Siglent SVA1015X and SVA1032X 1.5, 3.2GHz Spectrum & Vector Network Analyzers

[vsilves]

These are plots of the .s1p and .s2p files included with the LibreVNA that I received from Amazon US.  The info.txt file includes the serial number, firmware version (arrived with 0.2.3), and number of ports.  The standards downloaded online also include a time stamp and the equipment used to perform the characterization.

---
Serial:  ############
Firmware: 0.3.0
Number of populated ports: 4
---
Timestamp: 2025-04-25T10:11:08.036335

Calibration equipment:
Agilent Technologies,E5071C,MY46524282,B.12.04
---

Looking in GitHub I see the LibreCal described as "4-port, 9kHz to 6GHz eCal module for VNAs".  However, as seen in the plots, the included Touchtone files go all the way to 8.5 GHz, and use 1202 points.

The LibreCal arrived with the nice zipper case, USB-to-USB-C cable, and four 12-inch blue semi-rigid SMAm-to-SMAm coax cables (labeled using a white sleeve "nanovna. com ss405").  When I first saw the coaxial cables I thought they looked like Mini Circuits (because of the white sleeves), but of course not that lucky (haha), and missing the non-rotating nuts in the SMAm connectors.

[james38]

Hello Vicente,

Which program do you use to display the S-parameters?

Regards, Chris

[vsilves]

Which program do you use to display the S-parameters?

It is just a custom python script I made using scikit-rf (skrf) and matplotlib since I'm on a Mac and could not find any fancy software.  If you are not familiar here is a pointer...

https://scikit-rf.readthedocs.io/en/latest/tutorials/Introduction.html

Looks intimidating at first, but there are a lot of tutorials and you only need a few bits and pieces to make nice graphs.

[David]

I appreciate this has been asked before, but is there any chance the SVA1000x firmware can be modified to allow lower IFBW settings? When calibrating through attenuators the S11 response is very noisy making it almost useless.

[guidoted]

Here's an experience of a "stuck boot" with my SVA1032X, which might be helpful to others.

After turning it on and using it for about half an hour in spectrum analyzer mode, I switched it off and then on again after 2-3 minutes because I remembered I wanted to take a few more measurements.
The instrument powered on and seemed to go through the boot sequence normally, but then it always froze on the boot screen.
I tried all possible power-off and power-on combinations, including powering it on while holding down the "System" button, but without any luck. However, what I did notice was that the instrument was accessible via Web, Telnet (on ports 5024 and 5025), and USB.

Using the web interface, I was even able to successfully update the firmware to version 3.2.2.6.3R2.r2 (the installed version was 3.2.2.6.2R5), but there was no way to get the boot process to proceed correctly.

It then occurred to me to try experimenting with SCPI commands, and I verified that on my SVA1032X, I had set "Power On Type" to "Last", so it would always start up with the instrument's last configuration. I tried changing it to "DFT", and miraculously, after a reboot, the instrument started up correctly!

I'm not sure if it was a bug in version 3.2.2.6.2R5 or why it happened. Maybe I'll try setting "Power On Type" back to "Last" again and see what happens...
Ciao
Guido - ik2bcp

[james38]

It's generally advisable to perform a factory reset after every update.
This also prevents any potential problems.

Regards Chris

[KungFuJosh]

It's generally advisable to perform a factory reset after every update.
This also prevents any potential problems.

Regards Chris

Sometimes, 3 resets is better than 1.

[tautech]

Here's an experience of a "stuck boot" with my SVA1032X, which might be helpful to others.

After turning it on and using it for about half an hour in spectrum analyzer mode, I switched it off and then on again after 2-3 minutes because I remembered I wanted to take a few more measurements.
The instrument powered on and seemed to go through the boot sequence normally, but then it always froze on the boot screen.
I tried all possible power-off and power-on combinations, including powering it on while holding down the "System" button, but without any luck. However, what I did notice was that the instrument was accessible via Web, Telnet (on ports 5024 and 5025), and USB.

Using the web interface, I was even able to successfully update the firmware to version 3.2.2.6.3R2.r2 (the installed version was 3.2.2.6.2R5), but there was no way to get the boot process to proceed correctly.

