Siglent SSA3000X and SSA3000X-Plus Spectrum Analyzers

[leix_99]

Seems like rf-loop has some particularly good specimen of this analyzer - maybe some early hand-crafted/polished prototype

If I look at your measurements @ 2.1GHz and compare it to the specifications, then I get this:

RBW=10kHz, Preamp=on Measured: -119dBm; Specified: <-115dBm, typ. -118dBm; >typ.
RBW=10kHz, Preamp=off Measured: -101dBm; Specified: <-95dBm,   typ. -99dBm;  >typ.
RBW=1MHz,  Preamp=on Measured: -98dBm;   Specified: <-95dBm,   typ. -98dBm;  =typ.
RBW=1MHz,  Preamp=off Measured: -80dBm;   Specified: <-75dBm,   typ. -78dBm;  >typ.

As far as I can tell your unit is not only within the guaranteed specifications, but also meets or even exceeds the typical ones.

Okay, I just had the feeling something was wrong and then seeing those screenshots made me wonder even more. It's my first SA and by far the most expensive piece in my amateur set-up, I guess I'm just concerned about its state since my experience with this devices is somewhat nonexistent.

Nevertheless, I appreciate the reply, I was at work and felt extremely relieved until rf-loop's response 

I know this reply dates quite a while, but I received my SA last week and it "seemed" to be a previously used/returned unit (no screen plastic, really small dent on the casing).

After unpacking the unit, noise floor curve looked quite higher than the average ones I saw in videos/reviews but the image you posted makes me think there's something wrong with my SA.

I've used the same parameters you used to do a 2.1gig sweep, RBW/VBW 1kHz, all traces average (same colour/trace per measurement), sweep mode "sweep" and input 50ohm terminated. Check this out:

 

Also, the TG flatness (which I think it should be somewhere less or around 3dB) is quite off to. I'm not that concerned about that since the cables I had lying around are long and old but still, I have a bad feeling about this particular unit...

Has anyone had a problem like this with Siglent instruments?

I think it need look bit more deeply. 
I have one concern from what I see in your image. But I do not tell it yet before some further knowledge about your individual unit. (these traces are still in specs but....)

But first I need know do you have any reliable trusted signal source, pure sine wave and  example between 5 - 50MHz what level you really KNOW so that absolute error is least less 1 dBm. Something between -30dBm to 0dBm.

Measure some these known level signals with SSA and note if level error is inside this method reliable limits or clearly outside.


And note, RF input is extremely sensitive for EOS/ESD and it happen easy! You do not feel any "static electric" and still it can be damage level. This is common to all, not for just you.

Even if you do not have reliable known level source:


1.


Set  factory defaul. RF input terminated
Set trace A Clear write
Set trace D Clear write
Set Detector Trace A: "Average Log Pwr"
Set Detector Trace D: "Pos Peak"
Set RBW=VBW=10kHz
Set Attenuator 0dB
Set Ref level -50dBm

Take this kind of  image, SA continuously running






2.


Set SA for factory preset (defaults)
Connect SSA own Ref Out (10MHz) to RF input.
Set Atten 30dB
Set Ref level +10dBm 
Set center Frequency 10MHz
Set RBW 300kHz
Press  Peak (marker show now peak value, note this value)

Take this kind of image, SA continuously running

Then if you want at this time do more
Set factory defaults
connect again 10MHz ref out to RF in.
Set center freq 10MHz, Span 10kHz (I believe you see 1kHz RBW)
Set attenuator 30dB and reference level +10dBm
Press Peak and select also in this menu Cont Peak ON
Note marker level
Press System
In this menu select Calibration
There open selection AutoCal  and it is "Close"
Turn it "Open" and note how much your marker level change.
Write down "Open" value and "Close" value of marker level.
 
Tell these both values.

Quoting Mr. Snow - Everything before the word "but" is horse shit -

I have a couple signal generators but they are both too old and too slow, plus the output level is knob-set... That's a no-go.

So far I've barely used the device (just plugged it in, did the noise floor measurements, and some return loss testing on an antenna). Always at really low output levels and with protection (dc-block on the input, 6dB attenuator on the output of the TG), just in case.
Concerning an ESD event, I have the place all covered in ESD mats, and I regularly discharge myself, but there's always the possibility...

