Siglent SDM3065X Issues? RMA Time?

[KG7AMV]

  Sadly I had to RMA a Siglent SDM3065X 6.5 DMM voltage readings were way high and error kept climbing the longer it was powered on a week or so out of the box.

Unit Arrived at Siglent NA July 19th I Called on the 21st to get the plan of action as no response to e-mail i sent. They said they have a lot in the cue! They said about 10 days as for them to recalibrate it or swap the board if it cannot be recalibrated.

Anyone else have issues with the 6.5 DMM I found 1 posts on the forums and 4 other on a discord / fb group with similar issues.

My concern is getting the unit back with no explanation as to why and have a unit that will spend more time in the calibration lab and in shipping then in my lab!

I Recently dropped a few K$ on New Siglent equipment so far so good other then the DMM and the EasySDL software that seems incomplete and that it will uninstall when easywave is installed so you have  to install EasySDL rename EasyDSL dir then install easywave.

[bdunham7]

Just curious--how high?  Could it have been a reference burn-in issue or was it more than that?  I recall another member saying that their SDM3065X exhibited drift that resembled an LM399 initial aging curve (not that they all age identically) as if the references are just installed raw out of the box.

How long have you owned it?  And where did you buy it? If it is brand new out of the box, I'd want an exchange unit, not warranty service.

[tautech]

I Recently dropped a few K$ on New Siglent equipment so far so good other then the DMM and the EasySDL software that seems incomplete and that it will uninstall when easywave is installed so you have  to install EasySDL rename EasyDSL dir then install easywave.

FYI
EasySDL-V8.5 replaced an earlier version that had a conflict with EasyWave and EasyWave has a new and more powerful X version and those combinations of EasySDL-V8.5 and EasyWaveX fixed any conflict issues.

[KG7AMV]

In use for about a week, but past the amazon return window.

It was reading High. As volts went up so did the error.

Siglent PSU Display Output and Fluke 187 and a Generic DMM matched PSU display out. Siglent was +5mV to +15mV or more High!

[bdunham7]

Siglent PSU Display Output and Fluke 187 and a Generic DMM matched PSU display out. Siglent was +5mV to +15mV or more High!

Assuming you were below 20 volts, that's a big error, certainly more than reference drift would cause.  That's more like the error I would see in a meter that has lost its cal data.  Was it like that initially, did it get worse or didn't you notice until all the sudden you realized it was off?

[KG7AMV]

All Ranges had the error.  But only tested to 60V, 20V was about 7mV Hi.

[hpw]

My issues as already reported to this forum:

. time span resets to 60 seconds after reboot

. time span max as 3600 seconds

. dB math back to Volt after reboot

. dB math scale as xx.x format even for 6.5 digit gear

. copy of folder with xx screen captures BMP files to an external USB stick impossible

. using the given SW to capture remote .... 

all gear what you touch coming from the given land, as unfinished and even no deadline when fixed.

Looks like they lay off all engineers while those left open the hacking door.. 


 

[KG7AMV]

Wonder if the Cal setting and other settings are stored in same area?

[KG7AMV]

Tick Tock.. Day 9 not a word.

[tautech]

Tick Tock.. Day 9 not a word.

See #4
https://siglentna.com/service-and-support/warranty-information/

[KG7AMV]

tautech

Yea, I read that.

[hpw]

Tick Tock.. Day 9 not a word.

See #4
https://siglentna.com/service-and-support/warranty-information/

And how it goes with SW bugs?? Also 10 days??  while  with Siglent Engineers since months without any answers 

Even not getting an bug & tracking Nr. so acts as a student company. 

[wolfy007]

When I got my 3065X a few years ago now, it also arrived out of cal resulting in a sort of perturbed indignation which released my inner voltnut in the form of chaotic capital expenditure. Now I have compulsion to be able to self verify and possibly calibrate my own TE. 

In any case, soon after I sent it back under warranty, it was away for about two months including shipping (pre Covid) if I remembered correctly but the same unit came back with an added calibration sticker. I sent a few queries in between, with the same response each time "its away for calibration", but it came back and has behaved ever since.

I do still wonder if they kept it to burn in the reference then sent it off for a calibration?

[KG7AMV]

E-mailed Siglent today asking for status update went unanswered?

[KG7AMV]

Form Siglent:

I had a chance to take a look at your SDM3065X today and I put it on the calibration stand and it passed the DC performance verification. I know you had some measured values that also showed it was out of calibration, 6 - 15 mv on several ranges, so I also tested it with our Keysight B2901A PSU. All of the test equipment in our lab has been calibrated within the last year. I measured at the same values you did, 5, 10, 20, 30, 40, 50, 60 V on both the SDM3065x and our HP3458A 8.5 digit multimeter. I noticed that if I did not add the accuracy of the Keysight B2901A into my calculations the SDM3065X and the HP3458A would seem out of calibration. So I think this is where the problem is. You have to add the accuracy specification from the PSU you are using. This will then give you an accurate measurement.

I am planning on sending your unit back next week.


--------

So they are saying that the 3 meters I compared it to are not accurate and all have the same error?

[Kleinstein]

The response is a little odd: if you use a supply and than compare the reading from 2 meters (e.g.  DUT and 3458), there is usually no need to include the accuracy of the supply. At least this would not the normal accuracy, but only the stability for the time between the 2 readings, reduced by possible repititions.

There could still be a little tricky point. There could be some superimposed noise or ripple, that may effect both meters different. With a meter like the 3458 there is also some possible disturbance from auto zero switching. It is supposed to be not very large, but hard to tell how sensitive the supply reacts to this.

So they kind of confirm the meter is no in cal if the KSB2901A accuracy is not included (which is usually correct !).

[tautech]

Form Siglent:

I had a chance to take a look at your SDM3065X today and I put it on the calibration stand and it passed the DC performance verification. I know you had some measured values that also showed it was out of calibration, 6 - 15 mv on several ranges, so I also tested it with our Keysight B2901A PSU. All of the test equipment in our lab has been calibrated within the last year. I measured at the same values you did, 5, 10, 20, 30, 40, 50, 60 V on both the SDM3065x and our HP3458A 8.5 digit multimeter. I noticed that if I did not add the accuracy of the Keysight B2901A into my calculations the SDM3065X and the HP3458A would seem out of calibration. So I think this is where the problem is. You have to add the accuracy specification from the PSU you are using. This will then give you an accurate measurement.

I am planning on sending your unit back next week.


--------

So they are saying that the 3 meters I compared it to are not accurate and all have the same error?

You should ask for their accuracy verification report showing how it meets spec on P14 of the service manual:
https://siglentna.com/wp-content/uploads/dlm_uploads/2019/05/SDM3065X_ServiceManual_SM06036-E02F.pdf

[bdunham7]

So they are saying that the 3 meters I compared it to are not accurate and all have the same error?

You haven't listed your exact measurements here, but from what you did say, I'll assume you are saying that your SDM3065X read 5.00500V and 60.0150V when your Fluke 187 read 5.0000 and 60.000V, respectively, or something close to that, and that your PSU and 'generic DMM' pretty much agreed with the Fluke 187.  While the PSU and generic DMM (what is it, actually?) agreeing with the 187 probably means something statistically, you probably need something more concrete to go on.  Clearly one of the two precision meters is reading incorrectly or you have made an error.