It then occurred to me to try experimenting with SCPI commands, and I verified that on my SVA1032X, I had set "Power On Type" to "Last", so it would always start up with the instrument's last configuration. I tried changing it to "DFT", and miraculously, after a reboot, the instrument started up correctly!

I'm not sure if it was a bug in version 3.2.2.6.2R5 or why it happened. Maybe I'll try setting "Power On Type" back to "Last" again and see what happens...
Ciao
Guido - ik2bcp

Thank you, this is valuable info however you did not state the previous FW version you updated from. 
Was it 3.2.2.6.2R5 from new or a previous version ?

FWIW I always have any of these models SSA/SVA/SNA set to boot from last settings and dispatch them set in this mode/way also.
We if necessary also update FW before dispatch but never experience a failure to boot afterwards.

[guidoted]

Hello Tautech and thanks for your help in personal messages.
The 3.2.2.6.2R5 was updated from a previous version and with this version I used my SVA1032X for months without any problem.
I don’t figure out what is the glitch that locked my instrument.
Guido

[TERRA Operative]

Can these instruments do modulation domain analysis like the HP 53310A?

I'd love to not have to buy another instrument, especially as I just missed out on an auction for a 53310A recently.

[2N3055]

Can these instruments do modulation domain analysis like the HP 53310A?

I'd love to not have to buy another instrument, especially as I just missed out on an auction for a 53310A recently.

They have that option. Check datasheet for list. User manual has more details.

[vsilves]

Recently noticed that one of my scripts uses the SCPI command ":CORRection:COLLect:METHod:TYPE?" to retrieve the state of calibration for the SVA1032X, as I would like to record in my .s1p and .s2p files if the calibration was on or not, but I noticed that it always returns "NONE" whether the device shows "COR" (fully calibrated) on screen or not. Tried the SCPI command via web interface and same result.

If anybody knows a workaround I would like to hear it.

[tautech]

Recently noticed that one of my scripts uses the SCPI command ":CORRection:COLLect:METHod:TYPE?" to retrieve the state of calibration for the SVA1032X, as I would like to record in my .s1p and .s2p files if the calibration was on or not, but I noticed that it always returns "NONE" whether the device shows "COR" (fully calibrated) on screen or not. Tried the SCPI command via web interface and same result.

If anybody knows a workaround I would like to hear it.

Not possible.

Instead you save a Cor file and recall a Cal file which returns the calibration previously saved.
See the File menu and the save and load file types.

[vsilves]

Not possible.

Not possible with on device or web interface save .s1p file functionality, but my script pulls the S11 and S21 traces using SCPI and saves the data directly as .s2p in MacOS, so I can add any metadata I want in the header.  Works well with NanoVNA via USB (i.e., I can read calibration state and save in .s1p or .s2p metadata) but impossible with SVA1032X if the appropriate SCPI command that is supposed to return calibration state always returns "NONE" regardless of the calibration state of device.

[minsik]

Just an update re the new commands, I had been wating for calibrating the 10Mhz internal reference.
" local oscillator feedthrough self calibration, 10M out frequency accuracy Self-Calibration function and FTP function"
SVA1000X-V3.2.2.6.3R2.r2-EN.zip.
So after upgrade (bows head, spent 2 hours trying commands, result said success but nothing happened! Curses not the right bloody FW= idiot me!)
So after loading SVA1000X-V3.2.2.6.3R2.r2-EN its working great as it should do. 

Before = it was +227Hz error at 1.4G, very concerning! (must be a fussy bugger maybe!)

1) I set instrument as recommended, thought about the problem, did it my way in the end. 
2) SVA1000X with NO external reference. Centre freq set to 1.4G (as high as she will practically go!) Span 1Khz and lower.
3) Input signal generator locked to the Leo Bodnar 10Mhz GPS reference. Output 1.4G (no modulation)
4) Using the web page checked access ok.
5) Gave the SCPI command, yippee its moving. Tried a few values DEBug:HWACcess:FREQ:DAC:OFFSet +840 is perfect.
6) Locked value into place with DEBug:HWACcess:FREQ:DAC:OFFSet:CONFirm. 
7) SYSTem:RESTart rebooted and rechecked as spot on.

