Seeking Advice on DIY Calibration for 34401A Multimeter

[wmundstock]

Hello everyone,

I'm experiencing some discrepancies between my Rigol DM3068 and a 34401A multimeter, and given the Rigol is pretty new, I am thinking its a decent reference. I would like both meters agree on the readings.
I'm considering attempting a DIY calibration for the 34401A and have a few questions for those who might have experience or insights:

- Is DIY calibration a feasible task?
- The calibration menu, after disabling the security feature, is showing 000 for everything. Should I expect to see previous calibration values there? Does this mean the calibration was  "reset"?
- I understand that running the zero calibration might overwrite existing calibration values. Can someone confirm if this is the case? I am worried to start the zeroing and then make the readings worse.
- Is it necessary to calibrate every range of every function, or can I focus on the functions/ranges that need more attention?

Any advice or tips would be greatly appreciated as I navigate this process.

Notes:
- The HP is older model, with firmware 05-01-01.
- For DC reference voltage planning to use a chinece circuit bought from aliexpress.
- For resistors planning to use a chinece circuit bought from aliexpress.
- For AC voltage, planning to use a function generator with sine wave.
- I do not have means to generate DC voltage higher than 30v nor AC voltage higher than 10v or so. Would need to build/buy something.
- I do not know what I could use to generate current, maybe a power supply with low voltage and current limit?

Thank you in advance for your help!

[J-R]

IMHO, short answer: leave the 34401A alone.

Longer answer: none of the references you mentioned seem adequate for calibrating a DMM and I would suspect the 34401A could be better than the Rigol (do you mean the DM3068?).  So I would ship the 34401A out for calibration + data.

[coromonadalix]

the right answer is no,  34401a  and some others are software calibrated .... unless i see the contrary

Yes i would send the 34401a to calibration,  the lm399 in it is supposed to aged well enough,  and as a minimum  you need at least 6.5 digit precision reference or higher to do the tasks

What you are proposing mak absolutely no sense at all and you could destroy all the goodness of the 34401

I had 3x 34401a in the past, sent one to cal after a psu capacitos change (all of them),  and compared it with the 2 others once done, and they were still pretty good againt it for minumum 20 years old stuff

[wmundstock]

Thank you for the comments J-R and coromonadalix!

IMHO, short answer: leave the 34401A alone.

What you are proposing mak absolutely no sense at all and you could destroy all the goodness of the 34401

I thought I could get some feedback like this. Long live the goodness of the 34401! I am glad I asked before I messed it up.... sometimes I just cant leave things alone.

(do you mean the DM3068?)

Yes, fixed in the original post.

I would suspect the 34401A could be better than the Rigol

That crossed my mind at some point, but I have no reference to make such assumption.

[bdunham7]

I would like both meters agree on the readings.

Can you give some specific examples of how much particular readings are off and the source being used?  How long are you letting the 34401A warm up and what is the ambient temperature when you do these tests?

[coromonadalix]

34401a  need at least 1hr to be stable if not more

you have some teardown of the other dmm 3068, cant recall on wich reference it's based, maybe an lm399 too,  but it need to age  ...

wich 34401a  did over a looong period

[NoisyBoy]

Agree with all others comments, for a 6.5 digit meter, you need a stable source, which none of the ones you named are.  And if you plan to use a transfer standard, I for one will never trust a Rigol due to the lack of aging.  You should also complete each test point to record the calibrated gain constants, which the test equipment you listed cannot provide.

To do a proper calibration and adjustment, you are looking at equipment like a Fluke 5560A, which is not always within budget for most DIY.  If accuracy is important to you, send it out to get a professional adjustment, if not, like others said, leave it alone.  Trying to get two 6.5 digit meters agree on every measurement is a slippery slope, especially when one is relatively new. My guess is once adjusted, the 34401A will prove to be a far more stable meter despite the lack of newer features. 

