MegaSTE Repair Continued - Video Noise

[displaced]

Unfortunately yes, without any change. Poked and prodded the thing from all angles looking for some sort of change.

I suppose it’s worth reflowing the solder on the pins? It looks nice and clean, but you never know!

It scared the crap out of me just now - first power-on after re-fitting the shifter gave a whole screen of corrupted purple and green. Powered off and on again and it came back to life.

[exxos]

Yeah resolder the socket.. Though you may need to replace the socket.. bending pins does not always solve the problem. sometimes placing a thin layer or 2 of cornflake box card under the chip can lift it up and make better contact.

[displaced]

Ok - tried all the things, and no luck! But it still works, so I’ll take that as a win.

Removed the shifter then fluxed and reflowed each pin. Was a little taken aback by the fact that the pins move in the holes, so I had to nudge a few to re-centre them. First time soldering a PLCC socket. Possibly too brave, but seem to have got away with it. Tried the cardboard trick too, but no dice.

While underneath it, I took a closer look at the rest of the board. There were quite a few parts where a lot of flux residue was present - mostly around the modulator (the L510 pot in particular, and also the beige, rectangular thing (fancy cap?) at DL1)

Also the SN74LS06N at U406 up by the ACSI port has had a socket fitted - the solder was black with dried flux and it’s the only machine-pin socket on the board.

Looks like the keyboard socket had been at least resoldered if not replaced - yet more gobs of dried flux.

Cleaned up all the fluxy bits and reflowed them. Clean and shiny now.

After all that I thought I had a good chance of it coming good, but alas, no.

Still, I’m glad that other than those lines, it works and is stable!

[exxos]

What is it like on diagnostic cart on colour test ?

[displaced]

Afraid I’ve not picked up a diag cart yet - will pick one up in the new year!

Just researching the replacement of the PLCC socket. Looks like it’s a ‘funky’ one. Is your adapter for these available in an 84-pin version, or is my best bet to start looking for busted STEs to harvest one from?

[exxos]

Looks like it’s a ‘funky’ one. Is your adapter for these available in an 84-pin version, or is my best bet to start looking for busted STEs to harvest one from?

ah no, I only did 68 pin ones.. So harvest is probably the only source.. BEST may have them at high prices..

[displaced]

Well, I got a donor STE with the same shifter version.

Did a straight swap into the MSTE and no change - same interference is present.

Thought I might as well try swapping the sockets. Took it really slow and careful. Actually went fine and the machine started up first time. But still, very noisy video.

I suppose it’s worth recapping around the video circuitry, if not the entire board.

Oh and it turns out the weird corrupt-screen failures to boot were actually due to an earlier repair on the U11 GAL which had a broken leg. I re-did that repair and it’s now rock-solid. So the original shifter was probably fine.

[exxos]

Your not picking up interference from some motor, such as from a hard drive or fan or something ? Have you scoped the RGB lines to see if any noise pattern on any of them including syncs ?

[displaced]

Thanks, exxos -- the only thing with a motor would be the PSU fan, which is a new Noctua fan. I unplugged it briefly, but the interference continued.

There's no HD, just an UltraSatan and the internal floppy. The interference was present before the US was installed, and is still there when the floppy's not in use.

I've scoped out hsync, vsync and the RGB signals just now.

First up, hsync:

hsync.png

Looks more dramatic in action, but definitely looks noisy to me.

Now vsync:

vsync.png

This is an interesting one -- the pattern of noise travels 'crawls' along the signal as the scope samples it. This shows one 'set' of interference moving through the pulse on the right, while a new instance of the interfering signal appears on the right.

Finally red - green and blue were essentially the same:

red.png

This was tricky as my scope didn't know how to catch a trigger, so this is an action-shot. I think it's still not especially clean, showing bursts of noise similar to that on hsync's trace.

So, no 'ah-hah!' moments for me because I'm basically an amateur poking about! But would be very interested to hear if you see something here that might point to a possible source or fix.

Thanks again!

[exxos]

Odd.. Have you ran YAART to test the ram ?