I agree with Kleinstein that Siglent's response seems a bit murky, especially because they do not specify what their 'calibration stand' is nor can they apparently supply measurements from it.  It seems quite odd that you would take a meter off of a calibrator to check it with a PSU.  However, if they checked it against an HP3458A, that would trump your 187!  I suspect that their results are correct and your meter is actually in spec.

Rereading your posts, I cannot find any explicit mention that the meters were connected in parallel at the same time when you were comparing them.  If you connect the 187 and 3065X in parallel to a PSU, I would expect them to read the same within their specifications.  The interaction that Kleinstein refers to--which I have observed in other meters, even with a calibrator--is not too likely to be an issue in this case because the 187 will not inject any noise and it has extraordinarily good normal mode rejection.  Whether a 3458A would bother the 3065X or vice versa, I don't know.  In any case, you can use one of the logging or graphing functions on the 3065X to see how stable it is.  What would be an issue is if you are connecting them sequentially, then they might load the PSU differently--especially if the 3065X is in high-impedance mode by default.  The 187 (and likely your generic DMM) present a 10M load to the PSU, the 3065X would be virtually infinite in high-impedance mode.  I don't know how your PSU reacts at such small loads, but when you get it back, I would try again with all the meters in parallel and perhaps a load resistor of 1K or so.

[KG7AMV]

I tested them in parallel results were the same.

Power source is a Siglent SPD3303.

I repeated the tests with a friends Keysight 5.5 yesterday against my Fluke and yep the output reading form the PSU was the same as the reading on the keysight my fluke was spot on also my generic was not so good but acceptable.  So can rule out the PSU.

I will request their accuracy verification report.

[bdunham7]

I tested them in parallel results were the same. So can rule out the PSU.

I will request their accuracy verification report.

Then this might get ugly!  The PSU should really be irrelevant, except as a stable, low noise source.  The display is for reference only when it comes to this level of precision.

Their service manual specifies a Fluke 5522A, I'd want to know if that is what they use on their 'calibration stand'.  If not, then I'd ask the same question I asked you--did they test the SDM3065X and HP3458A in parallel, or just observe that both meters were within the tolerance of the PSU without noting whether they agreed with each other.  The lack of clarity there is a bit frustrating.  You might want to poke them and ask them all that before they send it back.

[KG7AMV]

The Fluke and Keysight 5.5 Agreed in parallel when hooked up to the Siglent PSU.

If I cannot place my faith in it then it is worthless on the bench.

[TheDefpom]

They should be testing this on a DC calibrator not a power supply, they are completely different beasts and have very different stability expectations, a calibrator will be much lower noise and have much higher stability, a power supply could drift constantly or contain high frequency noise which may or may not get filtered/captured by the meter being tested...

It sounds to me like they don't have the right gear there to do the testing... personally I have a few DC calibrators and I am just a nerd that fixes stuff, so I don't think they have an excuse for not having the right equipment to do the job.

[tautech]

I don't believe NA can do more than simple accuracy checks as they have partnered with Transcat as their calibration provider.
https://siglentna.com/service-and-support/calibration-certificate/

[TheDefpom]

If they are reliant on external calibration services that is fine, but if they are doing in house performance verifications for returned units they need to be using gear that is suitable for the task to ensure that they are getting the right results.

[bdunham7]

If they are reliant on external calibration services that is fine, but if they are doing in house performance verifications for returned units they need to be using gear that is suitable for the task to ensure that they are getting the right results.

I wouldn't assume that they don't have acceptable equipment, as they mentioned their 'calibration stand', whatever that is, and an HP3458A.  That and a stable PSU should be enough to spot any errors, especially a larger error like the one that the OP is reporting.  Their deal with Transcat may be primarily driven by the burden of establishing an accredited cal lab. 

We will have to wait and see how this turns out as we just don't have enough info, IMO.  The Siglent techs response was lacking enough info to convince me for certain that the process they have in place is valid.  The OP's story doesn't leave a whole lot of room for doubt, but there could be some weird error there somewhere.  If Siglent actually checked a meter with a specific claimed defect and returns it 'no problem found' and it turns out to actually be defective, that would be a huge black eye for them--but I'm not presuming that's the case.  The frustrating part, especially for the OP, is that you or I could look at the meter and verify or refute an error of that magnitude in a few minutes and clearly explain and demonstrate why in a manner that knowledgeable people would understand and agree with.  It is interesting to see the process play out.  5mV error at a 5V reading is in $100 handheld territory as errors go.

[KG7AMV]

I e-mailed Siglent the questions about the accuracy verification report and if the 3065X and The HP3458A agreed if hooked in parallel over the weekend have not received a response.

I have some photos of the 3065X hooked in parallel with my fluke. (I Did not take many but enough for my documentation)

Cold and off For 1 Hour or So Right After Firmware Update. The fluke and SDM3065X Almost Agree.


After an Hour or So. (The Error Kept Drifting Up on the Siglent) even with a shutdown and restart after an hour the error would be small but increase over time.

30V I know the Polarity on the silent is reversed but did the same the other way around.




10V


60V

[bdunham7]

So you're saying the error increases with power on time even after the meter is power cycled warm?  And even a short power cycle will reset or greatly reduce the error?  If the Siglent tech wasn't aware of this, perhaps the meter could pass a performance check cold?  The first photo shows no error, and the second, assuming the PSU drifted down and the Fluke 187 is more or less correct, an error of 500ppm, which is only possible if at least one of the meters is out of spec.  However, unless you subject the Fluke 187 to a similar hours-long test with a presumed good meter, I'm not sure how you can convince anyone that the Siglent is the one that is faulty.

[KG7AMV]

Short power cycle had no change.

Yes, i told siglent this went back 1 firmware revision the error was less. That is how I discovered that when you upgrade the firmware it agrees with the PSU Output and the Fluke.

Even if you shut the meters off for an hour and turned turned them back on the error would be present only time they would agree would be upon a firmware update and just for a few min.

Both meters were on and off the same amount of time.

The output displayed on the PSU agrees with the Fluke when psu is on for 30 or so min will read flat 10.000 on the fluke. 



Generic


I am sure if I had the Siglent in parallel it would have read 10.053 or higher.

[bdunham7]

Both meters were on and off the same amount of time.

The output displayed on the PSU agreed with the Fluke.

So if I was trying to look at the temperature coefficient and power-on stabilization time for a meter, I'd connect two meters and a PSU in parallel, like you have.  Then I'd leave the PSU and the known-good meter on overnight to let them completely stabilize, then turn on the test meter in the morning to see how it behaves.  If you turn them all on at the same time, its hard to know which one is doing what.  Your 187 and PSU are probably right, but that is a hard case to make when both of them are so much less precise (on paper) than the 6.5 digit DMM.  Hopefully Siglent can respond in a way that is convincing one way or another.

only time they would agree would be upon a firmware update

 

[KG7AMV]

We shall wait and see!

[Kleinstein]

If the reading changes with a FW update, this could mean there could be a change in the cal constants, like an additional constant used of possibly a different use of some data. This may effect some units more than others. So there could be FW bug to cause a changed reading that is not obvious on all units. Worst case a change in FW would need a new calibration, but this is usually something they would avoid, if possible at all.

Maybe they try to copy some feature of the KS34470 drift after ACAL. 

