Siglent SSA3000X and SSA3000X-Plus Spectrum Analyzers

[MRF237]

Are you sure it's not a problem of your 10 MHz source? As at 1 GHz you say it is ok...

Thanks for answering! Interesting, your device is above spec as well.
I am quite sure that it's not the source. I tried a few different sources with frequencies between 5 and 50 MHz, the image I posted is from an OCXO as well (from my GPSDO).

[1847123212]

can ssa3015x plus upgrade to sva1032x with software hack and add coaxial wire? or only ssa3021X plus can hack upgrade

[tv84]

can ssa3015x plus upgrade to sva1032x with software hack and add coaxial wire? or only ssa3021X plus can hack upgrade

Only 3021.

[HP4396A]

Hello,
Got me a SIGLENT SSA3015X Plus. The close in phase noise is quite good, about 100 dBc/Hz. And then there is a noise peak at about 70 KHz out, which seems to be how it is designed.
From there I would expect the noise to just decay, but I get a another peak at about 850 KHz offset. Is this how it looks for others too ?

Another thing is the internal 10 MHz reference clock. It has sidebands at 120-150 KHz out. Is this how it looks for others too ?
Is my analyzer faulty ?

I attached some pictures where these artifacts show up.

[RoV]

Got me a SIGLENT SSA3015X Plus. The close in phase noise is quite good, about 100 dBc/Hz. And then...
Is my analyzer faulty ?

You don't mention your source, that may be the true cause of the artifacts.
Also, try with a different source, e.g. any crystal oscillator.

[HP4396A]

Got me a SIGLENT SSA3015X Plus. The close in phase noise is quite good, about 100 dBc/Hz. And then...
Is my analyzer faulty ?

You don't mention your source, that may be the true cause of the artifacts.
Also, try with a different source, e.g. any crystal oscillator.

I get the same behavior with all sources I have used.
The source in the 10 MHz pictures is the 10 MHz internal reference output of the analyzer.
In the 0-1MHz picture there is no source. It just measures the analyzers own internal phase noise.

[szszjdb]

Hi,Sir,

I got a SSA3015X PLUS , and find the phase noise data as -99 dBc@1 GHz,+10 kHz, -98 dBc@1 GHz,+100 kHz, -120dBc@1 GHz,+1MHz in the datasheet. I want to  verify all this on my unit with the R&S SMX SG in hand, but  overload the input when input 0dBm 1GHz signal and enable the preamp and set the reference level to -50dBm. 
I would appreciate it if someone could give some advice on how to measure the phase noise of the SSA3015X with the SMX SG?
Below is the photo from N9020A, FYI. It has 210dB dynamic range, really?
Also a picture from the SSA5000 , it embed the phase noise measure function and I wonder how it does?

Many Thanks & Best Regards,
James

[tautech]

Hi! A few months back I upgraded my SSA 3021X Plus to SVA1032X and unlocked all the options. However I just realized that the top frequency is limited to 2.1GHz rather than 3.2 as I expected. I also tried to factory reset the unit to no avail. Do you have any advice? Thanks!

Anyone? 

Yup, wrong thread.
Try here:
https://www.eevblog.com/forum/testgear/hack-of-sigllent-spectrum-analyzer-ssa3021x/

[szszjdb]

Hi,Sir,

I got a SSA3015X PLUS , and find the phase noise data as -99 dBc@1 GHz,+10 kHz, -98 dBc@1 GHz,+100 kHz, -120dBc@1 GHz,+1MHz in the datasheet. I want to  verify all this on my unit with the R&S SMX SG in hand, but  overload the input when input 0dBm 1GHz signal and enable the preamp and set the reference level to -50dBm. 
I would appreciate it if someone could give some advice on how to measure the phase noise of the SSA3015X with the SMX SG?
Below is the photo from N9020A, FYI. It has 210dB dynamic range, really?
Also a picture from the SSA5000 , it embed the phase noise measure function and I wonder how it does?

Many Thanks & Best Regards,
James

Hi,Sir,
Could someone give some advice on it?

Best Regards,
James

[RoV]

I got a SSA3015X PLUS , and find the phase noise data as -99 dBc@1 GHz,+10 kHz, -98 dBc@1 GHz,+100 kHz, -120dBc@1 GHz,+1MHz in the datasheet. I want to  verify all this on my unit with the R&S SMX SG in hand, but  overload the input when input 0dBm 1GHz signal and enable the preamp and set the reference level to -50dBm. 
James

You don't need preamp and you can leave the default 20 dB attenuator with reference level at 0 dBm.
Set 20 dB/div and a few noise markers. Use an appropriate RBW (e.g. 1 kHz), a matching or lower VBW and something like 100 averages.
If the carrier is at 0 dBm, the readout is dBc/Hz.
This assuming that your SMX is in good conditions and has a much lower phase noise respect to the SSA.

