Siglent SDM3055 comparison

[mojoe]

Not too long ago, I ordered an SDM3055, because I wanted another bench meter. Other than a few firmware issues (which mostly seem to be fixed now), the meter got decent reviews here. I liked the one I had for the price, so I ordered a second one (at a sale price). Being a nut of various types, I had to compare the two meters.

For an initial test, I hooked both meters up to a MAX6350CPA 5V reference chip. I put this reference together over a year ago and have had it powered on lately for several days. The room temperature right now is 26C. The computer keeps this room a bit warmer than the rest of the house.

Both SDM3055 meters agree to within one LSD. The older meter is reading 05.0000 and the new meter is reading 05.0001. Assuming that both meters aren't off the same amount in the same direction, I would think that the MAX6350 is doing its job and both meters are doing theirs. The spec on the SDM3055 DC 20V range is 0.015% of reading plus 0.004% of scale.

My HP 3457A is currently out for calibration. When it comes back, I will of course see what it reads. Next, I will have to send my EDC VS330 (0-300 VDC) out for calibration, also.

[tautech]

Thanks for your vote of confidence in the SDM3055.
I've linked your comments and this thread in the SDM3055 thread.
https://www.eevblog.com/forum/testgear/siglent-new-bench-dmm-sdm3055/

[mojoe]

The HP 3457A just came back from calibration, so I thought I'd check that MAX6350 with the freshly calibrated 6.5 digit meter. With the 3457A displaying seven digits and NPLC = 100, I read 5.00036 VDC. Room temperature is 25C.

Looking at the 90 day specs on the 3457A, I could read from 4.99999 to 5.00073 (someone check my math).

The 6350 spec says 0.02% initial accuracy (which would be +/- 1mV) and 1ppm/C. It looks like the 6350 is performing much better than spec.

Now, back to the two SDM3055 meters. If I did the math correctly, DMM #1 is 0.0072% low and DMM #2 is 0.0052% low (asumming the 3457A is dead on). Several times better than the 1yr spec, which is all they publish.

[Andreas]

(asumming the 3457A is dead on).

How can you tell that?
You cannot assume that during calibration (=verification against specs) a adjustment of the meter was done.
For example at Keithley they usually adjust the meter only if it is not within 70% of the 1 year specs.

What does the outgoing protocol of your calibration tell for the 30V range?
And what is the uncertainity of the calibrator which was used for calibration?

With best regards

Andreas

[mojoe]

Let me preface this by saying that I am not a metrology expert. I'm just a hobbyist. If you are (or think you are) an expert, take this into account and don't jump down my throat if I don't explain what I am doing in the correct way. My intention is to share what I am doing so that I may learn more and perhaps others may learn from what we all post here. This was just an exercise in comparing the three bench meters that I have to see how closely they agree.

I "assumed" the 3457A as my standard so as not to over complicate my explanation. I'm sure some of the metrology experts would have explained it better. If you take the 30 VDC specs for the 3457A into account, the two Siglent meters still exceed their published spec for the voltage I measured.

As for my 3457A being within spec, I sent it in with all cal constants gone, as I purposedly didn't bother saving them when I changed the battery. In fact, I don't know if there were any valid cal constants in the first place. The battery had previously been changed by the last owner. The cal lab had to do the entire adjustment procedure in the service manual and not just a verification. The tech I spoke to had to spend an extra hour because of this.

Yes, it would be better if I had a 3458A and a Josephson Junction to calibrate it against. Can't afford those, so I use the best that I can afford (I spend too much as it is). The point I was trying to make was that the two Siglent meters seem to be well within spec at the DC voltage I measured. I have not checked other ranges yet, as my EDC VS330 needs to be sent off for calibration next.

And I know that the VS330 has a similar spec to the 3457A. Again, I should have a 3458A to check the VS330. Maybe if I win the lottery. In reality, the 5-1/2 digit meters are more than sufficient for any hobby work I do.

I'm going to buy some 0.01% resistors soon, so I can check my meters against those. And maybe I'll build a current reference, as Dave showed in some earlier videos.

I know that there are some here who have much better standards and a better understanding of metrology. I'd just thought I'd report the results of my testing. If I have made errors, I would welcome feedback, pointing out those errors.

[deadlylover]

I don't think Andreas was having a go at you or anything, it's just very common that meters aren't adjusted when they go for calibration. 

It sounds like your 3457A was adjusted to be right on the money, does the report state the uncertainties of the references? Usually they only tell you the equipment used, it's only when you get an accredited calibration that you get all the uncertainty data. This is all a biiit unnecessary but we are volt nuts after all, I just like seeing uncertainty data, no matter how big or small.

Just throwing it out there, I think my 3457A has never been adjusted and it reads roughly 20ppm high on 30V range, reference is about 4ppm uncertainty(95%). 

[mojoe]

I'll post some data from the report when I get home tonight.

[Andreas]

If you are (or think you are) an expert, take this into account and don't jump down my throat

Hello,

I do not regard myself being a expert.
And it is not my intention to go at your throat.
I am just curious (and asking a lot).
From the first post I did not recognize that the instrument has been adjusted/calibration data was lost.

But also if it has been just adjusted it is no guarantee that the reading after adjustment is perfectly spot on.
E.g. through noise or other effects as you can see in the thread of Dr. Frank from his 34465A.
https://www.eevblog.com/forum/testgear/keysight's-new-34465a-(6-5-digit)-and-34470a-(7-5-digit)-bench-multimeters/msg775998/#msg775998

I also still have to learn as most of us.
E.g. I did not expect that you can adjust a instrument just by changing orientation.
The big question is: where is this specced in the data sheet?

Most LM399 based instruments are affected, some more some less.
The HP34401A has its largest change between placing on the back side and the front side down.
The Keithley on left and right side. (As in the photo).

With best regards

Andreas

[mojoe]

I didn't have time to post it last night, but here is some data from the 5-page report I got back with my meter. Make of it what you will.

[quantumvolt]

(asumming the 3457A is dead on)

That is a rational preliminary assumption. Who would assume that one of the two 5.5 digit meters showing different values (and I assume with no recent calibration) would be more accurate than a newly calibrated 6.5 digit meter.

Of course, even if the calibration included adjustment, the meter might not be spot on for several reasons originating in calibration laboratory, transportation trauma, environmental change, intermittent instrument error or whatever.

That is not the point. The point is that - in lack of more information - your best estimate for an accurate volt reading is the newly calibrated 6.5 digit meter. And you rationally assume - until you gain suspicion, indications, evidence, 'proof' or you-name-it contrary to this assumption - that the meter is spot on.

Assuming, with no reason to do so, that the meter is 4.5 ppm low or 7 microvolt high when you receive it from calibration would imo be somewhat special.