Anyone can give me some information on a small form computer that will run allstarlink without skipping audio.
Right now i am using a raspberry pi3 with a URI-X and while trying to adjust my audio and running a vanilla system deside using the dummy dadhi app I can clerly hear small gap in the audio output that is received by my receiver. My spectrum analyzer will also give me a small bleep in the deviation of the signal.
I knew that it was possible that thing like that happen but I want this 900 mhz repeater to be perfect cause it will be the first one running for that club on allstarlink and this cant be ok.
If someone want to propose me to run the Hamvoip distro, no I wont. the day they will release all there source I could be thinking about it. But till then forget it.
I did try the rpi4 image that is in testing and I was running in the same small cut in the audio.
Dont want a embeded system that can tun games. 
Pierre
VE2PF
1 Like
OK I installed the latest version of allstarlink on a small thin client. a I686 from HP
It run on a USB stick.
It was a breeze to install and setup. still running a URI-X with simpleUSB. Still having some loss of small part of audio.
If I put my test set in tx and listen to the quantar speaker the tone is perfect no problem.
If I listem to 2 different radio at the same time both ear small cut in the tone.
wanted to test dadhi_speed.
the only thing I receive when doing this either with Sudo or as root is that:
dahdi_speed
Count: 0
Should I add the dadhy dummy extention?
Duh! I am dumb. I wanted to run dadhi_test. not dadhi_speed…
here is the result from test:
dahdi_test
Opened pseudo dahdi interface, measuring accuracy…
99.993% 99.985% 99.995% 100.000% 99.986% 99.995% 99.990% 99.994%
99.999% 99.987% 99.953% 99.940% 99.998% 99.991% 99.989% 99.995%
99.996% 99.992% 99.997% 99.995% 99.990% 99.996% 99.990% 99.996%
99.999% 99.990% 99.991% 99.993% 99.998% 99.988% 99.996% 99.995%
99.991% 99.999% 99.988% 99.994% 99.992% 99.994% 99.994% 99.998%
99.990% 99.994% 99.994% 99.996% 99.992% 99.992% 99.994% 99.996%
99.987% 99.998% 99.988% 99.992% 99.996% 99.992% 99.993% 99.995%
99.993% 99.994% 99.993% 99.993% 99.994% 99.995% 99.993% 99.999%
99.986% 99.993% 99.992% 99.995% 99.993% 99.981% 99.991% 99.996%
99.992% 99.995% 99.997% 99.986% 99.995% 99.994% 99.992% 99.995%
99.993% 99.991% 99.995% 99.997% 99.989% 99.995% 99.995% 99.994%
99.996% 99.988% 99.995% 99.994% 99.995% 99.996% 99.993% 99.991%
99.996% 99.991% 99.997% ^C
— Results after 99 passes —
Best: 100.000% – Worst: 99.940% – Average: 99.992325%
Cumulative Accuracy (not per pass): 99.993
Worst: 99.940%
Is that enough to cause audio glitch?
ok installed dummy.
root@va2dwe:/usr/src/asl-dahdi-linux-2.11.1/linux# dahdi_test
Opened pseudo dahdi interface, measuring accuracy…
99.992% 99.980% 99.994% 99.992% 99.993% 99.994% 99.994% 99.994%
99.993% 99.992% 99.993% 99.994% 99.994% 99.993% 99.993% 99.994%
99.992% 99.997% 99.985% 99.993% 99.993% 99.994% 99.994% 99.993%
99.993% 99.993% 99.994% 99.995% 99.994% 99.993% 99.993% 99.994%
99.994% 99.993% 99.994% 99.993% 99.995% 99.993% 99.993% 99.993%
99.994% 99.993% 99.994% 99.993% 99.993% 99.995% 99.992% 99.993%
99.993% 99.993% 99.994% 99.994% 99.994% 99.993% 99.993% 99.994%
99.994% ^C
— Results after 57 passes —
Best: 99.997% – Worst: 99.980% – Average: 99.993186%
Cumulative Accuracy (not per pass): 99.993
seem better. we will see
Pierre,
These motherboards should work.
Intel D201GLY2, D945GCLF, D945GCLF2, D945GCLF2D
I have personally used the Intel D525MW Atom motherboard.
Marshall
···
From: Pierre Martel via AllStarLink Discussion Groups [mailto:noreply@community.allstarlink.org]
Sent: Wednesday, May 20, 2020 8:46 AM
To: ke6pcv@cal-net.org
Subject: [AllStarLink Discussion Groups] [App_rpt-users] Known good small X86 fanless computer that wont skip some audio
Pierre_Martel
May 20
Anyone can give me some information on a small form computer that will run allstarlink without skipping audio.
Right now i am using a raspberry pi3 with a URI-X and while trying to adjust my audio and running a vanilla system deside using the dummy dadhi app I can clerly hear small gap in the audio output that is received by my receiver. My spectrum analyzer will also give me a small bleep in the deviation of the signal.
I knew that it was possible that thing like that happen but I want this 900 mhz repeater to be perfect cause it will be the first one running for that club on allstarlink and this cant be ok.
If someone want to propose me to run the Hamvoip distro, no I wont. the day they will release all there source I could be thinking about it. But till then forget it.
I did try the rpi4 image that is in testing and I was running in the same small cut in the audio.
Dont want a embeded system that can tun games. 
Pierre
VE2PF
Visit Topic or reply to this email to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these emails, click here.
Thanks Marshall
Will take a look at those.
My experience is that the Atom D5xx boards ran the old Centos 5 version just fine. But not the current version (ASL1.01). I now have a box full of retired Atom boards.
First - the dahdi score reportedly has to do with the ability to keep packets in order when processing multiple audio streams. This is of interest in DVSwitch where it is mentioned extensively. As I recall, 99.975 is the lower limit - so your HP would be fine. This is a tricky test - even very fast computers can fail this test if things are not set up right.
Second - drop-outs in the audio are related to USB updating and dahdi score is not directly related to that.
I tested a number of small mini-PC’s with the current version and found there is a loose association between CPU “horsepower” and gaps. More powerful machines don’t fix the problem (as you found out) but reduce it in two ways. One is that you are likely to get the gap only once every 3 minutes or so, which I think is acceptable. The other way is that the gaps become shorter and - importantly - you get single gaps instead of multiple ones in a row.
My tests on mini PC’s showed that a Passmark score of about 3000 was required for reasonable results. Higher score if you want to run more than one fob and get the same results.
For comparison, an Atom D525 scores 705. I don’t know about the D9xxx series boards - its not listed on Passmark under Intel Atom. There are some Atom boards that score above 3000 but they tend to be very expensive and not good value for money.
I know the ASL team is looking at the causes of the gaps and have made some progress at fixing it. Stay tuned, as they say …
Ken
Pretty good explaination. Thanks Ken,
Right now, after a reboot I have some cut them it stabilise and I get a cut out (very small) every few minute.
I have 20 board to build voters, I have all the stuff to build at least 5 full board.
Think I will be building some voters and will interface it to the Quantar.
The only thing I am not sure to do in building those board is loading the program into the micro controler.
Will read on this…
Pierre
VE2PF