Siglent Bench DMMs SDM3065X, SDM3055 and SDM3045X

[Kleinstein]

In the picture with the parts removed much of the reference circuit is quite clearly visible.

Pin 2 of the reference is the negative side and should go towards ground. Here the layout details are not visible and there may be an extra path for current compensation (e.g. resistor to -15V) to avoid much of the trace resistance effect. It would however not make much of a difference for the rest of the circuit.
RR5 should be connected to ground at the other side and would divider the 7 V down to 2.5 V. It would not make sense to have the precision divider towards 15 V. Pin5 of U11 is an input and thus provides essentially no current.
There is a path via R57 and D505. With a diode it would be less suitable to provide a stable current. Chances are this is the path for the start up from some 5 V. So it would not provide current in normal operation.
The visible circuit corresponds to the standard circuit around the LM399 and R5A would be the resistor to set the current. The only oddity is having 10 K and not some 5 K for RA5. Mistakes (especially simple ones) sometimes happen in complex designs. Borderline cases are not easy to catch in testing.

[Martin72]

I'd say that's a good reason not to attempt these sorts of mods without a reliable schematic.

You can see that from the fact that the layout on my SDM is different at U11; the pin assignment is not the same as on the 3065 from Korchahin.

[Korchahin]

Installed AD8672 instead of LT072. Returned LM399 back.
Recorded the readings of my standard on the old LM399.
07/16/2025 was 06.90943. Today after assembly and warming up the readings are 06.90945.

After a week of experiments, Siglent SDM3065X showed a deviation in the last digit of 20 μV. In turn, HP34401 demonstrates simply impeccable measurement repeatability. I want to replace RA5 with 5 kOhm (3 kOhm?). But I understand that changing the current can greatly deviate the readings. At the same time, the voltage / current on RR5 can also change. For now, I'm postponing the experiments until it becomes possible to calibrate it. Then I will continue torturing it.
Thank you all very much for your help and instructions!

[Kleinstein]

With a change in R5A there would be quite some change in the voltage readings. With 5 K I would expect something like 0.5 mV more for the 7 V reference voltage and thus connespondingly lower readings. So a new adjustment would be needed for most ranges (the ohms part may just get away without). The change in the current to RR5 would be just the small change in the scale factor.

[tautech]

Installed AD8672 instead of LT072. Returned LM399 back.
Recorded the readings of my standard on the old LM399.
07/16/2025 was 06.90943. Today after assembly and warming up the readings are 06.90945.

After a week of experiments, Siglent SDM3065X showed a deviation in the last digit of 20 μV. In turn, HP34401 demonstrates simply impeccable measurement repeatability. I want to replace RA5 with 5 kOhm (3 kOhm?). But I understand that changing the current can greatly deviate the readings. At the same time, the voltage / current on RR5 can also change. For now, I'm postponing the experiments until it becomes possible to calibrate it. Then I will continue torturing it.
Thank you all very much for your help and instructions!

User Adjustment/Calibration is a standard feature with modern firmware installed IF you have accurate Cal standards.

[Martin72]

After a week of experiments, Siglent SDM3065X showed a deviation in the last digit of 20 μV. In turn, HP34401 demonstrates simply impeccable measurement repeatability.

As a matter of principle, I wouldn't trust the last digit of a multimeter, neither on the Siglent nor on the HP.
Regardless of whether it fluctuates like with siglent or remains “nice and steady” like with HP.
Then the Siglent is set to 10 PLC and 10 G ohms – was this also done on the HP?
Final:
The supplied voltage: is this the “original” value, is the calibrated value known, is the calibration still valid?
Is the measured value still within the tolerance of the SDM3065X?
If so, there is no need to tinker with the SDM.

[hpw]

As I have two SDM3065x models as with AD8622 as SDM360GBD4Rxxxx serials.

Currently I do not have any calibration references, so did input shorted test on both with the settings:

DC: 200mV 100PLC, No Auto Zero, 10G
AC: 200mV 20Hz Filter

The goal was to check how this perform on shorted inputs. In other words to check how low levels could be measured.



Input Shorted      Cold               Hot (1h)
                     
[Device 1]              [mV]                 [mV]
                     
AC                      0.05               0.08               
DC                      0.003                 0.002                  
                     
[Device 2]                   [mV]            [mV]
                     
AC                      0.18               0.19               
DC                     -0.006            -0.01


Conclusions: The AC seams on device [2] as IMHO a bit off. As the DC may OK. How does your device(s) perform?

[tautech]

Would it not be better to Rel the short then check measurement repeatability ?

What did you use for the short ?

[hpw]

Would it not be better to Rel the short then check measurement repeatability ?

What did you use for the short ?

Used given SDM3065x cables. So please drop a line for standardized short cut so all uses the equal setup.. even one floating requires earth connections too or not.


 

[Kleinstein]

The AC reading with a short is tricky, as the analog RMS converter does not really work for near zero readings. For the reason the AC specifications are usually only for > 1% FS or so and below some 0.01% FS the actual readings can very quite a bit between units.