Okay enough, I'll cut to it:

STEP 1

Set  factory defaul. RF input terminated [DONE]
Set trace A Clear write [DONE]
Set trace D Clear write [DONE]
Set Detector Trace A: "Average Log Pwr" [DONE]
Set Detector Trace D: "Pos Peak" [DONE]
Set RBW=VBW=10kHz [DONE]
Set Attenuator 0dB [DONE]
Set Ref level -50dBm [DONE]




STEP 2

Set SA for factory preset (defaults) [DONE]
Connect SSA own Ref Out (10MHz) to RF input. [DONE]
Set Atten 30dB [DONE]
Set Ref level +10dBm  [DONE]
Set center Frequency 10MHz [DONE]
Set RBW 300kHz [DONE]
Press  Peak (marker show now peak value, note this value) 2.28dB



STEP 3

Then if you want at this time do more
Set factory defaults [DONE]
connect again 10MHz ref out to RF in. [DONE]
Set center freq 10MHz, Span 10kHz (I believe you see 1kHz RBW) [DONE]
Set attenuator 30dB and reference level +10dBm [DONE]
Press Peak and select also in this menu Cont Peak ON [DONE]
Note marker level 2.34-2.35dB Note marker freq. is slightly lower, it varies between RBW, i guess it's normal but worth mentioning
Press System [DONE]
In this menu select Calibration [DONE]
There open selection AutoCal  and it is "Close" [DONE]
Turn it "Open" and note how much your marker level change. [DONE]
Write down "Open" value and "Close" value of marker level.lOPEN: 2.35dB | CLOSE: 2.28dB

CLOSE: 2.35dB



OPEN: 2.28dB




Okay, that was it, I tried to eagerly follow as accurately as I could.

Thanks guys for the time, it's quite hard to get nice and elaborated replies when it's your first post 

[rfspezi]

Short (maybe stupid) question...
I would like to verify the DANL of my device.
In the specifications table, they write "detector = sample detector".
Wouldn't "detector = average Log Pwr" make more sense?

[rf-loop]

Short (maybe stupid) question...
I would like to verify the DANL of my device.
In the specifications table, they write "detector = sample detector".
Wouldn't "detector = average Log Pwr" make more sense?

In Siglent specifications there read bit more: sample detector, trace average >50
Exercise:
Set center freq 250MHz, Span 100MHJz. Atten 0, ref level -50dBm, RBW=VBW=1kHz
Set trace A for sample detector. set trace A first for Clear write
Set trace B for average Log Pwr, no trace average.
Set Marker 1 trace B (just normal marker)
Set Marker 2 trace B abs go to Marker Fn, set marker 2 function as Noise marker
Marker menu, set marker table on.
Look values and look also how Sample and Average LogPwr tracews differ.
Set trace A Average times 50 and leave it alone. After some time come and look result.

About visible noise level (trace on the screen):
THis what you see is after gaussian type RBW. But also it is displayed noise.
Now all know that 1000Hz RBW  and 1Hz RBW difference is 30dB  if look noise.
But it not right for noise dBm/Hz  because this mean 1Hz wide noise after square filter, not after gaussian shape filter. If normal marker there read  example -118dBm it can not convert for noise level just subtracting 30dB.

Now look again markers. Noise marker and normal marker difference is not 30dB, it is 27.5dB. Simple story can read in this Keysight pdf what also explain lot of how SSA3000X works.



Keysight AN-150
http://literature.cdn.keysight.com/litweb/pdf/5952-0292.pdf


-----------
@leix_99

This do not proof all is ok but in these images I can not see this kind of signs of damaged front end what I have previously seen. This tiny test result: all looks like ok. 

Why I previously slightly suspect something and like see bit more. You told that perhaps this your unit is not exactly factory new. Other thing was that base noise looks bit different what I have seen in many units - perhaps older ones.
Also there is not visible these Spur signals what have been typical and what I also bit wonder - perhaps they have done some improvements in HW.


If front GaAs RF switch is damaged it typically leads to wrong signal levels, perhaps noise level wrong but also I have seen big difference between AutoCal Open and Close.

What is your manufacturing date (or cal cert date). It looks like it have perhaps improved HW what I have.
Just do not look this base noise level with magnifier. It is less important what you perhaps think when we are talking this kind of  levels, example what can see in your images, in this kind of SA. There is lot of much more important things.

