DMMCheck Plus Multimeter Reference(and other References) - Experiences..

[KungFuJosh]

Maybe I'll make a test PCB for all the SMD stuff I got for calibration. 🤔

Remember my post ?

Yup. But for me, it doesn't matter. I only care to see if the accuracy drifts over time. And I don't need everything to match for 6 digits; if it's close enough, it's close enough. For you, I'm pretty sure you would want a cal lab to certify values first anyway. 😉

Though I don't think I'm going to limit this to SMT stuff. Especially for the inductors where THT gives a better range with more stability.

[KungFuJosh]

Something like these?

I'm going simpler than that. Similar to the DMMCheck's L/C board, but covering LCR. Test points for each device, and 1 common for use with kelvin leads in each section.

[Martin72]

Yup. But for me, it doesn't matter.

In most cases, you don't have the right test frequency, but what you are measuring is enough if it is always measured that way over time.
OK, then that's the way it is.

[KungFuJosh]

Yup. But for me, it doesn't matter.

In most cases, you don't have the right test frequency, but what you are measuring is enough if it is always measured that way over time.
OK, then that's the way it is.

Exactly. Check it out, I designed my PCB. 😉

[Martin72]

Why not ? 
Could be interesting to measure nH/pf on it.

Another voltage/current reference project:

https://www.elektormagazine.de/labs/voltage-current-calibrator-0-to-10v-and-0-to-40ma-0001

I must confess I was too fast and ordered the PCB by JLCPCB, they´re already on their way.
Too fast because did this before looking at the BOM..
Several ICs are not avaible before 2024 and then they cost..not less.

[KungFuJosh]

Several ICs are not avaible before 2024 and then they cost..not less.

What chips aren't available on the BOM? I haven't signed up yet, but I probably will. That looks like an interesting project.

I'll be ordering extra PCBs of mine too, since it costs less to buy more. 🙄

[Martin72]

Here the BOM of the reference itself and of the (additional) heater...Minimum 2 of the AD... OP-Amps are missed (Farnell)
In principle, you could still "tune" it to 1ppm, but that's another story.
You know there´s a 5pcs minimum order so I could send you a pcb, but I think, the shipping costs will be more expensive as if you order them directly..
Edit: Threadtitle changed..

[KungFuJosh]

I can't look through the whole BOMs right now, what specifically can't yet get? There's very likely plenty of substitutes.

For example, IH0515SH replace with: PEM2-S5-S15-S

[Martin72]

AD8676, AD5791 (expensive), AD8675, the rest I didn´t looked for.
Nevertheless, not a cheap project. When I decide to start with it, I´ll make a new thread for it.
One the cheapest parts would be the pcb...

[KungFuJosh]

AD8676, AD5791 (expensive), AD8675, the rest I didn´t looked for.
Nevertheless, not a cheap project. When I decide to start with it, I´ll make a new thread for it.
One the cheapest parts would be the pcb...

Yikes! The AD5791s in stock are $327 USD. The $85 version is next year, but the $135 version might be as soon as September. Any of those options are pretty expensive for a single chip on a test board.

The other two have SMD versions available for $4 to $7. Start the thread anyway. 😉

 

[Martin72]

OK, you talked me into it....

[DavidKo]

I got mine directly from Doug since 2019. There was a waiting list since it was not available in e-shop. The box I have modeled myself. It includes external power supply connector (the connector disconnect 9V battery when used) and let it run 24/7 from power supply. In case of the power outage it gets turned off, but the main reason is to run 24/7 to burn in the reference. I have also AD584 reference with printed values since 2016, but I have never trusted to the values and  when I had better multimeter it was already running 1+ year.

[DavidKo]

My measurements of DMMCheck and original values from calibration. 5,5 digit multimeter do not see any change .

[DavidKo]

When I measured the AD584 it is clear that something stinks, but I must say, that the values are stable.

[Veteran68]

When I measured the AD584 it is clear that something stinks, but I must say, that the values are stable.

Funny how that same cal sheet makes it around the world. Wonder how many references it's been shipped with? 