[szszjdb]

I got a SSA3015X PLUS , and find the phase noise data as -99 dBc@1 GHz,+10 kHz, -98 dBc@1 GHz,+100 kHz, -120dBc@1 GHz,+1MHz in the datasheet. I want to  verify all this on my unit with the R&S SMX SG in hand, but  overload the input when input 0dBm 1GHz signal and enable the preamp and set the reference level to -50dBm. 
James

You don't need preamp and you can leave the default 20 dB attenuator with reference level at 0 dBm.
Set 20 dB/div and a few noise markers. Use an appropriate RBW (e.g. 1 kHz), a matching or lower VBW and something like 100 averages.
If the carrier is at 0 dBm, the readout is dBc/Hz.
This assuming that your SMX is in good conditions and has a much lower phase noise respect to the SSA.

Many Thanks!
It works as your advice. I took that set the ATTN to 0dB and enable the preamp would lower down the DANL and the phase noise of the unit.
It is odd that the data vary in different RBW setting ,as -95.7dBc@10K in 10Hz RBW and  -103dBc@10K in 1KHz RBW.
Which might be the more accuracy?

Best Regards,
James

[MRF237]

It works as your advice. I took that set the ATTN to 0dB and enable the preamp would lower down the DANL

This is correct...

and the phase noise of the unit.

This shall not be the case! The phase noise you want to measure is caused by the local oscillators and their PLLs.

It is odd that the data vary in different RBW setting ,as -95.7dBc@10K in 10Hz RBW and  -103dBc@10K in 1KHz RBW.
Which might be the more accuracy?

This difference is suspicious  ! Here's a more detailed instruction...

0. Start with default power on settings.
1. Apply a test signal at an appropriate level (somewhere between -20dBm and +5dBm is fine, 0dBm is easiest, see below).
2. Set the frequency to your test frequency.
3. As RoV said, leave attenuator 'auto' and the preamp off!
4. Set the reference level approx. at the peak level of your test signal.
5. To measure phase noise at 10kHz, set the span to e.g., 50 kHz.
6. I prefer to set the RBW a bit lower, e.g., 100Hz.
7. Press 'Peak'. A marker appears, showing the peak level. Make a note of the level.
8. Press Marker Fn and select Noise Marker.
9. Set the Detector type to 'Sample' (this what I learned, once...) , or to 'Average video' (seems to be averaged sample). This is important, the reading with the Max detector is a few dB too high! Note that the peak might be lower now (or completely disappear), this is normal and does not affect the noise measurement.
10. Press Trace and select 'Average 100'.
11. Turn the knob and position the marker 10kHz to the right (or left) from your carrier and read out the noise level in dBm/Hz.
12. The phase noise is the difference between the peak your measured before and the noise marker value.

Try to change the RBW! On my machine (SSA3032X-R) the phase noise at 10kHz is -98dBc/Hz, and stays within a dB when I change RBW to any value between 30Hz RBW and 1kHz RBW.

Note: if you have software 6.2R5, you can use the 'delta marker' which is more convenient. Step 7a is then: press 'Marker', select 'Delta', turn the knob or type the desired offset. In earlier versions, the delta marker does not work correctly for noise measurements.

[szszjdb]

It works as your advice. I took that set the ATTN to 0dB and enable the preamp would lower down the DANL

This is correct...

and the phase noise of the unit.

This shall not be the case! The phase noise you want to measure is caused by the local oscillators and their PLLs.

It is odd that the data vary in different RBW setting ,as -95.7dBc@10K in 10Hz RBW and  -103dBc@10K in 1KHz RBW.
Which might be the more accuracy?

This difference is suspicious  ! Here's a more detailed instruction...

0. Start with default power on settings.
1. Apply a test signal at an appropriate level (somewhere between -20dBm and +5dBm is fine, 0dBm is easiest, see below).
2. Set the frequency to your test frequency.
3. As RoV said, leave attenuator 'auto' and the preamp off!
4. Set the reference level approx. at the peak level of your test signal.
5. To measure phase noise at 10kHz, set the span to e.g., 50 kHz.
6. I prefer to set the RBW a bit lower, e.g., 100Hz.
7. Press 'Peak'. A marker appears, showing the peak level. Make a note of the level.
8. Press Marker Fn and select Noise Marker.
9. Set the Detector type to 'Sample' (this what I learned, once...) , or to 'Average video' (seems to be averaged sample). This is important, the reading with the Max detector is a few dB too high! Note that the peak might be lower now (or completely disappear), this is normal and does not affect the noise measurement.
10. Press Trace and select 'Average 100'.
11. Turn the knob and position the marker 10kHz to the right (or left) from your carrier and read out the noise level in dBm/Hz.
12. The phase noise is the difference between the peak your measured before and the noise marker value.

