USB audio and telemetry timers
Hi everyone, I've been using app_rpt for well over a year now on a home UHF repeater that includes 2 remote bases (HF & VHF). For the hardware interface, I'm using DMK Engineering URI adapters. This config has worked fairly well, overall. But, so far I've used this system casually. Over the last few weeks, I've finally gotten "serious" about implementing some new repeaters, etc., for local ham community. So, I'm now trying to iron-out the remaining issues and, wondered if anyone else has already got solutions. To start with, I "upgraded" to the latest and greatest by installing the most recent ACID release on a new AMD X2 Athlon 64 6000+ box (MSI K9N6PGM2-V motherboard, 4GB RAM). I was, at this point, running the latest CentOS 5.3 and app_rpt version 0.174. I immediately hit problems. The telemetry timers were getting "stuck" (e.g. the idrecording would stop and the transmitter would hang keyed until timeout) and audio was sporadically moderately garbled.... So, I download the latest SVN version (v362). This helped, but, problems persisted....So, I download the most recent kernel (2.6.30.5) and compiled. I wanted to use the high-res kernel timer, and had to patch ztdummy.c to support the most recent (as of 2.6.28?) kernel interface API. This GREATLY improved everything! No more timer issues, etc., so far! ...zttest now reports timer consistancy of 99.997%! But, I still have 2 longstanding issues that I'd love to find resolution for. First, I'm still getting a fairly frequent stream of warning messages like: WARNING[xxx]: chan_usbradio.c:1640 soundcard_writeframe: write: used 3 blocks (1) And, these messages accompany somewhat choppy (or raspy) audio when they occur. I've seen these messages on several different boxes. So, it doesn't seem to be specifically hardware or kernel related. Second, I'm still seeing fairly frequent false DTMF decoding. This, too, causing a momentary choppiness (muting) in the audio. On some peoples voices, this false decoding (and chop) is severe. Last year I added some explicit code to chan_usbradio.c to disable DTMF decode, and, this problem was temporarily hidden. But, now I need DTMF decode--so, disabling isn't an option. Thanks in advance for any help! 73, David kb4fxc
Hi Everyone, Well, since I haven't heard of any solutions to my USB audio choppiness issues, I decided to have a go at it myself. I've added several hacks to chan_usbaudio.c that GREATLY improve the sound quality on my system. Now, false DTMF decode events are infrequent for me--as compared to sometimes occurring every few seconds. The trade off is that I'm letting a few frames of "real" DTMF tone pass thru the system without muting before deciding that a DTMF tone is really present. Also, I have experimentally modified the soundcard_writeframe() function slightly. Now, this source is raspyness is improved, too. I'll post code for evaluation if anyone is interested. 73, David kb4fxc On Thu, 3 Sep 2009, David McGough wrote:
Hi everyone,
I've been using app_rpt for well over a year now on a home UHF repeater that includes 2 remote bases (HF & VHF). For the hardware interface, I'm using DMK Engineering URI adapters. This config has worked fairly well, overall. But, so far I've used this system casually.
Over the last few weeks, I've finally gotten "serious" about implementing some new repeaters, etc., for local ham community. So, I'm now trying to iron-out the remaining issues and, wondered if anyone else has already got solutions.
To start with, I "upgraded" to the latest and greatest by installing the most recent ACID release on a new AMD X2 Athlon 64 6000+ box (MSI K9N6PGM2-V motherboard, 4GB RAM). I was, at this point, running the latest CentOS 5.3 and app_rpt version 0.174. I immediately hit problems. The telemetry timers were getting "stuck" (e.g. the idrecording would stop and the transmitter would hang keyed until timeout) and audio was sporadically moderately garbled....
So, I download the latest SVN version (v362). This helped, but, problems persisted....So, I download the most recent kernel (2.6.30.5) and compiled. I wanted to use the high-res kernel timer, and had to patch ztdummy.c to support the most recent (as of 2.6.28?) kernel interface API. This GREATLY improved everything! No more timer issues, etc., so far! ...zttest now reports timer consistancy of 99.997%!
But, I still have 2 longstanding issues that I'd love to find resolution for. First, I'm still getting a fairly frequent stream of warning messages like:
WARNING[xxx]: chan_usbradio.c:1640 soundcard_writeframe: write: used 3 blocks (1)
And, these messages accompany somewhat choppy (or raspy) audio when they occur. I've seen these messages on several different boxes. So, it doesn't seem to be specifically hardware or kernel related.
Second, I'm still seeing fairly frequent false DTMF decoding. This, too, causing a momentary choppiness (muting) in the audio. On some peoples voices, this false decoding (and chop) is severe. Last year I added some explicit code to chan_usbradio.c to disable DTMF decode, and, this problem was temporarily hidden. But, now I need DTMF decode--so, disabling isn't an option.
Thanks in advance for any help!
73, David kb4fxc
_______________________________________________ App_rpt-users mailing list App_rpt-users@qrvc.com http://qrvc.com/mailman/listinfo/app_rpt-users
participants (1)
-
David McGough