Easy way to test the calibration of a DMM (Fluke 45)?

[CalibrationGuy]

Yes, what Josh said.

[shabaz]

No, you're missing the point, when you stated:

People measuring the Chinese/Asian boxes do NOT (most likely) have the capabilities that my lab has, so PLEASE take their observations with a grain of salt.

Those cheap boxes are actually usable to 3 digits, for an identification of which instrument has the 20 mV discrepancy. It's a very successful way of seeing which instrument has a significant discrepancy. It makes zero sense to expend a lot more on such a check when you'd have to calibrate your device anyway, if you need accuracy.

Perhaps the OP has a lab full of instruments or wishes to go into the calibration business and willing to spend $2700 per year.

I was simply focussed on addressing the OP's problem. He doesn't need to expend anywhere near that amount to get measurement to the specification of his instrument, once it has been properly calibrated by actually shelling out for it (or a calibrated meter purchased, for a tiny fraction of that cost). Thanks for your perspective, but you missed my point regarding how low-cost it can be for an initial identification of which instrument is wildly out of spec. Only an idiot would assume it was in any way close to an actual calibration.

As for Josh, I'm surprised you can't find a way to get a half-price calibration?
You seemed pretty proud of the unethical practice of hacking instruments beyond what you'd paid for them, on another thread.

[CalibrationGuy]

No, I'm not missing the point. Without traceability, those numbers mean NOTHING. I have friends who have boards like these and the numbers were the same for every board. Your usability claims have no credibility because there is no verification of said values.

My point is that counting on these values is problematic at best.

TomG.

[shabaz]

Well now you're just trying to swing the discussion into something else.

"Without traceability, those numbers mean NOTHING"

If you need $2700 of traceability to figure out which instrument is off by 20 mV or so, then there's something wrong with the way you do engineering.

[KungFuJosh]

As for Josh, I'm surprised you can't find a way to get a half-price calibration?

I actually do have a half-price calibration waiting for me from the manufacturer of one of my DMMs. 🤣

[KungFuJosh]

Well now you're just trying to swing the discussion into something else.

"Without traceability, those numbers mean NOTHING"

If you need $2700 of traceability to figure out which instrument is off by 20 mV or so, then there's something wrong with the way you do engineering.

I don't know what your first language is, but there seems to be something lost in translation with you. He never said anybody else should spend that money on calibration. The dude's name is "CalibrationGuy" you might infer from that that he's somewhat into calibration.

[CalibrationGuy]

I'm going to reply because there are people who will read this thread and misunderstand what is being said. So, let me be clear.

1) I pay top dollar because I must be sure of my data.
2) Using a newly calibrated meter verified that a very reasonably priced voltage reference achieved excellent results.
3) You cannot verify with any certainty, 20mV which meter is off using a reference from unknown sources without proof of calibration.
4) To tell the OP that it's a good idea to rely on an unknown reference to test his meters is bad advice.

TomG.

[shapirus]

If you need $2700 of traceability to figure out which instrument is off by 20 mV or so, then there's something wrong with the way you do engineering.

The wonders of metrology. Yes you actually do need to spend a lot of money, if you want to have a definite, expressed in numbers, and certified amount of certainty uncertainty in the readings displayed by your test equipment.

But once you can be satisfied by qualitative expressions "most likely" and "good enough", or have a friend or someone who offers to use their calibrated (officially or unofficially) equipment out of their good will for you, then it can be done much, much cheaper.

[shabaz]

Regarding:
"Yes you actually do need to spend a lot of money, if you want to have a definite, expressed in numbers, and certified amount of certainty in the readings displayed by your test equipment"

No one is arguing with that - although with the slight caveat that some would say it's not a lot of money to pay to get a 5-digit DMM calibrated. Everyone has a different perspective on pricing, of course, and it would vary by region.

[shabaz]

"3) You cannot verify with any certainty, 20mV which meter is off using a reference from unknown sources without proof of calibration."

That's demonstrably untrue. Modern silicon comes factory-trimmed to a greater precision. You could pick up a $5 voltage reference (half the cost of the AliExpress one!) and with little care (a couple of capacitors) end up with a result accurate to within single-digit millivolts.

So yes, it is certainly possible to identify which of two multimeters (or both, or any number of them) are faulty, if their discrepancy is 20 mV.

