Siglent SSA3000X and SSA3000X-Plus Spectrum Analyzers

[noreply]

Thanks 

Its all VERY interesting to see how the FW is build and packaged 

Sure would be nice to get some links for appropriate tools - whereby we can unpackage the FW and then repackage 

New area for me 
 
- need to do some more research
- like where the .app(s) reside as a starter
- then we need to 'see' what is common to the different devices
- then identify how things differ

 so we can try to unify and see where the crossflash has a problem(s)

[tv84]

Looking at techneut uboot log, it surely seems the only thing that went bad was lack of space  to copy all app system files. The rest went pretty smooth and we almost achieved the crossflash goal.

If you look at  SVA's upgrade_app.sh which is responsible  for upgrading the siglent app filesystem, it contains:

Code: [Select]

echo "upgrade_app: starting"

if [ -f /usr/bin/siglent/ecomb ]; then
rm /usr/bin/siglent/ecomb
rm /usr/bin/siglent/libscpi*
fi

The only thing that was needed was to change that to:

echo "upgrade_app:      starting"

if [ -f /usr/bin/siglent/ecomb_p ]; then
   rm /usr/bin/siglent/ecomb_p
   rm /usr/bin/siglent/libscpi*
fi

With that we shouldn't have ran out of space!

Just an educated guess...
 
BTW: This adds credit to the fact that even the SVA's stock update script has a "feature" to replace a possible SSA app.   

[noreply]

Thanks tv84 - that's great news!

At least there is no brick wall - small steps in the right direction and we will get there

Perhaps to help those with appropriate SSA3000x Plus hardware who are game enough to 'try' for themselves, we could post a failsafe recovery process from the 11 nandbackup images??

Easy enough to generate the backup images - but not sure if the recovery process is just plain simple restore the backup with a 'reverse' command script?

My worry is that if the device will be totally 'bricked' - so no access to file system (boot , let alone telnet not possible), then there is no option other than direct console access via PCB header.

Not so keen to 'break-into' new device just now if I brick it 

Any 'passive' recovery method(s) would always be welcome - so more of us can 'have a go' at mastering the crossflash process

[Elasia]

official verdict....

it ran out of disk space lol

/usr/bin/siglent # df -k
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/root                63180     63180         0 100% /
devtmpfs                 69852         0     69852   0% /dev
none                     78140         0     78140   0% /tmp
ubi1_0                   30388     30256       132 100% /usr/bin/siglent
ubi2_0                    5848       708      5140  12% /usr/bin/siglent/firmdata0
ubi3_0                    5848        24      5824   0% /usr/bin/siglent/log
ubi0_0                   84752       408     84344   0% /usr/bin/siglent/usr

[techneut]

Not hindered with any knowledge I did a bit compare between my saved mtd's and the resulting mtd's after the not yet working change:
mtdo from line 005ed210 to line 005ed7fo sva = ffff ssa = 0000
mtd1 from line 002fa2fo to line 002fa7f0 sva = ffff ssa = 0000
mtd2 no difference
mtd3 no difference
mtd4 no difference
mtd5 no difference
mtd6 a lot of differences and ssa has a lot more data
mtd7 a lot of differences
mtd8 a lot of differences and ssa has a lot more data
mtd9 different in the end there is some data  but most is it ffff vs 0000 and that changes inthe end
mtd10 no difference
mtd11 no difference

[techneut]

Thanks tv84 - that's great news!

At least there is no brick wall - small steps in the right direction and we will get there

Perhaps to help those with appropriate SSA3000x Plus hardware who are game enough to 'try' for themselves, we could post a failsafe recovery process from the 11 nandbackup images??

Easy enough to generate the backup images - but not sure if the recovery process is just plain simple restore the backup with a 'reverse' command script?

My worry is that if the device will be totally 'bricked' - so no access to file system (boot , let alone telnet not possible), then there is no option other than direct console access via PCB header.