The Fluke meter and the supply are not really good to give a definitive answer. The Sigilent service should however be able to check this rather fast, with a suitable grade meter. The error looks so large, if should even show up comparing to a 2nd 3065 meter that is allready proper warmed up.

Which FW versions were used (before / after) ?

[KG7AMV]

Kleinstein,

Current 3.01.01.08R1 Large Error.

Previous 3.01.01.07 About 1/2 the error of firmware above.

Siglent told me a Firmware update should not change calibration.

[KG7AMV]

They should be testing this on a DC calibrator not a power supply, they are completely different beasts and have very different stability expectations, a calibrator will be much lower noise and have much higher stability, a power supply could drift constantly or contain high frequency noise which may or may not get filtered/captured by the meter being tested...

It sounds to me like they don't have the right gear there to do the testing... personally I have a few DC calibrators and I am just a nerd that fixes stuff, so I don't think they have an excuse for not having the right equipment to do the job.

TheDefpom, I think I have seen you on youtube.

When I spoke with Scott @ Siglent 2 days after DMM landed on their dock about 2 weeks ago he said they just got some new equipment but did not elaborate.

So far all the other Siglent stuff I purchased recently is preforming as it should rather impressed better then expectations. I am very happy with the Function Generator and Scope the DC Load I still has to prove itself but so far acceptable.







 

[mawyatt]

Been following this thread as we're considering a SDM3065X as a replacement for an aging HP34401A which needs new display and haven't the time to fix this now (in the middle of a complex project that needs this level of DMM performance).

I can vouch for the SPD3303X-E, we've been using one of these (have 3) for a variable reference. Did some simple tests using a KS34465A, and the SPD3303 (all 3) were very stable with little low frequency cyclic or random noise. An LM399 or LTZ1000, no, but pretty good PS IMO (sans issue with my improper calibration causing CC offset...another story, hopefully with a good ending).

If the OP SPD3303X-E is as stable and low noise as our 3, then using it as a short term reference for comparing the questionable SDM3065X against a known good DMM seems acceptable.

Good luck and hopefully Siglent will get back to you soon with an acceptable solution.

Best,

[KG7AMV]

mawyatt, I got my PSU used and agree it is very stable. I will be adding a 5A SPD1305X next prime/holiday sale.

SDG2042X is awesome the frequency counter on it has come in very handy recently. 

[KG7AMV]

Got a Siglent Reply..

As I mentioned in the first email I put your SDM3065X on the Fluke 5522A Mulit Purpose Calibrator and tested it and it passed all of the DC voltage performance verification. This alone is enough for us to verify your SDM3065X is within specification.

Then knowing you performed your tests without a Fluke 5522A I used our calibrated equipment to run the tests again using the accuracy specs for instrument in the test. All the equipment used here for testing has been calibrated within the last year. Here is my process of testing.

First I calculated the B2901A Accuracy from data sheet (% reading + offset) ±200 V range - ±(0.015 + 50 mV). Then I set the B2901A at 30 V and based on the setpoint accuracy the true ouput range will be between 30.0545 and 29.9455.

Here is my math.

30  * .00015 = .0045; .0045 + 50 mV = 54.5 mV; 30 + .0545 = 30.0545; 30 - .0545 = 29.9455; range is between 30.0545 and 29.9455

Then I took the SDM3065X data sheet on 200 V range specification which reads DC Voltage Accuracy±（% of Reading + % of Range） 0.0050 + 0.0005.

 
Here is my math.

Example measurement was 30 V on 200 V range – taking my B2901A source accuracy range which is between 30.0545 and 29.9455 I added and minused to get the measurement range.

30.0545 * .00005 = .00150272; 30.0545 * .000005 = .00015027; 30.0545 + .00150272 + .00015027 = 30.056153

29.945 * .00005 = .00149725; 29.945 * .000005 = .00014972; 29.945 - .00149725 - .00014972 = 29.943353


The measurement range is between 30.056153 and 29.943353.

The picture I have included shows your SDM3065X measuring the B2901A @ 30 V. It measured 30.0085. Which if you see the measurement range is between 30.056153 and 29.943353 so your SDM3065X is well within specification.

[bdunham7]

Which if you see the measurement range is between 30.056153 and 29.943353 so your SDM3065X is well within specification.

Yikes!  I think someone is lacking a fundamental understanding of how this all works, which is what I was afraid of.  Perhaps the best way to approach this is to reiterate that the meter is accurate at first, but after a warmup it starts to drift higher.  The calibrator should catch the error in that case,  but if they don't want to tie up their calibrator for hours, they can just connect their Keysight SMU, your meter and the HP3458A all in parallel and observe them for an hour or three.

The best part is that the Keysight SMU is probably dead accurate and they are actually sending you a photo that demonstrates the error, but instead of just measuring with the HP3458A, they are usign a calculated uncertainty and concluding that it is OK.  Perhaps they are conflating a Planck/Schrodinger type of uncertainty with the sort of uncertainty that can be eliminated by accurate measurement.   

[Kleinstein]

What I found he 1 year specs for the SDM3065 in the 200 V range are 50 ppm of the value + 5 ppm of the range ( = +-1 mV).
For 30 V input in the 200 V range this would be +-2.5 mV maximum error.

In the 200 V range the B2901 is a good supply, but to no surprise the specs (150 ppm+50 mV) are not good enough to test a 6 digt meter.
With 30 V this are 50 mV + 4.5 mV uncertainty, though the typical error is usually much smaller (e.g. 3 times for a well maintained instrument).
So the test with the supply does not tell that the meter is in spec, it only can not prove the meter is far out. The 8.5 mV are even not yet in the range to call it suspicious from this test.
It would need at least a second better meter.

The Fluke 187 should have specs of 300 ppm+3 mV (50 V range) and thus +- (9+3) mV. So it is slightly tighter specs for 30 V, but still not good enough to really call it out. The observed drift still makes the sigilent meter slightly suspicious, though the specs usually only apply after plenty of warm up. For an bench meter there usually are no specs for use after 1 minute warmup. For a Handheld this may be useful.
Some meters include corrections to improve the performance in the warmup period - this may change with FW version and maybe it got worse in this case.

However the part on the test with a 3458 still makes it a bit odd - so they may have done this test quite wrong in some way. Even than the stabilty of the B2901 should be consideratbly better than the accuracy specs.

The automated performance test may not include a point near 30 V - though it ideally should.

[rf-loop]

Afaik this is from Siglent:

"As I mentioned in the first email I put your SDM3065X on the Fluke 5522A Multi Purpose Calibrator and tested it and it passed all of the DC voltage performance verification. This alone is enough for us to verify your SDM3065X is within specification."

This is main thing. All other measurements and calculus they are more like entertainment stories.


---

Also there is important things on specifications.
SDM3065X Specifications are for 90-minute warm-up and 100NPLC integration time!

