New Siglent SDM4000A series Multimeter

[tautech]

8/19/2025 0.0.0.20
1. Supports SMB and NTP functions.
2. Optimized noise after AZ is enabled.
3. Optimized noise after 100PLC is enabled.
4. Supports saving readings, statistics, and histograms in the scanner card function.
5. Supports relative value settings, statistics, and histograms in the dual display function.

6. Bug fixes
a) System crashes when entering the wrong password during recalibration.
b) Statistics occasionally show 0 in math.
c) Fixed an issue where overload values were recorded during probe-hold.
d) The digitizer would freeze at certain sampling rates and readings.
e) Device malfunctions occurred when using the scanner card's universal remote control command function.
f) File path errors when browsing when holding readings.
g) Selecting uA in current mode did not work.

I think Siglent is following this thread! 

Did our PD checks on a unit today and when correcting the time clock from Shenzhen to NZ time I noticed the new NTP feature and as I had a LAN connection nearby, enabling auto DHCP and NTP time was just a couple of screen touches and hey presto the correct time was set. 
Too easy.....the NTP server address was already entered at the factory. 

[S13]

So, I have been complaining to Siglent about my SDM4065A DCI range being out of spec on the zero level, see my old post. Initially in that post i said the offset was still in spec, but actually it wasn't. By the way, me complaining to Siglent about this problem was back in november, over 2 months a go... Nice work Siglent support team...   
Ive told them that if they could finally get the user calibration function working on the 4065 i would at least have a solution. Im still baffled why the 4065 firmware is in such a poor shape at this point...

And then i came across a video from Defpom...
The 3065X appears to have nice user calibration function that can selectively be applied to individual ranges and modes. Exactly what i need! So why doesn't the 4065A have this function yet? Siglent, if you are paying attention, this is what we expect for the 4000 series multimeters as well!

[S13]

So what are your zero offsets in DCI mode?

Cant be as poor as mine i hope?

[2N3055]

Enable AutoZero..

[S13]

Enable AutoZero..

Done that, please read my original post explaining the situation.

[Kleinstein]

The picture looks like AZ mode on, but also Rel mode active. It may get better with Rel off or after a new manual zero via the Rel function.

[vsilves]

In that picture both AutoZero and Rel are off.  On my SDM4065 I get about the same reading.  However, turning AutoZero on improves by almost to orders of magnitude.  Separately turning Rel on, by itself, also gets to the last digit shown.

Not sure this bothers me at all.

[S13]

Yeah the picture is with Autozero and Rel off. Its kind of hard to see with the blue colors.

On my SDM4065 I get about the same reading.

So that is interesting, this could indicate the factory calibration process isn't sufficient to meet specifications? It's maybe a structural problem?


I suppose this isn't a big deal for everyone, we all use our equipment in different ways. But to me having a proper zero calibrated DCI mode is important. I like to measure battery powered microcontroller circuits, so having a large dynamic range with the ability to also measure very low currents accurately (sleep currents) and doing it with a fast sample rate is a requirement. So no autozero for noise and speed reasons, and preferably also no Rel mode.   

I mean, isnt it odd that my Brymen multimeter is better zero calibrated than the SDM4065A? In fact i cant remember ever having any multimeter (handheld or bench) that had any significant offset on the DCI range.

Now what everyone should care about is the lack of user calibration with the current state of the firmware. And that is what i like to see changed here. That would solve my problem at least and would help other people in the future as well. And Siglent has proven that they can do it, given the implementation on the 3065X.

I kind of wonder if they are holding back with SDM4065A firmware development until the SDM4075A sees the light of day?

[2N3055]

The picture looks like AZ mode on, but also Rel mode active. It may get better with Rel off or after a new manual zero via the Rel function.

 
ON is blue and OFF is gray, so on that photo it looks reverse.
Both Autozero and Rel are OFF on that photo.

[2N3055]

I mean, isnt it odd that my Brymen multimeter is better zero calibrated than the SDM4065A? In fact i cant remember ever having any multimeter (handheld or bench) that had any significant offset on the DCI range.

It is not. On Brymen you simply do not see those 2 last digits.

[vsilves]

On my Fluke 289 in mA setting shows "0.000 mADC", but also occasionally "0.001 mADC" and sometimes "0.002 mADC".   Whereas in µA setting shows "0.01 µADC" or "0.02 µADC" sometimes. All with leads laying on desk. I don't think much of these as a problem, for my use cases.

[electr_peter]

I mean, isnt it odd that my Brymen multimeter is better zero calibrated than the SDM4065A? In fact i cant remember ever having any multimeter (handheld or bench) that had any significant offset on the DCI range.

It is not. On Brymen you simply do not see those 2 last digits.

Small offset is always present, that is not news. Not seeing last 2 digits is fine.
Some multimeters deliberately show fake zero with low input AC V or AC current (IIRC, Hioki models). Not sure about this Brymen model, but at least compare both DMMs on the same AC/DC setting.

