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

[steve1515]

Steve, I suspect you are missing a small but very important step. 
Work through 1 Port Cal as you have then re-enter the Calibration menu and select Correction = ON.

Screenshots below just taken show first what a Load looks like with Correction not applied and then applied.

This is interesting... By default my Correction is on, so I never touched it. I will have to try switching it OFF before doing the cal and then switching it back ON after to see if that makes a difference. I'll try this out tonight.

Is your Correction defaulting to OFF?

No and why would one undertake a calibration process with correction ON ? Dunno if it makes a difference or not. it just seems logical to disable corrections. Anyways I'm working with a V5 beta and just got an update that's apparently soon to be released so I need get all my checks done with it today.

I just tried a 1-port cal again, but I turned Correction to OFF first (it always defaults on ON after power on). After the 1-port cal completes, Correction is automatically set to ON, but I get the same result. I see "---" instead of "Cor". If I toggle between Correction ON and OFF, I can see the indicator go between "---" when Correction=ON and "OFF" when Correction=OFF.

Do you think I'm still missing a step somewhere? I'm a little concerned because I would think it should work very similarly to the Open or Short cal modes. But, I'm not really where where I'm going wrong here. 

EDIT: One thing I just noticed... after an Open cal which seems work work correctly, toggling between Correction ON/OFF shows clear changes in the waveform, so I know it's doing something. BUT... when I toggle Correction ON/OFF after a 1-port cal, nothing obvious seems to happen. I also tried this out by doing a 1-port cal without actually putting the OSL ends on just so I could get a pretty bad waveform... Again, toggling Corrections ON/OFF shows no difference in the waveform. It's as if it's not actually doing anything after a 1-port cal. I'm a little confused here... 

[steve1515]

Can you post a picture of your screen? It looks you are missing a cable because I had the same problem but then  my Smith diagram screen was al over the place

I've attached some pictures showing the results after a 1-port cal in both log mag and smith charts. Do these show any noticeable issues?

[eloso]

Hi Steve,

Everything you have experienced is absolutely identical to my experience.

There are several folk going out of their way to be helpful and it is not for me to question them about whether they are using genuine SVA1032X or not but so far I am not aware of anyone with a cross flashed SVA1032X that is saying theirs works normally with respect to the 1-Port Cal routine.   i.e. those who are confident that their way of using the machine gives a proper result are all using genuine SVA1032X and so replicating their advice is not fixing the problem.

I've played around with this quite a bit now and my interpretation of events is that the one port cal is  buggy in my cross flashed SVA. There is no combination of button pressing will get you the little "Cor" label after a one port CAL.  It is necessary to follow it with an Open CAL or a Short CAL in order to do so. These latter two will of course provide you with the "Cor" label regardless of whether you did the 1 Port first. They should be superfluous with regard to a 1 port cal since the one port cal includes these CALs anyway.

The good news is that having done a 1 port CAL the machine does indeed seem to be calibrated, even though it doesn't say so.   Before a 1 port CAL with a short but fairly dodgy piece of coax attached to the port the Smith Chart seems to be all over the place with the 50 ohm load.   Afterwards, you get a nice dot in the centre of the chart with the  same 50 ohm load.

As long as you don't change the Frequency Span or any Amplitude settings the calibration seems to function well.

There are some obvious conclusions - one is that if you are serious about using the VNA for anything other than hobby educational  purposes  then you really need to buy the genuine article to feel confident.  The other is that until someone steps forward with a cross flashed SVA1032X to say that theirs works ok, it is clear that the behaviour of Cross Flashed SVAs is different to that of natural born ones (easy test - do a 1-port CAL, if the "Cor" label comes up at the end then it is behaving itself, if only -- or OFF comes up at the end of the CAL routine then it is behaving like the badly behaved cross flashed unit that Steve and I have been describing.).

Thanks everyone, for your contributions

Regards

Eloso

[steve1515]

Hi Eloso.

Thanks for your comments! I'm wondering if someone can figure out the mystery of what's missing from the converted one since the hardware should be the same. Maybe the upcoming firmware will magically fix things 

[tautech]

Hi Eloso.

Thanks for your comments! I'm wondering if someone can figure out the mystery of what's missing from the converted one since the hardware should be the same. Maybe the upcoming firmware will magically fix things 

Been balls out today and there's still a few hrs to go ! 
Installed new FW but work/business has got in the way of doing any more.
Hopefully I'll get back to it this evening ...........

[steve1515]

Hi Eloso.

Thanks for your comments! I'm wondering if someone can figure out the mystery of what's missing from the converted one since the hardware should be the same. Maybe the upcoming firmware will magically fix things 

Been balls out today and there's still a few hrs to go ! 
Installed new FW but work/business has got in the way of doing any more.
Hopefully I'll get back to it this evening ...........

I understand, and thanks again for helping out with this! 
(Note: I've also added some pictures in the post above... Image upload wasn't working for me earlier for some reason.)

[tautech]

Hi Eloso.