[leix_99]

Short (maybe stupid) question...
I would like to verify the DANL of my device.
In the specifications table, they write "detector = sample detector".
Wouldn't "detector = average Log Pwr" make more sense?

In Siglent specifications there read bit more: sample detector, trace average >50
Exercise:
Set center freq 250MHz, Span 100MHJz. Atten 0, ref level -50dBm, RBW=VBW=1kHz
Set trace A for sample detector. set trace A first for Clear write
Set trace B for average Log Pwr, no trace average.
Set Marker 1 trace B (just normal marker)
Set Marker 2 trace B abs go to Marker Fn, set marker 2 function as Noise marker
Marker menu, set marker table on.
Look values and look also how Sample and Average LogPwr tracews differ.
Set trace A Average times 50 and leave it alone. After some time come and look result.

About visible noise level (trace on the screen):
THis what you see is after gaussian type RBW. But also it is displayed noise.
Now all know that 1000Hz RBW  and 1Hz RBW difference is 30dB  if look noise.
But it not right for noise dBm/Hz  because this mean 1Hz wide noise after square filter, not after gaussian shape filter. If normal marker there read  example -118dBm it can not convert for noise level just subtracting 30dB.

Now look again markers. Noise marker and normal marker difference is not 30dB, it is 27.5dB. Simple story can read in this Keysight pdf what also explain lot of how SSA3000X works.



Keysight AN-150
http://literature.cdn.keysight.com/litweb/pdf/5952-0292.pdf


-----------
@leix_99

This do not proof all is ok but in these images I can not see this kind of signs of damaged front end what I have previously seen. This tiny test result: all looks like ok. 

Why I previously slightly suspect something and like see bit more. You told that perhaps this your unit is not exactly factory new. Other thing was that base noise looks bit different what I have seen in many units - perhaps older ones.
Also there is not visible these Spur signals what have been typical and what I also bit wonder - perhaps they have done some improvements in HW.


If front GaAs RF switch is damaged it typically leads to wrong signal levels, perhaps noise level wrong but also I have seen big difference between AutoCal Open and Close.

What is your manufacturing date (or cal cert date). It looks like it have perhaps improved HW what I have.
Just do not look this base noise level with magnifier. It is less important what you perhaps think when we are talking this kind of  levels, example what can see in your images, in this kind of SA. There is lot of much more important things.

Thanks again, having the opinion of someone who knows this instrument makes me feel better

Regarding the HW revision:

Calibration sheet dates november 2017, HW revision on the SA is 07.03.00 and FW version 1.2.8.5a.

Time to do some experiments then!! 

 

[KevinNZ]

Hi,
New forum member here.
Just received my 3021x yesterday. Repaired instruments including Spec ans early in my working life. But used them for the last 30+ years.
Getting my head around some of the features of the unit.
Did some IMD tests with my SG1062Z (Rigol..why does the Siglent catalogue show something similar...lol)
The use of attenuation around the mixer tracks quite nicely..and the RF switches (Macom) just start to show their performance limitation at the top end of the attenuation (10 dBm tones and 45 to 50 dB attenuation), which is right on their spec.
I reduced the attenuation by 10 dB and put external 10 dB attenuator in and saw several dB improvement.
Still getting my head around using the Siglent user interface.
I notice when using the tracking generator and normalising the display..when leaving the TG mode to look at noise levels..it retains the normalised offset, which is a little confusing..so have to remember to switch normalizing off, also when switching The TG off.
But impressed with the package.
Also have an early Signalhound SA44 which has some useful features.

[tautech]

Hi,
New forum member here.
Just received my 3021x yesterday. Repaired instruments including Spec ans early in my working life. But used them for the last 30+ years.

Still getting my head around using the Siglent user interface.
I notice when using the tracking generator and normalising the display..when leaving the TG mode to look at noise levels..it retains the normalised offset, which is a little confusing..so have to remember to switch normalizing off, also when switching The TG off.
But impressed with the package.

Welcome to the forum Kevin.

After our short emails it seems you've got your head around the UI some. 
This is the right place to ask if you have other questions that I can't answer directly/fully/promptly.
Glad you like it as they've been quite popular.

[tautech]

New firmware for SSA3kX models.