[vsilves]

Got a used Fluke 85RF probe and testing on Siglent SDM4065A to make sure the probe is not blown.  Supposed to measure only 0.25-30Vrms from 100KHz-500MHz, but was able to go down to -5dBm and right above 10KHz fairly accurately.  Detects peaks (diode rectifier) and outputs VDC equivalent to sine Vrms, so will be off for non sinusoidal signals (but still good for a ballpark).

Nice that SDM4065A does dBm math on VDC!  My Fluke 289 only does dBm math on VAC.  Extends SDM to see beyond its 300KHz rated (1 MHz tested) AC bandwidth.

[KungFuJosh]

So what are your zero offsets in DCI mode?

Cant be as poor as mine i hope?

You can't compare any 6.5 digit bench meter to a lesser digit DMM, especially not a handheld.

How are you shorting the inputs? 4-wire short?

[S13]

How are you shorting the inputs? 4-wire short?

For DCI and ACI zero measurement, you leave the connections open without any leads attached, as shown in the picture. This prevents any antenna or inductive loop that could cause unwanted current flow in to the input terminals. Someone else pointed out to me that shorting the input leads is not good practice for measuring the zero point of DCI and ACI.


Im going to quote Siglents specifications here, just to point out the obvious. Its the 90day spec for the DCI 2A range:
0.080+0.005 Accuracy ± (% of Reading + % of Range)

Nowhere in the specification does it say this specification requires Autozero to be turned on, as one would expect.

So doing the math for my previous example: (0.08% * 380u) + (0.005% * 2A) = 100.304uA

So the specification says the accuracy should be within 100.304uA. But I'm measuring 380uA when it should theoretically be exactly zero. So my meter is out of spec by a factor of 3.8
And so if Vsilves says his meter is showing the same offset, then it is clear that this is not an isolated case, but it could be a structural problem.

[Kleinstein]

For accurate measurements one should have the AZ mode on. If one needs to use the non AZ mode one should at least use the REL function before the measurement to get a valid zero.

To some degree the meter chould do at least an internal zero (like in AZ mode) when changing range or function in the non AZ mode. Some meters do this, but the SDM4065 seems to not do this (or has it disabled by default).

Usually the accuracy specs are for the AZ mode - they may just have missed that point here.

Some multimeters deliberately show fake zero with low input AC V or AC current (IIRC, Hioki models). Not sure about this Brymen model, but at least compare both DMMs on the same AC/DC setting.

A few (usually cheap) meters (even more common with balances) indeed fake a zero or have an enlarged zero bin that is split as +0 and -0. In part this was an issue with dual slope ADCs, that have very good DNL, except at 0.

The zero with AC measurements is a different topic. The analog RMS converters just don't work at very low voltage like < 0.1%. There is also the issue with meter internal noise and subtracting that in a correct way. Users would complain of the meter sometimes shows negative RMS readings. To be fully honest they should show the 0 in a different way than normal, like as 00 or 0 0 .

[Martin72]

Hi,

Results from my SDM4065A-SC, 10PLC, auto-zero on, input short-circuited.
All manual ranges, once without, once with rel function.
I don't see any problems with the meter.

Martin

EDIT:

Auto-zero was off, measurements will be repeated.

[Martin72]

Here's the same thing, now with auto-zero....

[S13]

Results from my SDM4065A-SC, 10PLC, auto-zero on, input short-circuited.
All manual ranges, once without, once with rel function.
I don't see any problems with the meter.

Thank you for the detailed measurements. You say you dont see any problems, but your meter is also out of tolerance when not using Autozero. It confirms the measurements i made earlier.

Obviously you can Rel out the zero point, but that is bad practice for current measurements, since you are now just shifting the entire calibration (so basically everything else is measuring wrong now). Rel should only be used when you want to make a true relative measurement, not as a band-aid for a bad calibration.
And sure, Rel is typically used to subtract unwanted probe / lead capacitance or resistance when measuring Capacitance or Resistance, so effectively zeroing out those measurement modes. But in those cases, the calibration isn't affected, you are just making a true relative measurement.

Autozero is not something i like to use, because it has it own drawbacks as mentioned earlier. Even as Kleinstein said, Autozero is a tradeoff; less drift vs more noise and lower speed. So for accurate measurements i'd rather rely on having Autozero switched off. 
Now what is interesting from the Defpom video i posted earlier is that the SDM3065X has a separate calibration for each mode and range with and without Autozero. So there is no reason why the zero point calibration would be out of tolerance when not using Autozero. Take a look at the 4:26 mark in the video:
https://youtu.be/8gZZg605ioI?si=3r1mO64nqXR1L-BL&t=267
Interesting right? Assuming the same applies to the 4065A...

[Martin72]

Hi,

According to the calibration protocol I have for my model, everything is fine.
If the measurements were performed with auto-zero (not specified), this shows even more clearly that there is no problem with using this function.