Try to change the RBW! On my machine (SSA3032X-R) the phase noise at 10kHz is -98dBc/Hz, and stays within a dB when I change RBW to any value between 30Hz RBW and 1kHz RBW.

Note: if you have software 6.2R5, you can use the 'delta marker' which is more convenient. Step 7a is then: press 'Marker', select 'Delta', turn the knob or type the desired offset. In earlier versions, the delta marker does not work correctly for noise measurements.

Many Thanks!
Following your advice and when narrow down the span , the phase noise data was arround -103dBc,which is within the spec ,and nearlly the same in different RBW (30Hz-3KHz).
Is there any possibility to improve the phase noise performance by some special setting/method or further hardware modification?

Best Regards,
James

[MRF237]

Is there any possibility to improve the phase noise performance by some special setting/method or further hardware modification?

THAT would be a great hack  ! Phase noise is one of the key specs of such equipment, I hope that the engineers at Siglent did their very best to keep it as low as possible !

Phase noise for lower frequencies (say, below a few 100 kHz) mainly depends on the PLL, its reference oscillator, and its loop filter settings.
The PLL loop filter parameters might be somewhere in the calibration files and in theory one could experiment a bit with it...

But, 103dBc is not bad, I think you've to spend a _lot_ more money to find something significantly better. Did you measure it at different RF frequencies? Siglent only specifies it at 1GHz, it might be very different at other frequencies.

[szszjdb]

Is there any possibility to improve the phase noise performance by some special setting/method or further hardware modification?

THAT would be a great hack  ! Phase noise is one of the key specs of such equipment, I hope that the engineers at Siglent did their very best to keep it as low as possible !

Phase noise for lower frequencies (say, below a few 100 kHz) mainly depends on the PLL, its reference oscillator, and its loop filter settings.
The PLL loop filter parameters might be somewhere in the calibration files and in theory one could experiment a bit with it...

But, 103dBc is not bad, I think you've to spend a _lot_ more money to find something significantly better. Did you measure it at different RF frequencies? Siglent only specifies it at 1GHz, it might be very different at other frequencies.

Many Thanks!
The 1K/10K offset phase noise are as below, both with 100Hz RBW and 11Khz span.
@1G 1K=-81.4dBc, 10K=-103.3dBc, @500M 1K=-87dBc, 10K=-104.5dBc, @100M 1K=-98.8dBc, 10K=-106.7dBc, @50M 1K=-102.5dBc, 10K=-106.1dBc, @10M 1K=-99.7dBc, 10K=-105.9dBc, @1M 1K=-98.7dBc, 10K=-105.4dBc.
Yes it is not so bad as my CUM200 got -109dBc @10K. But at 1M , CMU200 is -133dBc vs -113dBc of 3015x. It seems there are something could be improved above the 100KHz offset. ATTached are the photo and the datasheet of the PLL.

Best Regards,
James

[Conrad Hoffman]

I'm loving those filter/coupling structures etched into the circuit board above. Can anybody give more details on exactly how they work?

[MRF237]

ATTached are the photo and the datasheet of the PLL.

Oh, interesting! The phase noise measured is quite close to the numbers as specified in the datasheet. Keep in mind that for higher frequencies (more distance to the carrier) the VCO noise is leading. I guess that the VCO is also constructed using some a Hittite/AD chip? Maybe there's a better one available  ... I always wondered why those RF generators are so expensive (compared to e.g., a SA), maybe low phase noise is one of the key challenges?

[tautech]

I'm loving those filter/coupling structures etched into the circuit board above. Can anybody give more details on exactly how they work?

The Signal Path review/teardown of SVA1032X (and SSA3000X Plus to 3.2 GHz) has a look at these.

https://youtu.be/ToVJTKCyIU8?t=455

[szszjdb]

ATTached are the photo and the datasheet of the PLL.

Oh, interesting! The phase noise measured is quite close to the numbers as specified in the datasheet. Keep in mind that for higher frequencies (more distance to the carrier) the VCO noise is leading. I guess that the VCO is also constructed using some a Hittite/AD chip? Maybe there's a better one available  ... I always wondered why those RF generators are so expensive (compared to e.g., a SA), maybe low phase noise is one of the key challenges?

Hi,Sir,
The VCO is from ZCOMM, attached FYI.  The phase noise between 100Hz-100KHz seems similar with the Keysight or R&S ,but it does not ramp down so much from 100KHz to 1MHz.  Anybody could modify it ?

Many Thanks & Best Regards,
James

[szszjdb]

Hi,Sir,
It looks like the PLL and VCO is quite good and should be ramp down to arround -130dBc. Is there any possible the total phase noise degradation is due to the charge pump design of the 1st PLL ?  Or due to the 12bit ADC, ad9235-40?

Best Regards,
James