The Siglent SDG2042X Thread

[tautech]

Very rare for these to give problems and if/when they do we have recovery packages for them now.

Here's the FW history:
https://siglentna.com/wp-content/uploads/dlm_uploads/2019/12/SDG2000X-Firmware-Revise-History-Update-Instructions.pdf

[Blue]

I have a SDG2042x liberated to a SDG2122x.

There is a bug regarding the phase between the two channels:
Using a sine wave on both channels @ 120MHz @4Vpp there is a difference between the phase of around 12deg or 3.8ns as shown on my scope (Bandwidth 200MHz).
i used two cables with exact the same length to my scope.

Is this normal for this instrument? Can you check it on your instrument and report back here.

[exe]

I have a SDG2042x liberated to a SDG2122x.

There is a bug regarding the phase between the two channels:
Using a sine wave on both channels @ 120MHz @4Vpp there is a difference between the phase of around 12deg or 3.8ns as shown on my scope (Bandwidth 200MHz).
i used two cables with exact the same length to my scope.

Can you post a picture of your setup? What happens if you move wires around? If you swap inputs, is it still the same lag (checking if there is a problem with oscilloscope)?

PS I wonder if scew can be calibrated. May be the siggen was not calibrated at such high freq.

[Blue]

I swapped the cables to different channels on my scope to no avail.

There exists a calibration procedure using python. I have not done that yet. I'll try to do that if others report better results than I have.

Here is a picture of the setup:
Oh, If you wonder about the cables. I use Lemo 00 the NIM / CAMAC standard.

[tautech]

Please spend some time in the Utility menu and learn how to phase lock channels etc.

[nez]

I think you've got the skew calculation backwards (reverse to be CH 1-2), so the real skew measurement is about 257 ps.

Edit:  Also, I'm not sure if I'm reading the Keysight scope screenshot correctly, but should you change the channels in the skew (Delay) measurement to be both rising or both falling edge?

Regardless of the way of calculating, my unlocked SDG2042 has a similar phase deviation between it's outputs as yours (@120 MHz with channel locking on).

I'll have to look into that phase calibration procedure you mentioned too.

I'm not sure what the spec is (or if one exists) for output phase alignment between channels on the AWG...

[Blue]

Please spend some time in the Utility menu and learn how to phase lock channels etc.

You mean track on in the Channel coupling? Well channels are coupled with track on and zero phasedev.
It still gives the same result. (About the resolution, one can enter 0.1 mdeg and is a bit way too optimistic.)

If I set the phasedev to -12 degegree, the two sine outputs are completely in phase. Note that this is at 120MHz.
When I go down in frequency e.g. to 50MHz, the outputs do not stay lined up. The phase difference is -5 deg @50 MHz.
In other words, phase difference between the two channels is frequency dependant.

I think you've got the skew calculation backwards (reverse to be CH 1-2), so the real skew measurement is about 257 ps.

Edit:  Also, I'm not sure if I'm reading the Keysight scope screenshot correctly, but should you change the channels in the skew (Delay) measurement to be both rising or both falling edge?

You are right, trap for young players I presume 
See figure :

Regardless of the way of calculating, my unlocked SDG2042 has a similar phase deviation between it's outputs as yours (@120 MHz with channel locking on).

Thanks, so at least two instruments have this problem reported. Would be nice if more people test their equipment. Better yet, with way better scopes. Mine is only 200MHz bandwidth with only 30 ps time resolution. (Perhaps I will liberate it into a 500MHz scope in the future. It involves some component changes on the frontend.)

I'll have to look into that phase calibration procedure you mentioned too.

Perhaps Tautech could elaborate on this....
What does it exactly, does it need to be performed after a firmware update etc.

I'm not sure what the spec is (or if one exists) for output phase alignment between channels on the AWG...

I've been looking at the spec sheet. I cannot find anything about this. Perhaps I do not quite know the correct definition of this discrepancy.

[ebastler]

If I set the phasedev to -12 degegree, the two sine outputs are completely in phase. Note that this is at 120MHz.
When I go down in frequency e.g. to 50MHz, the outputs do not stay lined up. The phase difference is -5 deg @50 MHz.
In other words, phase difference between the two channels is frequency dependant.

Sounds like the phase difference is a fixed time (0.28 ns), which of course translates into varying phase angles at the different frequencies. But the generator's channel coupling menu only lets you specify a phase deviation, not an absolute time difference between the channels. So I see that it is a bit inconvenient to correct the deviation on this level.

