SSA3021X - big distortions for frequency below 10MHz

[dpenev]

Hello,

I am measuring a sine wave generated by a generator I am developing and
I have observed very big second harmonic distortion around -30dBc for 1MHz around 0dBm level.
Initially I thought that my generator is introducing this distortions
but I have tested with another 3rd party generator (with specification second harmonic distortion < -65dBc @ 0dBm)
and I saw similar -30dBc. See the attached measurement.
At 10 MHz and above the distortions decreases. 

SSA3021X specifies:
Second harmonic distortion: -65 dBc (fc≥50 MHz, mixer level -30dBm, attenuation = 0dB, preamp off, 20 ℃ to 30 ℃)
So they specify frequency above 50MHz only.

At work we have another SSA3021X which shows much better distortion for the same settings as on the attached plot (around -60dBc)

So I am wondering if something is damaged on my analyzer?
Also I have observed that with my analyzer the distortions for the settings (0dBm sine wave, Att 30dB) -30dBc is very different compared to (-30dBm sine wave, Att 0dB) -65dBc
Assuming the distortion is not in the signal itself it is strange, for both settings we should have the same level at the mixer

Any ideas how to investigate more?
Anyone willing to check same measurement on his SSA3021X?

Thanks
Dimitar   
   

[rf-loop]

Before I make some measurements I need ask. (I know why I ask so do not wonder why)

This your image:




What is your input signal true level? In your text there is something about 0dBm signal but... in my eyes it looks like lot of less... so, what is your 1MHz signal true input level.

Do you have any suspects that SSA levels are wrong? Any other measurement with known input signal level.

(I will later show my test (what have lot of more low harmonics), but first please confirm that your true input signal level (in image) in test is 0dBm@50ohm.)

[rf-loop]

Due to time table I can not wait your answer.

Here tests. Input 1MHz sine levels 0dBm (exactly) and -30dBm (exactly). (levels not HP8642B generator displayed output set levels but true levels)

Image names tell...

[dpenev]

Thank you rf-loop for the measurements

It seems my analyzer is seriously sick
Observations:

- Generator set to 50 Ohm, 0dBm, 1 MHz sine wave
- If Generator terminated with 50 Ohm load I measure with oscilloscope 656 mVpp => 0.31dBm
- If Generator is loaded with SSA3021X  I measure with oscilloscope 500 mVpp => -2dBm (the SSA3021X is not 50 Ohm)
- SSA3021X shows 5.46dBm for the main tone (7.46 dB more compared to what it should be)   
- See the attached measurements

 

[rf-loop]

Thank you rf-loop for the measurements

It seems my analyzer is seriously sick
Observations:

- Generator set to 50 Ohm, 0dBm, 1 MHz sine wave
- If Generator terminated with 50 Ohm load I measure with oscilloscope 656 mVpp => 0.31dBm
- If Generator is loaded with SSA3021X  I measure with oscilloscope 500 mVpp => -2dBm (the SSA3021X is not 50 Ohm)
- SSA3021X shows 5.46dBm for the main tone (7.46 dB more compared to what it should be)   
- See the attached measurements

Please do some more cross tests just for check.

You can also try next test if you have available good cables and adapters and so on.
THis is not good test but this can do without more things and equipments.

Connect generator out with cable to T (or if you have calibration grade splitter use it). Then other T out to scope using probe/adapter or coaxial.
Connect other T out to cable what go to SSA input. Then you need accurate terminator what you can connect to cable end when you disconnect this cable end from SSA. (for SWR reasons keep all cables same, only swap SSA input with terminator.)
Now look with scope how much level differ if load is SSA input or load is 50 ohm terminator.
This is best to do using some low frequencies, example 1-10MHz.
This is just for keep all things unchanged, (cables lenght and so on so that only part what change is load. In other case load is terminator and other case load is SSA input (SSA powered on!)

Also if possible compare same signal with other spectrum analyzer.
If there is something like 1-6dB difference my suspect is SSA input failure, perhaps first RF switch.

[dpenev]

Thank you rf-loop,
Confirmed I measure with oscilloscope 656mVpp if 50 Ohm load and 500mVpp if SSA
So the SSA is behaving as 40 Ohm 

I tried to investigate a bit using Dave video.
The input RF switches are marked 955C.
I don't know what this may be. Siglent seems to use Hittite a lot.
So far my best bet is HMC595A but the diodes which they put on pin 4 and pin 6 are not very logical to me.
Anyone having a clue?

If I know what component it is I would rather try to replace the first switch than ship it back and wait for the service.

Hmm this 40 Ohm vary a bit from the attenuation settings.
It may be something related with the input attenuator.
I think I better contact Batronix.       

[rf-loop]

Thank you rf-loop,
Confirmed I measure with oscilloscope 656mVpp if 50 Ohm load and 500mVpp if SSA
So the SSA is behaving as 40 Ohm 

I tried to investigate a bit using Dave video.
The input RF switches are marked 955C.
I don't know what this may be. Siglent seems to use Hittite a lot.
So far my best bet is HMC595A but the diodes which they put on pin 4 and pin 6 are not very logical to me.
Anyone having a clue?

If I know what component it is I would rather try to replace the first switch than ship it back and wait for the service.

In repair work (RF GaAs switch repair) there is also recommended to add diodes D31 - D34 if they are not allready installed (with these zeners and pin diodes it is better ESD/EOS protected but perhaps small very marginal drop in performance in some parameters)

I think best is call Siglent and tell you have suspected front end failure in new unit if it is quite new.
This is reason for warranty service if you have not burned it with over limits.

But then ( if you really want do it yourself and have good soldering tools including (example Metcal MX-500 or equivalent)  and skills to do it. Also 4x - 8x stereo microscope for soldering is good to have)
 
you can ask Siglent if you can try repair it and if in this case Siglent can offer free components.
You also need have available equipments for check it is inside specifications after work. If you look some tear down videos these are not how professionals do works and handle things. In some tear down videos peoples handle these like potatoes and shoes or PC components in car garage. First, this main PCB is very fragile and it need handle with care.

My recommendation No 1. is: Send it to Siglent for repair.

ETA:
Oh: Now I see this:
"Hmm this 40 Ohm vary a bit from the attenuation settings.
It may be something related with the input attenuator.
I think I better contact Batronix.    "

Yes, contact Batronix and tell your SSA is defective.