Page 1 of 1

An interesting experiment using linux. (TX power vs freq)

PostPosted: Wed May 01, 2013 12:52 am
by hansvi
I've done a fun experiment today: I measured the TX power versus the frequency.

Tools needed:
  • Peaberry (duh!)
  • RF power meter.
  • Dummy load
  • Gnuradio
  • Spreadsheet
  • usbsoftrock, working with peaberry (see forum for instructions)
  • Files provided in the attachment

The setup:
Connect the peaberry to the rf power meter. Connect the rf power meter to the dummy load.

The experiment:
  1. Start the gnuradio file, either by running test_peaberry_audio.grc in gnuradio-compainion or by running test_peaberry_audio.py. Keep this running throughout the experiment! This will send a continuous tone to the transmitter.
  2. set the frequency to where you expect you get an optimal tx.
  3. turn the transmitter on (usbsoftrock ptt on)
  4. Adjust the slider of the gnuradio program until you get about half a Watt output (I had to set the slider to 36 for that).
  5. turn the transmitter off (usbsoftrock ptt off)
    So right now, we can assume that the transmitter is sending about half a Watt. Now we have to edit freq.csv, containing the frequencies you want to test. Make sure each frequency contains a period, or it won't be recognized correctly. I used a spreadsheet to create the file, it is included so you can take a look how it is supposed to look. Change the cell format to have one decimal place to force a period.
  6. Run the provided script run.sh.

The script will set the PTT on, and test all the frequencies. At each one it says the current frequency through the internal speakers of your computer (should work if you have the command spd-say, it's installed per default on debian) and waits 3 seconds.

At this point you just have to write down the power for each frequency, and plot it out.

After the test, double check that ptt is off. You might want to do a "usbsoftrock ptt off" manually to make sure, or even unplug it all together. This is because you don't want to transmit too long on non-optimal frequencies. It will heat up the transmitter.

peaberry_tx_power.tgz
experiment files
(2.21 KiB) Downloaded 367 times


Results
After entering the data into a spreadsheet, I got a nice response diagram of the transmitter.

tx-power.png
Results
tx-power.png (12.21 KiB) Viewed 3450 times


(Sorry the titles of the axes got dropped. X = freq in MHz. Y= RF power measured * 100 (so '50' is 1/2 Watt measured))