How to verify my DVOM?

[John Frum]

I searched the web but could not isolate signal from noise.
I guess I would need references for voltage and resistance.
Is there any reference for either that is cheap and easy to source?
Back in the day It was fairly common knowledge at least in some circles that an American nickel is a pretty accurate weight reference(5.000 g).
I expect there are more of these little touchstones of civilisation.
Is their something like that for voltage and/or resistance?

 

[RCinFLA]

I don't know if you can find them any more but a mercury battery open circuit was very accurate at 1.34 v. I used to keep "N' cells for calibration.
There may be some mercury button cells for hearing aids still available.

Maxim has some reference source integrated circuits better then 0.1% accuracy.

 

[offgriddle]

I plucked this article from the noise floor of the Wild World Web; seems to address your question and then some.

I used to send my RF analyzers to a certified calibration lab yearly to assure FCC compliance.

Me, if I would just compare the readings of two good quality meters if I had any DVOM accuracy doubts.

 

[John Frum]

Is crossing the leads and getting 0.00 volts or 0.00 ohms useful?

 

[svetz]

Is crossing the leads and getting 0.00 volts or 0.00 ohms useful?

That's how mine self calibrates at startup.

 

[BiduleOhm]

We don't use mercury cells anymore (ecology, etc...) but there's chips who does the same job. You can also have a zener biaised as close as possible to it's 0 tempco but you'll need to calibrate it with a reference meter as zener are pretty good at maintaining the same voltage but not at giving a nice precise value out of the factory; but it's very cheap and simple (a LM78xx, a resistor, a zener). Then of course if you want more you can go very deep in the accuracy rabbit hole and it gets very expensive real quick...

For resistance it's easy as you just need a precision resistor with a low tempco and low drift. But it can get expensive if you want crazy high precision. For most people a 10-20 $ Vishay one is plenty enough tho, so not too bad.

Is crossing the leads and getting 0.00 volts or 0.00 ohms useful?

Not really, it removes the offset but that's all, your meter can still read completely wrong at other voltages than 0.

 

[grizzzman]

Start with a voltage calibration board. The AD584-M is cheap and has a track record.

 

[BiduleOhm]

Not sure I would trust a standard from aliexpress...

 

[gnubie]

It's pretty hard to fubar the operation of single chip ADCs and the same goes for the voltage references too. Pretty much everything that matters is internal to the chip with usually only supply stabilisation capacitors being on the outside and even then the current draw of the chip makes it possible to get away with crap there too. The only real concern is fake chips but the price these things are it makes it a difficult proposition even for the fakers.

 

[BiduleOhm]

Well, if you power it from a super noisy DC/DC converter or from a LDO with the capacitors not quite right so it oscillates then you can have very bad results.

But my main concern would be counterfeit chips. I'd recommend to buy the chip from a trusted source (Farnell, Mouser, Digikey, ...) instead

 

[grizzzman]

Or...... buy one that has the power supply. As an example the one I bought has an on board Rechargeable power Supply (lithium )

 

[gnubie]

Well, if you power it from a super noisy DC/DC converter or from a LDO with the capacitors not quite right so it oscillates then you can have very bad results.

Yes, that is possible but have you actually looked at this stuff on the usual sites? It's so stupidly easy and cheap to do it right that even the cheapies get it right enough.

 

[BiduleOhm]

Or...... buy one that has the power supply. As an example the one I bought has an on board Rechargeable power Supply (lithium )

I meant that about the onboard PS, not the user external PS.

And you just need the chip and two 9 V batteries in series (simple and you can be sure the PS is clean), nothing more if you just want to have a single output (selectable at 2.5, 5, 7.5 or 10 V) so why risking having a fake chip?

Yes, that is possible but have you actually looked at this stuff on the usual sites? It's so stupidly easy and cheap to do it right that even the cheapies get it right enough.

Yea, but as I said my main worry would be getting a fake chip, the PS is a secondary problem

//

Edit: I checked the prices and there's no way they use a real metal package AD584 as they have in the pictures. The plastic DIP version one (and in the 0.3 % grade, not the 0.1 % one, too expensive) would be possible however.

 

[gnubie]

I can't say for that particular item but the cheapie I got off ebay is the metal can variety. I don't know if the part is fake, but it doesn't appear to be salvaged. As far as accuracy goes, it's in spec.

 

[BiduleOhm]

My guess is fake or factory reject.

It's good it's in spec tho

 

[gnubie]

Factory reject perhaps, but if it's a fake they went to far too much trouble. It's in spec, according to a local cal lab, ultimately that's all that matters.

*edit*
Before anyone asks no, I didn't pay for it to be tested. That'd be stupid. FWIW none of my gear has been professionally recalibrated, there's just no need with what I do with it. I do know someone that works at this particular lab, about 15 minutes walking distance from me, and he offered to put my cheapie reference board through the mill and it came back good.

 

[offgriddle]

Buy or borrow a good quality DVOM and compare the two and be done with it already so you can move your project along. ~ You do have things to design, build, test, troubleshoot and confirm ... right?!