I don't think that I am decoding some signals on FT8 that I should be able to. The basic system configuration works, and I am making quite a few contacts. But there are a lot of DX stations that my neighbors are copying that do not decode for me, and sometimes I will see a 15 second period with half a dozen decent-looking signals in the waterfall, but nothing decoding at all. (This is on 6M, by the way.)
One thing I wonder about is the impact of strong local signals within the passband that drive the AGC; could they be crushing weak signals?
I've tried changing VAC audio levels and played a bit with AGC settings, but nothing has made much difference. My local noise level is typical suburban -- not horrible, but turning on a noise blanker will drop the displayed S-meter power level by 5 to 10 dBm in a 3 kHz passband. Usually NB1 seems to be best.
Are there any tricks to optimize the VAC configuration (or other settings for that matter) to work best with WSJT-X?
Thanks,
John
Optimizing WSJT-X
- Tony EI7BMB
- Posts: 653
- Joined: Sun Apr 09, 2017 2:31 pm
- Location: Dublin
- Contact:
Re: Optimizing WSJT-X
Good question , I'd be interested also in what guys think.
Re: Optimizing WSJT-X
A couple of comments from my experience. I had drop out sequences where nothing would decode - frustrating especially the few times when it was someone I needed and I was working. Turned out to be my fault. I had Dimension4 time program set to update once every 10 minutes or something like that - not every time it updated but a good share of the time - WSJT would not decode! I changed this to an hour and proved this was the case - since I've gone to manual updates - I find if I update the time once a day I'm fine and I no longer have drop outs.
Strong signals will most definitely take out weaker ones - the difference in strength and their closeness are what determine how much. I use the manual notch system a good deal - I've had 4 notches going at once. I can tell you just as an example it's not unusual for a station that is say 200 Hz from the offending one (strong one) where it might be a -20 and I turn on the notch and it is all of a sudden -8. This happens all the time. You have to take care when setting the notch though - too close and you'll impact the signal itself - keep the width down if possible. I run typically 110hz width.
I haven't found any other settings that impact decoding in general. And last year in ClubLogs DX Challenge I was #1 in the world for DIG having worked something like 243 countries across all bands 160 - 6.
Gary
K9RX
Strong signals will most definitely take out weaker ones - the difference in strength and their closeness are what determine how much. I use the manual notch system a good deal - I've had 4 notches going at once. I can tell you just as an example it's not unusual for a station that is say 200 Hz from the offending one (strong one) where it might be a -20 and I turn on the notch and it is all of a sudden -8. This happens all the time. You have to take care when setting the notch though - too close and you'll impact the signal itself - keep the width down if possible. I run typically 110hz width.
I haven't found any other settings that impact decoding in general. And last year in ClubLogs DX Challenge I was #1 in the world for DIG having worked something like 243 countries across all bands 160 - 6.
Gary
K9RX
Re: Optimizing WSJT-X
Are you running PowerSDR or Thetis?
Re: Optimizing WSJT-X
Running PowerSDR, Scott.
- Tony EI7BMB
- Posts: 653
- Joined: Sun Apr 09, 2017 2:31 pm
- Location: Dublin
- Contact:
Re: Optimizing WSJT-X
Using a notch is great info, never thought of that
Re: Optimizing WSJT-X
If you can see the signal you want to decode on the WSJT-X waterfall display, and assuming you and the other station are in time sync, then it should decode. Decoding problems are rarely audio quality related and are more often timing related.
The first thing I would do is listen to the audio coming from the VAC output. This is easy to do if you are using Voicemeeter, just enable a route that sends VAC audio to your PC speakers. With Muzy VAC you have to use his Audio Repeater application. If it doesn't sound right to your ears then you've found that the problem lies with the quality of the audio. However, this is rarely the issue, especially if you are not using the VAC resampler function. Try turning the resampler off.
When running digi modes all narrowband receiver audio processing should be turned off. This is largely done for you when selecting DIGU except for the NR modes, which you should not use. NB processing is wideband in nature and some people have reported improvement using it; I don't have that kind of noise around here so can't report one way or the other.
FWIW Meinberg NTP is the best performing NTP software for Windows. It is much more complex to setup than Dimension 4, but it will smoothly and transparently keep your system perfectly in sync with the universe.
The first thing I would do is listen to the audio coming from the VAC output. This is easy to do if you are using Voicemeeter, just enable a route that sends VAC audio to your PC speakers. With Muzy VAC you have to use his Audio Repeater application. If it doesn't sound right to your ears then you've found that the problem lies with the quality of the audio. However, this is rarely the issue, especially if you are not using the VAC resampler function. Try turning the resampler off.
When running digi modes all narrowband receiver audio processing should be turned off. This is largely done for you when selecting DIGU except for the NR modes, which you should not use. NB processing is wideband in nature and some people have reported improvement using it; I don't have that kind of noise around here so can't report one way or the other.
FWIW Meinberg NTP is the best performing NTP software for Windows. It is much more complex to setup than Dimension 4, but it will smoothly and transparently keep your system perfectly in sync with the universe.
- Tony EI7BMB
- Posts: 653
- Joined: Sun Apr 09, 2017 2:31 pm
- Location: Dublin
- Contact:
Re: Optimizing WSJT-X
Using net time here for sync without any problem http://www.timesynctool.com/