For a slow reading of the small DC voltage one would normally use the AZ mode. So 100 PLC and non AZ mode is an odd combination, that is hardly relevant. The point with the low voltage performance is not so much one reading, but looking at the noise and drift over a few hours.

[tautech]

Would it not be better to Rel the short then check measurement repeatability ?

What did you use for the short ?

Used given SDM3065x cables.

Oh, the 2x 1m antennas ?

So please drop a line for standardized short cut so all uses the equal setup.. even one floating requires earth connections too or not.

No Volt Nut here other than knowing when down in the uV region every little thing matters where buddy Defpom gets out his shielded cables and Banana plug shorts.




The Performance Versus Integration Time chart on P9 in the Service manual is worth a study.
https://int.siglent.com/u_file/document/SDM3065X_ServiceManual_SM06036-E02F.pdf

[hpw]

>> Oh, the 2x 1m antennas ?

YES, as both SDM3065x stacked as using the same antennas and one gear is may more sensitive to the other, how knows

For me as the DC looks as differential and currently do not know whether the AC is differential too. While 200mV get anyway amplified to the internal ADC.

IMHO, it requires any shielded differential connections as to avoid the antennas & GND at the backside to connect.

[TheDefpom]

One thing to consider is when checking with shorted inputs...

The calibration isn't done at 0V on DC, it is done at 1% of full scale, so for 200mV range it is at 2mV, 200V range it is 2V etc, with the values below the cal point being calculated, if the - and + lower cal points are not quite set correctly you will see a jump as it goes between them around the 0V reading depending upon where it calculated the 0V to actually be.

Also the DC has two sets of calibrations, one for negative voltages and one for positive, so if you are sitting at 0V it could easily be toggling between the - and + calibration points if they overlap the 0V very slightly (which can happen), if the calibration was done well enough it should be almost impossible to see where it is, you may only see a couple of counts change.

AC does use a 0V cal point for all ranges, both filtered and unfiltered.

[FordRanger1998]

Hello. I have a SDM3045X that after a long time, I was able to fix it's boot freeze issue. In that thread, I included the full protocol and file link from Siglent. If you want me to provide it directly to you, just ask.   Now my question. I need a replacement bail handle for my meter. Siglent support USA is no help. Can't even provide any links, overseas or anywhere. I'm having trouble finding anything myself. If some kind soul has any insight, I'd really appreciate it. Thanks.

[tautech]

Now my question. I need a replacement bail handle for my meter. Siglent support USA is no help. Can't even provide any links, overseas or anywhere. I'm having trouble finding anything myself. If some kind soul has any insight, I'd really appreciate it. Thanks.

 
The tilting bail is a real spare part and any from SDM and all the cheaper SDG models are interchangeable.

Siglent HQ don't provide us parts lists but we can show the image of it on P41 of the SDG1000X Service manual.
https://int.siglent.com/u_file/download/25_03_07/SDG1000X_ServiceManual_SM0201X-E01C.pdf

Last time I sourced one a few years back from HQ they were so cheap it wasn't worth the paperwork to invoice a valued client.
Keep pestering Ohio until you find someone with an ounce of sense.

[FordRanger1998]

Could you provide a link for Siglent HQ. Again, thanks..

[FordRanger1998]

I just emailed Siglent HQ in China. Hopefully it will lead somewhere.

[tautech]

I just emailed Siglent HQ in China. Hopefully it will lead somewhere.

Hope you told them NA wasn't much use.
Still you'll get referred back to them and hopefully with someones email.

HQ will be on deck in a couple hours and you should get a reply within 8.

[FordRanger1998]

I just received a response from a Siglent engineer in China. He asked for the meters serial number, so I'm crossing my fingers.

[tautech]

I just received a response from a Siglent engineer in China. He asked for the meters serial number, so I'm crossing my fingers.

What ever for ?
Maybe he just needed to check you had one. 

Most of the time they are pretty good but always make you needs are very clear as English is not their 1st language.

[FordRanger1998]

He typed English very well, but I'll make sure. And I'll keep you posted. Thanks.

[FordRanger1998]

I'm still waiting. I also sent him a cropped snip of the handle from the manual PDF you pointed me to. Thanks again.

[FordRanger1998]

After initially not having any luck with Siglent support in the US, I contacted Siglent in both Europe and China. But what happened was that they both sent my requests back to Siglent USA where they found a handle in their "junk" pile. They are sending it to me free of charge. Persistence pays off, sometimes anyway. I appreciate ALL the help I've received from the kind folks here on EEVBlog.!

[slavoy]

Freshly bought DMM6500 (cal'd 2024) vs SDM3065X (cal'd 2022) after warmup. They stay off by 26-27ppm all the time, stable at least.

[KungFuJosh]

Freshly bought DMM6500 (cal'd 2024) vs SDM3065X (cal'd 2022) after warmup. They stay off by 26-27ppm all the time, stable at least.

You can adjust the SDM3065X to match the DMM6500 if you want. The stability is key, and the accuracy adjustment takes care of the rest.