Calibrate out Attenuator from VNA measurement

[dmowziz]

Hi..

Using a SVA1015X

Want to measure Open, Short and Load but with an attenuator on channel 1.

Like in this video. from 10:20



I've tried port extension but not working


Thanks for any help

[joeqsmith]

Not sure what the port extension would have to do with it.   He explained what he was doing at about 10:00.  Did you skip through the video?

[dmowziz]

Hi Joe,

Actually skipped the video this time...

But last year, I watched it and tried to do what I'm trying to do now, without success.


Please can you try with your VNAs and confirm?
and also, please what's the minimum power from your LiteVNA?

[joeqsmith]

I've cal'ed with an additional attenuator before and not had a problem.  You don't say what the problem is your are having and I wonder if maybe you're understanding of what was shown is wrong.   Consider providing more details about what you are seeing and what you are expecting and I'm sure someone here can help.   

For the LiteVNA's minimum power output, you would need to define the frequency range and how you would like it measured.   With the LiteVNA supporting harmonics and pulsing the output, I just want to make sure I understand what you are asking for.   

[dmowziz]

Thanks

This is after calibration: SOLT (including the attenuator).

Showing for a Short
For an Open, it is similar but concentrated on the right


The power level from CH0...
If I recall, on the first NanoVna, the power level could be set at 2mA, 4mA, 6mA or 8mA

From the video in #1, the reason for the attenuator is to reduce the power (This is the power i'm asking about) so to keep the amplifier operating in small signal region

[tautech]

Select much higher points ....10001 is max. Turn off averaging.
SVA1032X here.

[dmowziz]

Getting the same thing...

Please did you calibrate with an attenuator?


And I saw from the Siglent thread that you are connected to Siglent.
There was once a firmware update that made the TG Level being able to drop down to -20 dBm

Maybe it's possible to reduce the level below this in the next update?

[tautech]

Getting the same thing...

Please did you calibrate with an attenuator?


And I saw from the Siglent thread that you are connected to Siglent.
There was once a firmware update that made the TG Level being able to drop down to -20 dBm

Maybe it's possible to reduce the level below this in the next update?

Datasheet spec for 1015X and 1032X is -20 dBm ~ 0 dBm
1075X uses different HW for -40 to 0dB.

I can do examples tomorrow as it is late here now but always when using a wide span use more points otherwise their are insufficient for proper interpolation. With early firmware 751 points was max but maybe 2 years back it was increased to 10k.
You can get the latest firmware here:
https://int.siglent.com/download/firmwares/?ProId=30

[joeqsmith]

Thanks

This is after calibration: SOLT (including the attenuator).

Showing for a Short
For an Open, it is similar but concentrated on the right


The power level from CH0...
If I recall, on the first NanoVna, the power level could be set at 2mA, 4mA, 6mA or 8mA

From the video in #1, the reason for the attenuator is to reduce the power (This is the power i'm asking about) so to keep the amplifier operating in small signal region

As he says in the video, the attenuator is to get the system (VNA + Amp) into a region where the VNA would normally be used.  But, yes, you don't want to overload the amp.  When you consider that these low cost VNAs use squarewave drive, harmonic modes, are not a flat response.....  If you want to measure amplifiers with them, it's best to understand what you are doing before you let the smoke out or waste a lot of time making bad measurements.

He shows a 13dB attenuator and sweeping 1M to 30M.   You don't state what value attenuator you are using but you show sweeping from 50M to 1G.   Assuming you are not testing an amplifier, you could start with something smaller.  I assume with a 0dB attenuator, things appear normal. 


20dB_TestAttn:  Showing a VERY low cost 20dB attenuator from China that came as part of a set.  I will use this as my test attenuator that will be attached to Port1.  Much more difficult than what was shown

20dBAttn_1M_1p5G:  Just to show that the attenuator is good,  here sweeping from 1M to 1.5G.  Note, I'm using interpolation rather than recal.   Good enough.

AfterSOLT_Short:  After inserting the attenuator on Port1 (actually at the end of the cable) then running a SOLT at the attenuator, I reinsert the short.   Not great but again, his video is sweeping at much lower frequencies and I have added another 7dB.   Basically my thought was to show you in your particular case of sweeping to 1GHz, how my LiteVNA compares.   Also, note that I am using minimal IFBW.   The LiteVNA defaults are actually lower.

peak_min_power:
I am not aware of a way to change the output level on my original NanoVNA outside of adding attenuators.  The V2Plus4 had hardware to support it but I don't believe they ever implimented the firmware for it.   The LiteVNA actually does allow you to change it.  I suspect you didn't elaborate on what you wanted to see as far as the minimum power as you just don't know.   So, let's assume because you are only running up to 1GHz, we don't need to be concerned with harmonics.   And lets assume you just want to know the peak levels with the gain of both synthesizers set to their lowest values.   This image is with my spectrum analyzer using the max hold.    I have an external 10dB attenuator at the SA's input port.  Sadly, I could never get the people at Signal Hound to add a feature to allow the software to compensate for this. 