[Blue]

Sounds like the phase difference is a fixed time (0.28 ns), which of course translates into varying phase angles at the different frequencies. But the generator's channel coupling menu only lets you specify a phase deviation, not an absolute time difference between the channels. So I see that it is a bit inconvenient to correct the deviation on this level.

I think you are right.
Oh, stupid me, how did you calculate that number?

Could this be an option for a improved firmware?

[ebastler]

how did you calculate that number?

1/(12 MHz) * 12°/360° = 1/(5 MHz) * 5°/360° = 2.8 ns.

Oops, seems that I was off by a factor of 10 when doing that in my head...
Thanks for asking!

1/(120 MHz) * 12°/360° = 1/(50 MHz) * 5°/360° = 0.28 ns.

Sorry, I got the frequencies wrong in this post. I need more coffee!

[Blue]

Thanks.

Did you mean?

absolute timing difference at 120MHz : 1/(120MHz) *  12°/360° = 277.8 ps

computes to a phase difference at 50 MHz:  (360 * 2.77.8E-12) / (1/50E6) = 5 deg

Same as I have measured.


I think that it can be solved using a firmware update:

either changing a parameter in the hardware DAC (I have to look into this)
or using a (interpolated) lookup table and setting the phase angle according to the requested frequency.

[ebastler]

Yes indeed. Posting before my second cup of coffee is no good... 

[tautech]

Thanks.

Did you mean?

absolute timing difference at 120MHz : 1/(120MHz) *  12°/360° = 277.8 ps

computes to a phase difference at 50 MHz:  (360 * 2.77.8E-12) / (1/50E6) = 5 deg

Same as I have measured.


I think that it can be solved using a firmware update:

either changing a parameter in the hardware DAC (I have to look into this)
or using a (interpolated) lookup table and setting the phase angle according to the requested frequency.

Only with a lookup table as the phase deviation varies with frequency.
SDG1032X @ 30 MHz and 2 identical channels into SDS5104X = 9^o (1ns) phase difference that is much less at lower frequency.

I thought maybe FreqCoup in the Channel Coupling menu might null the channel phase difference but no only Phase Coupling> PhaseMode>Deviation seems to work as required however haven't played with PhaseMode>Ratio as yet....getting late now. 

P123 Channel Coupling
https://int.siglent.com/upload_file/user/SDG2000X/SDG2000X_UserManual_UM0202X-E02C.pdf

[ebastler]

Only with a lookup table as the phase deviation varies with frequency.
[...]
I thought maybe FreqCoup in the Channel Coupling menu might null the channel phase difference but no only Phase Coupling> PhaseMode>Deviation seems to work as required however haven't played with PhaseMode>Ratio as yet....getting late now. 

Seems to me that it would need a third phase coupling mode, where you specify the deviation as absolute time rather than degrees of phase. (Assuming that the timing offset indeed stays constant across all frequencies.)

[Blue]

Problem solved:

There is a around 278 ps timing difference at 120 MHz measured. When corrected by this amount, the phase difference seen and measured is zero.
This 278 ps influence on the phase difference deminishes with decreasing frequency.

So I created a time delay of 278 ps (approx) by using two additional lemo connectors. The cable length increased by 36mm. It is not ideal since it used up two connectors.

I've checked the phase difference between all the frequecies in the range of 1..120 MHz. No errors are seen.

It would be nice if Siglent came up with a solution. Perhaps a programmable delay line in the asic?

[nctnico]

Only with a lookup table as the phase deviation varies with frequency.
[...]
I thought maybe FreqCoup in the Channel Coupling menu might null the channel phase difference but no only Phase Coupling> PhaseMode>Deviation seems to work as required however haven't played with PhaseMode>Ratio as yet....getting late now. 

Seems to me that it would need a third phase coupling mode, where you specify the deviation as absolute time rather than degrees of phase. (Assuming that the timing offset indeed stays constant across all frequencies.)

'Channel skew adjustment' you mean

[Blue]

'Channel skew adjustment' you mean

Yes, if that's the term.
I did not find a menu item for that.
Perhaps a fixed delay line timing in a calibration menu would be useful. Then one can still use a phase of zero in the main display.

[hpw]

Have the SDG2082X since some days as: FW 2.01.01.35R3B2, HW 02-02-00-40-00

and run into a Square Wave & FM modulation bug....

