Hi Erik and all

I updated the report se delete the one I published yesterday.
I now get the same OIP3 figures as you. The levels must be much higher than the -30dBm I used to generate the better OIP3 figures .
However by using a 10dB internal Atten the figures raises 10dB now the TinySA is acting more linear.

The TinySA.exe also show good result on OIP3 but it is needed to toggle the RBW as repeating the measurement is chooses a very strange wide RBW so there is a bug

Kind regards

Kurt

?

?

On Mon, Jul 27, 2020 at 11:22 PM, Erik Kaashoek wrote:

The value of +4.7dBm is surprisingly low. It should be above +15dBm.

Here are my measurements. Only settings changed from default where RBW set to manual 10kHz in second measurement both with attenuation set to manual and 0dB

First two signals at 0dBm with combiner loss of 3dB



Then the same signals set to -10dBm to check if the measurement is OK



The IP3 signals drop indeed 20dB when lowering the input with 10dB so what you see is internally generated.
Adding 10dB input attenuations drops them into the noise floor

I'm very curious why you measure a much different value

Hi DiSlord

Thank you very much for clarifying these details

Kind regards

Kurt

?

?

TinySA internal for calculation use measure?data * 32, for not send negative values additional add 128
This 32 multiplier need for not use slow floating point calculations, and give error 1/32 ~ 0,03 in dBm (measure step is 0.5dBm, and other is calibrations error)

And this values used for get correct data in program, need substract 128 and divide on 32

Hi Herb

Thank you for the comments. DiSlord has also clarified why these two settings are made

I am sure Erik will update the ini file accordingly als the weird behavior with log frequency settings

Kind regards

Kurt

?

?

?

On Mon, Jul 27, 2020 at 02:06 PM, Kurt Poulsen wrote:

Hi Erik

I slowly learn to treat the TinySA the right way.
One funny detail is that the vertical grid lines for the Log sweep show up when Center frequency is set to 0 and the span to max frequency.

The ini file need and update like Herb has pointed out as the Level per dB must be 32 and not 2 and Herb mention the -125 should be changed to -128 why I have not figured out.

Kind regards

Kurt

Kurt,
? When I originally ran TinySA.exe the levels displayed in the application were 3dB different than those on the tinySA display.? I questioned Erik about the difference and he said it was caused by the "zero level" value under the 'Settings'.

? Per Erik, "If you go to the setting tab in TinySA,exe the "zero level" should be set to 128".

?Also per Erik, "zero level' values other than '128' are an error in the initial automatic configuration of TinySA.exe.?

My understanding is the 'zero level' setting will be depreciated in a future TinySA.exe release.? Until then I've just make it a habit of changing the 'zero level' setting to '128' whenever I update the TinySA.exe application.

Further comments on TinySA.exe from Erik:

1.? Is there a tinySA.exe software manual available??

? ? ? ?"No tinySA.exe was my swiss army knife SW for SA experiments, too complex for anyone else. I still need to simplify and create some wiki pages"

2.? When I right click on the tinySA.exe screen and "select marker", I don't see a marker display.? I have yet to figure out how to turn markers on.

? ? ? ?"Only automatic markers are implemented. No manual markers yet"

Automatic markers are implemented by setting a threshold for marker display.? Manual placement is a much needed addition to TinySA.exe.

- Herb

Hi Erik

I also wondered about this low value. The -27dBm – 3dB is far to low a level but I could not produce a higher level of time of writing. ?That is simply that my R&S CRTU port 3 which produces such high level was destroyed during the TinySA testing because my HP power meter as the only one did not by my mistake had an AC plug with ground, and when I wanted to measure the output level on port 3 the N center conductor was getting contact prior to ground and “puf” port 3 killed. The old HP instruments has AC filters with large filter capacitances and are bloody dangerous. Shit happens.

I will take my Stabilock and use that and update the document and also try out using the TinySA.exe

Kind regards

Kurt ???

?

?

On Mon, Jul 27, 2020 at 02:15 PM, Kurt Poulsen wrote:

, and also two amplifiers OIP3 and there was a considerable difference.

Kurt

In your document you state: "The combined signal tested with the R&S CRTU-RU And the IOP3 lower than -75dB relative carrier levels so it seem like the TinySA has an build in IOP3 of 4.7dB which is correctly calculated as rule of thumb is Pout + delta P/2 . That means The two testsignal with level -27dBm and combiner loss 3dB."

I guess IOP3 should be IIP3?
The value of +4.7dBm is surprisingly low. It should be above +15dBm.
See?
Can you check what happens when you increase the mixer drive?
CONFIG/EXPERT CONFIG/MIXER DRIVE
and try +10dBm and +13dBm.
This should increase the noise floor a bit but IIP3 may increase.