The output starts around -10dBm.  You can see as the LiteVNA selects the second synthesizer, we loose a bit more power.   Also note that the level at 1.5GHz is not the same as at 200MHz. 

Attached video showing an amplifer I made specifically to go unstable when tested with these low cost VNAs.   All things you would need to be aware of.   

[joeqsmith]

Just for comparison with the video, I repeated the test using a 13dB attenator sweeping 1M-30M.  I left the IFBW and also left the output level set to its minimum.  Changing these two setting would improve it but I wanted to show worse case condition as I really don't know anything about the VNA shown in the video. 

[dmowziz]

@Tautech

I think there was a time that the TG power level was fixed at 0 dBm ?
If you look at the revision note :
https://siglentna.com/wp-content/uploads/dlm_uploads/2022/06/SVA1000X-Firmware-Revise-History.pdf

Version 2.2.1.2.7       2020/05/12
"Port 1 power (TG) adjustable in VNA mode"

Anyways... Thanks!! I saw your antenna thread some weeks ago and was shocked to see you could get about -40 dB S11. My connectors were not great and I didn't know.





Thank you Joe for the tests, it has cleared up some things... and the video... and it was/is a 20 dB attenuator like yours

About calibration (and de-embedding?). For example, the VNA's sine wave output is at 0 dBm. (this will overload an amplifier and not give the correct small signal s parameters of the amplifier).

If the s parameters of an (attenuator + transistor) is measured. and the s parameters of the attenuator is known. Is it possible to use math (T matrix?) to get the small-signal s parameters of the amplifier?)

Btw, the error terms you mentioned, guessing they are in the CAL file. I tried opening a CAL file with notepad but gibberish. Please how can it be opened?
Thanks

[tautech]

@Tautech
Anyways... Thanks!! I saw your antenna thread some weeks ago and was shocked to see you could get about -40 dB S11. My connectors were not great and I didn't know.

Which one ?
The antenna project or the pissing around characterising a few ?

I normally use Port 1 (TG) on default stimulus settings so for a -40dB Log Mag display indicates great antenna emissivity. Then pop a continuous peak Valley marker on it and you can see instantly if it meets the target frequency.
When tuning them I turn Continuous to OFF so as you move around it tuning you're not always fighting with proximity effects.

I could still so a SOLT Cal if you want using a 10dB N type feedthrough and our F603FE SMA Cal kit.

[joeqsmith]

Thank you Joe for the tests, it has cleared up some things... and the video... and it was/is a 20 dB attenuator like yours

I would have included that detail in your first post. 

About calibration (and de-embedding?). For example, the VNA's sine wave output is at 0 dBm. (this will overload an amplifier and not give the correct small signal s parameters of the amplifier).

Guessing you're not suggesting this is true for all amplifiers but just the one you are attempting to work with.  Normally, I would have some idea about the amplifier and add attenuators to its output and a step attenuator on the input.  Manually sweeping the amplifier's input to get a feel for where it starts to compress and what the gain is.  I've not yet damaged any of my VNAs but you can imagine playing with active circuits certainly increases the odds.   

If the s parameters of an (attenuator + transistor) is measured. and the s parameters of the attenuator is known. Is it possible to use math (T matrix?) to get the small-signal s parameters of the amplifier?)

That was pretty much the point of the video you linked in your first post.  Keeping in mind, you may end up having attenuators on the output of the PA as well.  I've seen posts from hams who have connected their VNAs directly to their radios ending in a dead VNA and then wondered why.   I've made a few videos showing the low cost VNAs to look at amplifiers.  The one linked below covers a lot of ground.  Amplifiers start around 30 minutes in.

Btw, the error terms you mentioned, guessing they are in the CAL file. I tried opening a CAL file with notepad but gibberish. Please how can it be opened?
Thanks

Someone has a similar question about the recorded data file format.  For that, I now support converting the data to CSV.   Normally I do not use ASCII formatted data.  The CAL file is binary.  The raw data that is collected during the calibration is stored in this file.   The error terms are calculated using this data.   

[dmowziz]

Hi tautech,

It's this : https://www.eevblog.com/forum/rf-microwave/antenna-project-log/

Please what simple antenna can I make for 88 - 108 MHz that will show a low return loss in this band?

[tautech]

Hi tautech,

It's this : https://www.eevblog.com/forum/rf-microwave/antenna-project-log/

Please what simple antenna can I make for 88 - 108 MHz that will show a low return loss in this band?

Just a quick Google search returned this:
https://www.matchmaster.co.nz/digital-tv-antennas/03mm-1efm/

Recommend you spend some time asking Mr Google.

[dmowziz]

That's a product to buy

Something small to make, maybe with microstrip?, solenoid?


Thanks thanks. Great recommendation

[tautech]

That's a product to buy

Yes of course however its description lets you hunt and learn more about the design: ring-type dipole antenna