P25RX-II P25RX-II Firmware Updates / Testing

[pcman67]

@btt , I just read what @maus92 had posted above and I just tried 2022-02-02_1333. The EVM in this release is noticeably better in this version -- I am seeing resting control channel EVM of a constant 2% now, P1 transmissions are all at 2% EVM and P2 transmissions are improved at 2-4%. This is better than the _1107 version from earlier today.

I am trying the 2022-02-02_1352 release now.

 

[btt]

Thank you. I changed the filter for P2 in _1352. I haven't been able to observe P2 since the change in this version, but I think this might improve the EVM for P2 slightly.

 

[pcman67]

@btt , release 2022-02-02_1352 is even slightly better than the _1333 version! Now seeing resting control channel EVM's going between 1-2%, P1 transmissions are now varying from 1-2% in EVM, and P2 transmissions are holding at a nice, steady 3% EVM with the occasional fluctuation down to 2%! By far the best I've ever seen to date! Audio sounding extremely good on both P1 and P2! Just fantastic work, Todd!

 

[btt]

Thank you sir! Glad to hear it is getting better!

 

[maus92]

2022-2-2-1352 is closer to what we had a few days ago with 2022-1-31-0919, so back on the right track. P2 traffic channel is 4-5% EVM, CC is back to 2% EVM. Graphs have also returned to being mostly flat. This is with fixed 18dB VGA.

 

[btt]

2022-2-2-1352 is closer to what we had a few days ago with 2022-1-31-0919, so back on the right track. P2 traffic channel is 4-5% EVM, CC is back to 2% EVM. Graphs have also returned to being mostly flat. This is with fixed 18dB VGA.

Thanks. Are you seeing a difference in performance with the VGA gain set to AUTO vs manual 18dB?

 

[maus92]

Thanks. Are you seeing a difference in performance with the VGA gain set to AUTO vs manual 18dB?

RSSI suffers about a -5dBm degradation when VGA is set to AUTO. AUTO likes to set VGA Gain at over 40dBm variable. Still readable, but not realistically representing signal quality.

 

[btt]

@pcman67, @maus92
If either of you (or both) have time to continue, would you evaluate 02-02_1440 for P1/P2 EVM and see if this is an improvement or otherwise?

 

[maus92]

I’ll try it when I get back to my office in the next hour or so.

 

[btt]

Just saw a few P2 transmissions here with version 02-02_1440. 2% EVM! Looking forward to hearing what others see.

 

[btt]

New version up. 2022-02-02_1703. This should give good EVM for both P25 P1 and P2 based on what we have observed today. The one exception that I'm observing here is what I believe is a Harris control channel. The EVM has improved with the latest version and is now 3% (was 4-5 %). I think the difference must be stubble, but I'm certain that it can be improved. If someone has a suggestion, I would appreciate it.

 

[freqseeker]

Here is 02-02_1440 on P2 system. Not too shabby for a voice channel.
Downloading 1703 now.

 

[btt]

Thank you @freqseeker! That does look good.

 

[pcman67]

@pcman67, @maus92
If either of you (or both) have time to continue, would you evaluate 02-02_1440 for P1/P2 EVM and see if this is an improvement or otherwise?

@btt, I am trying the latest version now. Results will be posted shortly.

 

[pcman67]

@btt , version 2022-02-02_1758 was tested against P1 and P2 transmissions. Results - which were excellent - were very similar to the last version I tested, which was 2022-02-02_1352. CC EVM was a steady 2%, P1 EVM is almost always 2% with the occasional 3% EVM noted. P2 EVM's ranged from 2 to 4% with the majority of the transmissions being 3% EVM.

One thing I noticed right away is that the constellation in the I/Q Symbol Plot for P2 transmissions seem tighter, more closely resembling CC/P1 transmissions in the I/Q Symbol Plot. There were no P1 or P2 reception issues noted -- all sounded great and as they should.

 

[btt]

Thank you sir! I think we are getting a good idea of the viable range of channel filter values for optimal results. In the next advancements, the "scanning profiles' will allow for selecting optimal channel filters per frequency. We are getting closer!

 

[pcman67]

Thank you sir! I think we are getting a good idea of the viable range of channel filter values for optimal results. In the next advancements, the "scanning profiles' will allow for selecting optimal channel filters per frequency. We are getting closer!

@btt , wanted to let you know that I emailed you screenshots of resting CC activity, as well as P1 and P2 activity and TG logs showing mixed P1/P2 transmissions. Hoping this additional info might be useful in your fine tuning of the hardware/software.

 

[btt]

@btt , wanted to let you know that I emailed you screenshots of resting CC activity, as well as P1 and P2 activity and TG logs showing mixed P1/P2 transmissions. Hoping this additional info might be useful in your fine tuning of the hardware/software.

Thank you again sir! P1 and P2 are both looking really good now. Time to move on to the next improvements/features.

 

[maus92]

Running the 2022-2-2-1758 version has returned to poor RSSI on both the cc and traffic channels. a degradation of almost -10dBm over -1352 and varies quite a bit with both AUTO VGA and set at 18 dB. EVM on the FDMA cc is steady at 25, and 3-4% on the TDMA traffic channel, an improvement.

 

[btt]

@maus92,
Thanks. That is pretty close to what I'm seeing here for EVM. 1-2% for the control and 2-3% for the P2. I'm happy with that. Regarding the inaccurate RSSI, I will look into this again. Keep in mind that the AGC doesn't care what the "RSSI" is at all. It tries to maintain the voltage levels to the ADC around the target level. The RSSI is then a calculation of the measured voltage levels combined with all the gain levels. I think the issue is related to something I'm still figuring out with the VGA gain. I made a change today that might get it back in the range you were seeing before, but the RSSI really should be within 1-2 dB for those levels. I'll check on the calibration again and see if I can find anything.