Thanks for your comments! I'm wondering if someone can figure out the mystery of what's missing from the converted one since the hardware should be the same. Maybe the upcoming firmware will magically fix things 

Been balls out today and there's still a few hrs to go ! 
Installed new FW but work/business has got in the way of doing any more.
Hopefully I'll get back to it this evening ...........

I understand, and thanks again for helping out with this! 
(Note: I've also added some pictures in the post above... Image upload wasn't working for me earlier for some reason.)

Smith charts look mint 
Pro tip for Log Mag is to change the vertical scaling to see the linearity of the Cal load. 

[RoV]

My unit is an upgraded SSA3021 and 1 port calibration works perfectly. I get "Cor" at the end of the procedure, I can save the calibration and subsequently restore it. If I change the frequency settings, it changes to "C?" as it should.
The only difference I see is that I am using an 85033 cal kit.

[luudee]

I've attached some pictures showing the results after a 1-port cal in both log mag and smith charts. Do these show any noticeable issues?

I wonder if you are masking the same mistake I was after cross flashing.
Port 1 if the TG port on the SSA, and Port 2 is the RF Port.
When doing calibration, you need to use Port 1 (TG Source).

I'll do some test on my SSA3032X+ converted to SVA1032 later on today
and will post some pics too.

rudi

[steve1515]

I've attached some pictures showing the results after a 1-port cal in both log mag and smith charts. Do these show any noticeable issues?

I wonder if you are masking the same mistake I was after cross flashing.
Port 1 if the TG port on the SSA, and Port 2 is the RF Port.
When doing calibration, you need to use Port 1 (TG Source).

I'll do some test on my SSA3032X+ converted to SVA1032 later on today
and will post some pics too.

rudi

For my testing, I have been using the TG port as port 1.

[tv84]

WARNING:

It seems that crossflashed SVAs once in "pro mode" (hacked with no S/N) can't be correctly calibrated. That's the problem.

[tubularnut]

....  The other is that until someone steps forward with a cross flashed SVA1032X to say that theirs works ok, it is clear that the behaviour of Cross Flashed SVAs is different to that of natural born ones (easy test - do a 1-port CAL, if the "Cor" label comes up at the end then it is behaving itself, if only -- or OFF comes up at the end of the CAL routine then it is behaving like the badly behaved cross flashed unit that Steve and I have been describing.).

I had already posted previously that my cross flashed SSA3021X-plus to SVA1032X works OK. However, mine is not a 'pro-mode' conversion, it used a different technique and kept the serial number.

[steve1515]

Can you elaborate on how you did this? The closest I've gotten was verifying that the hash key used in other siglent products matches this product by inspecting the memdump. It just seems that the algorithm used to generate the keys is different, so I'm kind of stuck on where to go. I initially started looking at this because I was wondering about adding my serial number back in and adding the appropriate keys. Any tips?

[tv84]

For reference, here are some tests made by an anonymous member:

Some more tests have been carried out which show that in Factory Mode the VNA exhibits quirks surrounding one port calibration function.  There are work arounds and the author has continued to use the VNA for hobby S11 measurement.  The key is in being methodical and checking the calibration status frequently by means of looking at a range of loads on the Smith Chart.

But having reapplied the serial number the VNA appears to work impeccably.

In S21 similar results were observed but the author is a little hazier on the details and doesn’t want to flip the machine between different states since this is a non-trivial task.  This is left undocumented for the moment therefore.

I would suggest that for anything other than educational use, a genuine SVA1032X should be purchased.   It is the only way to be sure of having an instrument in calibration. But for educational use only the results seem sound enough.

Test	Expected Result	Actual Result in Factory Mode	Actual Result after reapplying original S/N
Carry out 1 Port Calibration 1Mhz – 500Mhz	Should operate over SPAN requested.  On completion it should show “Cor”. On completion a Smith chart with a 50ohm load on Port 1 should result in a small spot in the middle of the chart	As soon as Calibration routine is entered, frequency  SPAN annoyingly changes to 100kHz – 3.2Ghz. Can be changed back only with mouse pointer on screen. On completion there is no “Cor”, just “—“.   But note that  results on the Smith chart  appear to show a calibrated status	No problem – does exactly as requested
Power cycle machine after One Port Cal	Machine should come up into VNA mode with appropriate settings and with “Cor” highlighted to show status is calibrated	Problem. Switches on into SA mode.  Putting into VNA mode there is only “—“ where there should be “Cor”	No problem – does exactly as requested
Change Frequency Span after Calibration	Calibration status should change from "Cor" to  “C?”	Problem because calibration status is never shown as “Cor” in the first place.	No problem – does exactly as requested
Save Calibration File	Should create  file.	Creates a file.	Creates a file.
Load Calibration file after having made changes to frequency span.	Should show C? until frequency span is adjusted to be the same as when the calibration file was created	Problem. Calibration status shows “—“ all the time. i.e. loading a calibration file never appears to work as far as the "Cor" indicator is concerned	No problem – does exactly as requested

[luudee]

For reference, here are some tests made by an anonymous member:

Some more tests have been carried out which show that in Factory Mode the VNA exhibits quirks surrounding one port calibration function.  There are work arounds and the author has continued to use the VNA for hobby S11 measurement.  The key is in being methodical and checking the calibration status frequently by means of looking at a range of loads on the Smith Chart.

