Panadapter Insertion loss and other side consequences

What's Insertion Loss (IL)?

Insertion Loss is the loss of signal power resulting from the insertion of a device in a transmission line. By tapping the IF line with the Panadapter, you are automatically causing some kind of disturbance on the signal line.

For example, if you split the signal equally between two outputs, you are effectively halving the power - 50% goes to one side; 50% goes to other side. In radio frequency, that 50% loss translates to a -3dB loss (-3dB equals to half of the power).

PAT-991A findings

Design and initial findings

While G4HUP came out with a clever design that minimizes the insertion loss of the Panadapter board by using a JFET + LNA design, that should have translated to a minimal losses, surely within 1 dB or less.

During the PAT-991A initial iterations, I tested the PAT-991A losses with less than empirical modes - since my (then) preferred band was 40m, I limited to watch the S-Meter in a known signal and looking for any kind of signal fading in the 991A built-in band scope, and of course, noticing the signal that was being outputted from the speakers. I found no S-Meter, Scope or Audio influence - and deemed that as kosher and moved along.

BG0AUB findings

Feng Zhao, and avid QO-100 operator gave PAT-991A a try and reported in QRZ Thread that the PAT-Board was consuming a full S-Unit (in FT-991A that translates to 3 dB, or, again and for extra drama, eating half the power of the received signal).

Feng was absolutely right - I was able to reproduce this issue: When unplugging the Panadapter board, I got one more S-Unit in signal strength in both VHF and UHF bands.

Testing

While my findings are consonant with Feng's findings, I went to investigate further and I found that while this issue was faithfully reproduced in VHF and UHF bands, it did not happen in HF band - I wasn't able to find any S-Meter difference in 40m band.

Back to VHF and UHF testing, I wanted to ensure if was it some S-Meter imbalance caused by the Observer Effect of the board over S-Meter calibrations - or if was it really some non-trivial signal loss.

Listening test

I used my PocketVNA to generate an VHF signal in 143 MHz (I went offband because a contesting was happening at the time I was testing) and used the Full Version PAT-991A DIS (Disable) signal to toggle the PAT-991A RF switch to Reflecting mode, pretty much like what the TX9 signal does, essentially causing the board to not tap any signal - and paid attention to the received / perceived signal to any degradation.

I could not perceive any significant difference. This test was recorded and can be watched in this video

Insertion Loss tests

Feng got me worried - was it indeed causing a degradation in the radio? My Rev. P boards arrived this week - and I went to test its insertion loss figures. It was done in S21 mode, using the IF In and SCP Out ports - any signal absorption that could affect / desensitize the radio should be happening between these two ports.

The inherent Insertion Loss

In the below picture, I measured the board's minimum insertion loss - meaning, the connector loss. I have removed the first 1 nF capacitor before the RF switch. At this point it is essentially only the connectors:

PAT-991A Connector insertion loss - click to enlarge

Around -0,1 dB of loss. Moving on.

No-Power Board Insertion Loss

In this test, the first 1 nF cap is soldered back again, and we are measuring the board insertion loss - with the catch that the board is not powered. While this goes against PSemi recommendation in their RF switch (where they require the RF switch to be powered all the time), we can find here some non-neglectable losses:

PAT-991A non-powered board insertion loss - click to enlarge

In this test we can find around -0,5 dB of signal loss. Remember this is not a valid setup.

Powered board, nominal operation

Okay - This got weird and might be what can be playing fun with the S-Meter. In this test we have the board powered and in normal receive mode. After 60 MHz the board actually puts some gain back to the IF line - don't ask me, I'm clueless too.

PAT-991A Powered board in Receiving Mode "Insertion Loss" - click to enlarge

Well... Approximately + 0.05 dB gain around 69 MHz (grins)

Powered board, TX9 / DIS signal on

In this test, I'm verifying the TX9/DIS signals effectiveness on isolating the board. The RF switch does an superb job, with losses similar to connector-only losses.

PAT-991A Powered board, TX9 or DIS signal up - click to enlarge

Results on par with the connector-only test.

Conclusion

I'm feeling much more comfortable now using the board - while this can cause some side effect in S-Meter, I found no noticeable problem in the actual received signals - and if desired, I can use the DIS signal to move the board outside the receiving chain.

Acknowledgment

I would like to thank BG0AUB for giving this project a try - and for his diligent tests and reporting this issue.