Siglent SSA3000X and SSA3000X-Plus Spectrum Analyzers

[bozidarms]

rf-loop thanks for answer  i'll be careful!
I agree with DL4RAJ:
till -130dBm everything is precise,
and work es it should be! 

[rf-loop]

New FW 8.01

TG  is now bit improved.



First some older test images from this thread (around page 13).
OLD FW TEST

@TurboTom,  Siglent OLD FW




OLD FW TEST

@ rf-loop Siglent OLD FW
Used Xtal name "Tedford 2-71"
With external sweep (trace B) and then using SSA-TG with 2 different setting.




@ TurboTom at this time compaared with Rigol  815


NEW FW TEST

rf-loop Siglent NEW FW 8.01
Used Xtal name "Tedford 2-71" (same as used in previous test, but other probes )

It is clear that accuracy improvement is remarkable.
But adverse effect is - slow. 
With this kind of Xtal or filter (dB/Hz) it can run without any visible problems around double speed (optimal can find using RBW and VBW) Minimum RBW is 30Hz. Minimum VBW is same as normally 1Hz.

Sidenote. Also this .png image is now directly from SSA and original without any image adjustments. It is fast and it is now default. (if want, there is still crap jpg and then .bmp)

[rf-loop]

Here some other example. One 14.2MHz filter. (bit off)

At this time used RBW 1kHz 

Upper and lower limits and delta cursors are there only just for example without any special purpose.
(also as I'm lazy I use only minimal amount of points in limit lines. Btw, limit lines use real frequency and level points. If change span or other freq settings and level these lines follow.

[DL4RAJ]

Here some other example. One 14.2MHz filter. (bit off)

Yes,this filter has not been designed for 50 Ohm,hence the strong ripple.


Regards

[TurboTom]

Today had a closer look at the new frequency scanning scheme of the SSA3000X with the 8.01 firmware: I set up the SSA with a center frequency of 1MHz, Span 200kHz, RBW 100Hz and TG enabled. This results in a sweep time of 161.56s. I then connected the TG output to an MSO2xx2A-S scope and started tinkering with the settings while the TG was sweeping. It's obvious that finding infrequent events in a relatively continuous stream of sine waves may be challanging. But fortunately, at some frequencies of the TG, the aliasing moire patterns are well enough defined to visualize the change, see the attached screenshot.

The sweep isn't continuous but one frequency is held for a certain time (apparently something like e/RBW) and then the PLL is adjusted to the next frequency. With the parameters I selected, there are approx. eight individual frequencies scanned for each resolution point on the screen.

It appears that one of the PLLs that are in use to generate the TG frequency (not sure whether 1st LO or the offset PLL oscillator in the TG box) isn't locking too well (read: directly, quickly) to the new frequency since between the "measurement intervals", there's a considerable "chirp" with frequencies well outside the span range. Maybe that's what caused the problems with sweeping high Q factor filters/crystals with the TG (with 7.07 and before firmware that is) , and now the input is gated to prevent "seeing" the wrong signals.

Anyway, time is limited (family duties...  ) -- otherwise I would have had a less superficial view on the TG detais. I hope this information is of some interest anyway.

Cheers,
Thomas

[DL4RAJ]

New FW 8.01
TG  is now bit improved.

Here again my 8-pole crystal filter,once with 30kHz RBW (yellow) and once
with 30Hz RBW (pink).
Blue is noise floor in SA mode (30Hz RBW/1Hz VBW) with input terminated in 50 Ohm.
It can be seen that there are still around 20dB left down to the SA's noise floor.
For comparison once more also the screenshot from my VNA2180 with the same filter.
The Siglent now comes quite close.
Re sweep time:
I've noticed that the sweep time ist set to VERY long times in TG mode when one is playing with
VBW,e.g. 1Hz.When the VBW is reset to higher values,e.g. 30Hz or more,the sweep time
seems to stay at the extremely long value from before.
So it is worth checking sweep time if it can be cranked down bevor finally starting the
sweep.

Regards

[DL4RAJ]

New FW 8.01

TG  is now bit improved.

Here some screenshots from my SDR showing the old TG version vs the new version.
TG1 is old version with 30Hz RBW of the SDR.
TG2 is old version zoomed in with 1.9Hz RBW of the SDR.
TG3 is new version also zoomed in with 1.9Hz RBW of the SDR.