Also whole test setup need be fully reached thermal equilibrium including contact interfaces where the material changes at the joint, including coatings. (thermal emf, example one worst case is copper-copper oxide joint... 1mV/celsius. In normal serious cal I have used pure copper wires without any plating and just before tight contact all the oxide layer removed mechanically, but with 6.5digit meter it is not so critical specially when can wait enough for thermal equilibrium between all joints)

They have checked  it using Fluke 5522A, if I understod right, and if with it this SDM3065X meet its specs it is well enough said.
Fluke calibrator specs. All other things are then nonsense and mainly entertainment.
https://download.flukecal.com/pub/literature/3931295f-en-5522A-ext-specs-w.pdf

Measuring SDM3065X accuracy using some old fluke 187 and comparable instruments is more like joke if use this for claim nearly ten times more accurate instrument with fresh cal is out of specs. Or with some friend old Keysight 5.5digit unknown DMM. 
First there need need use very stable low noise cal grade ref source... you do not have it at all here but just used normal PSU.
When there is fluctuation and noise different meters may handle it bit different.
For claim SDM3065X is out of specs you need even some ground under legs... other ways it is waste of time.

Even if you connect many DMM parallel and wait until all are stable and in thermal equilibrium so that also environment temp is enough stable...  even then if do not have cal grade voltage reference it is like an entertaining hobby for spending free time. (because diffeerent meters handle DC fluctuation and noise differently - comparing nearly like apples and shoes or least oranges)

I do not understand at all why Siglent NA did this playing with this B2901A and some calculus just for generate just extra mess more  than give answers specially (perhaps they was trying to do teaching but imho is better to leave for teachers).

If they have just Fluke 5522A and HP3458A and these together tell it is OK then it IS ok if these tools have valid certified cal (it was told they was in cert cal.) and measurements done as need do.

[bdunham7]

This is main thing. All other measurements and calculus they are more like entertainment stories.


First there need need use very stable low noise cal grade ref source... you do not have it at all here but just used normal PSU.
When there is fluctuation and noise different meters may handle it bit different.

With respect for the expertise you have brought us in other areas, you are making some errors of thinking here.

Your basic premise is that you need calibration-grade sources to determine that at meter is out of spec.  This is false.  You need calibration grade sources to properly calibrate the meter, and possibly to verify that it is in spec.  If your Siglent DMM is reading 2X actual values, I can demonstrate that with a Simpson 260 or a Harbor Freight freebie.  In this case, the customer complaint is that it starts out reading correctly and then drifts high.  It may be the case that the calibration procedure may not catch this because we don't know the cause of the drift nor the exact conditions under which it will occur.  It may be, for example, that it only happens when you take a continuous reading of a stable value for an hour straight, in which case the presumably automated calibration procedure won't catch the error, but any idiot with a stable PSU and a decent DMM--not necessarily one rated as well as the DUT--would catch the error the OP complains of.

The issue of the low-noise 'cal-grade' source is a bit trickier.  Calibrators are not necessarily very low noise, in some cases a PSU or SMU might actually be better than a given calibrator.  While it is true that a significant amount of noise might affect the actual calibration process, we don't buy expensive meters just to measure 'low-noise sources' and batteries.  Any reasonable DMM needs a fair amount of normal mode rejection (the 100NPLC is there for that exact reason) and the difference in effect of PSU noise on various multimeters will be fairly low, most certainly much lower than the error reported by the OP.  This is especially true of an SMU like the Keysight model used here.

For claim SDM3065X is out of specs you need even some ground under legs... other ways it is waste of time.

I do not understand at all why Siglent NA did this playing with this B2901A and some calculus just for generate just extra mess more  than give answers specially (perhaps they was trying to do teaching but imho is better to leave for teachers).

If they have just Fluke 5522A and HP3458A and these together tell it is OK then it IS ok if these tools have valid certified cal (it was told they was in cert cal.) and measurements done as need do.

Attitudes like this are why we have return policies and customer protection laws.  I do not know whether the OPs meter is defective or if he has made some usage error that just hasn't come to light yet.  However, Siglent has not addressed his actual stated issue (at least as stated to us after some clarification) and unfortunately, the 'playing' you referred to reveals at least a communication error and at most a gross misunderstanding of the fundamentals of uncertainty.  The correct way to handle this would be to simply take the OP's meter, the Keysight SMU and the HP 3458A, connect them all in parallel, set the SMU to 30.0000 volts and go to lunch for an hour.  If both the HP3458 and OP's meter agree to within the specs of the SDM3065X (the SMU specs are completely irrelevant) then the complaint has not been verified.  If they differ by more than allowed in the SDM3065X specs, the meter is broken.  Not out of cal, broken.  Now if you insist on using 'cal-grade' sources, you could simply substitute the 5522A calibrator and set it manually to 30 volts and then go to lunch with that running--the HP3458A would be optional.

[2N3055]

Your basic premise is that you need calibration-grade sources to determine that at meter is out of spec.  This is false.  You need calibration grade sources to properly calibrate the meter, and possibly to verify that it is in spec.  If your Siglent DMM is reading 2X actual values, I can demonstrate that with a Simpson 260 or a Harbor Freight freebie. 

That would be true for a CALIBRATED Simpson260 or a Harbor Freight freebie. And then it would be only possible to do so if error was larger than what the guardbands are for Simpson.

Answer from Siglent is correct but not written in language easily understandable for those that are not clear on all the concepts.

Let's break it down:

1. Factory specification for meter is after 90 minutes of warmup (1 hour 30 minutes). Anything before that is IRRELEVANT to specification verification. By the way, that is somewhat similar for all long scale meters. I believe someone mentioned than new Keithely meters are bit faster in warmup, less than 45 min to good accuracy. My old Yokogawa takes 2-3 hours to settle (no forced heating), Rigol DM3068 60-90 minutes, depending on temp in a lab.
2. They tested meter with Fluke calibrator and factory calibration fixture and it was verified in spec. That alone does mean it is passing verification.
3. Then they went to try and demonstrate how people can get confused by using equipment that is not good enough to test accuracy. OP tested with bench PSU, and 20+years old Fluke 187 that was never calibrated and a noname cheapo multimeter.  Why would you proclaim brand new 6.5 digit meter is wrong when tested against 20 year old handheld instrument of inferior specification, whose temperature coefficient in 1°C  contributes more than TWICE complete accuracy specification the instrument ??

Let me repeat: 1°C of temperature difference for Fluke contributes to error TWICE as much as full specified error for that range. 1°C of difference and your error in that range is not anymore 0.03% but a full 0.08%.. Few more degrees and all bets are off..

So they went and tried to demonstrate how you CANNOT do that with inferior instruments, but that got misunderstood. They didn't try to demonstrate how you can do performance verification with B2901A, but opposite, how you cannot do that, even with such an expensive SMU, because even THAT is inferior to what is needed to verify 6.5 digit meter.. He was probably confused how OP doesn't understand those basic facts.

All this is just a demonstration what kind of bias exists out there. These people are not amateurs. They know their stuff.

So let's do some math for worst case scenarios..

Quick calc for Fluke  187 (providing it is FULL in spec):

At calibrated temperature                        What can it really be:      
      Accuracy                          Measured                      Min              Max                P-P error
      0,0300%    +   0,003   .   30,04100E+0   .   30,02899E+0   30,05301E+0   24,02460E-3
      0,0300%    +   0,003   .   10,01700E+0   .   10,01099E+0   10,02301E+0   12,01020E-3
      0,0300%    +   0,003   .   59,99000E+0   .   59,96900E+0   60,01100E+0   41,99400E-3
                              
At 1°C+calibrated temperature                              
      Accuracy                         Measured                Min              Max                P-P error   
      0,0800%    +   0,003   .   30,04100E+0   .   30,01397E+0   30,06803E+0   54,06560E-3
      0,0800%    +   0,003   .   10,01700E+0   .   10,00599E+0   10,02801E+0   22,02720E-3
      0,0800%    +   0,003   .   59,99000E+0   .   59,93901E+0   60,04099E+0   101,98400E-3


So WITHOUT even adding SDM3065X specification, what is shown is included in error specification of THE Fluke 187. And by that I mean that nothing is wrong with it, it just how accurate(inaccurate) it is. Which, by the way, is excellent for 60000 count handheld instrument that old.

