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

[NY2KW]

You may already be aware but using external software is the  best way to get full control over the cal - and the models used for the standards.

Thanks, Not familiar with Python but I do some C programming so open to learning Python basics.  I downloaded latest Windows Python then Anaconda and then installed the scikit rf module.  I went through the early examples including the plots - very cool!  It says scikit rf only supports a few VNA s as virtual instruments.  Next steps for me is to learn how to set up my calibrations and load SVA1032X data for manipulations in scikit rf.   Too bad there isn't a scikit rf GUI interface to simply things... I rather be using the SVA1032x than getting on the scikit rf learning curve.

Jerry NY2KW

[hendorog]

You may already be aware but using external software is the  best way to get full control over the cal - and the models used for the standards.

Thanks, Not familiar with Python but I do some C programming so open to learning Python basics.  I downloaded latest Windows Python then Anaconda and then installed the scikit rf module.  I went through the early examples including the plots - very cool!  It says scikit rf only supports a few VNA s as virtual instruments.  Next steps for me is to learn how to set up my calibrations and load SVA1032X data for manipulations in scikit rf.   Too bad there isn't a scikit rf GUI interface to simply things... I rather be using the SVA1032x than getting on the scikit rf learning curve.

Jerry NY2KW

There kind-of is a basic IDE for it, in the form of Jyputer notebook. Its been a while, but I think you will already have that installed due to using Anaconda.
It all takes a bit of time and effort to get your head around, but it does help being able to see the results quickly.

I wrote some docs for myself here:
https://github.com/hendorog/SVA1015X

If you store your code on GitHub, then the Jyputer notebook is rendered nicely and you can see the results just like in the scikit docs.
I did some tests here using a faked short measurement, don't read anything into the results as I was just messing around.
https://github.com/hendorog/SVA1015X/blob/master/docs/SDR%20Cal%20Test1.ipynb

I keep meaning to write a driver for the SVA, but haven't done so yet. The above data was extracted manually, which was a pain.

[hendorog]

Will this VNA do amplitude sweeps?

Not at the moment. IIRC there is no amplitude control in VNA mode right now, but there is in SA+TG mode. Hoping this will be fixed soon.

[NY2KW]

I scanned through your code. Looks like fun.   What is the file format for your standards that you are reading in?  For example, the file "data/ideals/KirkbySN0092" ?

I see there is a VI file that Siglent provides.  I know even less about Labview, etc but wounder if the Siglent VI file could be easily modified for use in scikit-rf?

Jerry

[hendorog]

I scanned through your code. Looks like fun.   What is the file format for your standards that you are reading in?  For example, the file "data/ideals/KirkbySN0092" ?

I see there is a VI file that Siglent provides.  I know even less about Labview, etc but wounder if the Siglent VI file could be easily modified for use in scikit-rf?

Jerry

I'm not sure - I don't know much about Labview either and don't have any desire to learn it anymore
Regardless, there is a good doc on the SVA remote SCPI API so you could just use that instead of reverse engineering the VI.

The files are touchstone format (blah.S1P), which is a standard format for exporting/importing VNA data. To create those from the SVA, I think I just exported mag + phase into a csv and then massaged it by hand using a text editor.

The cal kit I have, from Dr Kirkby, came with its own touchstone files which describe the standards. I use that method instead of the C0, C1, C2 etc model, as it is easier. It's one less approximation so should be more accurate. The 'ideals/SN0092' directory contains these files - i.e. the ones which came with the cal kit.

Of course with the Rosenberger cal you don't get the touchstone files, but you could theoretically create one which reflects the values they provide, or alternatively you could get the kit measured on a properly calibrated VNA and export them.

I have the Rosenberger kit too, and I must have swept them at the same time. (Note that these measurements are uncorrected, and they need to be corrected using the scikit cal before they mean anything)
https://github.com/hendorog/SVA1015X/tree/master/docs/data/measured/SVA/SDRKits

[tautech]

Discussion of possible improvements to the measurement suite within VNA mode:
https://www.eevblog.com/forum/testgear/sva1032x-vna-disappointment/

[tautech]

Latest whisper is a SVA1075X will be released quire soon. 