[CalibrationGuy]

You quote me but don't understand my words. I said unknown sources. Do you know what factory seconds are? How about counterfeit goods that pervade the sources you mentioned.

Where in anything I said did you get the idea that I meant you cannot get an idea of approximate voltage using parts from a reputable manufacturer?

In fact, I went out of my way to mention a device that is reasonably priced yet has remarkable performance.

I mentioned my equipment and cost of calibration to emphasize that the result I posted was in fact true, NOT that the OP should waste his money for no reason.

Go take your distortions elsewhere.

TomG.

[shapirus]

"3) You cannot verify with any certainty, 20mV which meter is off using a reference from unknown sources without proof of calibration."

That's demonstrably untrue. Modern silicon comes factory-trimmed to a greater precision. You could pick up a $5 voltage reference (half the cost of the AliExpress one!) and with little care (a couple of capacitors) end up with a result accurate to within single-digit millivolts.

So yes, it is certainly possible to identify which of two multimeters (or both, or any number of them) are faulty, if their discrepancy is 20 mV.

If, and only if, the reference has been measured with a trusted voltmeter!

Otherwise, you have three devices, DMM 1, DMM 2, and voltage reference, showing the same or different readings, and you have no way of telling which reading is true, because none of the devices have been verified against a known-good reference.

That's exactly the problem with the Chinese references. They aren't that bad in terms of design and parts. But how do we know whether their output voltages match what's written on the sticker, and if it does, then with what exactly degree of uncertainty?

Oh and regarding the modern silicon... What did you have in mind? If we take, for example, LM399, then its initial tolerance is specified at 2%. That's 0.139V at its nominal 6.95V output. Definitely not good for checking DMMs out of the box without measuring and adjusting. What's the highest precision single-chip reference? Is there anything that has at least a 0.05% tolerance?

[J-R]

Recommending that someone buy the cheap/unknown references to check their equipment is an indefensible position.  You will go down with that ship...  It's been well established here on the forums that those cannot be trusted and that is essentially their only job.  Someone who is new to this might try to calibrate their TE or apply attributes to the reference that it cannot support.  So it's more dangerous to have one than not.  Fine, if you're bored and want to play around with them go ahead, but don't suggest anything further.

[shabaz]

Oh, so a $5 reference chip is totally fine?
Now you agree!

There's every chance that there could be a fault with a reference from a reputable manufacturer too.
If it was drifty, or in fact had an error greater than 20 mV, you'd easily be able to detect it.

You literally created an argument that wasn't there. You dreamt it up. No-one stated that a $10 board should be relied on (or a $5 one for that matter) for trusting your multimeter is performing to spec. That would be an idiotic thing to do.

It was simply to determine if a multimeter had a significant discrepancy or not, so that you'd know which one to send for calibration.

[CalibrationGuy]

"3) You cannot verify with any certainty, 20mV which meter is off using a reference from unknown sources without proof of calibration."

That's demonstrably untrue. Modern silicon comes factory-trimmed to a greater precision. You could pick up a $5 voltage reference (half the cost of the AliExpress one!) and with little care (a couple of capacitors) end up with a result accurate to within single-digit millivolts.

So yes, it is certainly possible to identify which of two multimeters (or both, or any number of them) are faulty, if their discrepancy is 20 mV.

If, and only if, the reference has been measured with a trusted voltmeter!

Otherwise, you have three devices, DMM 1, DMM 2, and voltage reference, showing the same or different readings, and you have no way of telling which reading is true, because none of the devices have been verified against a known-good reference.

That's exactly the problem with the Chinese references. They aren't that bad in terms of design and parts. But how do we know whether their output voltages match what's written on the sticker, and if it does, then with what exactly degree of uncertainty?

That's exactly my point. There were numerous discussions about certain suppliers who were writing random numbers on the included "calibration" sheets. You need to start with a known good reference. It doesn't have to be as accurate as ours, that wasn't the point, the point was that it has to be of known values which meets your needs, tolerance wise.

TomG.

[shabaz]

"but don't suggest anything further."

I don't see that I suggested anything dangerous. I caveated each time that one cannot rely on it as a calibration.

