SSDR V 1.7 DAX Distortion (continues)

  • 4
  • Problem
  • Updated 2 years ago
  • Acknowledged
  • (Edited)
SSDR V1.7 appears not to have addressed the ongoing DAX Distortion Problem that has been reported on over the past year or so.  This problem is seen here at K3VAT when running the JT modes, usually after a period between a few hours and several days.  Other user postings have reported DAX distortion with other modes - screen captures have additionally been included to show exactly what is occurring to the DAX output when this problem occurs.  I'm also including three screen shots from the last couple of days showing 1) clean signal 2) 'start' of the distortion 3) full distortion.  Other important notes include that Flex Engineering has been made aware of this problem and that mentioned earlier this year reported, "... that this is probably a memory leak issue ...".  There are dozens of past postings on this topic that provide background and additional info (primarily from Tim).  HelpDesk folks (Dudley and of course Tim) have been very understanding and supportive; and I realize that this is a very unique problem affecting only a small cohort of specialized users; nevertheless, it would be great if this issue is resolved (so I don't have to restart my system in the middle of a JT QSO [essentially loosing the QSO]).  73, Rich, K3VAT (please see QRZ.COM page for more info on my station and operations).

Top screenshot: 'start' of DAX distortion;  Middle: 'full' distortion; Bottom: clean signal



Photo of rich

rich

  • 37 Posts
  • 3 Reply Likes

Posted 2 years ago

  • 4
Photo of Drax

Drax

  • 99 Posts
  • 13 Reply Likes
The problem I always have along these lines is that the transmit audio goes distorted or more often it just "goes silent" where the radio keys up but audio won't transmit until I reboot DAX and the radio.    This happens consistently when I leave things running and monitoring JT65 for a long time.
(Edited)
Photo of rich

rich

  • 37 Posts
  • 3 Reply Likes

Drax, thanks.  Yes, I've also seen these "going silent" instances.  For me, it happens a bit less often and on about 1/2 the occasions I get a fix from restarting the DAX (which if done quickly can 'save' the QSO). 


Both these problematic conditions (distortion and no keying) indicate that the DAX needs fixing.


73, Rich, K3VAT

Photo of Bill W2PKY

Bill W2PKY

  • 448 Posts
  • 78 Reply Likes
Still on 1.61; with this release have been noticing problems with the receive side. Will see signals on the panadapter but no decodes. Restarting DAX restores decoding on the problem slice.

Have tried running DAX as RealTime priority with no help. My computer is 4ghz I7 8 execution units with 16 gigs memory. Usually run my radio 24X7, have to restart DAX now and then.
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9130 Posts
  • 3454 Reply Likes
Based on your problem description, I suspect you have a long duration DPC latency issue.  The What are DPCs and why do they matter? HelpDesk article explains what they are, how to measure DPC latency and the types of things needed to reduce DPC latency.

If your issue is intermittent, you may need to run LatencyMon for a while until you trap the event.
Photo of Rob Fissel

Rob Fissel

  • 270 Posts
  • 47 Reply Likes
FWIW Tim, I have the same issue. Happens in the same way as others seem to experience - only after several days where DAX is running 24/7 doing things like monitoring WSPR or JT65/9. A quick shutdown/restart of DAX fixes the problem for me. 
Photo of Drax

Drax

  • 99 Posts
  • 13 Reply Likes
My wish is some day that software can detect that a killer DPC occurred and just drop the audio for a moment and reset itself without my intervention.

If I can click close and re-open DAX to fix it then why couldn't it be done programmatically?  Perhaps Flex doesn't have fine grain enough control over Microsoft's audio driver?

My PC doesn't seem have DPC issues when I'm using it.  It seems to happen when I walk away for awhile then perhaps the backup or anti-virus decides to start or the video is powered down automatically.  Then I come back to find I have no TX audio.
Photo of Ed, K0KC

Ed, K0KC

  • 268 Posts
  • 19 Reply Likes
Rich,

I have seen problems similar to this with PowerSDR and my Flex-5000, but not with my Flex-6700 which I have had for almost three years now. This might point to the DPC latency issue that Tim mentioned since PowerSDR does not use DAX.

By the way, I noticed that you are set to USB rather than DIGU, but I doubt if that is related to your problem.

Ed, K0KC
Photo of rich

rich

  • 37 Posts
  • 3 Reply Likes

thanks Ed

Eliminated any DPC latency issue about 1.5 years ago with all new hardware and optimization by Neil. 

Yes, I run USB with NO EQ or compression so I can use the 4 kHz bandwidth to cover all of the JT band (usually xxx.076 to xxx.080).

Finally, thanks for the four JT QSOs (160, two on 80, and 40M) !!

73, Rich, K3VAT



Photo of Al K0VM

Al K0VM, Elmer

  • 573 Posts
  • 93 Reply Likes
Rich,
  I think you could use DIGU and cover the 4khz for JT    I do..

AL, K0VM
Photo of rich

rich

  • 37 Posts
  • 3 Reply Likes

Al, thanks.  On my SSDR selections under DIGU are either 3.0 or 5.0 K.

Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9130 Posts
  • 3454 Reply Likes
Those are just filter presets.  Adjust the filter to what you need for the digital mode you are operating.
Photo of Jay / NO5J

Jay / NO5J

  • 1467 Posts
  • 219 Reply Likes
I created per band memories for WSJT-X operation, and then edit the filter high and low values for a 4kHz filter width. When I open WSJT-X and it selects the frequency the memories auto select the 4kHz filter width, or to return to a 4K width after trying either a 3 or 5 kHz mode preset, i just reselect that bands memory.

I also save and store the same as favorites in FRStacks which is usually also running.

73, Jay - NO5J

 
Photo of Al K0VM

Al K0VM, Elmer

  • 573 Posts
  • 93 Reply Likes
Rich,
   You can set up DIGU, RX pass band, TX High cut, 14076 and other JT unique things once and then save it to a SSDR memory.    Recall to active slice as needed and use WSJT CAT to move to other bands..

AL, K0VM
Photo of Larry Loen  WO7R

Larry Loen WO7R

  • 187 Posts
  • 23 Reply Likes
This may not be as restricted to the "JT" modes as it appears.  Rather, we can think of the JT modes as being more of a "key down" mode than most.  It would be interesting to see if the lab could recreate it with a simple, basic brick-on-the-straight-key approach.

I know that I never permit myself to run the "JTs" more than 30 watts myself, even though it "could be" 50.  It seems to be just plain hard on the rig with this kind of duty cycle.  Stresses all my equipment, not just the Flex.
Photo of Rob Fissel

Rob Fissel

  • 270 Posts
  • 47 Reply Likes
FYI Flex has done key down tests at 100 watts with no issue. I believe one of the engineers said on these forums that they forgot about it overnight and came back in the morning to find it still chugging away at 100 watts. I've never had any issues running various "key down" digi modes at 100 watts before - those transistors are rock solid.
Photo of Andrew Russell

Andrew Russell

  • 257 Posts
  • 33 Reply Likes
In the past I had trouble with distortion producing multiple spots for strong WSPR and other digital signals. In my case under Win8.1 DPCs were largely eliminated with DAX still going feral after a while. In my case and against most predictions upgrading to Win10 has fixed it. It seems that there were some corrupt system files that needed a complete refresh or flushing. Or the drivers under Win10 are better.
Andrew de VK5CV