[hpsdr] about latency growth with HPSDR

[Phil Harman]

Hi Riho,

You may care to do some tests on you system using KISS Konsole since I suspect you are dropping packets between the PC and HPSDR hardware.  If you are only losing the odd packet then you are not going to hear them.

The Metis Ethernet code has a packet number included in each data frame both to and from the PC. KK will indicate if a  packet is lost and what the expected and received  packet numbers are.  Similarly Metis will light a LED if it misses a packet.

Using KK with Metis at 192ksps I have never seen a missed packet in either direction - in fact I had to deliberately send incorrect packet numbers to check the code was working!

I'm also currently doing some work that involves time stamping the data from Mercury via Metis - again no dropped packets. 

The PC has absolutely no idea what the sample rate is of the data it is being sent. It's not acceptable that an SDR can't be used for DRM, SSTV etc - this needs to be investigated and corrected.

73 Phil...VK6APH 




  ----- Original Message ----- 
  From: Riho B., ES7AAZ 
  To: hpsdr at lists.openhpsdr.org 
  Sent: Saturday, October 01, 2011 6:00 AM
  Subject: Re: [hpsdr] about latency growth with HPSDR


  ***** High Performance Software Defined Radio Discussion List *****




------------------------------------------------------------------------------


  Steve,

  It may be right that SDR is not capable to run time sensitive applications.
  I was scoping 10Mhz WWV today, output shows that real time/SDR time
  varies, escpecially while minimizing and maximizing PowerSDR window
  which causes higher CPU load. IIt was @ 48kHz sampling rate  all over
  the path, biggest buffers, no audioble drop-outs, max DPC latency 170?s.
  Anyway, it's now clear why's there are drop-outs while listening DRM, why's
  there are out of sync line blocks in SSTV, why's digital SSTV misses the blocks.
  I'm not struggling with it anymore. All the other, CW and SSB works just fine.





  73's,
  Riho, ES7AAZ.


  29.09.2011 23:56, Steve Bunch kirjutas: 
    Riho, 


    I can't answer your question exactly, but this can happen because of buffering somewhere in your sample chain, when the sampling rate of your system is slightly slower than realtime so the samples aren't being consumed as fast as they're generated.  The unread samples are saved up somewhere, and accumulate there.  I had this happening to me in an ALSA-based Linux system -- I would over time be listening several seconds behind realtime when using a USRP to listen to FM radio.  Restarting the chain will clear out the buffered data; when you restart it will grow again.  


    I don't know if the Ozy/Mercury sample chain is capable of buffering this much data (someone familiar with it should be able to tell you exactly how much can be buffered there), but in my case it was in my audio system.  Unfortunately, I don't know the various Windows audio systems well enough to give you any help, but that is where I would look first.


    Steve, K9SRB


    On Sep 29, 2011, at 3:07 AM, Riho B., ES7AAZ wrote:


      ***** High Performance Software Defined Radio Discussion List *****


      Hi,

      I have a program called Faros which is a beacon monitor.
      Faros gets his time sync from ntp GPS time servers.
      NCDXF network beacons have GPS dicsiplined controller
      which allows path delay calculation and Faros does it.

      After 2h running I see 120ms latency growth (linear graph)
      and it's growing on. Which means that radio waves are traveling
      about one  more circle around the world. It can't be so. Restarting
      PowerSDR brings the latency back to its initial state. Exactly the same 
      effect with CWSkimmer console. It's almost same vith VAC and with
      analog audio cable.

      The question is not even about latency itself or how much it is.
      The question is why's latency growing and how much it does?
       Is it somewhere in USB connection ?  I'm Ozy user.



      73,
      Riho, ES7AAZ

      _______________________________________________
      HPSDR Discussion List
      To post msg: hpsdr at openhpsdr.org
      Subscription help: http://lists.openhpsdr.org/listinfo.cgi/hpsdr-openhpsdr.org
      HPSDR web page: http://openhpsdr.org
      Archives: http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/






------------------------------------------------------------------------------


  _______________________________________________
  HPSDR Discussion List
  To post msg: hpsdr at openhpsdr.org
  Subscription help: http://lists.openhpsdr.org/listinfo.cgi/hpsdr-openhpsdr.org
  HPSDR web page: http://openhpsdr.org
  Archives: http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/


------------------------------------------------------------------------------


  No virus found in this message.
  Checked by AVG - www.avg.com
  Version: 10.0.1410 / Virus Database: 1520/3929 - Release Date: 09/30/11
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20111001/84710bc8/attachment-0003.htm>