Using Direwolf, adjusting receiver volume vs microphone gain vs I really should be using demodulator output ???

[needairtime]

Anyone using direwolf here?

I've successfully decoded exactly ONE APRS packet so far with Direwolf. The packet I received is very reasonable comparing to data that aprs.fi is reporting. But listening to another radio I should be getting much more than just that one packet --I have a feeling it's a problem getting the data from my radio (Kenwood TR-2500 speaker output) to my computer (microphone input). Any suggestions on why I'm not able to decode packets?

I suspect the radio should be kept with loose squelch so there are no SOT delays. However, how "loud" should the receiver be? Should I minimize gain on the PC side? Or is this simply not going to work unless I tap straight off the demodulator output of the receiver prior to any audio voodoo the receiver may do? If I need the demod output I guess I will have to break out the solder iron (but people have been successful using Baofengs?)

After this works I'll have to see if I can transmit something...then once again I wonder what volume settings is needed for the transmitter to accurately produce the tone pattern that others can decode.

I'm running on Linux that uses PulseAudio (versus straight ALSA). I think most distributions are using PulseAudio, but I hope that this has nothing to do with things and identical in operation to Windows.

 

[KE5MC]

I would turn squelch all the way down to make sure it not cutting off the front of the reception. It's not clear if you are acoustically coupling or wire coupling. I'm assuming wire. Speaker out to mic in is the least likely to work. Speaker signal to big and mic input very sensitive. You are likely overdriving the input. Use some resistors to create voltage divider of at least 10 to 1 and go higher like 20 to 1. Line input is also a better choice. On the mic input make sure there are no PC applied sound effects introduced.

 

[needairtime]

Strange... but yes, I'm using wire coupling, not audio (had to use another radio to hear when transmissions were occurring.)

The nice thing about PulseAudio is that I can see the audio record level as a bar graph with a program that's outside of Direwolf. So I can actually see the level go up and down as I adjust the volume control. I also can see the volume level drop off if I tighten squelch, so even if I don't hear the noise, I can see something -- which confused me why this isn't working. Whether the apparently internal attenuation is causing distortion I don't know.

Perhaps I feed it through Audacity and see how Audacity sees the signal... or even run both at the same time....

I can't seem to find the line input jack, will have to try that, though I thought these newer machines the microphone jack is dual purpose and can be software set to microphone or line level input, letting the manufacturer cheap out and remove a hardware jack. grr...

 

[KE5MC]

Looking at the levels is helpful, but it doesn't show the quality of the signal. Can you hear the signal from the other radio and play the packet on your counter's speaker to hear the difference? Likely there might be a little time delay between them.

Nothing wrong with wire coupling, but it does take some steps to isolate the radio and computer from each other while still passing the signal without distortion. One I talked about the resistor attenuation pad and the other is adding a capacitor inline with the signal. This will decouple the DC difference between the radio and PC. DC bias into the mic input could be playing havoc with the AC signal.

I had to Google Direwolf to find out what it was and on the GitHub page it looked like several files were available that might have setup/configuration for the audio. I didn't take the time to open them, but just viewed titles and comments at the bottom of the page.

Good Luck,
Mike

P.S. If you have a PC mic or headset with mic boom give the acoustic coupling a try. Many have used that method for ham digital modes with some success and I would believe APRS would work too.

 

[needairtime]

Okay, I ran the signal through audacity.... turns out, THE SIGNAL appears really WEAK! It never got higher than about -34db the whole time (0db = 65535 as a PCM value I suppose?) and there's a very large 60-cycle hum component int he recorded signal... something wrong with grounding, and I wonder why the recorded signal level is so low...

However there even appears to be some clipping when I normalized the signal so I could see what's going on, so there's more to it than this... Hmm...

Unfortunately my PC doesn't have a microphone so I can't try that, but it would be strange that a microphone coupling would be less distortion, though one may never know...

 

[KE5MC]

Find a 600 ohm 1:1 audio isolation transformer used to break physical connection of your common ground between radio and computer. That’s a frequent fix for those running audio feeds of Public Service on the internet. See if that doesn’t improve the level of the signal of interest.

 

[needairtime]

Not sure what's the bigger problem at this point: the hum or the overall signal value. Granted the SNR is currently really bad, but both are bad.

Incidentally I have the microphone gain at +30dB so the hum is really -60dB. and the signal is not much better... why speaker output into a microphone input is so weak is the next order of business. Perhaps I should stick in another sound card and see what it does...

 

[KE5MC]

Compound problems for sure. Fix the hum as that will cause distortion that the decoder will have problems with. Once the hum is gone and you have a clean signal into the decoder you might be surprised at how low the signal can be and still decode. The audio isolation transformer will likely eliminate the hum and break any DC difference that the mic input will not like. If you have any old modem or phone answering machines you will find isolation transformers in them. Transformers are easily found online too. Shipping likely to cost more than the device.

 

[needairtime]

Now the question is why am I seeing this, basically I just have an HT to the computer... it's not a Baofeng...

 

[needairtime]

Okay, I tried the Baofeng too. Same deal.

I also tried sampling broadcast radio from the Baofeng and played it back on my speakers... As expected, it sounded awful after normalizing the samples (the passthrough sounded okay however). The 60 cycle hum was probably there but probably was as audible as CTCSS.

Probably need to try another machine or perhaps need a voltage amplifier?

 

[needairtime]

Okay... I think I got it!

I tried hooking up the transceiver to another machine - and this machine I was able to get the sampling all the way up to -6dB (according to Audacity) which is a HUGE dynamic range improvement. I still had to increase the speaker output of the transceiver almost full but Direwolf was now able to receive and decode many packets just fine. This is still a direct connect from the speaker jack to the microphone input much like the other machine.

So it turns out that there's hardware problems or hardware incompatibility.