V1.2.9.1
~7.5 MB
https://www.siglentamerica.com/download/6019/

Changelog.
1. This firmware must be upgraded from V1.2.8.1 or later. If your SSA3000X has an earlier version, please update
    to V1.2.8.1 first and then proceed to update to the latest version.
2. Fixed bugs on peak search and next peak.
3. Optimize SCPI commands for peak table, screenshot, etc.
4. Optimize the LOG X-scale for EMI test.
5. Optimize the Quasi-Peak detector accuracy for EMI test.

[Pinkus]

Also a new EasySpectrum 1.0.5.0 (usually a firmware update comes with a new Easy Spectrum version):
https://www.siglentamerica.com/download/6022/

Changes fox SSA3000X:

• Add log scale for pre-scan process in EMI test
• Optimize the save and recall functions for .cor file and .lim file
• Optimize the interpolation for Correction and Limit
• Optimize some UI

[Ghislain]

Upgrade from 1.2.8.5 worked like a charm 

[Orange]

It seems that with this 1.2.9.1 version, one annoying 'feature' is also solved.

There have been several reports, that the TG would switch off after some hours of use. The only way to get it ON again was to power down the SA completely.
I had this myself as well. After a few days of letting it run in several modes this does not happen any longer.

With this new SW the thing keeps working 

[tautech]

It seems there might be some problem with the latest V1.2.9.1 firmware.
Been reported and soon to be checked on a new unit as well.

hello.

I update from 1.2.8.5a  to 1.2.9.1  and lost the service info to.??  Bugs?   Never tried to hack the SSA3021.
Get it few days ago.

Someone tried to only activate the the options on this last firmware? It was delivered with TG

Anyone else ?

[markus_jlrb]

Dears,

I have updated my SSA3021 box to FW version 1.2.9.1 (from FW 1.2.8.3)
While checking the BW and RBW I observed an issue.

When switching BW from 10Hz to 3Hz I observed a significant drop of noise
level that not fit to the reduces new bw. I check this at 1GHz center and
span of 1MHz and 100kHz.


See the attached screen shot.

Can you confirm this observation or point me the reason for this curiouse
behavior.

Many thanks in advance.


Markus

[markus_jlrb]

Dears,

FYI

Yesterday I was doing the frequency calibration task according to the thread listed in the link below

March 25, 2018, 10:02:10 pm
https://www.eevblog.com/forum/testgear/siglent-saa30xx-frequency-calibration/msg1461231/#msg1461231

After two modifications of the file cali_freq_dac in the firmdata0/calib directory
I was able to match the displayed peek frequency to 2Hz precision related to my
HP  53181A reference.

The attached pictures show the original factory settings ref freq. from the SSA on my HP counter display
and the HP counter internal ref frequency on the SSA screen after peek search as well as the reduced
freq. delta after two adjustments inside the cali_freq_dac file.

Markus

PS.: Counter and SSA running for warming up over two hours.

[markus_jlrb]

Any ideas concerning my observation with the new firmware  in my posting #761 above.

Markus

PS.: The drop in the noise level below -148dBm after changing the RBW was visible on the
whole display. To understand the effect better I posted the attached screen shot im may previous
posting imediatelly after changing the RBW to demostrate this effect. But when the sweep is
on the right side of the screen the entire x-Achse show the low level and no ripple anymore.

I hope you could understand my explanation about the observer issue.
 

[Dubbie]

Marcus,

I think I got lucky.
Mine came exactly dead on to matching my 53181A locked to my GPSDO. (after 30 mins warmup for everything). no need for further calibration.

[markus_jlrb]

Dear Dubbie,

could you please verify the issue I described in my previous thread (from May 20, 2018, 06:56:05 am),
means the behavior of the SSA when changing the RBW from 10Hz to 3Hz if you have a SSA with the
newest FW (1.2.9.1).

Thanks

Markus

 

[Dubbie]

I will do so tomorrow.

[Dubbie]

could you please verify the issue I described in my previous thread (from May 20, 2018, 06:56:05 am),
means the behavior of the SSA when changing the RBW from 10Hz to 3Hz if you have a SSA with the
newest FW (1.2.9.1).

Hi Marcus,

I have done as you requested, and can confirm that I see the identical behavior to your unit.

[rf-loop]

Not so seriously becausae this is.....and so on... :


Sad, There in SSA is not enough noise...     
(or over 16bit ADC if it is 16bit)

Perhaps there need inject more noise for more nice cosmetics on the screen and for avoid wonderings...