[Kleinstein]

Using the REL function is not effecting the calibration. It only substracts an offset, that can especially be present in the non AZ mode. So if there is a good way (open input for current) to get a zero, using the rel function is a good way. Especially with old meters it was standard practice to check and maybe adjust the zero before use.

If the meter has so much offset as shown, one should use the rel function (or the AZ mode).

It depends on the meter, function and time frame whether one wants the AZ mode or not. Some meters kind of need AZ mode (e.g. HP34420, 3458), while others can get away without (e.g. Datron 1281). Usually shorter time windows for an experiment prefer non AZ mode, while over a long time AZ mode is usually better.
From the little I know about the front end, I would expect the SDM4065 to be relatively good in non AZ mode, but it would still profit from AZ for longer times (e.g. > 5 min).

The rather high offset in non AZ mode points to a poor calibration / drift since then. Many meters automaticly correct at least what they can internally with a function change. This is kind of a nice feature that the SDM40xx is missing.
If not using the REL function is kind of needed for the non AZ mode.

It is a bit odd that the specs are not more specific about AZ mode and use of REL for the offset.

[S13]

Using the REL function is not effecting the calibration. It only substracts an offset, that can especially be present in the non AZ mode. So if there is a good way (open input for current) to get a zero, using the rel function is a good way. Especially with old meters it was standard practice to check and maybe adjust the zero before use.

I strongly disagree. That offset that you are subtracting when using REL mode is applied everywhere. So all those calibration points that Siglent so delicately measured and programmed are now also out by that same offset. And having all the calibration points out by 380 counts (as in my example) is not up to standards if you ask me. Sure, percentage wise it is better to have that offset at a high level calibration point than at the zero point, but it is still just shifting the problem.
Rel should never be a replacement for a proper calibration.

The rather high offset in non AZ mode points to a poor calibration / drift since then.

Exactly!

According to the calibration protocol I have for my model, everything is fine.
If the measurements were performed with auto-zero (not specified), this shows even more clearly that there is no problem with using this function.

Thank you, this is actually quite insightful. From what i gather from your calibration report and the Defpom video, it appears that Siglent doesn't use a zero level calibration point for their calibration procedure. I'm not sure which calibration points Siglent uses exactly for the DCI ranges, since the defpom video only shows DCV and your calibration report only shows a +40% of range and a -107.5% of range verification (which i doubt are the only calibration points for the a single DCI range, i'd expect at least 4 points like the defpom video shows). But if they are not using a zero level calibration point, then that might explain why the meter could be measuring a wrong zero level, even when the calibration was performed correctly at Siglent. That could imply that there is some kind of non-linearity in the system somewhere...? Or it could be drift?

Well, im curious to see what Siglent's response is. If they could just implement that user calibration function in the 4065A, just like the 3065X has, i would be happy.

[vsilves]

The technical data sheet for SDM4065A states "After 90 min warmup and at 100 PLC integration time" and the shown picture had 10PLC.  I think I matched your 10PLC so will have to try again with no leads and 100PLC when I get home.

With "90 days at TCAL ± 5C" the expectation is Error: 0.080% reading + 0.005% of range = 0.0001 A according to my calculation.  But I think that may not apply to the zero point based on Google Gemini (see below).

Gemini: "Yes, a small offset in the DCI (DC Current) ranges—especially high-current ranges like 2A—is perfectly normal for 6.5-digit multimeters like the Siglent SDM4065A, even immediately after a professional calibration. At the 6.5-digit level, even microscopic temperature differences between the shunt resistor and the internal circuit board create tiny voltages (Thermal Electromotive Forces). The absolute "zero" is the most volatile point on a 6.5-digit meter. It can shift slightly due to changes in ambient temperature (even by 1C), humidity, or the magnetic environment of your lab bench."

[vsilves]

Have not had 90 minutes warm up time to try the 100PLC DCI yet.

Overall, I find the UI of the SDM4065A rather rudimentary, in line with a 0.4 release.  The little use of the rotary knob, and the inability to zoom horizontally bug me slightly.  I did find auto vertical scale, which saves me having to press "rescale once" repeatedly.  However, I am quite happy with the performance and value.  Of course, there is nothing wrong with the SDM4065A.  The only thing wrong would be my expectation that it does everything everybody may ever want or thinks they need.

Ultimately, having SCPI interface like other Siglent devices, allows one to build custom interfaces for custom testing scenarios, and that matters a lot more to me than the actual UI in the nice screen that I find myself squinting on most of the time.

Here is an example of what I mean by custom test interface using a few lines of Python (using Fluke 289 and Fluke 85RF probe).  When I have time, I plan to see how many samples per second this puppy can provide (SDM4065A).

https://youtu.be/s2ykCpWX0KA

[KungFuJosh]

Here is an example of what I mean by custom test interface using a few lines of Python (using Fluke 289 and Fluke 85RF probe). 

What did you use to connect the 289 to your computer?