SOLVED-- SVA1032X -- Why is VNA noisier than SA-TG same measurement?

[NY2KW]

I just got a nice deal on a Eagle RLB150x5 rated beyond 3GHz for my SVA1032X and wanted to check it out.  First test is always to make sure previous owner didn't zap it, so using the SA TG I wanted to check the specs on the internal loss for a maximally reflected source.  The specs say if not zapped, with a good 50 ohm load (mine is a Narda to 12GHZ) , the reflected signal should be -35 to -40 dB or better from normalized.  So TG (0 dbm) to Source port, SA to Reflected port normalized with 50ohm load shown in first two pictures.




Then I did the same in VNA mode with S21 Log Mag format, which should be the same signal paths, keeping everything the same:

[ Specified attachment is not available ]

Why is the trace for the same load so much noisier in VNA mode, this is averaged - it as even noisier than SA-TG mode given all conditions are otherwise equal?   There seems to be better sensitivity but more noise with VNA mode.  Aren't the internal paths for this measurement identical?

[bson]

Is it actually noise, and not ripple?

[tautech]

I think it's an interpolation issue with just 201 data points (default). Change it to the max 701 751 points and the interpolation result becomes more accurate.
Post another screenshot with 751 points please.

Edit, corrected max # of data points.

[NY2KW]

Makes sense, will change points and post new image.  Any way to change the default # points?

Jerry

[tautech]

Makes sense, will change points and post new image.  Any way to change the default # points?

Jerry

Not near a SVA but IIRC it's in one of the top row of buttons or in the Meas menu. If you do a full Preset to factory default it returns to 201 points.

[hendorog]

To compare the paths while using the same amount of averaging in both cases, try this:

SA Mode:
Set attenuator to zero (Edit)
Leave RL Bridge test port open
Set averaging to 100
Normalise once averaging reaches 100
Attach 50 ohm load and restart averaging
Set trace to View once averaging reaches 100

Switch to VNA mode:
Remove load from test port
Set averaging to 100
Save trace to Mem once averaging reaches 100
Select Math, Data/Mem
Attach 50 ohm load and restart averaging
Once averaging reaches 100 compare the two traces.


Edited to add attenuator setting.

[NY2KW]

This wont work with Attenuator set to zero.  I need to set the SA TG to 0 dBm because I believe the VNA sets the TG output is fixed at zero dBm.   I get a over power beep with TG=0 dBM and Att=0.  I will try it with Att=5.

[hendorog]

This wont work with Attenuator set to zero.  I need to set the SA TG to 0 dBm because I believe the VNA sets the TG output is fixed at zero dBm.   I get a over power beep with TG=0 dBM and Att=0.  I will try it with Att=5.

No problem - that was just a fine tuning.

[NY2KW]

OK here is the SA, TG=0 dBm, with DUT open (full reflection), averaged 100x then used this to normalize, then added precision 50 ohm load to DUT



then did same with VNA, and looks different



Any ideas what might be going on?  Can you try this on your SSA or SVA with an RF bridge?

[NY2KW]

for some reason, the images are displayed in wrong sequence but are correctly labeled

[NY2KW]

I have an older RF/SWR bridge from Amtronics up to 1.5GHz.  I will do the same comparison in morning and post.  Need to know if this is a quirk of the SVA or the Rerun Loss Bridge

[NY2KW]

I am no expert, but by process of elimination it seems most of the S21 noise in VNA mode comes from just having an open port near the bottom of the noise floor.   

Here are my comparisons:

First, in SA mode, TG= 0 dBm, RG405 4” cable directly from Port 1 to Port2 and the SA displays a very low noise trace showing signal at 0 dBM for 1 MHz RBW, full span.  Doing same in VNA mode gives near identical trace.

