Tvheadend

Found out it seems to fail only if there are different satellite mixed AND mixing DVBS with DVBS2. When selecting only multiplexes of DVBS over different satellites it works OK.
Strange thing; when on the same satellite mixing DVBS and DVBS2 work also fine?!

so let me get this clear;

you have one sat-disc, with a monoblock that can receive 19.2 and 23.5. With that you can only receive one sat at the same time. You need to send diseqc signals to have the monoblock switch between the 2 sats. So if you switch from 19.2 to 23.5 the muxes on 19.2 can no longer be received. So to me it would make sense that tvheadend reports that there is no signal anymore on the 19.2 mux because you switch the the other sat.

What is the real problem? only that line in the logs? or can you also not watch/record the new channel that you switch too?

Hein

It's the 'idle scanning' failing, so only one of both sats is active at the time. If you look at the logs you see, one of the transponders, the status alternates between "OK" and "No signal". Also notice it fails when a DVB-S transponder is scanned after a DVB-S2 transponder.

feel free to look at my config : http://hoc.xs4all.nl:9981

Jeroen

Your remark about the idle scanning trigger my memory. There was a bug with idel scanning and s2api that should be fixed in the latest git version of tvheadend. So could you please try a latest version by either compiling it yourself, or by using a ubuntu repository that keeps a recent git version of tvheadend available (for example lard op den kamp's repository)

Hein

I tried the latest git, same problem there. I'm thinking this is most likely a driver problem. Are there more people on this list with a Tevii S660 ?

Jeroen
ps. oscam, or probably my CD-card, stopped working. I'll be debugging this first.

increase msleep into diseqc code + add msleep after "FE_SET_TONE" + send "FE_DISEQC_SEND_MASTER_CMD" 2x and all works...

RoboSK,

Maybe you can supple a patch with these changes in it ... The he can try this himself.

Hein

.

Thanks,

Applied the patch but the problem remains. Also tried to increase the msleep() even higher.

I've attached another log; notice the scanning fails when a DVBS channel is scanned after a DVBS2 channel, I hope this will trigger a developers mind.

Jeroen

Hi,

I can confirm this problem.

Problem exist on latest git (git-d66746c) and on same latest git with diseqc patch.
Current configuration:
HTS Tvheadend git-d66746c
Tuner: Tevii S660.

Some tests:
D0 - Diseqc port with Linear universal LNB
D1 - Diseqc port with Circular 10750 LNB
H - Linear Horizontal
V - Linear Vertical
R - Circular Right
L - Circular Left

Switching tests results:

BAD (this mean that transponder locked after one-time switching R->L->R or L->R->L):
V->R
H->L

VERY BAD (this mean that transponder locked after many times switching R->L->R or L->R->L):
V->L

OK:
H->R

Second test (after moving of Circular LNB from port2 to port3):
D0 - Diseqc port with Linear universal LNB
D2 - Diseqc port with Circular 10750 LNB
H - Linear Horizontal
V - Linear Vertical
R - Circular Right
L - Circular Left

BAD:
V->R
H->L

OK:
H->R
V->L

It's very strange, because switching not work as expected when voltage is not changed. (Vertical or Right circular - 13V; Horizontal or Left circular - 18V)

Third test:
Voltage measurement on DVB-S output and on diseqc switch output's.

Voltage at the output of the tuner is always right (regardless of outcome).
On DiSEqC result is very strange: when FE can lock transponder, voltage and port is correct; when FE unable to lock transponder - voltage is correct, but port incorrect!
When port should be P2, after switching it can be P0 ir P1 !

I'm afraid it's a driver problem. Not sure though..

I also thought so, but then in the russian dvb forum I found a solution to this problem in the linux-vdr:

For some DiSEqC required to be repeated command with E1.

Example:
S53E 11700 V 9750 t v W15 [E0 10 38 F0] [E1 10 38 F0] W15 t

This solved 90% problems with diseqc switching (some times switching V->L not work, but this can be solved by switching V->R->L or V->L->R->L).

So, now I don't know where is issue: DVB-S driver, DISEqC switch or tvheadend diseqc implementation

My current patch attached. It is temporary solution, because I am not sure where is the problem. At next month I get a new diseqc switch from dvbshop - then I can eliminate (or ensure that I have problematic diseqc) the switch from the list of problems.

I can't test your patch at the moment. When my setup is available again I will test your patch.
Please keep me posted on your findings here. Thanks !

Hi,

some updates:
A new diseqc switch don't solve issue.
New dvb-s2 card, which I try to use, have same issues, but i think, that it is due same driver. My new card - Mystique SaTiX-S2 Sky Xpress DUAL, USALS, DiseqC 1.2. I bought it due a different chipset (should be STV0900B/STV6110A), but in new hardware revision, this card uses almost the same chipset as Tevii. (Tevii - ds3000/ts2020, Mystique - ds3103/ts2020).

Hi,

Tried the above patch without luck. Seems the behaviour changed a bit but still fails.

Jeroen

I bought a Technisat Skystar USB HD : http://www.linuxtv.org/wiki/index.php/Technisat_SkyStar_USB_HD
Diseqc problems are gone.