[robert.rozee]

Is DIY calibration a feasible task?

as others have said, no.

were you to send your 34401A off for calibration there is a good chance it would come back with a report of how close to 'perfect' each range was, but no actual adjustments made. for instance, measuring precisely 10vDC, and given a 0.0015% basic 24-hr dcV accuracy (https://www.keysight.com/us/en/assets/7018-06774/data-sheets/5968-0162.pdf), the meter would be within specifications for any reading between and 10.00015 and 9.99985 and thus not warrant any adjustment being made.

the above also tells us that the last digit, in this example, on its own has no meaning. it only becomes meaningful if the 34401A is used to measure and compare two values. but when combined with a series of calibration reports of a single meter going back over some number of years, then greater meaning can be attributed to that last digit. the calibration reports give us confidence that the meter reading may be stable over time. in general, adjustment makes those calibration reports far more difficult to interpret.


cheers,
rob   :-)

[iMo]

The "adjusting" of the 34401A is a feasible task, provided you have the references against which you want to make the adjustment.

The "Calibration" menu shows first the password (the standard one is "034401" you may change !! therefore be careful what you do) and deeper in the menu, when you are for example in the DCV 10V range, it shall show you "+10.00000" at the beginning of the adjusting process.

You can adjust a single range only, I did it several times with my DMM (inspired by TiN's video on his xdevs.com).

I did 10V and 1V ranges in my case and it worked fine.

For example - Set the DCV 10V range in your typical setup, wire your reference voltage into the DMM's inputs (it shall be around 10V), go to Calibrate, unlock when necessary, navigate to "+10.00000" on the display, and then with the arrow buttons dial in Your reference voltage (for EXAMPLE "+09.99772"), then press Enter. It will show something like "Full Scale.." on the display during the internal adjustment process running.
Note: you may dial in negative numbers too, therefore take care with the sign..

It will do the internal adjusting for some time (like 10-30secs afaik), WAIT (!!), and DO NOT press anything.

After a while it returns to the normal DCV 10V mode automatically and you shall see Your 9.99772 (EXAMPLE only) reference voltage on your display.

Disclaimer: Do it at YOUR OWN RISK, NO WARRANTIES of any kind are provided..

PS: Your 34401A has to be powered on for some time, I did with at least 1 week powered on. The reference voltage shall be also powered on for some time, of course..
Before the adjustment double check your Zero reading is 0V at the 10V range with the above example (with all 4 inputs shorted together properly!)..
For more details on proper wiring, temperatures, etc. do consult the Metrology Section..

[Kleinstein]

The Rigol meter is not a good reference, even if brand new. So best keep the 34401 as is.
The HP meter has a good reputation in keeping it calibration even over a long time, while the DM3068 is more likely to drift. With the age difference we don't know which meter to trust more.


It could still make sense to check the zero reading and the difference to the other meter. Some difference it normal.

Another point could be to give the 2 meter a quite check on a few points, that can be done without much instrumentation.This can be
1) check the zero
2) turn over error (so measure the same voltage positive and negative, maybe with a few repeats)
3) input current  (e.g. via the drift rate with a 1-10 nF PP capacitor)
4) noise with a shorted input (if you have a PC interface)

[alm]

To me the Rigol is more likely to drift, particularly if it's fairly new, since Rigol and Siglent don't do the burn in and selection of their references that the established brands do. So I'd agree with sending out the 34401A rather than the Rigol. It will likely also be easier to find a lab that will calibrate the 34401A than the Rigol since it's a very popular meter and calibration and adjustment procedures are well published.

I would be very hesitant to do any adjustments at home unless you are confident you are going to improve the accuracy. This generally requires a reference that is better (not just equally good) as the meter. So for DCV you'd need a reference that is much better than 35 ppm. Otherwise you are more likely to make the situation worse than better.

I have done calibrations with very limited means like function generators and power supplies, but not adjustments except in the one case the DMM was way off: always reading 0V in a particular ACV range, which I did adjust with the max 7 Vrms my function generator would output on the 30V range.