Next I placed two Mini-circuits 18GHz 20 dB attenuators onto each Port and connected Port 1 to Port 2.   I did this in SA mode and VNA mode and the Thru traces look pretty clean and almost the same in both modes

VNA 40dB THRU

I wanted to see the noise floor under very low signal, so I added an additional 10 dB attenuator to each Port (total 60 dB)
The SA Thru trace looks good though for the BW setting, pretty flat, low noise across the full span not quite getting down to -60dBM. 

On the other hand, doing the same in VNA yields that very noisy trace (even averaging 100x with 700 points). 
This looks like the noisey trace I saw when checking out the reflection on my Eagle RLB:

VNA 60dB THRU 700pts


Next I just removed the cable from Port 1 and left the 60dB of attenuators connected to Port 2 and you can see most of the noise is still there even without a input signal:

[ Specified attachment is not available ]VNA 60dB NO TG

The SVA1000x datasheet says there should be a minimum dynamic range for the SVA 60 dB between 10MHz – 1.5 GHz and 55 dB above 1.5GHz.  The Trace noise should be < 0.1 dB rms with 50 averages. 

So I am puzzled if this is considered out of spec.  Can anyone else with a SVA1000x try to duplicate my observations?

[NY2KW]

I don't know why my uploaded images are getting scrambled in sequence and some missing  Here is the VNA with 40dB attenuation Thru.



As you can see, the noise only begins at the lowest limits of TG input

[eb4fbz]

S21 ripple due to reflections, because analyzer ports doesn't present 50ohm.

Check placing a good 10dB attenuator right at the nanoVNA port 2 and calibrating with it in place. If my theory is right, the measurement ripple will be lower. However the S21 dynamic range will be 10dB lower too.

[NY2KW]

Port 2 (input) in VNA and SA mode is specified as being 50 ohm input.   The easiest way to replicate my observations is:

1) In SA mode (no TG connected !)  - connect a precision 50ohm load rated > 6 GHz load directly to the Port 2 connector.  Set SA full span, 1 MHz RBW, TG OFF and average about 10 sweeps.  You will find a relatively smooth trace that goes from -60dBm at low end up to about -55dBm on the 3 GHz end.   

2) Now switch to VNA mode, S21, with same 50 ohm load on Port 2 and average the traces.  You will see a trace spiking up and down ranging from -50 to -80 dBM.  Calibration doesnt help and when using a bridge or coupler and connect the TG if the reflected or Thru signal you are trying to measure is less than -40 or -50 dBm then all you see is this noise.  Doing the same identical type of measurement in SA mode, even with TG attenuated down to -60 or more show a nice clean trace.

I dont know if its a problem with my SVA or something intrinsic to the design but I ahve sent it all to Siglent Tech support and waiting to hear back.

[TheSteve]

I've never used an SVA1032X but in VNA mode what are you using for IF bandwidth or equivalent? I don't think you can expect the same trace in both SA and VNA modes. You should also be terminating port 1 for any testing.

[tefe]

The TG OPEN floor is actually the SA noise floor, even if it is normalized.
But VNA S21 OPEN floor is a ratio of SA noise floor/TG output full reflection, so it will be the shape of TG output full reflection.
If this box has an access of Port1 reflection, I think you will get the truth.

[NY2KW]

But VNA S21 OPEN floor is a ratio of SA noise floor/TG output full reflection, so it will be the shape of TG output full reflection.

Can you elaborate that a little more?  Why does the SA+TG mode Thru to Port 2 traces compared to VNA mode S21 traces look virtually identical in terms of noise up to attenuating the TG (Port 1 ) signal to just about below -50dBM - at that point in VNA mode the noise increases dramatically with flux +/- 30 dBM ( from -50 to -80).  In SA+TG mode the Thru trace looks same low noise well below -60dB attenuation.     

[NY2KW]

I've never used an SVA1032X but in VNA mode what are you using for IF bandwidth or equivalent? I don't think you can expect the same trace in both SA and VNA modes. You should also be terminating port 1 for any testing.

