[pvrusb2] Black screen, device stuck, need to power cycle to clear

[Bill Kristan]

Wow, Mike, thanks for the long, detailed response (over something that's not 
likely related to your driver!). I have used the device in Windows just to 
check if it was working, but I didn't run it long enough to see whether a 
transient problem like this could be cropping up. Heat could definitely be an 
issue - it's not a well ventalted space, and we're having a heat wave this 
week here (So. California). I'll rearrange the hardware this evening to 
improve ventiation and try again. So close...

Cheers,
Bill

On Monday, September 27, 2010 08:22:12 am Mike Isely wrote:
> On Mon, 27 Sep 2010, Bill Kristan wrote:
> > New problem... my wintv-hvr-1950 now works only briefly, then stops. The
> > device node is still there, and I can play it (with mplayer or mythtv),
> > but it gives a black screen. The first time I saw this, dmesg gave a
> > "device stuck" error from pvrusb2. Removing and loading the module
> > didn't fix it, nor did rebooting the computer, but power cycling the
> > capture device did. I was able to play /dev/video0 with mplayer, but
> > when I tried to record with mythtv I got a black recording again. No
> > error message was reported the second time.
> > 
> > I'm using the firmware extracted from the driver CD.
> 
>    [...]
> 
> > Not sure what to try ... maybe go back to the firmware from isely.net?
> > Seems the version shipped with the device would be best, but it would be
> > something to try.
> 
> Software normally doesn't "break" by itself - if it was working before
> and you didn't change anything then it should keep working.  (Well this
> isn't quite true; MS software seems able to break by itself...)
> 
> If it was working before and now it isn't you have to ask yourself what
> has changed?  Also, generally if it works at all - even briefly - then
> the software is probably OK and you have some other sort of problem.
> 
> Based on your description it sounds like the FX2 in the device itself is
> crashing.  Of course that should *never* happen and there's really no
> way for the pvrusb2 driver to cause that, short of downloading corrupted
> FX2 firmware - but even then if it's working briefly then the firmware
> is probably just fine.
> 
> I'd probably turn my attention to other possible causes for the device
> to crash.
> 
> Is it getting too hot?  (Having it work for a short time before crashing
> each time is something that can be caused by heat.)  I've never heard of
> an HVR-1950 actually overheating, but the other PVR-USB2 models
> sometimes had this sort of issue.  I learned of one case years ago where
> a user had stacked 4 PVR-USB2 devices inside a closed box with no
> ventilation, and that eventually caused the hardware to die from
> overheating.
> 
> Is it possible that it's getting hit by power spikes?  Remember that
> your device is getting power from its own little power brick and that
> brick isn't anywhere near as resilient against power line flakiness as
> your PC's internal power supply will be.  I have heard of numerous cases
> where power glitches in the nearby environment (e.g. A/C compressor
> turning on in the house) will crash your capture device.  You could try
> moving it to a different circuit, putting it behind a surge protector,
> or (best choice) plugging it into a UPS.  If any of that helps then the
> problem is dirty power.
> 
> Another possibility is that the power brick itself is failing, drifting
> out of tolerance.  This was exactly the problem a while back with the
> Silicon Dust HD HomeRun tuners - they had a bad batch of power bricks
> which would fail after a few months.  Now, I have *NOT* heard of that
> happening here with Hauppauge, but power bricks frequently are a weak
> spot with externally powered devices.
> 
> One big lesson I learned years ago when I was messing with TTL circuits
> (yes it was a LONG time ago) - if the circuit is misbehaving then always
> suspect the power first.
> 
> Can you run the tuner successfully under *cough* Windows *cough*?  If
> there really is some kind of external problem that's killing the tuner,
> then the answer to that question should be "no".
> 
>   -Mike

-- 
Department of Biological Sciences
California State University
San Marcos, CA 92096
voice: (760) 750-8096
fax: (760) 750-3063
wkristan at csusm.edu
http://www.csusm.edu/wkristan