Maybe then this thread topic can be updated to include this 7.5 GHz model.

[jayk]

Looks to be 2.5x the price of the 3.2GHz model....

https://www.siglent.eu/vna-spectrum-analyzers

[tv84]

Here is the FW to open the telnet session over port 10101 on a SVA1032X.

It's different from this one for the SVA1015X.

[taxilian]

This isn't working for me on my SVA1032X w/ V2.2.1.2.5.

I loaded SVA1032X_telnet.ADS onto a FAT32-formatted thumb drive (one which I've been able to save PNGs onto from the unit) and put it in. I went to System, System Info, Firmware Update, selected SVA1032X_telnet.ADS, and selected "Open/Load". I selected "Enter" on the warning about rebooting and stuff. It went into the "Upgrading now, please wait..." screen and stays there.
 
I then tried to run "telnet (ip address here) 10101" and got a "connection refused". I used nmap to map the ports on the unit and there is no new port open. The only open TCP ports are 21, 80, 111, 864, 5024, 5025, and 5900.

Am I missing any steps?

[taxilian]

It seems not. We need another member to test.

You tried the connection while the SVA remains in "upgrading..." mode?

Yeah, multiple times :-/ No luck so far. I don't lose connectivity to the device, it just goes into the "ugprading..." mode and nothing happens, no matter how long I wait

[borjam]

Curious, @tautech didn't mention it.

New firmware.

V2.2.1.2.7 2020/05/12
Notice:
This firmware must be upgraded from V2.1.1.1.12a or later. If your SVA1000X has an earlier version, please update to V2.1.1.1.12a first and then proceed to update to the latest version.
New Functionality:
 Add Limit mask offset in SA mode
 Frequency LOG scale in VNA mode
 Add OPEN、SHOT、enhanced response in VNA mode
 Add 85033D/E Cal kit in VNA mode
 Save/recall .cal file in VNA mode
 Port 1 power (TG) adjustable in VNA mode
 Zero span in VNA mode
Improvements:
 Smooth the glitch in Phase format in VNA mode
Solved Issues:
 Fixed AM and FM symbol rate error, Lpf coefficient error, in AMA mode
 Fixed Filter bandwidth error in DMA mode
 Fixed Freq counter error in SA mode
 Fixed Harmonic error in SA mode
 Help Display error

[jjoonathan]

>  Add OPEN、SHOT、enhanced response in VNA mode

My cal kit lives in a pistol case, but it doesn't include an actual pistol.

"Calibration step 5: remove the NIST calibrated pistol and shoot the bastard that munted the good connectors."

[tautech]

Curious, @tautech didn't mention it.

New firmware.

V2.2.1.2.7 2020/05/12

Versions for SVA1015X and SVA1032X models, they are each different !
~16 MB
Available here:
https://int.siglent.com/download/firmwares/?CateIdss=16

[borjam]

I have updated and it's great to be able to control the generator level in VNA mode.

Of course the noise level will be higher, but it feels much better if you want to, for example, investigate a recever front end filtering

[noreply]

Hi, I've been lurking arownd - had to chime-in to let you know that the file SSA3000X_Plus_telnet.ADS created on Saturday, ‎7 ‎December ‎2019, ‏‎10:05:34 AM does indeed boot-up Telnet without need for login 

Thank You for this TV-84

I was brave enough to 'try' on a new ssa3021x plus - did the appropriate homework;-

1. read all of the relevant forum posts
2. did nand backup
3. got latest sva1032x firmware
4. was about to flash
5. got your 'hint' re: product ID
6. now  to figure out how to change the firmware file to match

Keep up your great contributions - many would be lost without them 

[tautech]

Teaser for you guys with hacked SSA's to SVA's.

Real stock SVA1032X factory cal, open ports and latest FW.
Smith chart scaling.