The difference between TG2 and TG3 is considerable.
Signal is MUCH cleaner now and also stable.
The zoomed in view at 1.9Hz RBW of TG3 looks even better (small fft window) than the 30Hz RBW view of TG1
where you can't see so much details.

Edit:In all three screenshots the *SA* was set to 30kHz RBW in TG mode.
When VBW or RBW of th SA are set to smaller values the "spikes" seen on the SDR
move more and more together,i.e. the discrete frequency steps of the TG become
smaller and smaller.

Regards

[mike509]

Hi Guys, I am new to all this,

I was wondering if there was an amplitude Auto-Correction setting to normalize the trace down. Due to external interference, I am getting 3 spikes (+18dB) on my screen when my DUT is not connected. The "Add point"  from the Amplitude correction seems to be tedious as you need to add the adjacent point ( ex 94.281, 94.282,94.283...)

Also, I saw a few post from people using Perseus software, does it work?

I will be doing EMI measurement on different frequency range (~5 ranges) with different limit for about 10 different items. Is there a way to automate the process?

[rf-loop]

Also, I saw a few post from people using Perseus software, does it work?

Perseus software works with Perseus receiver. Using Perseus or what ever radio  receiver  (conventional or software defined)  you can "listen" what ever transmitter. In this case  @DL4RAJ  have used Perseus for listen SSA Tracking Generator signal.

For test automation there can use SCPI commands, in this case PC control what SSA do.
You can look SSA3000X Programming Guide

[Bicurico]

Hi,

Does anyone know the SCPI port for the SSA3000X?

I am considering writing my own software,  but don't want to friend on NI to interface.

Thanks,
Vitor

[alexo]

I apologize for asking such a simple question : how do I connect a parallel resonant circuit ( capacitor and coil ) tuned to around 1Mhz to a Siglent spectrum analyzer with tracking generator ? Do I need some sort of impedance adapters ?

[Bicurico]

Does anyone know the SCPI port for the SSA3000X?

I guess my question doesn't make sense: I used Wireshark to find out the correct port and I run a port-scanner, too.

The SSA3000X listens to ports:

23 -> Telnet
111 -> Apparently to RECEIVE SCPI commands
897 -> Apparently to RECEIVE SCPI commands (* one of both 111 or 897 seems to be used to setup the ports for reception on the PC side)

The PC is receiving responses on ports:

3648
3649
3650

I don't know much about network protocols and I know nothing about SCPI, but I guess that the Siglent uses two different ports: 111 and 897 to receive data and sening data to ports 3648-3650 on the PC.

This is too much for me to fiddle with, without having any documentation explaining what is going on.

Seems much easier on the Rigol DS1054Z, where you just telnet to port 5555.

I guess I will base any possible software development on the NI driver framework, which kind of sucks, due to the HUGE software package that has to be installed, with all kind of services ending up running on the system...

Anyway, I stand corrected - if someone can tell me how to "easily" connect to the SSA3000X, please tell me!

Regards,
Vitor

[nctnico]

Just read about TCP/IP. You'll learn the incoming port is fixed and the 'return' port is always somewhere above 1000. This is negotiated by the TCP/IP protocol so no need to worry about that. Try to telnet to port 111 and 897 and see what works (or not).

[DL4RAJ]

I apologize for asking such a simple question : how do I connect a parallel resonant circuit ( capacitor and coil ) tuned to around 1Mhz to a Siglent spectrum analyzer with tracking generator ? Do I need some sort of impedance adapters ?

Just put 50Ohm 10dB pads before and behind the circuit to avoid reflections and/or
impedance transformation on the cables.
This costs you 20dB of dynamic range but you can afford this easily as there are around
90dB DR if you set the SSA to 30Hz RBW.

regards

[nctnico]

I apologize for asking such a simple question : how do I connect a parallel resonant circuit ( capacitor and coil ) tuned to around 1Mhz to a Siglent spectrum analyzer with tracking generator ? Do I need some sort of impedance adapters ?

Just put 50Ohm 10dB pads before and behind the circuit to avoid reflections and/or
impedance transformation on the cables.
This costs you 20dB of dynamic range but you can afford this easily as there are around
90dB DR if you set the SSA to 30Hz RBW.

