setting up a tevii s660
Quoted 7 lines · Tim Coote
Hullo
I've been struggling with this for a couple of days. I have checked archives, but missed anything useful.I've got a tevii s660 (dvbs2 via usb). It works with some limitations on windows xp (I cannot get HD signals decoded, but think that's a limitation of the software that comes on the CD).
I downloaded version from tevii.com and it worked for me on laptop running win 7. Had problems with HD on an XP machine but I assumed it was because the video card was old/slow.
Quoted 10 lines
I'm trying to get this working on Linux. I've tried VMs based on fedora 12 and mythbuntu (VMWare Fusion on a MacBookPro, both based on kernel 2.6.32), using the drivers from tevii's site (www.tevii.com/support.asp). these drivers are slightly modified versions of the v4l tip - but don't appear to be modified where I've not yet managed to get the drivers working :-(. Mythbuntu seems to be closest to working. Goodness knows how tevii tested the code, but it doesn't seem to work as far as I can see. My issues could just be down to using a VM.
I tried on Ubuntu 9.10 but had problems which I documented here on 16 april. Loading the firmware worked fine but there were problems with remote control messages being logged continually as well as stability problems. The card would tune (with scan) and worked with mythtv (for a day or so)
Quoted 2 lines
I believe that I need to load up the modules ds3000 and dvb-usb-dw2102,
+ add a rule to/etc/udev/rules.dand a script to/etc/udev/scripts.
I didn't touch rules.d
Quoted 9 lines
I think that I must be missing quite a lot of context, tho'. When I look at the code in
dw2102.c, which seems to support the s660, the bit that downloads the firmware looks broken and if I add a default clause to the switch that does the download, the s660's missed the download process. This could be why when I do get anything out of the device it looks like I'm just getting repeated bytes (the same value repeated, different values at different times, sometimes nothing). I'm finding it non-trivial working out the call sequences of the code or devising repeatable tests.
Had no problem with Ubuntu recognising the device and the correct .fw file being downloaded.
There are various versions of dw2102.c from tevii, etc. I think I tried all of them. I did change the timeout on RC messages (dw2102.c?) which helped but did not cure the problem.
Also tried the s2-liplianin library as well which seemed promising but also did not cure the problem.
Quoted 8 lines
Can anyone kick me off on getting this working? I'd like to at least get to the point where scandvb can tune the device. It does look like some folk have had success in the past, but probably with totally different codebase (there are posts that refer to the teviis660 module, which I cannot find).
Any pointer gratefully accepted. I'll feed back any success if I can be pointed at where to drop document it.
I didn't have the knowledge (or time) to decide if the problem was in the firmware or the v4l drivers or perhaps some strange interaction with my dvb-t usb box.
Exchanged some emails with tevii guys who had posted here however they appear to take the view that if it works under windows then the device is fine and offer no other solution.
In the end I bought a Nova S2 PCI card which works fine but would be interested in trying to get the S660 working.
paul
Quoted 4 lines · Paul Shepherd
[snip]
I downloaded version from tevii.com and it worked for me on laptop running win 7. Had problems with HD on an XP machine but I assumed it was because the video card was old/slow.
that seems a fair test. you've got to know that the device itself's working first
Quoted 1 line
[snip]
Quoted 6 lines
I tried on Ubuntu 9.10 but had problems which I documented here on 16 april. Loading the firmware worked fine but there were problems with remote control messages being logged continually as well as stability problems. The card would tune (with scan) and worked with mythtv (for a day or so)
I get this output, too. I think that it's debugging code that's been left in. I'd guess that the driver polls the device periodically and spits out the results. If that's right, then you can remove the messages from dmesg and /var/log/messages by commenting out the relevant info lines in the source (should be lines 1131, 1137 and 1155 of ./linux=tevii-ds3000/linux/drivers/media/dev/dvb-usb/dw2102.c). I think that line 84 implies that there was an intent for these messages to be debug rather than info, but I'm not familiar with the coding standards enough to fix that. I'd guess that the crash that you experienced comes from something else, but goodness knows what or how to debug it.
Quoted 5 lines
Quoted 3 lines
I believe that I need to load up the modules ds3000 and dvb-usb-dw2102,
+ add a rule to/etc/udev/rules.dand a script to/etc/udev/scripts.I didn't touch rules.d
Since my initial post, I've tried a vanilla xubuntu 10.04, (at the suggestion of a member of the list who'd got it working with this build) which still only populates /dev/dvb/adapter0/{demux0 drv0 net0} and not frontend0. See below. [snip]
Quoted 5 lines
Had no problem with Ubuntu recognising the device and the correct .fw file being downloaded.
Quoted 7 lines
There are various versions of
dw2102.cfrom tevii, etc. I think I tried all of them. I did change the timeout on RC messages (dw2102.c?) which helped but did not cure the problem.Also tried the s2-liplianin library as well which seemed promising but also did not cure the problem.
I'm less sure now about my conclusions about downloading the .fw. Certainly on xubuntu *something* is happening as I get a blue led flash on plugging in the tevii.
Quoted 6 lines
Quoted 3 lines
[snip]
didn't have the knowledge (or time) to decide if the problem was in
the firmware or the v4l drivers or perhaps some strange interaction with my dvb-t usb box.
Quoted 7 lines
Exchanged some emails with tevii guys who had posted here however they appear to take the view that if it works under windows then the device is fine and offer no other solution.
In the end I bought a Nova S2 PCI card which works fine but would be interested in trying to get the S660 working.
I don't have a pc that's big enough to take a pci card, and the hardware clearly works. However, to debug what's going wrong, there needs to be a repeatable test process, and a baseline system where the different failure modes can be tied down. That's why I set up the xubuntu system and used a .config file thats known to work. However, there's still other stuff to load on a vanilla box, including:
build-essentials
linux-headers-2.6.32-22-generic
and possibly libncurses5-dev and a complete kernel source tree, although I've avoided this by reusing the .config file.
I'm still assuming that 'working' means that I can use dvbscan (which needs the frontend0 file, I think, although the reported error on xubuntu is different from mythbuntu and it just says "Failed to open frontend".
I'm afraid that in a situation like this, I don't know what I don't know. And I don't know what I can rely on or what success is. Not a great recipe for getting anything to do what you want.
I can understand tevii's point of view, but I need to get this working on Linux. I don't think that Tevii wrote the code and it is slightly modified from the mercurial code, so there needs to be some sort of merge back if and when it is ever made to work.
Quoted 1 line
paul