But having reapplied the serial number the VNA appears to work impeccably.
...

Hi TV84 !

did your Anonymous friend say how to restore the serial number ?
And can it be the original SSA one, or should I make one up for SVA ?
I tried the obvious places (NSP_system_info.xml), but that did not work.

Many Thanks !
rudi

[tv84]

The model type has nothing to do with the S/N.  You can use the SCPI command to re-insert the S/N or rollback the files that you changed.

Remember that there is a binary replica of the XML file that doesn't get updated when you manually play around with the XML.

[Bicurico]

Sum up:

1) Factory mode: implies not having a serial number, by editing/removing/deleting the respective XML file
2) With factory mode, all options are active
3) It has been found out that factory mode has bad/weird behaviour regarding VNA calibration/measurements
4) Enabling/restoring the serial number will deactivate factory mode and options need to have valid license again ("key") - cross graders will have no option at all
5) Owners of an SSA3021X+ can do the cross-grade to a SVA1032 but have to live with the mentioned issues in VNA mode if they use the factory mode to get all options activated
6) There is no public keygen

[luudee]

Sum up:

1) Factory mode: implies not having a serial number, by editing/removing/deleting the respective XML file
2) With factory mode, all options are active
3) It has been found out that factory mode has bad/weird behaviour regarding VNA calibration/measurements
4) Enabling/restoring the serial number will deactivate factory mode and options need to have valid license again ("key") - cross graders will have no option at all
5) Owners of an SSA3021X+ can do the cross-grade to a SVA1032 but have to live with the mentioned issues in VNA mode if they use the factory mode to get all options activated
6) There is no public keygen

I believe there is a way:
1. We need someone with a genuine SVA1032X, to share the following files:
/usr/bin/siglent/firmdata0/NSP_system_info.xml
/usr/bin/siglent/firmdata0/NSP_sn_bandwidth.xml
/usr/bin/siglent/firmdata0/nsp_data_b1

2. Once we have those files, we should be able to change the S/N and (if needed) regenerate the licenses ...

Anybody wants to try ?
We can do it in private ...

Thanks,
rudi

[antenna]

I believe there is a way:
1. We need someone with a genuine SVA1032X, to share the following files:
/usr/bin/siglent/firmdata0/NSP_system_info.xml
/usr/bin/siglent/firmdata0/NSP_sn_bandwidth.xml
/usr/bin/siglent/firmdata0/nsp_data_b1

2. Once we have those files, we should be able to change the S/N and (if needed) regenerate the licenses ...

Anybody wants to try ?
We can do it in private ...

Thanks,
rudi

What post explains how to copy those files? I've never had mine connected to the computer or done any telnet stuff. 

[luudee]

I believe there is a way:
1. We need someone with a genuine SVA1032X, to share the following files:
/usr/bin/siglent/firmdata0/NSP_system_info.xml
/usr/bin/siglent/firmdata0/NSP_sn_bandwidth.xml
/usr/bin/siglent/firmdata0/nsp_data_b1

2. Once we have those files, we should be able to change the S/N and (if needed) regenerate the licenses ...

Anybody wants to try ?
We can do it in private ...

Thanks,
rudi

What post explains how to copy those files? I've never had mine connected to the computer or done any telnet stuff.

There are several posts with a telnet loader. You'd have to connect your SSA to your LAN and then telnet to it.

Then insert a Flash Drive and copy the above files I mentioned to the flash drive.

rudi

[ExaLab]

I'll receive a genuine SVA1032X shortly.

I know that the various SW options can be permanently enabled. Where can I find a trace of the procedure for doing this?
Given what has been reported in previous posts, does this in any way compromise the functionality of the analyzer?

Thanks everyone!

[luudee]

I'll receive a genuine SVA1032X shortly.

I know that the various SW options can be permanently enabled. Where can I find a trace of the procedure for doing this?
Given what has been reported in previous posts, does this in any way compromise the functionality of the analyzer?

Thanks everyone!

You need to download the python keygen ...  then enter the keys as you normally would on the System Submenu ...

Since you enter the keys like real keys you would receive from Siglent, it should not have any negative impact at all ...

rudi

[tv84]

The public keygen doesn't work for SSA or SVA.

[luudee]

The public keygen doesn't work for SSA or SVA.

Sorry TV84, but that is not correct. Before I liberated by SSA to an SVA, I was able to enable all features.

So I think the public keygen DOES work. You just have to see, if you need to use the SCOPID or SN ...


Cheers,
rudi

[tv84]

So I think the public keygen DOES work.

You MAY think BUT no.