Settings:

- 1MHz or 10Mhz square wave
- 2Vpp or on 50E 1Vpp
- 50%
- FM 1kHz
- Freq. Dev 1Hz
- FM shape sine or square, equal behavior

After may 1 minutes running, the modulation terminates greyly = dead 

Sine wave is OK, but square wave is the issue. Checked with a SSA as seen on my pictures


Nice would be:

- to have TTL/HCMOS/AHC output level in one snap, than dealing with the output levels & offset, as on pulse too

- also better precision on 0.010 Hz deviation, as none symmetric side peaks or lower  is may a limitation on 16bit DAC, but lower is even possible and hard to check as on SSA

- 25MHz is really a poor mans limitation so higher than 25 MHz please

- use always last settings after boot

[Roger Need]

Just purchased a Siglent 2042x.  It works quite well and I was able to calibrate it to a GPSDO to well under 1 Hz. at 10 MHz using the service menu.  Very impressive AWG for the price.

One thing that is disappointing is the level of fan noise when working near the unit. .  Not an annoying whine but just much louder than my Siglent 1202X-E scope.  Maybe I have a bad fan or that is just the way it is.   

Can anyone else comment on the noise level they have experience especially if you have a 1202X-E to compare it to.  Hopefully Tautech will comment since he seems to know a lot about this product..

Thanks Roger

[tautech]

Just purchased a Siglent 2042x.  It works quite well and I was able to calibrate it to a GPSDO to well under 1 Hz. at 10 MHz using the service menu.  Very impressive AWG for the price.

One thing that is disappointing is the level of fan noise when working near the unit. .  Not an annoying whine but just much louder than my Siglent 1202X-E scope.  Maybe I have a bad fan or that is just the way it is.   

Can anyone else comment on the noise level they have experience especially if you have a 1202X-E to compare it to.  Hopefully Tautech will comment since he seems to know a lot about this product..

Thanks Roger

SDS1202X-E is the quietest fan cooled Siglent product other than those that use smart fans (PSU's and SDL E-load) and SDG products are ~ 2x (my guess) the noise level of SDS1202X-E.
So yes SDG are noisier than SDS1202X-E by some amount and what you hear is normal.

[Blue]

I bought a SDG and it has a very noisy fan.

have a look here:

I installed the Noctua NF-A6x25 FLX  and it has two working fan settings. I used the lowest one and it works great!

Nailig many thanks.

[hpw]

regarding: SDS1202X-E is the quietest fan cooled Siglent product other than those that use smart fans (PSU's and SDL E-load) and SDG products are ~ 2x (my guess) the noise level of SDS1202X-E.

OK, all is now shoe boxed alike and needs cooling! Shaking fan's is not the best for crystal oscillators, so we have free PN or jitter added. may will not see much of them, while performance may not that state of the art, this means enough for home workers

Having a Siglent SDG2082x, SSA 3000X Plus & DSO 2000x Plus on the table and one will but some beans in the ear so you can accept those noise. While world went into fan modding as noise less at it get's. Even with a SSD HD, noise level is reduced..

Also, having a Rigol PSU DP832... fan noise with all to gather, loader than inside a car 

Otherwise Tinnitus will your next friend  but all likes it cheap, factory to collect each saved penny 

[Blue]

Hi HPW,

I tried to do the same with your waveform modulation test. I do not have a spectrum analyser but a handy fft in my scope.
I have no problems with the modulation yet. However, more testing needs to be done.
Perhaps you could show it with a more faster modulation? 1 Hz is quite slow you know and superimposed on a square wave is hard for me to visualise on a scope.
Can you modulate a sine wave properly?

[hpw]

Can you modulate a sine wave properly?

Just read: Sine wave is OK, but square wave is the issue

Cheers

Hp

[tautech]

regarding: SDS1202X-E is the quietest fan cooled Siglent product other than those that use smart fans (PSU's and SDL E-load) and SDG products are ~ 2x (my guess) the noise level of SDS1202X-E.

OK, all is now shoe boxed alike and needs cooling! Shaking fan's is not the best for crystal oscillators, so we have free PN or jitter added. may will not see much of them, while performance may not that state of the art, this means enough for home workers

Having a Siglent SDG2082x, SSA 3000X Plus & DSO 2000x Plus on the table and one will but some beans in the ear so you can accept those noise. While world went into fan modding as noise less at it get's. Even with a SSD HD, noise level is reduced..

Also, having a Rigol PSU DP832... fan noise with all to gather, loader than inside a car 

Otherwise Tinnitus will your next friend  but all likes it cheap, factory to collect each saved penny 

Well you may say that but these instruments require sufficient cooling for stability and long life....a customer needed to quieten his SDS1104X-E and I cut out the grille in my unit.....no change. He fitted a Noctua fan to his unit....no change. We concluded the fans Siglent use are perfectly fit for purpose and the only option to quieten them further was to slow them down....I don't advise this without detailed analysis of the changes to airflow and its effects on instrument thermals and reliability.