Not so keen to 'break-into' new device just now if I brick it 

Any 'passive' recovery method(s) would always be welcome - so more of us can 'have a go' at mastering the crossflash process

If you are worried about the nice shiny sticker, I applied some ipa and a hairdryer and put a thin wire in the slit between case and front bhind the sticker. With soft pulling it in the direction of the front it came loose.

[tv84]

Elasia (and soon techneut) has good news for all!   

My respect for these two guys and the quest they undertook. 

[Elasia]

Successful upgrade

Also note.. some of the options vanished?

After restoration i just changed the version config file to the SVA code and ran the .7 ads file

[Elasia]

Also here is the updated file to change the product id

its in /usr/bin/siglent/config

[noreply]

Thanks tv84 - that's great news!

At least there is no brick wall - small steps in the right direction and we will get there

Perhaps to help those with appropriate SSA3000x Plus hardware who are game enough to 'try' for themselves, we could post a failsafe recovery process from the 11 nandbackup images??

Easy enough to generate the backup images - but not sure if the recovery process is just plain simple restore the backup with a 'reverse' command script?

My worry is that if the device will be totally 'bricked' - so no access to file system (boot , let alone telnet not possible), then there is no option other than direct console access via PCB header.

Not so keen to 'break-into' new device just now if I brick it 

Any 'passive' recovery method(s) would always be welcome - so more of us can 'have a go' at mastering the crossflash process

If you are worried about the nice shiny sticker, I applied some ipa and a hairdryer and put a thin wire in the slit between case and front bhind the sticker. With soft pulling it in the direction of the front it came loose.

Yeah, I hear what you say
 
- but this device is my new toy and I have not played with it at all
- thinking why not just 'upgrade it' to an SVA before the playing starts

Unfortunately easier said than done 

You and Elisia are way ahead of the curve (with a helping hand from tv84) and rightfully deserve your successes 

I have lots of catching up to do before doing something useful with my device - hoping to get enough knowhow first - so to avoid the 'brick' problem (if it happens I will use your advice about the sticker ) and at least tread safely in the area you guys bravely explored alone with the high risks taken all for the interest of those who might follow your steps.

Well done - I love it when it all comes together after some clever FW detective work 

 

[noreply]

Successful upgrade

Also note.. some of the options vanished?

After restoration i just changed the version config file to the SVA code and ran the .7 ads file

Congrats Elasia - especially your quick recovery from a 'bricked' device.

Looking at your SVA system Info page - I drool with envy and fully appreciate the taste of the sweet nectar of success you must be enjoying now 

[tv84]

Also here is the updated file to change the product id

its in /usr/bin/siglent/config

Exactly. Maybe in the coming days I'll pack these 2 little changes in a new SVA1032X_2.2.1.2.7_ProdID_11405.ADS:

- correct removal of the ecomb_p
- add the SVA's NSP_config_upgrade_info.xml

to fully automate the crossflashing.

 

[Elasia]

Successful upgrade

Also note.. some of the options vanished?

After restoration i just changed the version config file to the SVA code and ran the .7 ads file

Congrats Elasia - especially your quick recovery from a 'bricked' device.

Looking at your SVA system Info page - I drool with envy and fully appreciate the taste of the sweet nectar of success you must be enjoying now 

Thanks, yeah wasnt really bricked to the point of needing to hit them with jtag... it was my computer being a horses ass earlier and making all the serial port data garbage... switched to the tablet i use normally for reprogramming network switches... worked fine... pos comp lol

And yeah, tv knows what to do now to repack the firmware to convert it.. HOWEVER  you need to buy the calibration kit, they do not come with SVA calibration data loaded, you must use the sva cali menu to do this yourself

I'll pick it up one of these days myself unless someone wanted to send me one ~

[tv84]

How to crossflash a SSA3000X Plus into a SVA1032X:

1. Telnet into the equipment
2. rename
   /usr/bin/siglent/ecomb_p
   to
   /usr/bin/siglent/ecomb
3. in file /usr/bin/siglent/startup_app.sh change the line
   /usr/bin/siglent/ecomb_p &
   to
   /usr/bin/siglent/ecomb &
4. in file /usr/bin/siglent/config/NSP_config_upgrade_info.xml change the line
   <upgrade_static_id>11405</upgrade_static_id>
   to
   <upgrade_static_id>11403</upgrade_static_id>
5. Reboot
6. Now you can update your SSA with the Siglent's SVA stock FW.
   (should work with SVA1032X_2.2.1.2.5 or SVA1032X_2.2.1.2.7, at least)
7. After having flashed your first SVA stock FW, the SSA has become a "true" SVA.