AFAIK that is called a pi-network. Another option is to use a directional coupler / directional bridge but either way this is getting into network analyser territory.

[alexo]

I apologize for asking such a simple question : how do I connect a parallel resonant circuit ( capacitor and coil ) tuned to around 1Mhz to a Siglent spectrum analyzer with tracking generator ? Do I need some sort of impedance adapters ?

Just put 50Ohm 10dB pads before and behind the circuit to avoid reflections and/or
impedance transformation on the cables.
This costs you 20dB of dynamic range but you can afford this easily as there are around
90dB DR if you set the SSA to 30Hz RBW.

regards

But given that the LC circuit is in parallel with the TG and the SA input, wouldn't this crash the Q of the circuit ?

[TurboTom]

Depends what you want to measure. If it's just the resonance frequency, I would go the route with a return loss bridge and maybe an inductive probe (if your DUT permits magnetic coupling). Otherwise, you can also use two very small capacitors (in comparison to the resonance capacity) to couple into and out of the LC resonator. Or just use an additional winding on the inductor. It's all a question of what parameters you need to figure out and which options / tools you've got. My recommendation would be to just start experimenting with the system and gain experience.

Have fun,
Thomas

[DL4RAJ]

I apologize for asking such a simple question : how do I connect a parallel resonant circuit ( capacitor and coil ) tuned to around 1Mhz to a Siglent spectrum analyzer with tracking generator ? Do I need some sort of impedance adapters ?

Just put 50Ohm 10dB pads before and behind the circuit to avoid reflections and/or
impedance transformation on the cables.
This costs you 20dB of dynamic range but you can afford this easily as there are around
90dB DR if you set the SSA to 30Hz RBW.

AFAIK that is called a pi-network. Another option is to use a directional coupler / directional Bridge...

An attenuation pad can be made either in pi or T-configuration.
A directional coupler/bridge is useless in this application as a parallel resonant circuit
has >>1kOhm (depending on its Q),so the return loss is near zero.
A directional coupler makes only sense when you deal with DUTs not too far from 50 Ohm.

but either way this is getting into network analyser territory.

Yes,but only into scalar network analyser territory which is far less complicate
than vector network stuff.


regards

[Siglent America]

Hi,

Does anyone know the SCPI port for the SSA3000X?

I am considering writing my own software,  but don't want to friend on NI to interface.

Thanks,
Vitor

Hello.

We will have this feature in our next FW (V1.2.8.2) and it is currently being tested. One will then we able to use SCPI to control the SSA3000X directly through LAN without having to use VISA.

We expect the FW update to be tested and ready in January.

[Bicurico]

Great!

Could I then make a wish?

Using i.e. Rigol, one can interface to SCPI through one single port. This even allows using telnet as a client.

With Siglent, the port on the PC side is constantly changing. This may be documented and part of the TCP/IP protocol, but it does make life difficult for people like me, who do not master TCP/IP and want to start in VB .net with basic networking code samples.

The easiest is to just listen to a port. But I do need to know what port to listen to! It should always be the same port!

I think this could lead to some interesting tools to be developed. Take a look at my "VMA Simple Spectrum Analyser" software (http://vma-satellite.blogspot.pt/) and all the cool features I implemented so far. I would LOVE to be able to interface to the SSA3000X in the same manner, to be able to develop my custom measurents.

Thanks,
Vitor

[Siglent America]

I will check on this with the factory.

[nctnico]

Great!

Could I then make a wish?

Using i.e. Rigol, one can interface to SCPI through one single port. This even allows using telnet as a client.

With Siglent, the port on the PC side is constantly changing. This may be documented and part of the TCP/IP protocol, but it does make life difficult for people like me, who do not master TCP/IP and want to start in VB .net with basic networking code samples.

What happens if you telnet to the spectrum analyser? Does that work? If yes, then you shouldn't worry about the return port because that will be handled by the OS.

[Bicurico]

Hi.

With telnet I can send commands and the SSA will switch to remote.
But I won't receive any return data, because it is sent to a different port.

Regards,
Vitor

[nctnico]

Hi.

With telnet I can send commands and the SSA will switch to remote.
But I won't receive any return data, because it is sent to a different port.

So it is a bug!

[Bicurico]

I don't think it is a bug.  The port is previously indicated in TCPIP but I don't know how to handle that.  Wireshark shows that well. And VISA works, too.