages, and the newest was last calibrated in 2013.

A quick check shows they all agree within better than 0.01% to each other....

[bdunham7]

If you do get calibration, make sure you get before and after data (and post the results!). I don't think it will increase the 'value' of your DMM by more than the price of the calibration.

When Fluke started including cal certificates with some meters, they published a note essentially saying that you could ignore the cal date on new meters that had been sitting on the shelf for a while (months, typically, not years) and presume that the meter would be accurate for one full cal cycle from the date it was put in service.  They claimed this was based on evidence and testing.  I don't know HPAK's position on this, but I would guess that if your meter was stored properly (not at extreme temperatures) in its sealed bag with a desiccant, it will likely still be spot-on.

I have one that is 25+ years old and has never had the cal adjusted and it is spot-on on the DC ranges after a warmup, but it does have some short term variation with temperature.  I suspect the long-term drift issues are minimal and mostly related to power-on time and power-on cycles ageing the LM399 reference.

[YetAnotherTechie]

I wouldn't bother getting it calibrated. They don't drift much - I have 3 of them, of varying ages, and the newest was last calibrated in 2013.

A quick check shows they all agree within better than 0.01% to each other....

All diferent readings?    Well, you definitely need another (better) one to trust!   

[bdunham7]

I wouldn't bother getting it calibrated. They don't drift much - I have 3 of them, of varying ages, and the newest was last calibrated in 2013.

A quick check shows they all agree within better than 0.01% to each other....

I'm not sure you make the best case for not checking calibration--yours are spread 49ppm apart and no two agree to even 5 digits...  Are they fully warmed up?

[Keith956]

I suspect it's noise from the power supply - the last 2 digits are constantly changing, and they are triggering at different times. Maybe I'll try a battery and shorter leads...

[mc172]

I wouldn't bother getting it calibrated. They don't drift much - I have 3 of them, of varying ages, and the newest was last calibrated in 2013.

A quick check shows they all agree within better than 0.01% to each other....

Nice!

[guenthert]

I wouldn't bother getting it calibrated. They don't drift much - I have 3 of them, of varying ages, and the newest was last calibrated in 2013.

A quick check shows they all agree within better than 0.01% to each other....

I'm not sure you make the best case for not checking calibration--yours are spread 49ppm apart and no two agree to even 5 digits...  Are they fully warmed up?

     They differ by 11ppm and 37ppm of each other respectively.  Quite possibly they are still within spec.  Of course, it's impossible to tell by this measurement alone.

[bdunham7]

     They differ by 11ppm and 37ppm of each other respectively.  Quite possibly they are still within spec.  Of course, it's impossible to tell by this measurement alone.

Well, yes, if you take the 1-year uncertainty of 47 counts at 1 volt, you can conjecture that if the input is actually within a certain range, all 3 meters are within 47 counts of it.  However, I would expect 3 such meters on the same source, in the same room, all warmed up, to be much closer.  The entire uncertainty budget includes the standard distribution (2 sigma), guardbanding and most importantly, +/-5C worth of tempco.  His explanation about the PSU being noisy actually is more plausible than those 3 meters disagreeing that much, IMO.

[HighVoltage]

In my experience they do not drift much at all.
I think I took this picture in 2017 and these shown instruments had not been calibrated in years before this picture was taken.

[mendip_discovery]

If I was buying a meter for a lab then history is nice, much like service history for a car.

Get a UKAS calibration on it. A good lab will give you before and after readings if they adjust it.

Here are the plots for mine to give you an idea of stability.

[Keith956]

I'm not sure you make the best case for not checking calibration--yours are spread 49ppm apart and no two agree to even 5 digits...  Are they fully warmed up?

Here's a pic of the 4 I have, all connected to a battery. 2 have just come back from calibration, the other two have not been in cal for almost 10 years. Can you tell which two are in cal?

[DH7DN]

I'm not sure you make the best case for not checking calibration--yours are spread 49ppm apart and no two agree to even 5 digits...  Are they fully warmed up?

Here's a pic of the 4 I have, all connected to a battery. 2 have just come back from calibration, the other two have not been in cal for almost 10 years. Can you tell which two are in cal?

Which value and uncertainty has your voltage standard? 

[Keith956]

3041A calibrator. Think that's 25ppm?

[DH7DN]

Hmm, those Agilent 34401A measurements are very difficult to interpret without stated uncertainties 
25ppm = 1 year accuracy of your calibrator, not its uncertainty! If has a traceable calibration, it should also have an uncertainty for certain voltage values! 

[Keith956]

Hmm, those Agilent 34401A measurements are very difficult to interpret without stated uncertainties 
25ppm = 1 year accuracy of your calibrator, not its uncertainty! If has a traceable calibration, it should also have an uncertainty for certain voltage values! 

The cal certificates I have state the uncertainties (for DC voltage):
2V: 5.6ppm +/- 10uV; 20V: 4.4ppm +/- 73uV; 200V: 5.1ppm +/- 0.7mV; 1kV: 5.7ppm +/- 7mV +/- 1LSD

[guenthert]

If I was buying a meter for a lab then history is nice, much like service history for a car.

Get a UKAS calibration on it. A good lab will give you before and after readings if they adjust it.

Here are the plots for mine to give you an idea of stability.

      Well, I do not work in a calibration lab and need help interpreting those graphs.  On the abscissa the range (or calibration set point?  w/o uncertainty?) is plotted, but what is on the ordinate?  The deviation (in ppm)?

      Am I wrong to have expected some time series instead, if those graphs are meant to demonstrate the instruments stability?  Stability over time was meant, wasn't it?

[DH7DN]

I made some calculations and actually three of four of your DMM measurements agree within manufacturer's specifications (see diagram).

DMM: Keysight 34401A, DC Voltage, 1 year Spec. 
Range 10 V: +/- (0.0035% of reading + 0.0005% of Range)

I'm not sure about the values of your calibrator. Usually, in a calibration certificate, one gets the deviation from a nominal value (e. g. 10.00000 V +/- Deviation) and a stated expanded uncertainty of the standard.

Well, just to compare the Keysight 34401A measurements: I assumed you set the calibrator to a nominal value of 7.757000 V (+/- 0.000107 V). If so, then only two of your DMMs agree with the measurement against the voltage standard.

Just an estimate, I could be wrong uncertain 

[Keith956]

I'm not sure about the values of your calibrator. Usually, in a calibration certificate, one gets the deviation from a nominal value (e. g. 10.00000 V +/- Deviation) and a stated expanded uncertainty of the standard.

I have the docs from the cal lab specifying the expected deviation etc. if you're interested. Only one needed adjustment to meet the cal spec, and that was on the 100V/1kV AC volts ranges only. So as said before, they don't drift much over a 10 year period, and calibration only makes sense if you need to prove it...

[Kleinstein]

The usual calibration does not include an adjustment. So in essentialyl all cases the DMM will give the same reading after calibraiton as it would heve shown without the calibration (excluding the extra stress from transport). So the meter itself does not get better, one only knows the meter was checked and in some cases one gets actuall readings of the cal soures, so one could numerically correct the deviation to some degree.
The 4 meters shown are still close enough togehter, that all of them could be still in spec.
The difference between the extremes is still relatively large. So likely at  least 1 of the meters seem to have drifte quite a bit.

[bdunham7]

Here's a pic of the 4 I have, all connected to a battery. 2 have just come back from calibration, the other two have not been in cal for almost 10 years. Can you tell which two are in cal?

If your meters have all been warmed up for at least an hour, I sincerely hope it is the bottom 2.