[pvrusb2] B&W video problem...

[Mike Isely]

On Wed, 4 Jan 2006, Helmut Toplitzer wrote:

>
> Have you tried some "non software" tests?
>
> Temperature instability?
> Mechanical shock?

When this happens, the device is not being touched.  And I can power 
cycle-clear the problem without touching it.  So I don't think a 
mechanical disturbance is doing this.  That doesn't necessarily rule out 
temperature.


>
> On NTSC color info is coded and modulated on seperate
> frequencies to keep backwards compatibility to B&W.
> So they are normaly filtered out and amplified/decoded
> in a seperate way.

Right, that's the 3.57MHz colorburst frequency, whose relative phase and 
amplitude provides hue and chrominance information, respectively, IIRC. 
If it isn't filtered out, then you can see the overlaid colorburst 
frequency in the B&W video - which I am definitely seeing.


>
> Maybe your tuner has got a problem with a cold soldering
> point or a dieing electrolytic capacity?
>
> Cold solderings are sometimes detectable by turning the device
> upside down, capacity problems with ELKOs happen during time
> and are also temperature sensitive.

I suppose that is possible.  Knowing that somebody else has this model of 
tuner and either is or is not seeing this problem would provide a strong 
datapoint here.


>
> Does this problem also happens on MS-W?

Duh.  Never thought to try it!  My brain is so locked into Linux these 
days.  It will take some time to set that up, but it is a worthy test.  So 
I'll try it.


>
> My 2c.
>

Just got another very interesting datapoint.  After the device flipped to 
B&W again, I plugged a DVD player into its composite-in and switched the 
device to that input.  Result: Still B&W video.  I also tried the S-video 
input (which should NOT involve colorburst frequency demodulation) and it 
was still B&W video!  This would seem to exonerate the entire tuner 
section of the device.  The only common element that makes sense there I 
believe is the saa7115.  Input selection is internal to the saa7115 and 
the output is all digital bits at that point.  That would seem to narrow 
down the cause to be in the silicon of that chip.  Perhaps this is a 
different saa7115 rev which is causing trouble.

Time to get some sleep...

   -Mike


-- 
                         |         Mike Isely          |     PGP fingerprint
      Spammers Die!!     |                             | 03 54 43 4D 75 E5 CC 92
                         |   isely @ pobox (dot) com   | 71 16 01 E2 B5 F5 C1 E8
                         |                             |