You're gonna love this one;

I'm working on this 53XBR45 without a print. I'm not that unfamiliar with this
type of unit, so I didn't scream "print, bloody print" when I saw a pretty
strange symptom.

Every second, or a bit over, for about 1/10th of a second the geometry would
change. In the "normal" state, the side pin is off, but this is right for about
100mSec every second.

I looked around in there and I find a board with a micro, the jungle IC, and an
EPROM on the back.

After fiddling around in the menus (at first just the customer ones) I found
that changing the color setting might make the set go green, that is shutoff
red & blue, not slick green. Other abberations exist when I get into the
service menu. None of the values "look" corrupt, changing them frequently has
the wrong effect, or no effect, or an effect until the thing cycles again.
Changing certain things changes the effect when it cycles, but it always cycles
at the same rate.

Now I got a micro, jungle and an EPROM on the same board, so I decide to take a
shot at the EPROM. Sony EPROMs are not known to be a big problem, at least to
me, but I know the higher failure rate of them versus other chips, except for
power devices. Reasonable, order the part, especially since it's not expensive.

I change it expecting one result, but what I got was the same cycling, but now
what it does is flash green. Not slick green, but as if the drive was say
tripled during the short part of the cycle. Let me add this, the picture
appears as though the picture and sharpness are turned all the way down, but
the OSD dosn't confirm this. This was true before as well, it did not change in
that respect. The new chip of course isn't set to the set, and now, why bother
? I did save the old one :-)

Now I am a bit fuzzy on the data structure and the architchture of this set,
but it wasn't built on Mars. After talking to Sony tech ($$) they said a
different IC (EPROM) than that which I had changed. I changed IC3002, the EPROM
they want me to change is over in the assembly with the tuners and PIP circuit.

Now after a bit of thought I wonder if we're BOTH WRONG.

Well known fact: these EPROMS and all the chips that run off them are serial
data. different data are seperated by timing. These chips are externally
clocked. They are clocked by the micro right ? as is everything else ? (not
certain parts, especially of the jungle IC)

While all the other chips might actually generate data, they don't generate the
timing (CLK) right. Only one chip times the system, everything else just reads
that. I know that's not true of an RCA DTV series, but it should be true of
this set. Although now I may have to regroup, you see, the data are not
changing, the abberation is only affecting reads, obviously, so either we have
a read that is being obliterated by something else on data & clk. It has to be
a chip out of sync. It has to be because it corrects itself, but the it's
screwed up again for 9/10th of a second. How come those EPROM values aren't
changing, and by now would've changed possibly to values that would not allow
the set to run.

Here's the kicker: When I changed that EPROM, the set remember it's channels
and all settings (or so it thinks), it was still programmed. This leads me to
believe the other EPROM is OK. It has to be an IC that can output data, some
don't. It has to be something with it's own clock, and THAT is what's out of
sync. Of course an EPROM does put out dat . . . . .It could be, I will change
it.

If anyone ( I am at a loss for words ), thanks in advance.

JURB