To lower the risk you can do an additional step which is to activate the telned daemon before the line reference in step 3.
(for example, place there: /sbin/telnetd -l /bin/sh -p 10101  &     )

As always: do it at your own risk. (if you don't feel comfortable with linux commands ask somebody else to help)

EDIT Jun 27th:
ATTENTION: You MUST do a "sync" command before rebooting to flush any pending file operation.

[noreply]

Successful upgrade

Also note.. some of the options vanished?

After restoration i just changed the version config file to the SVA code and ran the .7 ads file

Congrats Elasia - especially your quick recovery from a 'bricked' device.

Looking at your SVA system Info page - I drool with envy and fully appreciate the taste of the sweet nectar of success you must be enjoying now 

Thanks, yeah wasnt really bricked to the point of needing to hit them with jtag... it was my computer being a horses ass earlier and making all the serial port data garbage... switched to the tablet i use normally for reprogramming network switches... worked fine... pos comp lol

And yeah, tv knows what to do now to repack the firmware to convert it.. HOWEVER  you need to buy the calibration kit, they do not come with SVA calibration data loaded, you must use the sva cali menu to do this yourself

I'll pick it up one of these days myself unless someone wanted to send me one ~

Good to hear not ‘bricked’
 
– hate it when this happens
– it drains all my mental energy,  because I need to go to basics, make a cable, get working serial port on pc, get correct baud speed, bits , etc, etc.

Not something you want to do
 
– especially when you just had a big letdown because you failed with your flashing of the device


With regard to Calibration Kit, I’m not sure you need to get a Siglent Kit??

My take on the calibration is not very strict, because there are two ways of looking at it ;-

1.   Unless you are going to use the VNA of the SVA for calibrating the actual DUT for some certification requirements, then it’s not necessary to have a precisely certified calibration kit.
2.   You can make very good ‘design measurements’ when the VNA is calibrated with an alternative non-certified calibration kit.
 
If you are ok with 2 above – then there is a very good low cost solution.

You probably heard of the $35 NanoVNA device??

If not, check it out, it’s a great device (think there are some forum threads on it here at eevblog)

It uses an AD chip good to 900Mhz (there is a V2 out now with 3GHz ability)

– it’s a must have toy / tool 

The NAnoVNA comes with its own SMA calibration kit and cables

– see attached photos of my device
– the SMAs are 50 Ohm load, Open, and Short with associated cables.

All for $35 it’s a no brainer 

Unless I totally missed the level of calibration accuracy required for the VNA of the SVA, then for non-critical work this kit should be fine!

Hope this is useful info to get you going with you new (he says this with some envy) SVA – well earned by your patience and risk taking  – all good fun I guess as we are ‘locked away’ during the covid-19 pandemic   

[Elasia]

oh thats a nice find noreply.. i'll have to get one of those

and yeah more than likely... you just need a similar kit to at least give it basic cal data

[NY2KW]

SDR-kits.net in UK has excellent SMA calibration kits up to 12GHz for about $65 US. Each set (male and female) is individually calibrated and comes with full calibration specifications that  should be able to be entered manually into the SVA1032X 


Jerry NY2KW

[NY2KW]

here's the link:  https://www.sdr-kits.net/calibration-information-for-DG8SAQ-VNWA-3-3EC

[tautech]

SDR-kits.net in UK has excellent SMA calibration kits up to 12GHz for about $65 US. Each set (male and female) is individually calibrated and comes with full calibration specifications that  should be able to be entered manually into the SVA1032X 
Jerry NY2KW

From a few days ago:

Those are good parts, but they are not properly characterised.
But probably good enough for the low freq range of the vnwa based on what I remember of their analysis. And so could be usable for the 1.5GHz SVA. However, the load resistance they provide cant be used for a custom cal kit, but the offsets can.

hendorog and I have checked a few Cal kits against the factory SVA calibration and unless you're characterizing DUT's at some detail the factory calibration is quite good enough .....but it's only to the port so adapters and cables do add some error. DTF could give you some ps delay measurements to add into Port extensions that might improve basic results without the need for Cal kits.
I have the Siglent 4.5 GHz SMA kit coming in a few weeks so hendorog and I will be running against several Cal kits he already has with Nano's, HP VNA and Signal Hound stuff he has.

Edit
Oh and how could I forget, also with a SVA1075X ! 

[noreply]

How to crossflash a SSA3000X Plus into a SVA1032X:

1. Telnet into the equipment
2. rename
   /usr/bin/siglent/ecomb_p
   to
   /usr/bin/siglent/ecomb
3. in file /usr/bin/siglent/startup_app.sh change the line
   /usr/bin/siglent/ecomb_p &
   to
   /usr/bin/siglent/ecomb &
4. in file /usr/bin/siglent/config/NSP_config_upgrade_info.xml change the line
   <upgrade_static_id>11405</upgrade_static_id>
   to
   <upgrade_static_id>11403</upgrade_static_id>
5. Reboot
6. Now you can update your SSA with the Siglent's SVA stock FW.
   (should work with SVA1032X_2.2.1.2.5 or SVA1032X_2.2.1.2.7, at least)
7. After having flashed your first SVA stock FW, the SSA has become a "true" SVA.

To lower the risk you can do an additional step which is to activate the telned daemon before the line reference in step 3.
(for example, place there: /sbin/telnetd -l /bin/sh -p 10101  &     )

As always: do it at your own risk.

tv84 - thanks for your great mentoring and guidence for newbees like me on this subject


I could not resist in 'closing the loop' with your crossflashing proof of concept - to reverse the process 



How to crossflash a SVA1032X into a SSA3000X Plus:

1. Telnet into the equipment
2. Rename

    /usr/bin/siglent/ecomb
   to
   /usr/bin/siglent/ecomb_p

3. in file /usr/bin/siglent/startup_app.sh change the line
 
   /usr/bin/siglent/ecomb &
   to
   /usr/bin/siglent/ecomb_p & 
   
4. in file /usr/bin/siglent/config/NSP_config_upgrade_info.xml change the line

   <upgrade_static_id>11403</upgrade_static_id>
   to
   <upgrade_static_id>11405</upgrade_static_id>
   
5. Reboot
6. Now you can update your SVA with the Siglent's SSA stock FW.
   (should work with SSA3021/3032X Plus_V2.2.1.2.5 or SSA3021/3032X Plus_V2.2.1.2.7, at least)
7. After having flashed your first SSA stock FW, the SVA has become a "true" SSA.
8. If you already ‘own’ a SVA1032X – there may not be a good reason for crossflashing the FW for this device to turn into a SSA3021/3032X Plus.
    It’s just proof of concept as proposed by tv84

To lower the risk you can do an additional step which is to activate the telned daemon before the line reference in step 3.
(for example, place there: /sbin/telnetd -l /bin/sh -p 10101  &     )

As always: do it at your own risk.

**** The startup_app.sh for the SVA is different to the startup_app.sh for the SSA rendering the 'reversing' instructions destructive - DO NOT PERFORM STEP 2 and STEP 3 , if you decide to crossflash from an original SVA to become a SSA

[Elasia]

SDR-kits.net in UK has excellent SMA calibration kits up to 12GHz for about $65 US. Each set (male and female) is individually calibrated and comes with full calibration specifications that  should be able to be entered manually into the SVA1032X 
Jerry NY2KW

From a few days ago:

Those are good parts, but they are not properly characterised.
But probably good enough for the low freq range of the vnwa based on what I remember of their analysis. And so could be usable for the 1.5GHz SVA. However, the load resistance they provide cant be used for a custom cal kit, but the offsets can.

hendorog and I have checked a few Cal kits against the factory SVA calibration and unless you're characterizing DUT's at some detail the factory calibration is quite good enough .....but it's only to the port so adapters and cables do add some error. DTF could give you some ps delay measurements to add into Port extensions that might improve basic results without the need for Cal kits.
I have the Siglent 4.5 GHz SMA kit coming in a few weeks so hendorog and I will be running against several Cal kits he already has with Nano's, HP VNA and Signal Hound stuff he has.