Now whats the local osc feedthrough adjustment?

[vsilves]

Ah, interesting "new" SCPI commands. Ran around the house mad for a while because I could not find them in the programming or service manuals, which apparently are cost prohibitive for Siglent to update. But then found the Word doc that @tautech posted.

Tried DEBug:HWACcess:FREQ:DAC:OFFSet? which returned 0 (my original offset).  Then measured using 120MHz sine out of SDG2042X to SA (converted SSA3021X).  Using Leo Bodnar GPSDO input to SG, SA, both, or none.  See pictures 1..4 before any adjustments.  Key is picture 3, showing SG with GPSDO and SA using internal 10MHz, which shows that the SA was about 11 Hz too high.

After DEBug:HWACcess:FREQ:DAC:OFFSet 530 see picture 5, which shows SG with GPSDO and SA with internal 10MHz reference after the SCPI command, and a roughly perfect 0 offset without needing external GPSDO input.  The value to the SCPI command is clearly not in Hz.  The DEBug:HWACcess:FREQ:DAC:OFFSet:CONFirm does make it stick even after reboot.  This is perfect.  Ended up settling on a value of 610 after things warmed up a bit more.

This is a variation on the Siglent cal procedure which I think is more logical and easy to use (in my case).

Now, it would be cool if we could also adjust the 10MHz reference in the SDG2042X, or the SDS804X HD that I also own.  Should be theoretically possible with a similar firmware update.

[james38]

Hi,
you said, "I found the Word document that @tautech posted."

Where did he post it?
Couldn't find it.

Regards, Chris

[vsilves]

Where did he post it?
Couldn't find it.

Towards the end of this thread

https://www.eevblog.com/forum/testgear/siglent-ssa3000x-spectrum-analyzers/2600/

[james38]

Thanks, I found it ;-)

https://www.eevblog.com/forum/testgear/siglent-ssa3000x-spectrum-analyzers/msg6007659/#msg6007659

Regards Chris

[AndyC_772]

Hi everyone,

I have a new toy, a Siglent SVA1032X with all the options, and I'm using it for EMC pre-compliance testing.

I've been using the EMI mode to search for and quantify conducted emission spikes using a current probe. I understand that in EMI mode the analyser can be set to follow a sequence whereby it scans a part of the frequency spectrum to locate peaks, then homes in on each one and measures it with different types of detector to give an amplitude figure which can be compared against a test limit.

Unfortunately it appears there's a firmware issue with this mode. I've attached my unit's settings (.STA) and a few screen shots that illustrate the problem.

With the settings as per the attached file, the instrument is in EMI mode and I can press Start to run a scan. It correctly scans from approx 72 to 272 MHz, and my UUT generates some spikes including one at 172 MHz with an amplitude of around 8dBuV. The SVA identifies the spikes, measures their amplitudes and presents a list of 6 that are significant. This is all perfectly OK. (See screen shot PNG1.png)

Now, if I press Mode > Spectrum Analyser, it shows a live scan centred on the spike at 172 MHz with a total span of 10 MHz. This is also perfectly OK. (PNG2.png)

Without changing anything, press Mode > EMI Measurement and then press Start. The scan shows an apparent signal level that's way off the top of the screen, and the list of spikes is also nonsense. (PNG3.png).

Also, on multiple occasions I've also had the EMI mode scan just completely fail to start; the SVA just sits there apparently doing nothing.

Switching off and on again temporarily resolves these issues, though of course, this isn't a fix.

Firmware is 6.3R2, Uboot-OS 3.2.2, FPGA 20240828-1001, CPLD D4, HW 01.00.00.

Anyone else come across problems with EMI mode on this device?

[vsilves]

This is a follow up, several days after adjusting the SSA3021X+ internal TCXO offset to 610 using new SCPI commands added on the last firmware update.

Graph shows data pulled via python script for marker 1 peak with counter enabled on a 100MHz input from Siglent SDG2042X with GPSDO.  No GPSDO on SSA3021X+ so the frequency counter shows the warm up of the internal TCXO. 

