[vdr] Re: burned DVB-s

[Thomas Breuer <tb@tb-electronic.de>]

On 23/10/2002 01:17 Roberto Deza wrote:

[Potential differences]
> Measuring the voltage between two *un-referenced* AC devices, many times
> I've saw 70~90 V AC, and this *voltage* will break electrostatically many
> semiconductors.

Yes, ungrounded metal cases may be under voltage due to stray
capacitances in transformers or Y-capacitors at the mains input etc.
However, the current potential is quite low, so that series resistors
plus clamping diodes usually protect inputs and outputs. Unprotected
semiconductors, of course, could easily break.

The outputs of the AV7111 seem to be quite robust, but the developers
could have at least added some series resistors. A buffer would have
been even better, but probably too expensive. Anyway, I don't feel very
well knowing that output pins of one of the most important parts of my
DVB card go directly to other mains powered devices ...

[Hitting BGA balls]
>>Hitting the correct balls might be possible, but it is a multilayer PCB,
>>so you might hit an inner supply layer and end up with a short.
>
> Sure, but one could over-drill, fill with epoxy or ceramic compound and
> redrill after cured. I agree that is too risky and not to be done "at
> hand".

I also thought about this, but we have to remember that the balls are
arranged in a 1.27 mm grid. There's not that much space to drill large
holes ...

>  About the layout (~coordinates) I offered a high resolution scan of a
> X-ray picture, of a v2.1 board (now, it could be also of a v1.3).

I remember that I saw your PCB drawings (no x-ray scans), but the link
http://www.baycom.org/~hannej/dvb21-rgb/ is no longer available. Could
you sent me the scans by PM? Not that I need the info, but I'm curious
to see some x-ray pictures of the DVB card.

[extracting RGB/CVBS of compensation components]
>>Well, before I applied the brute force method to the dead card, I
>>checked all pads and vias around the AV7111 to find at least h and v
>>sync signals, but as far as I can say there is nothing!
>
> Very likely, but it could work on a non_damaged card and solve the
> problem of rev>=2.1 cards :-).

This was a bit unclear: I did indeed check the pads and vias on a
*working* card.

Best regards,

Thomas







-- 
Info:
To unsubscribe send a mail to listar@linuxtv.org with "unsubscribe vdr" as subject.