[Martin72]

While I will probably repeat my series of measurements....
In the Siglent SDM thread, one criticized that I had contacted the test points on the DMMCheck with probe tips.
Better would be, at least for the voltage / current measurement, to solder shielded wires to the test points and connect the multimeters.
For fun I had done that yesterday and just quickly connected a multimeter - Hmmm...That is actually a difference,
And that's why it will probably be repeated.

[J-R]

I disagree with the idea of soldering anything to the reference.  The first issue would be heat into the PCB from soldering impacting the reference output, and the second would be simply invalidating the current calibration.  So naturally there would be a high probability of a difference in the readings after performing this procedure.

Currently I have 5 calibrated voltage references and two calibrated 6.5 digit DMMs and I have not observed any issues whatsoever from my choice of probe or connection.  The only major factor impacting measurements has been time since power on and ambient temperature.

Only with my 7.5 digit DMM am I able to start to begin to observe some of the effects that you have to pay close attention to in the area of metrology, of which I admit I only lightly dabble.

[KungFuJosh]

I agree, modifying the reference doesn't make sense to me. I would use good quality probes with clips. If possible, I'd use the same probes and clips with every meter you're comparing.

[J-R]

To clarify, even the quality of the probes and clips really doesn't matter at this level.

Consider a 5V reference connected to a 10M Ohm input DMM.  Current flow is 0.000,5mA. Add 10 Ohms to that circuit to represent a very bad connection and the voltage drop is only 0.005mV.  So you would need a 5 million count or better DMM to even see this (typically a 7.5 digit DMM).  If your DMM has a > gig Ohm input option at this voltage, then it truly doesn't matter.

Consider a 1mA current source from the DMMCheck Plus, which in my sample maxes out just above 500mV.  The resistance in the circuit would have to exceed 500 Ohms before it would not be able to generate 1mA.  Common DMM shunt resistances are around 2 Ohms for mA and 100 Ohms for uA.  Again in this case, adding 10 Ohms for a bad connection will not make any difference.

[Martin72]

  The first issue would be heat into the PCB from soldering impacting the reference output

You shouldn't fry it for minutes, that's right.
But I know that too, accordingly short it happened.
We will see, if the values have become worse, it was due to the soldering.

[DavidKo]

Using the probes is not ideal, you can create the Seebeck effect simply by heating one probe more when holding it.

[J-R]

Using the probes is not ideal, you can create the Seebeck effect simply by heating one probe more when holding it.

It's just not really visible on a 6.5 digit DMM @ 5V.  Test leads are well insulated, so the metal tip is not going to be heating up much at all, and if it does, once you start holding it against the terminals the temperature differential will be absolutely miniscule.  Maybe if you have soldering irons for fingers.

For sure the way to easily see this though is to drop into the mV range and heat the probe tips with your fingers directly.  Then the effect is immediately visible.

Otherwise as I mentioned you mostly start to worry about such things at 7.5 digit DMMs and above.  And really one of the top rules at this level is that your hands are in the next room...

[KungFuJosh]

Using the probes is not ideal, you can create the Seebeck effect simply by heating one probe more when holding it.

And really one of the top rules at this level is that your hands are in the next room...

This is why I saw to use probes with clips/grabbers, so your hands aren't anywhere near the test.

[J-R]

I experimented and researched a bit more and will still postulate that holding the probes makes no difference in the context we are discussing.  The reason this is the case is that you are heating up both probes at the same time, so it cancels out the effect.

Furthermore, 5V on most bench DMMs is going to push you up into the next range, so you lose a digit.  So even at 7.5 digits, whether you hold the probes or not really makes no difference.  I was hard pressed to even see the effect at 1V on my Keithley 2010 when holding one probe tip directly.

[KungFuJosh]

I experimented and researched a bit more and will still postulate that holding the probes makes no difference in the context we are discussing.

For the sake of heat, you're probably right. However, relying on hands to hold probes in place consistently gives inconsistent results. Grabbers force contact consistency, and you can run the test 100 times without change to the results based on contact changes.