So OP's findings are just fine. But conclusions were wrong.  Errors made:

1. 90 minutes to warm up.
2. Not calculating what specifications really mean
3. Cognitive bias that 20 years old inferior specification Fluke must be "more right" that Siglent's modern, new 6.5 digit meter...
4. Not being suspicious and going back to it, until all the math is right.

Those are all human failings. Nothing unusual to make mistakes. But they have to be opportunities for learning. And if we put in effort, it can be opportunity for all of us to learn new things.. It makes us better..

[bdunham7]

That would be true for a CALIBRATED Simpson260 or a Harbor Freight freebie. And then it would be only possible to do so if error was larger than what the guardbands are for Simpson.

No, no, no and no.  You, Siglent and rfloop are completely missing the point and all of this talk about calibration is just irrelevant.

The facts as stated by the OP, if all correct, seem to indicate the possibility that the meter is BROKEN.  Not 'out of cal', BROKEN. 

If I have a BROKEN meter that reads a 9 volt battery as 18.0678 volts and I confirm that my many-decades old Simpson reads that same battery as 9-ish volts, I have demonstrated that the meter is BROKEN.  Of course there is always some possibility of a mistake on my part or some freakish mishap, but those aren't calibration issues per se.  In this case, the error is not 2X, but still much larger than the specs of the meter.  Whether it is large enough to be definitively affirmed by the instruments the OP had at hand may be an issue, but that is easily resolved by simply using better instruments.

The fact that the meter has 6.5 digits and some spec or another does not prevent you from concluding that it is wrong using a lesser-spec known good meter.  What you are willing to accept as 'known good' is debatable, but it has nothing whatsoever to do with the relative precision of the devices.  The only other issue is by how much the suspect meter is wrong.

It is possible for an instrument to go through calibration, pass its performance verification and still not meet its specifications.  I've seen it happen.  I'm not concluding anything about this situation yet, but IMO a proper analysis starts with determining whether or not the customers observation is true or not, not calibration.  Calibrating a broken meter solves nothing, doing a calibration check on a broken meter proves nothing. 

Factory specification for meter is after 90 minutes of warmup (1 hour 30 minutes). Anything before that is IRRELEVANT to specification verification.

That's the procedure for calibration or the achievement of maximum accuracy.  If the SDM3065X normally has errors this large for more than a few seconds after power-up, then it is unlike any other 6.5-digit DMM I have seen.  In any case, the OP reported just the opposite--the meter seems accurate at power-on and then drifts away.  A lot.

Let me repeat: 1°C of temperature difference for Fluke contributes to error TWICE as much as full specified error for that range. 1°C of difference and your error in that range is not anymore 0.03% but a full 0.08%.. Few more degrees and all bets are off..

  Where do you get this?  The 187 DC specs are valid from 18C to 28C with no temperature correction.

They didn't try to demonstrate how you can do performance verification with B2901A, but opposite, how you cannot do that, even with such an expensive SMU, because even THAT is inferior to what is needed to verify 6.5 digit meter..

The SMU with the 3458A are plenty to confirm or refute the OP's complaint.  To repeat, I'm not making any conclusion as to the OP's complaint.  Siglent had the tools at hand to make a pretty clear determination one way or the other, but they didn't do it.  They're stated reasoning, if I read it correctly, is that you can't use a precision DMM to accurately measure the output of an SMU/PSU because of the 'uncertainty' in the SMU/PSU.  That is nonsensical.  Many will try to reinterpret what was said into something that makes sense, but reread exactly what he said, including the ending "Which if you see the measurement range is between 30.056153 and 29.943353 so your SDM3065X is well within specification."

[wraper]

The facts as stated by the OP, if all correct, seem to indicate the possibility that the meter is BROKEN.  Not 'out of cal', BROKEN.

but reread exactly what he said, including the ending "Which if you see the measurement range is between 30.056153 and 29.943353 so your SDM3065X is well within specification."

Yep, I had a similar issue with a 2 weeks old Rigol DM3068 about 7 years ago. It had fluctuating readings on AC ranges which technically were barely within calibration spec. After meter was kicked back for 2 times with very long turnaround, with no repair and stupid arguments over email (even though Rigol person had seen fluctuating readings), I became tired with their BS. Took it apart and made that meter really no longer measured AC at all and got it swapped. Surprise, no more fluctuating readings.

[KG7AMV]

I just got off the phone with Dave the GM at Singlet I explained to him whats going on.

They will be sending me a new replacement.

[2N3055]

That would be true for a CALIBRATED Simpson260 or a Harbor Freight freebie. And then it would be only possible to do so if error was larger than what the guardbands are for Simpson.

No, no, no and no.  You, Siglent and rfloop are completely missing the point and all of this talk about calibration is just irrelevant.

The facts as stated by the OP, if all correct, seem to indicate the possibility that the meter is BROKEN.  Not 'out of cal', BROKEN. 

If I have a BROKEN meter that reads a 9 volt battery as 18.0678 volts and I confirm that my many-decades old Simpson reads that same battery as 9-ish volts, I have demonstrated that the meter is BROKEN.  Of course there is always some possibility of a mistake on my part or some freakish mishap, but those aren't calibration issues per se.  In this case, the error is not 2X, but still much larger than the specs of the meter.  Whether it is large enough to be definitively affirmed by the instruments the OP had at hand may be an issue, but that is easily resolved by simply using better instruments.

The fact that the meter has 6.5 digits and some spec or another does not prevent you from concluding that it is wrong using a lesser-spec known good meter.  What you are willing to accept as 'known good' is debatable, but it has nothing whatsoever to do with the relative precision of the devices.  The only other issue is by how much the suspect meter is wrong.

It is possible for an instrument to go through calibration, pass its performance verification and still not meet its specifications.  I've seen it happen.  I'm not concluding anything about this situation yet, but IMO a proper analysis starts with determining whether or not the customers observation is true or not, not calibration.  Calibrating a broken meter solves nothing, doing a calibration check on a broken meter proves nothing. 

Factory specification for meter is after 90 minutes of warmup (1 hour 30 minutes). Anything before that is IRRELEVANT to specification verification.

That's the procedure for calibration or the achievement of maximum accuracy.  If the SDM3065X normally has errors this large for more than a few seconds after power-up, then it is unlike any other 6.5-digit DMM I have seen.  In any case, the OP reported just the opposite--the meter seems accurate at power-on and then drifts away.  A lot.

Let me repeat: 1°C of temperature difference for Fluke contributes to error TWICE as much as full specified error for that range. 1°C of difference and your error in that range is not anymore 0.03% but a full 0.08%.. Few more degrees and all bets are off..

  Where do you get this?  The 187 DC specs are valid from 18C to 28C with no temperature correction.

They didn't try to demonstrate how you can do performance verification with B2901A, but opposite, how you cannot do that, even with such an expensive SMU, because even THAT is inferior to what is needed to verify 6.5 digit meter..