[shabaz]

LM399 has low drift hence the price, but there are other ICs with higher initial accuracy (higher drift). REF50xx could be an option (sub-$5) that would easily determine if a meter had a 20mV discrepancy (at the spot(nominal) voltage of the chip).

[shapirus]

So it's more dangerous to have one than not. Fine, if you're bored and want to play around with them go ahead, but don't suggest anything further.

As usual, they may be useful if you know what you're doing and understand the limitations.

For example, I bought one of those Chinese LM399 boards and measured it with my then freshly purchased Brymen BM869s and used that measurement as a reference point. They still stay in agreement today, after 3 years or so, which may mean one of: a) neither of them have drifted; b) they both have drifted by the same amount. By comparing this to my other meters I conclude that the former is much more likely, and that's more than good enough for my hobby. Of course I'd like to re-measure that LM399 thingy with a more precise meter, but I don't have access to any.

The important point here is that I had a known-good meter to compare that reference against.

In the case of two meters, neither of which is known-good, such a board wouldn't work, especially given that there are reports of bogus values on the stickers that some people received, especially when those stickers are printed. Mine was hand-written, which kinda invites to trust it a little more.

[KungFuJosh]

Here's some simple solutions.

Buy a $20 ref, and pay $100 (plus shipping) to have it calibrated (after hundreds of hours of runtime to confirm stability).

Or buy a $140 ref that comes calibrated and reliable.

Or spend a lot more money, build a ref yourself, and still need to send it out for calibration (after hundreds of hours of runtime to confirm stability).

Or just send both meters out for calibration. Simple. 😉

[Fungus]

No, I'm not missing the point. Without traceability, those numbers mean NOTHING. I have friends who have boards like these and the numbers were the same for every board. Your usability claims have no credibility because there is no verification of said values.

OTOH lots of people have a good one where the numbers are good.

eg. Me. I've measured mine on a few different high-end multimeters and the numbers seem good to at least three decimal places.

Traceability or not, that's very unlikely to be luck.

[shabaz]

"Or spend a lot more money, build a ref yourself, and still need to send it out for calibration (after hundreds of hours of runtime to confirm stability)."

Or, spend $5! and not send it for calibration, nor wait hundreds of hours (a ridiculous suggestion) because it will immediately tell you which of the two multimeters has the 20 mV discrepancy. Then, you know which multimeter might actually be functioning, and you can send it for a cal, at less than the cost of a calibrated reference.

[shapirus]

REF50xx could be an option (sub-$5) that would easily determine if a meter had a 20mV discrepancy (at the spot(nominal) voltage of the chip).

Now, this lengthy discussion has finally produced a second result that may be useful not only to the OP.

To answer this:

Are there any reliable around-the-house ways I can test these multimeters?  Reliable resistors?  Pencil lead?  The nicest resistor I have laying around?

REF50xx is specified (standard grade) at +/- 0.1%, which is, for REF5050 that outputs 5V, +/- 0.005V, which is almost good enough for the initial problem. If we get a high grade type (the "I" suffix), then precision is 0.05% which I think gets rid of the "almost".

This can be considered, I think, an acceptable solution for an in-house basic verification of DMMs to what, say ~10 mV uncertainty? Doesn't sound too bad!

Is there anything better than REF50xx and still reasonably cheap?

[shapirus]

OTOH lots of people have a good one where the numbers are good.

eg. Me. I've measured mine on a few different high-end multimeters and the numbers seem good to at least three decimal places.

Traceability or not, that's very unlikely to be luck.

Yeah but still you don't know if you received a good one or a bad one until you measure it with a known-good meter. That was the point. Once you do such a measurement, whatever its accuracy is, you have a decent reference point (of a long-term stability that remains unclear though) with about that accuracy. But if you don't have a known-good meter to do it, then the reference is useless.

[shabaz]

I've not kept up-to-date on the latest reference chips unfortunately, but I do have a few REF50xx parts here (not kept sealed though, they are a few years old), I could check a few samples of them, if a lashed-up circuit would in practice have that accuracy, I think there's a chance they would, those are excellent chips for ADC references. I could check at 4.5V and possibly 2.048V (I think those are the ones I've got). But there may be better chip options than this at a similar price.

[Fungus]

Yeah but still you don't know if you received a good one or a bad one until you measure it with a known-good meter. That was the point.

Correct, but what's the worst thing that could possibly happen? That you have to return it?