Siglent SDG1032X mini teardown (hidden self adjust)

[diodak]

A few photos from inside SDG1032X (new model).
I think that the sign "DMMD-150300" could mean that it is a common part of the new multimeter SDM3045X.
Sorry for the quality - made in low light.
Telnet is active (port 23) in my model but I do not know the login and password.

[diodak]

This model has a hidden menu self adjust/calibration:
(Also applies to models of the SDG2000X series) Utility->System->System Info->Enter Password “123654”->SelfCal

[janekivi]

https://www.eevblog.com/forum/testgear/siglent-sdg1000x-waveform-generators/msg1057306/#msg1057306
If you want in... you already in... ok, with telnet then, I can only replace this with known password.

[tautech]

https://www.eevblog.com/forum/testgear/siglent-sdg1000x-waveform-generators/msg1057306/#msg1057306
If you want in... you already in... ok, with telnet then, I can only replace this with known password.

Which suggests this 30 MHz model might be able to be improved to 60 MHz. 

Who's game to try ?

[janekivi]

I don't know anybody...
SDG1000X_V100R001B01D01P22_eevblog.ads

[diodak]

How safe is it to upload the modified file from the link?

[janekivi]

Who knows...
SDG2000X files I can test before by myself. There is ZIP packed shadow file replaced with HexEdit too.
You can compare 2 files side by side. Actually SDG2000X has very similar firmware and some files are
the same so I was using there part of SDG2000_eevblog_P22R5.ads what I did before and what is in
my SDG2042X. You can find those replaced parts viewing attached zip files with HexEdit.

[diodak]

Wow ... after uploading telnet already works What's next?

[janekivi]

Now... we looking at you,
can you do there something interesting like with SDG2000X is done here:
https://www.eevblog.com/forum/testgear/the-siglent-sdg2042x-thread/25/

[diodak]

File "NSP_system_info.xml" has the same syntax as in SDG2042X. There is an interesting file called "NSP_trends_config_info.xml":

Code: [Select]

(...)
<language>english</language>
<product_type_1>SDG1012X</product_type_1>
<product_type_2>SDG1032X</product_type_2>
<product_type_3>SDG1062X</product_type_3>
<manufacturer>Siglent</manufacturer>
<Support_Touch_Flag>0</Support_Touch_Flag>
<Support_VXI11_Flag>1</Support_VXI11_Flag>
(...)Whether it means the ability to activate eg. a touch screen?

[diodak]

There is one thing that is bothering me. Why is the command in SCPI to transmit/set serial number?

[rf-loop]

There is one thing that is bothering me. Why is the command in SCPI to transmit/set serial number?

If you do not like serial number what it transmit, why you do not change it so that it transmit what ever you want instead of serial. But of course when change what ever inside there - after then all is entirely at your own risk, of course.

btw, there is now available FW version 22R5. No one know how it works with all previous "things". 
changelog is extremely tiny and there have been long long time after last update. It is strange because there is several small things what have been waiting improvements, specially some quirks in UI ).

[diodak]

Downloading the new version 22R5 changes the Software Version fields from "1.01.01.22" to "1.01.01.22R5" and Hardware Version from "02-01-00-17-00" to "02-01-00-19-00". Modified login to telnet no longer works ...

[klausES]

This model has a hidden menu self adjust/calibration:
(Also applies to models of the SDG2000X series) Utility->System->System Info->Enter Password “123654”->SelfCal

I picked this up out of interest.

When setting sine 1KHz, 1,000V RMS (rest standard) measured the output AC voltage.
Of course, the DMM was not accurate to the fifth digit (it was 1,00208 V).
Now run self calibration.
When it was over, the voltage had dropped to 1,00114 V.
Let run a second time self cal. Then the voltage dropped to 0,99992, the third time to 0,99876...

Ok, the output voltage output is slightly lower with each self cal attempt. 

Measured again this morning (FG "cold"), the tension was still lower than at the beginning by the amount of the three cuts from yesterday.

A new self cal with now lower device temp (I had to risk it to test, because the deviation annoyed me) achieved a higher output voltage.
It is (coincidence, the right temperature caught?!?) Almost as before the very first self cal attempt.
Even when it is warm, this is true as with the origin.
A little easier ... to know that you can get it back. 

What exactly does the SDG1062X do during self calibration?

[diodak]

In my opinion, this calibration eliminates thermal drift. I guess it is especially about the offset - so that the DC output is correct. For example, the SDS 1104X-E series oscilloscopes have from some version of the FW the option of auto-calibration (fast) based on temperature changes. Here you have to do it manually and can be called via the SCPI command.