The SMU with the 3458A are plenty to confirm or refute the OP's complaint.  To repeat, I'm not making any conclusion as the OP's complaint.  Siglent had the tools at hand to make a pretty clear determination one way or the other, but they didn't do it.  They're stated reasoning, if I read it correctly, is that you can't use a precision DMM to accurately measure the output of an SMU/PSU because of the 'uncertainty' in the SMU/PSU.  That is nonsensical.  Many will try to reinterpret what was said into something that makes sense, but reread exactly what he said, including the ending "Which if you see the measurement range is between 30.056153 and 29.943353 so your SDM3065X is well within specification."

Broken meter was his claim based on fact that cold it agrees with 20 year old, never calibrated Fluke 187. That is no verification for all the reasons explained. As far as I know it could have been Fluke drifting..

You are correct about temperature coefficient, I did misread,  it is outside 18-28°C, unlike what I understood. I made a mistake there. Corrected calculations are:

   At calibrated temperature                        What can it really be:      
         Accuracy   .          .   .   Measured   .          Min              Max             P-P error
      .   0,0300%    +   0,003   .   30,04100E+0   .   30,02899E+0   30,05301E+0   24,02460E-3
   .   .   0,0300%    +   0,003   .   10,01700E+0   .   10,01099E+0   10,02301E+0   12,01020E-3
   .   .   0,0300%    +   0,003   .   59,99000E+0   .   59,96900E+0   60,01100E+0   41,99400E-3
                                 

So my calculations for the tempco were very much off. I apologize for that. 
But still, basic errors for F197 are in the order 12-44mV P-P for measured values. Providing it is perfect. Which there is no reason to believe it has to be.

They also said that SDM3065X and 3458A agreed between themselves while measuring B2901A. Rest is trying to explain how large errors can be if you only use B2901A as reference .. It is written a bit confused..

Anyways, I'm not saying meter was perfect. I cannot know that, short of testing it myself.
But whole thing went about kinda wrong way.

If I buy two meters: Keysight 34465 and SDM3065X, connected them in parallel, and they differ 20 mV while measuring 30V, which one do you send back ?
Even in that case proper answer is you verify both. You might say that Keysight might be less likely to be wrong (years of experience etc..) but that is wrong. If error between those two instruments is so large, there must have been some damage in transport or something like that. Those can happen to any vendor.

But if you take 60000 count meter that is very old and never calibrated, you can maybe start suspecting that your brand new SDM3065X is not doing well, but you cannot be sure which one is wrong.
Would you be so eager to say the same it if it was Uni-T 60000 count meter.. I personally have experience those are very accurate, albeit bad mechanical quality so they start falling apart after some time... But as long as they are new they are very accurate.

So yes, OP could have suspect something was not OK with his new SDM3065X, but I would try to compare it at least with another 5.5 digit meter (that one could have been good enough to verify such a large claimed error). Fluke 187 is only good enough to rise suspicion and prompt an additional action.

But good news is that OP will get a replacement one, just to make sure this is solved for him in satisfactory manner. That is all that matters. And that is, after all the effort, good support.

[bdunham7]

They also said that SDM3065X and 3458A agreed between themselves while measuring B2901A... It is written a bit confused..

I didn't actually read that, I think it is just what most people would assume happened because it makes the most sense.  It's what you or I would do.  What the tech appears to have done is noted that both the SDM3065X and the HP3058A separately met their respective combined uncertainties, a calculation that is obviously dominated by the SMU.  If he had simply connected them in parallel and sent a photo of them agreeing after an hour, I would be pretty convinced.

If I buy two meters: Keysight 34465 and SDM3065X, connected them in parallel, and they differ 20 mV while measuring 30V, which one do you send back ?

Of course the OP could have theoretically done more to verify the issue, but he did what he did with what he had at hand and didn't consult all of us ahead of time.  Once he identified an issue, he expected Siglent to investigate and take care of it, all perfectly reasonable.  As for what to do in your example, it just depends on what you have available.  Strange errors like this can actually be quite difficult to nail down.

But if you take 60000 count meter that is very old and never calibrated, you can maybe start suspecting that your brand new SDM3065X is not doing well, but you cannot be sure which one is wrong.
Would you be so eager to say the same it if it was Uni-T 60000 count meter.. I personally have experience those are very accurate, albeit bad mechanical quality so they start falling apart after some time... But as long as they are new they are very accurate.

FWIW, my experience with the Fluke 187/189 types is that there is an 80% or better chance of them being accurate to within an absolute 2 counts, 3 on a bad day, regardless of what the maximum uncertainties in the specs are.  That includes old and very beat ones--they're all old at this point.  They're much more likely to break than go out of cal.  A 20-count discrepancy would concern me, specs or no.  It's true that if the SDM3065X were on trial, that evidence alone would leave some reasonable doubt, but it would also not be something I would dismiss out-of-hand.

But good news is that OP will get a replacement one, just to make sure this is solved for him in satisfactory manner. That is all that matters. And that is, after all the effort, good support.

Well, lets see what the new one does...I hope the OP does the same routine and posts it either way.

[mawyatt]

I just got off the phone with Dave the GM at Singlet I explained to him whats going on.

They will be sending me a new replacement.

This indicates to me Siglent did the right thing for the customer, and what one would expect from them. Sure it took a little effort but in the end they supported the customer, so hats off to Dave and Siglent

Been following this for awhile to see how things turned out. We need another quality DMM for our on-going project, one of the 2 HP34401As displays is fading and needs replacement (will try and replace the display when time permits). We were considering another KS34465A (we have one), but that is overkill for the need and over budget, so please tell Dave that his and Siglent's handing of this issue made the decision for us get the SDM3065X. Just ordered from Saelig in NY, the last one they had

Keep us up to date when the replacement arrives and how it behaves.

Best,   

[KG7AMV]

bdunham7,

I will for sure be doing the same checks with the new replacement when it arrives and will post results.

mawyatt,

Yea, Dave handled it well he said it would ship tomorrow 08/05 or the next day. 

I just checked their site it is showing (1 week lead time) so I will just have to wait and see if thats the case.

[KG7AMV]

Got a package shipment notification from FedEx today. 

[mawyatt]

Good news

Ours is due tomorrow

Best,

[TheDefpom]

Measuring at the LOWER END of a range is not good to see if the entire range is OK, they should be checking at both the bottom end and upper end to check linearity, if it is reading out by 0.0085V at 30V, at 200V it will could be out by 0.0566V.

The calibration procedures for these are done at 0V and +- 200mV, 2V, 20V, 200V, 1000V (the centre and ends of each range).

BUT it is good to see that Siglent are going to be replacing your unit, I am sure they just had to be sure first that there really was a problem.

[KG7AMV]

Good news

Ours is due tomorrow

Best,

Cool, let us know how ya like it.

[mawyatt]

Arrived a coupled hours ago. Got it setup and running now for quick evaluation, in parallel with KS33465A, Agilent 34401A & HP 34401A. Using a 10V Source from a LM399 that been aging for ~year, or a LTZ1000 that's been running for a ~month. Also have some other references, one is based on an LM399 (aged over a year) that give 10, 1 and 0.1V outputs. I'll work on this later, in the middle of some other things at the moment.