Edit
Oh and how could I forget, also with a SVA1075X ! 

Teardown? ^^ pcb pics? ^^

[tautech]

SDR-kits.net in UK has excellent SMA calibration kits up to 12GHz for about $65 US. Each set (male and female) is individually calibrated and comes with full calibration specifications that  should be able to be entered manually into the SVA1032X 
Jerry NY2KW

From a few days ago:

Those are good parts, but they are not properly characterised.
But probably good enough for the low freq range of the vnwa based on what I remember of their analysis. And so could be usable for the 1.5GHz SVA. However, the load resistance they provide cant be used for a custom cal kit, but the offsets can.

hendorog and I have checked a few Cal kits against the factory SVA calibration and unless you're characterizing DUT's at some detail the factory calibration is quite good enough .....but it's only to the port so adapters and cables do add some error. DTF could give you some ps delay measurements to add into Port extensions that might improve basic results without the need for Cal kits.
I have the Siglent 4.5 GHz SMA kit coming in a few weeks so hendorog and I will be running against several Cal kits he already has with Nano's, HP VNA and Signal Hound stuff he has.

Edit
Oh and how could I forget, also with a SVA1075X ! 

Teardown? ^^ pcb pics? ^^

Sorry no, it has an NDA attached to it. 

[noreply]

oh thats a nice find noreply.. i'll have to get one of those

and yeah more than likely... you just need a similar kit to at least give it basic cal data

I guess we can always use the SSA part of the SVA to get some basic calibration of the VNA calibration kit's SMA connectors and associated cables 

We can always normalize the basic 'test' setup and then 'plug-in' the various Kit elements - and get a decent calibration - so any discrepancies in the connectors can be factored into the final calibration of the VNA 

This way we might as well use the US$1600 SSA we already have -  to get some decent measurement and save some $$ rather than buy an expensive calibration kit

I bet after using SSA to calibrate the NanoVNA calibration kit's connectors - SVA VNA calibration will be close to the Siglent Kit's result 

[tautech]

oh thats a nice find noreply.. i'll have to get one of those

and yeah more than likely... you just need a similar kit to at least give it basic cal data

I guess we can always use the SSA part of the SVA to get some basic calibration of the VNA calibration kit's SMA connectors and associated cables 

We can always normalize the basic 'test' setup and then 'plug-in' the various Kit elements - and get a decent calibration - so any discrepancies in the connectors can be factored into the final calibration of the VNA 

This way we might as well use the US$1600 SSA we already have -  to get some decent measurement and save some $$ rather than buy an expensive calibration kit

I bet after using SSA to calibrate the NanoVNA calibration kit's connectors - SVA VNA calibration will be close to the Siglent Kit's result 

When you become familiar with SVA's UI you might find the setting to reset to the factory cal so to check any discrepancies between another Cal kit. 

Start exploring into 0.01dB/div and you'll find differences alright. 


Anyways, SVA discussions might be best here:
https://www.eevblog.com/forum/testgear/siglent-sva1015x-1-5ghz-spectrum-vector-network-analyzer-(coming)/

[noreply]

When you become familiar with SVA's UI you might find the setting to reset to the factory cal so to check any discrepancies between another Cal kit. 

Start exploring into 0.01dB/div and you'll find differences alright. 

Sounds good - once I get my 'SSA' SVA up and running - looking forward in exploring the UI 

@ 0.01db/div - just touching the connectors will definitely generate differences 

I guess getting some connector torque wrenches will come in handy - if you are really serious about calibration.