[klausES]

In my opinion, this calibration eliminates thermal drift. I guess it is especially about the offset - so that the DC output is correct...

Yes, that was also my guess and what you write absolutely coincides with my further attempts
(now brave since I know that this is not a dead end from which you would never come back ...). 

That is why today I carried out another calibration for testing.
Also one on purpose at almost exactly 23 degrees Celsius (temperature of the moving air in the case).
This calibration gave the tension almost exactly as it did before the very first attempt by self-cal (as from the factory).
If one assumes that the FG was also calibrated in the factory at approx. 23 degrees C, then it is plausible with the influence of the temperature.

You can play with this menu item if you e.g. intends to compensate for drifting values.
If you want to go back to factory values, you can also do this, if you note that this has to be done at almost exactly 23 degrees.

[diodak]

There is one more question - why is the self-calibration menu hidden? In the oscilloscope, this is a specific menu. This option has just been hidden in this generator. Maybe something more than offset is corrected? But in that case it would require some system (inside) that it would measure. ADC is available on the oscilloscope. Maybe the ADC from the microprocessor is used in this case?

That would require further investigation ... Because the RMS has changed after calibration, it means that the gain / amplitude has changed. The offset itself in the RMS AC measurement mode would have no effect ...

[klausES]

Yesterday continued to test and find out more.

In addition to exploring this, it was also my goal to test, it was my wish that the FG should rather be as warm as it is with me after a long time
(more than 23 degrees factory setting, namely for me about 29 degrees) to adjust the AC output voltage and offset.

It looks like you can enter corrective values for the self-cal process.

I came across it by accident and initially thought it was a coincidence.
If it turns out to be true (although tested more than a dozen times with successful results, still amazed),
I don't want to announce it so loudly before it's absolutely safe.

Of course, one can argue about whether it makes sense in an FG case to try to use three or four decimal places.
With all this comes the aspect under which conditions exactly you measure it and how exactly you can measure.
Fortunately, I recorded all the original values and the circumstances beforehand. Otherwise you could be unsure where everything could end up.

[diodak]

A new FW(1.01.01.33R1B6) for SDG1032X has appeared. In addition, there is a new program for creating waveforms and downloading them from the oscilloscope. Interestingly, the maximum number of samples for SDG1032X was set at 8388608 ... it's quite strange.

[klausES]

Have the new firmware since yesterday to test on an SDG1062X.
According to log, it should only fix a change regarding problems with the button bounce (at least I had never noticed a problem about that).

Did you notice a problem with the buttons bouncing before?

I could not find any change in any other way with the new firmware.
Did you find anything else?

[tautech]

Siglent are just rolling a snippet of FW to address this somewhat rare keypad bounce issue.
A ~2013 SDG1025 I had had this issue and someone's SDG2000X model had it too.
What causes it I'm unsure but at least there's a simple SW patch now to fix it. 

I suspect this snippet of FW is now in all factory new SDG's and this FW is so current owners can be protected from this annoying issue too.

[klausES]

The bouncing of buttons increases with age.
Perhaps it is only at an advanced age or after a large number of times that the button is pushed that a threshold is exceeded that the software no longer properly interpreted.

It is possible that this has now been recognized in the older series SDG1025 and that the algorithm has now been improved to prevent all SDGs.

It may well be that I could not notice any of this with my relatively new SDG.

[tautech]

It may well be that I could not notice any of this with my relatively new SDG.

Only in the older SDG1025 is when I have seen it and not in any other model I have had as demos or ones I have sold. No reports from my customers and only the one unit I had.....that I did discuss with Siglent at the time.

Yes it seems age related but why ?
Is it the membrane keypad degrading or some other cause like a noisy PSU or the existing algorithm wasn't sufficient to suppress key bounce ?  What really matters is it's been addressed. 

As mentioned earlier one or maybe two members here have had this problem but now there is FW to fix it should it every become a problem.

[klausES]

..Is it the membrane keypad degrading...

Exactly this.
All buttons, regardless of whether the surface of the contacts in the case of mechanical frogs or a conductive layer of rubber coating,
increase the contact resistance with time and frequency of use, making the signal less clear.
An eternal grief with remote controls and device keyboards as they get older ...

What helps if it gets too bad is first cleaning the contacting surfaces.
(If it is worse, other means need to be taken).

[Ringmodulator]

Hi,

it looks like there is a possibility to solder in a TCXO like on the SDG1025.
This would give the possibility to fine trim the frequency accuracy.

Did someone already try this?

Chris