Seems the SDM3065X was factory calibrated 2/2020 

So 1 1/2 year old calibration, had asked for the latest available, but Saelig said this was the last one they had.

It's now showing ~80uv above the other 3 DMMs at 10V. One thing I did notice it fired up and displayed ~70uv high initially, then settled into 80uv for the past hour, so little drift from the initial startup.

I haven't even read the manual yet, and don't know if it has a built-in cal capability, so don't take this as any anything other than it appress to be working OK.

Best,

[mawyatt]

Better Image

[kcbrown]

Better Image

You definitely don't have enough test gear.   

[tautech]

Seems the SDM3065X was factory calibrated 2/2020 

Right when the world ground to snails pace and shipping took so much longer and port clearance too. 

How it ages against your references will be interesting if you can take the trouble to make a few notes following it.
Drop me mail about adjustments if you think you need it.

[mawyatt]

Better Image

You definitely don't have enough test gear.   

Not nearly enough for what I'm doing 

Need to monitor and display 128 analog channels from the device we are developing, which can produce up to +-75V independent outputs with 15 bit precision. It's like a HV 15 bit 128 output AWG, the outputs must be perfectly balanced in that if one phase is +68.657 volts, the next phase must be -68.657 volts within a couple millivolts as the output waveform bi-phase toggles back and forth. The slew rates must also match to this level. The waveforms are created from high resolution DACs, then bi-phase modulated and amplified up to the +-75V levels.

Needed variable supplies that's precise and stable,  and with low noise voltages outputs with precise low level current limits. Selected 3 of SPD3303X. These produce +-7V, +-60V, +5V, +3.3V, another supply provides +5.5V and two precison voltage reference sources provide 2.500V and 0~10V variable. If we need the supplies for the higher ranges then a couple switch mode 120V supplies are used, but these produce some switching noise which masks getting very accurate measurements, so only used when the highest output voltage are required. The SPD3303X are very quiet and stable supplies, really like them

A custom low noise SMPS design is underway to supply all the required voltages. Eventually this may lead to a custom chip design.

Thinking about displaying 16 channels simultaneously, could use ADC and such, but would need 16 input stages (they need to be very low capacitance, like one gets with a quality scope probe), high impedance, variable voltage ranges, and such. 8 Channels scopes seem expensive, so maybe using 4 4 channel DSOs synched up, but that's down the road some, no time for this now.

Anyway, the SDM3065X seems stable at 80uv high on 10V compared to the other DMMs.

SD is showing 2uv for 35K samples

Now back to design work

Best,

[mawyatt]

Seems the SDM3065X was factory calibrated 2/2020 

Right when the world ground to snails pace and shipping took so much longer and port clearance too. 

How it ages against your references will be interesting if you can take the trouble to make a few notes following it.
Drop me mail about adjustments if you think you need it.

I suspected so, lucky we got the last one Saelig had in stock

The parts supply shortage is wreaking havoc with the design process. I've had to redesign things so many times, just for the available parts, only to find that they evaporated 

Now we've gone to a mode if we think a component will be useful we buy them right away then start designing around them. If the design seems OK, then design the PCBs

Yeah, probably should start a log on the SDM3065X. Maybe give it a few weeks to settle in, it's rock steady at 10.00008 volts now. The AC is cycling ON/OFF and air blows right over my bench setup, so not exactly a thermally stable environment.

Best,

[2N3055]

Better Image

You definitely don't have enough test gear.   

Not nearly enough for what I'm doing 

Need to monitor and display 128 analog channels from the device we are developing, which can produce up to +-75V independent outputs with 15 bit precision. It's like a HV 15 bit 128 output AWG, the outputs must be perfectly balanced in that if one phase is +68.657 volts, the next phase must be -68.657 volts within a couple millivolts as the output waveform bi-phase toggles back and forth. The slew rates must also match to this level. The waveforms are created from high resolution DACs, then bi-phase modulated and amplified up to the +-75V levels.

Needed variable supplies that's precise and stable,  and with low noise voltages outputs with precise low level current limits. Selected 3 of SPD3303X. These produce +-7V, +-60V, +5V, +3.3V, another supply provides +5.5V and two precison voltage reference sources provide 2.500V and 0~10V variable. If we need the supplies for the higher ranges then a couple switch mode 120V supplies are used, but these produce some switching noise which masks getting very accurate measurements, so only used when the highest output voltage are required. The SPD3303X are very quiet and stable supplies, really like them

A custom low noise SMPS design is underway to supply all the required voltages. Eventually this may lead to a custom chip design.

Thinking about displaying 16 channels simultaneously, could use ADC and such, but would need 16 input stages (they need to be very low capacitance, like one gets with a quality scope probe), high impedance, variable voltage ranges, and such. 8 Channels scopes seem expensive, so maybe using 4 4 channel DSOs synched up, but that's down the road some, no time for this now.

Anyway, the SDM3065X seems stable at 80uv high on 10V compared to the other DMMs.

SD is showing 2uv for 35K samples

Now back to design work

Best,

So with that "long" calibration, it differs 8-9 ppm from Keysights. An has 0.2 ppm RMS noise. That is not bad at all.

LM399 probably didn't move much at all since cal, because they practically don't drift when off.
But I know you'll be keeping the records  so let us know how it moves with time.
Take care.

[mawyatt]

All the references are powered all the time, except during a power outage. So these have a number of "burn-in" hours on them

Still at 80uv high and stable!!!

Best,

[2N3055]

If there will be drift on meter, first 3 months will be the most, than after 6 months it will settle to tempco level variations and hard to keep track of.....

[mawyatt]

If there will be drift on meter, first 3 months will be the most, than after 6 months it will settle to tempco level variations and hard to keep track of.....

Thought you were talking about the LM399 in our references, not the DMM.

My bad


Best,

[KG7AMV]

Replacement Scheduled for delivery tomorrow last Scan was Chicago.

[mawyatt]

We haven't seen much drift since this new SDM3065A arrived last Saturday. Started out with about ~80uV high, last couple days moved up to ~100uV high relative to KS34465A and a pair of Agilent & HP 344041As monitoring a 10 volt reference.

Hopefully your new DMM will be as good or better

Best,

[KG7AMV]

Receive a package today, it seems like they ship old stock.

Per Siglent Site: https://int.siglent.com/account.php?module=warranties

The one I sent in had a Warranty expire date of 01/22/2024


The Replacement has a Warranty Date of 10/28/2023


I know I am splitting hairs here but time is time! Right now will keep it boxed up!

[bdunham7]

Receive a package today, it seems like they ship old stock.

It doesn't surprise me that SiglentNA would have slightly older stock then Amazon.  The have to travel by boat, so those units may have arrived more recently than you think given the state of global shipping lately.  Anyhow, I'm pretty sure they start the warranty from the purchase date and those dates are just there as the default if you don't have the receipt.  You can clarify with SiglentNA, but I'm pretty sure your warranty and cal are OK and you should just unbox it.

[mawyatt]

Wow, you have more will-power than I do

I couldn't wait to get the SDM unpacked, hooked up and running

If you decide to open the box and check, please report the cal date certificate.

BTW, agree with bdunham7, think the warranty starts running from the time of purchase/receipt, and should not be any less than this. Your new SDM should be warrantied for the span of the original SDM I believe.

Best,

[tautech]

Receive a package today, it seems like they ship old stock.