Took about 90 mins to match the SG to +/- < 1 Hz.  The offset setting holds well.  Still more practical to use GPSDO but good to have the option to adjust TCXO offset!

[tautech]

Hi everyone,

I have a new toy, a Siglent SVA1032X with all the options, and I'm using it for EMC pre-compliance testing.

I've been using the EMI mode to search for and quantify conducted emission spikes using a current probe. I understand that in EMI mode the analyser can be set to follow a sequence whereby it scans a part of the frequency spectrum to locate peaks, then homes in on each one and measures it with different types of detector to give an amplitude figure which can be compared against a test limit.

Unfortunately it appears there's a firmware issue with this mode. I've attached my unit's settings (.STA) and a few screen shots that illustrate the problem.

With the settings as per the attached file, the instrument is in EMI mode and I can press Start to run a scan. It correctly scans from approx 72 to 272 MHz, and my UUT generates some spikes including one at 172 MHz with an amplitude of around 8dBuV. The SVA identifies the spikes, measures their amplitudes and presents a list of 6 that are significant. This is all perfectly OK. (See screen shot PNG1.png)

Now, if I press Mode > Spectrum Analyser, it shows a live scan centred on the spike at 172 MHz with a total span of 10 MHz. This is also perfectly OK. (PNG2.png)

Without changing anything, press Mode > EMI Measurement and then press Start. The scan shows an apparent signal level that's way off the top of the screen, and the list of spikes is also nonsense. (PNG3.png).

Also, on multiple occasions I've also had the EMI mode scan just completely fail to start; the SVA just sits there apparently doing nothing.

Switching off and on again temporarily resolves these issues, though of course, this isn't a fix.

Firmware is 6.3R2, Uboot-OS 3.2.2, FPGA 20240828-1001, CPLD D4, HW 01.00.00.

Anyone else come across problems with EMI mode on this device?

No expert on EMI mode but a couple of questions.

We can see by your screenshots PA is engaged and attenuation is no longer Auto but 0dB flagged by a star at the top of the display, why ?
Also you don't have the default dB scaling selected, why ?

Attached is a screenshot of the default CISPR A band with zero signal and open RF In port to compare.

[AndyC_772]

I'm doing pre-compliance conducted emissions testing for Land Class A - see https://www.xppower.com/resources/blog/def-std-59-411-and-the-challenge-of-land-class-a

The limit is 0dBuA, as in, just 1uA RMS from 2 MHz upwards. Without setting the attenuator to 0dB and engaging the pre-amp, the emissions from the device are below the noise floor. My current probe has a transfer impedance of 20dBOhm, so I'm looking at noise spikes below 20dBV.

Thankfully, with the setup as I have it, and a suitably low RBW, I can see the very low level spikes I'm looking for, which is great - but I really do need all the sensitivity I can get!

[AndyC_772]

Today I've found a way to reliably trigger the problem with the EMI mode scan.

Start by recalling the factory default preset, then go into EMI mode and change the sweep start frequency to 1 MHz and the end frequency to 300 MHz. With Sequence set to 'scan only', press 'start' and it works fine.

Change the mode type to Spectrum Analyser, press BW and change the filter to "EMI" and the RBW to Manual, 9 kHz. As expected, the noise floor drops because of the narrower filter bandwidth.

Now, switch back to EMI mode and run a scan. The trace is way off the top of the screen, totally wrong. Or it might not even start at all, I've seen both symptoms occur.

Switch back to SA mode, press BW and change RBW to 120 kHz.

Go back to EMI mode and scan again, and it works perfectly.

It's definitely something to do with the bandwidth setting in SA mode, even though that shouldn't even be used in EMI mode.

[sm5uiu]

SVA1015X with the latest sw 3.1.2.6.3R2.r2, not that it matters, the issue has been from the start. I have never ever got the video trigger to work in zero span mode.
My unit is from 2019 with HW 02.01.00. Here is a video of how it should work:
Is it just me, the software or a hardware fault? Have done reset of the unit more than once, no change. Everything else seems to work normally.

As a sidenote, my older Rigol DSA815-TG work exaclty as expected in zero span and video trigger.