[Elasia]

Teaser for you guys with hacked SSA's to SVA's.

Real stock SVA1032X factory cal, open ports and latest FW.
Smith chart scaling.

Can you go into the cal menu and show a screen with the calibration in off mode? <3

[hendorog]

Teaser for you guys with hacked SSA's to SVA's.

Real stock SVA1032X factory cal, open ports and latest FW.
Smith chart scaling.

Can you go into the cal menu and show a screen with the calibration in off mode? <3

In general on the SVA you can't turn off the factory cal.

Speculation, but I think on recent FW you do get a glimpse of the native performance as you do an OSL cal. It shows very different traces as you step through Open, short and load than what we normally see.

[Elasia]

Mode > VNA > Calibration Menu > Correction Off

[hendorog]

Mode > VNA > Calibration Menu > Correction Off

That will not remove the factory cal, it only removes a user cal.

[Elasia]

Mode > VNA > Calibration Menu > Correction Off

That will not remove the factory cal, it only removes a user cal.

That's the point, does that screen match the prior?

[hendorog]

Mode > VNA > Calibration Menu > Correction Off

That will not remove the factory cal, it only removes a user cal.

That's the point, does that screen match the prior?

Ah sorry, whoosh.

[tautech]

A recent afternoon with hendorog, inspecting his Covid lockdown lab rebuild  and investigations of SVA's and Cal kits............my own 4.5 GHz SMA F603FE Cal kit coming in July from Siglent:
https://siglentna.com/product/sma-type-vna-calibration-kit/

We wanted 2 good calibrations to save with my SVA1032X while good and warm, one on Port 1 and another after an N-SMA adaptor from Siglents UKitSSA3X Utility Kit:
https://siglentna.com/product/ukitssa3x-utility-kit/

The calibrations done were the normal SOL and each was saved into the file system where they can be selected and reapplied at will.
Cal kits were a SMA Kirby (7 GHz) and 18 GHz rated N type....didn't record the brand. 

Shot of the Calibration Save/Load file manager and menus.


A saved calibration is a Correction file and we need select, load and apply it for it to be of any use.
Once turned ON, indication the correction has been applied is flagged on the upper left of the display as COR.

Factory Cal

Correction added


But Corrections need be applied to the right setup and as some example, this is the Kirby SMA correction on just an open N port 1 whereas it's intended to be used with an N-SMA adaptor and shows a massive difference to the previous screenshot.


The same sweep with an open N port instead of with a N-SMA adaptor but in Smith chart mode.


hendorog taught me to work with Log Mag/Return Loss mode rather than a Smith chart until you have a Cal sorted as more detail is available over the full sweep.
As some example of this here is a quite good 50 ohm load hendorog gave me a while back in Log Mag and Smith chart modes.


Log Mag looks awful but some of what we see is system noise and applying some averaging tidys the trace up some.

I also came home with a spare NanoVNA Short and Open from hendorog so curiously swept them as to how they might align to the Kirby Cal in the SVA and as can be seen below their Short is an absolute POS ! 



Next I had a look at a 51 Ohm 0805 SMD resistor soldered onto a SMD bulkhead fitting that are very useful for component tests.
Log Mag looks pretty ordinary and the Smith chart too.





A few of these are handy to have and just scalpel the dielectric short and solder on components to be tested.

[OwO]

I'm suspecting you can't turn off factory cal because the raw directivity of the coupler isn't very good and they don't want you to see that

I also see significant trace noise around 3GHz, it looks just like the nanovna V2. That is because it doesn't have a simultaneous reference receiver and can not cancel out synthesizer phase noise.

Don't bother comparing the original Nano to this, look at the V2 instead which has far better calibration standards and the VNA itself performs basically the same as the SVA.

All SVAs and the nano V2 currently are hampered by the magnetics based coupler. It performs poorly at lower and upper frequency extremes. I'm experimenting with balun-less active reflectometer designs that can potentially give you decent directivity from DC to 6GHz. Siglent has some catching up to do.