One of the limitations of the SVA1032X is you cannot adjust the TG output level in VNA mode, its fixed at 0 dBm and as far as I can see the VNA bandwidth seems the same as the default SA (1 MHz) when full span, at least by sweep speed and trace noise -- up to about -50to -55 dBm input when the trace falls apart with noise flux ranging from -50 to -80 dBm. 

[TheSteve]

I've never used an SVA1032X but in VNA mode what are you using for IF bandwidth or equivalent? I don't think you can expect the same trace in both SA and VNA modes. You should also be terminating port 1 for any testing.

One of the limitations of the SVA1032X is you cannot adjust the TG output level in VNA mode, its fixed at 0 dBm and as far as I can see the VNA bandwidth seems the same as the default SA (1 MHz) when full span, at least by sweep speed and trace noise -- up to about -50to -55 dBm input when the trace falls apart with noise flux ranging from -50 to -80 dBm.

The datasheet mentions an IF bandwidth of 10 kHz for the VNA. It also lists a dynamic range of 10 kHz - 10 MHz @ 75 dB, 10 MHz - 1.5 GHz @ 60 dB and 1.5 GHz - 3.2 GHz @ 55 dB. So what you're seeing could be normal. The dynamic range really isn't too great.

[tautech]

You should also be terminating port 1 for any testing.

Nope, SVA models have just a single active port in VNA mode; Port 1 (also TG for SA mode)
DUT reflections are then passed back through port 1 to an internal coupler that's not in the signal path in SA and TG ON mode.
In VNA mode just port 1 is used for basic antenna and DTF measurements.

Examples of just the single VNA port 1 usage are here:
https://www.eevblog.com/forum/rf-microwave/what-really-is-this-antenna/

[TheSteve]

You should also be terminating port 1 for any testing.

Nope, SVA models have just a single active port in VNA mode; Port 1 (also TG for SA mode)
DUT reflections are then passed back through port 1 to an internal coupler that's not in the signal path in SA and TG ON mode.
In VNA mode just port 1 is used for basic antenna and DTF measurements.

Examples of just the single VNA port 1 usage are here:
https://www.eevblog.com/forum/rf-microwave/what-really-is-this-antenna/

How do you measure S21 with only a single port?

[tautech]

You should also be terminating port 1 for any testing.

Nope, SVA models have just a single active port in VNA mode; Port 1 (also TG for SA mode)
DUT reflections are then passed back through port 1 to an internal coupler that's not in the signal path in SA and TG ON mode.
In VNA mode just port 1 is used for basic antenna and DTF measurements.

Examples of just the single VNA port 1 usage are here:
https://www.eevblog.com/forum/rf-microwave/what-really-is-this-antenna/

How do you measure S21 with only a single port?

You don't.
Watch TSP from 33 mins in:
https://youtu.be/ToVJTKCyIU8?t=1979

Take note of the changes from VNA to SA mode for some measurements.

[TheSteve]

You should also be terminating port 1 for any testing.

Nope, SVA models have just a single active port in VNA mode; Port 1 (also TG for SA mode)
DUT reflections are then passed back through port 1 to an internal coupler that's not in the signal path in SA and TG ON mode.
In VNA mode just port 1 is used for basic antenna and DTF measurements.

Examples of just the single VNA port 1 usage are here:
https://www.eevblog.com/forum/rf-microwave/what-really-is-this-antenna/

How do you measure S21 with only a single port?

You don't.
Watch TSP from 33 mins in:
https://youtu.be/ToVJTKCyIU8?t=1979

Take note of the changes from VNA to SA mode for some measurements.

I know, lol
I thought he was looking at the S21 noise floor with port 2 terminated. If that is the case port 1 should also be terminated as RF is being generated there when looking at S21.

[tautech]

I know, lol
I thought he was looking at the S21 noise floor with port 2 terminated. If that is the case port 1 should also be terminated as RF is being generated there when looking at S21.

This is where I reach the limits of my VNA understanding with my somewhat limited single port experience however as I understand it a S21 measurement is obtained from the stimulus generated from port 1.
Am I wrong ?