[mythtv-users] PCI, S-Video, VDPAU, SDTV, interlaced CRT: what card?

[f-myth-users at media.mit.edu]

    > Date: Wed, 16 Dec 2009 17:31:38 -0500
    > From: sdkovacs <sdkovacs at gmail.com>

    > I don't know what that reviewer is talking about with regards to
    > svideo not working on Linux. That review is the first I've heard of
    > any problems on Linux and I'm running on CentOS 5.4. I have the
    > fanless Sparkle 8400GS PCI 512MB linked below. Running via *s-video*
    > (and using vdpau advanced 2x) to a 27" Sony CRT.

    > http://www.newegg.com/Product/Product.aspx?Item=N82E16814187042

Aha!  That's -extremely- encouraging!  Okay, in that case, it sounds
like that's the card to get, and I should just ignore the review that
claimed it didn't work with Linux---sounds more like he was confused
about connector issues.

I assume you either found a much-longer 4-pin cable -without- locking
tabs, or modified one, or used a female/female to join their short
cable with a longer one of your own?  (I can't believe how short a
cable they shipped; I'd basically have to butt the CPU up against
the center back of the TV to be able to use it. :)

Thanks!

P.S.  Sounds like several reviews were claiming they tend to die
early.  I'm not sure what to make of that.  Inspection of the PCIe
Sparkle I have here shows an enormous heatsink that overhangs several
electrolytic caps, which wouldn't be so bad except that in every
motherboard/case I've ever seen, cards go in with the component side
-down-, and hence that heatsink is -upside down- and air has to waft
up through it and then work its way past those caps and along the
(horizontal) board until it finally spills up past the edges.  Hardly
ideal thermal design.  Same problem with fans, of course, except then
at least there's some air pressure to force the hot air sideways until
it can leave the vicinity of the card...  And fanless cards should run
cooler, whereas this one stabilized at about 55-60C at the heatsink
whether idle or loaded (measured with an IR thermometer); the top side
of the board (e.g., opposite the components, or what I'd call the
"trace" side in a one-layer board) was at the same temperature (so I
guess -some- of the heat is being dissipated that way as well).  This
was in a wide-open case; a closed-up case with a case fan might lower
those a little bit, or might not.