16 bit adc theoretical FS range is 96dB

Perhaps there Siglent can do some adjustments for optimize ADC autoranging for more nice cosmetics with under 10Hz RBW ...... or (joke) just reject user for make settings what go somehow  over dynamic range...

Perhaps this image helps smiling...

[markus_jlrb]

Dear Dubbie,
dear rf-loop,
dear forum,

let my summarize my observation again to make things more clear.
Maybe rf-loop misunderstood my way of thinking and the way I
report my observation.

After the SW update to version 1.2.9.1 I mounted a 50 Ohm Shuner terminator
at the SSA input in order to watch the spectrum. At the center of 1GHz I found
a -110dBm carrier and wonder where he came from.

I decrease the RBW and SPAN and while I set the SPAWN to 1MHz or 100kHz
and change the RBW from 30Hz over 10Hz and finally to 3Hz I observed the drop
of more then 25dB in noise level. But the noise level was to smooth and the
droping delta did not fit the formula 10*log(10Hz/3Hz) as   -5,2dB that should be
and what I would expect.

So I understood the digital IF RBW response as present in the Keysight picture
as a technical behavior due to the measurement technique but not the amount
of noise drop.

@rf-loop
Please give me a more datailed explenation to understand this observation if
you still insist that this SSA behavior is in the range of spec and not a FW feature ;-).

Many thanks in advance for your effort as well as to Dubbie for the reproduction
and confirmation of ihis issue.

Markus

[Safar]

Looks like bug. Some guys report of disappearance of Last Cal Date. May be something wrong with calibration data also (not date) with 1.2.9.1?

May be try to feed some signal to SA and look for level on different RBW.

[kerouanton]

Hello,

As mentioned here*, I'm currently renewing my hobby equipment (old Hameg 203, etc.) with Siglent gear. My first piece is a SSA3032X, which I like very much.

I noticed a minor annoyance that could be easily fixed on a future release of the firmware.

I like the "Display / Screen Text=Off" feature, but when selected, it doesn't display at all anything we enter on the numeric dialpad.
It would be very convenient to display what is typed, and once validated hide it, as it would avoid typing mistakes for being blind... Currently I need to switch on the Screen/Text feature, type my digits, and switch back off the text, which is a bit long.

My 2 cents.

However this is a great piece of hardware.


* https://www.eevblog.com/forum/testgear/siglent-sds1204x-e-released-for-domestic-markets-in-china/msg1568293/#msg1568293

[videobruce]

Has this finally fixed that RBW/VBW manual adjustment problem the keeps slowing down manually set speed speeds?? 

[kerouanton]

Hi,
I'm thinking about adding the RBSSA3X20 reflection bridge to my new SA3032x, but I have a few novice questions:

- Has anyone done a review of it? (I've seen the good TG introduction video here https://youtu.be/0Am3itqZ5LY but it doesn't mention the bridge).

- What are the main uses ?

- Is is anyway worth buying one considering I have a DG8SAQ VNWA, which I'm pretty happy with?

[hendorog]

Hi,
I'm thinking about adding the RBSSA3X20 reflection bridge to my new SA3032x, but I have a few novice questions:

- Has anyone done a review of it?

- What are the main uses ?

- Is is anyway worth buying one considering I have a DG8SAQ VNWA, which I'm pretty happy with?

The RBSSA3X20 bridge, is mechanically very good in the way it fits the SSA, however it does not have very good directivity.

IIRC the directivity spec is only 20dB. I have had a play with one and I am not convinced it meets that spec across the entire range.

If the directivity is 20dB, then you can measure return loss to 10dB to about +/- 3dB accuracy which is not great.
https://www.markimicrowave.com/assets/appnotes/directivity_and_vswr_measurements.pdf

The directivity of the bridge is not corrected in a scalar system (i.e. Spectrum Analyser + TG). This is different to a vector system (i.e. VNA).
So for use with the SSA + TG, for the best accuracy you want the best directivity. If you are just testing antennas for go/nogo then it's probably going to be fine. Otherwise you want something better.

Since you already have a VNWA so I'd just use that instead.

If you do want a good bridge, Mini Circuits have some which are better and cheaper, and there are plenty on eBay. Or you can DIY. If you do get a better bridge, then you need good cables and adapters as well as a good termination to test it of course Return loss nuttery is beckoning