Re: [Svxlink-devel] DTMF fine tuning and every_minute event
Brought to you by:
sm0svx
From: SM0SVX <sm...@us...> - 2009-09-22 11:05:01
|
On Tuesday 15 September 2009 14:13:17 st...@ar... wrote: > Hi Tobias, > > sorry, really forgot the patch. It fixes also a minor issue where the > compiler complains about unititialized variables in the new LocationInfo > code. OK. Most of it have been applied as you may have noticed. I didn't apply the one minute stuff since I thought I should rework it a bit. Your code looked cleaner than mine though. I don't know why I had made it so complicated :-) 73's de SM0SVX / Tobias > > vy 73 de Steve, DH1DM > > ----- Original Nachricht ---- > Von: Tobias Blomberg <sm...@us...> > An: Discussions about development issues > <svx...@li...> Datum: 13.09.2009 23:26 > Betreff: Re: [Svxlink-devel] DTMF fine tuning and every_minute event > > > On Sunday 06 September 2009 07:54:23 st...@ar... wrote: > > > Hi Tobias, > > > > > > enclosed you can find a small patch for the SwDtmfDecoder, which > > > slightly narrows the detection bandwidth. This is required, since the > > > two lowest frequency tones need to be separated by at least one DFT > > > bin. The TRX > > > > delay > > > > > line is still long enough for the inclreased detection time. > > > > I think you may have forgotten to attach the patch. > > > > > You can safely ignore the splitstr stuff in the patch, which has been > > > turned out a real hot spot of development compared to its simplicity. > > > However, there is an additional problem in the Logic::everyMinute timer > > > value calculation, which happend only a few times per year. See the log > > > below at 3am. The timer event is fired slightly too early > > > in this case, might by only a few microseconds. > > > > That one I've never seen before. I have seen hourly identification being > > missed > > totally though so there is something wrong with the every minute code. > > > > > This happens probably because of time jumps when using > > > the gettimeofday function (limited resulution). I'd suggest > > > to align and round up the to full 10ms intervals. > > > > Sounds like this could be it. > > > > 73's de SM0SVX / Tobias > > > > > Sat Sep 5 00:00:00 2009: Tx1: Turning the transmitter ON > > > Sat Sep 5 00:00:02 2009: Tx1: Turning the transmitter OFF > > > Sat Sep 5 01:00:00 2009: Tx1: Turning the transmitter ON > > > Sat Sep 5 01:00:02 2009: Tx1: Turning the transmitter OFF > > > Sat Sep 5 02:00:00 2009: Tx1: Turning the transmitter ON > > > Sat Sep 5 02:00:02 2009: Tx1: Turning the transmitter OFF > > > Sat Sep 5 03:01:00 2009: Tx1: Turning the transmitter ON > > > Sat Sep 5 03:01:02 2009: Tx1: Turning the transmitter OFF > > > Sat Sep 5 04:00:00 2009: Tx1: Turning the transmitter ON > > > Sat Sep 5 04:00:02 2009: Tx1: Turning the transmitter OFF > > > Sat Sep 5 05:00:00 2009: Tx1: Turning the transmitter ON > > > Sat Sep 5 05:00:02 2009: Tx1: Turning the transmitter OFF > > > Sat Sep 5 06:00:00 2009: Tx1: Turning the transmitter ON > > > Sat Sep 5 06:00:02 2009: Tx1: Turning the transmitter OFF > > > > > > vy 73s de Steve, DH1DM |