Per Siglent Site: https://int.siglent.com/account.php?module=warranties

The one I sent in had a Warranty expire date of 01/22/2024


The Replacement has a Warranty Date of 10/28/2023


I know I am splitting hairs here but time is time! Right now will keep it boxed up!

Splitting entirely the wrong hairs too !
Siglent's factory warranty conditions DONT apply in the US. Period !
Where instead US warranty is from proven date of sale.
https://siglentna.com/service-and-support/warranty-information/

[KG7AMV]

Right now I am disappointed.. my thoughts have been confirmed!
 
Really want to open it but have to hold off! More info soon!



Maybe Tonight, Tomorrow, Next Week / Month? I will fill ya in just waiting for the conclusion!

 

[wraper]

Right now I am disappointed.. my thoughts have been confirmed!
 
Really want to open it but have to hold off! More info soon!

What thoughts, what confirmed? Is receiving replacement meter which was produced less than 3 months earlier (and that does not affect actual warranty) something worth to fuss about?

Maybe Tonight, Tomorrow, Next Week / Month? I will fill ya in just waiting for the conclusion!

As if it can be something worthwhile.

[KG7AMV]

I have very little trust these days with company's especially with ones as young as singlet and the prior customer service I received. The first after sale service interactions crossed off to many of the never do or say to a customer items fermenting the distrust.   

What I did not mention publicly was that the box had already been opened and resealed! I was Awaiting the Reply from Siglent which I receive this morning.



On the call this morning From David. 

He explained they only have "old stock" they all have expired cal-certs after 6 mo on the shelf and they have to open them and check the calibration on them prior to shipping reason the 1 week lead time. He understood my concerns of an open box (I was worried I receive B stock or another customer return). He told me that all they did is check calibration units was not a return  / refurbished etc. (He understood my concern about the calibration procedure and explanation on the returned one) He wrote me an e-mail reassuring me if I am not happy (even being old stock) just use it and they would replace with an unopened current stock when it arrives.  That was good enough for me! 



Have a Coffee or two!! Grab the blade and carefully open the box. There is an envelope containing a Cal Cert Dated 08/05/21 Scott Rocco



I fire up the PSU and the DC Load and Let it warm up! Plug in the 3065X power it up let it burn in a bit go about my day.

Many Hours Later! 
I hook up the Fluke, DC Load, and 3065X In Parallel and Proceeded to check volts up and down. See if they Agree!!

Side Note: The Siglent PSU is 3.5 years old and was calibrated using the procedures in the SPD3303 service manual with a friends Keysight EDU34450A. I hope I can talk him in to bring it back down with him one day.



The big moment do they agree after a few hours? I checked 1v to 60V. Here is 25V!
 
Fluke and SDM3065X


Siglent DC Load a Little disagreement but acceptable. (Leads are shorter to the DC Load)


So Why is it that this 3065X agree's with the PSU, Fluke, and DC Load and the other one did not? This one seems to be stable also after a few hours will keep it on for (week) 168 hours and recheck.

David is very reasonable person and listens to logic. If you ever need assistance from Siglent talk to David. Scott in my opinion needs a big  lesson in customer service (pink slip type) I have other words I could use but will not express them publicly.

[bdunham7]

So Why is it that this 3065X agree's with the PSU, Fluke, and DC Load and the other one did not? This one seems to be stable also after a few hours will keep it on for (week) 168 hours and recheck.

Obviously the other one was broken.  It would be interesting to figure out what the exact issue was with it, but I guess we'll be deprived of that particular learning experience.  At first glance it might appear to be a tempco issue, but I actually think it is more likely to be an issue with some sort of tempco compensation scheme and related stored values, as Kleinstein suggested earlier.  I think a the undocumented (publicly at least) nature of many modern instruments is a huge drawback.  We aren't likely to ever know exactly how this system works. 

David is very reasonable person and listens to logic. If you ever need assistance from Siglent talk to David. Scott in my opinion need a lesson big in customer service (pink slip type) I have other words I could use but will not express them publicly.

Frustration at this sort of debacle is understandable, but I'm jaded enough to actually prefer an honest but wrong opinion to carefully rehearsed PR bullshit.  The tech in question tried an experiment to try and confirm or refute your assertion of a defect, apparently got the concepts wrong and I suppose things went downhill from there.  Perhaps he felt strongly that he was right--it's not uncommon for support people to think that their customers are idiots because often times they are.  That's why I predicted that things might get ugly--and I'm glad to hear that everything was quickly resolved, although not perhaps at a technical level or for technical reasons. 

And for those that doubted the F187, don't ever bet against ancient Fluke or HP meters, even if they have tire tracks on them.   

[KG7AMV]

Yea, for-sure something was not working right with that other unit.

The mathematical dribble Scott spewed trying to justify the error.

It would be nice to know why Serial # SDM36GBQ4R0577 was that far off and drifted with time. I was going to post the SN earlier then deleted.  I figure the SN should be public just in case it's resold as a factory second / refurbished unit and someone has the same issues with it they may find this post.

Quickly? that is subjective.. It Seemed like forever!

My fluke is a Fricken reliable tank very few issue the last 20 years. It sat in a drawer for a good 12-13years being used off and on while I ventured in my Aviation Career traveling this planet solo never knowing what county I would end up in next!

I plan on grabbing a Fluke 287 soon and pass this 187 down to the kid I's sure it's got another decade of life left in it.

I will put it thru its paces.. I got a few Venkel Engineering sample kits with hundreds of values of caps and resistors. So far this dmm seems faster at resolving the values! 

[wolfy007]

Yea, for-sure something was not working right with that other unit.

The mathematical dribble Scott spewed trying to justify the error.

It would be nice to know why Serial # SDM36GBQ4R0577 was that far off and drifted with time. I was going to post the SN earlier then deleted.  I figure the SN should be public just in case it's resold as a factory second / refurbished unit and someone has the same issues with it they may find this post.

Quickly? that is subjective.. It Seemed like forever!

My fluke is a Fricken reliable tank very few issue the last 20 years. It sat in a drawer for a good 12-13years being used off and on while I ventured in my Aviation Career traveling this planet solo never knowing what county I would end up in next!

I plan on grabbing a Fluke 287 soon and pass this 187 down to the kid I's sure it's got another decade of life left in it.

I will put it thru its paces.. I got a few Venkel Engineering sample kits with hundreds of values of caps and resistors. So far this dmm seems faster at resolving the values!

Glad to hear you got it sorted out, nothing worse when you get new equipment and it doesnt work as described.

BTW, there are many in these forums which very much like the 187/189 meters. I have both the 189 and 287 (both of which I got very good deals on), but the one that usually sits on the bench is the 189 mainly because it starts up quicker. Be interesting to see which you prefer after you start using the 287.

[KG7AMV]

It has powered on over 24 hours stable so far.

I have some Venkel Capacitor and Resistor engineering kits I was testing some values today and it seems to be accurate.



 

[KG7AMV]

48 Hour Update.. Lets Try a Video..

https://youtu.be/RehWxR3P4bE

[MathWizard]

I have a SDM3065X and I'm very happy with it so far. But I've never actually tried to test it, or compare it to anything. I want to make or buy a uCurrent source and reference voltage source like I've seen on the EEVblog channel. But still this is the 1st and only 6.5 D meter I own.