SE/30 dashed horizontal line

[terminalcancerman]

This SE/30 didn’t have any battery leakage, but the aluminum electrolytic were leaking. Recapped with polymers. I only found verdigris on the legs of one IC southwest of the video chip and cleaned that up. I did run the board through ultrasonic without first removing the crystals. At first I thought this may be vertical collapse as I haven’t touched the analog board, but other examples I’ve seen aren’t dashed/broken lines like this. It gives a happy Mac chime and displays drive activity. Can anyone point me in the right direction? I’ve performed some searches and didn’t find anything that matched this.

 

[finkmac]

perhaps vertical collapse together with wonky video memory muxes?



these guys often get damaged by the leaking cap juices... and they can cause the "simasimac" screen. a vertical collapse might compress that and cause a dashed line?

 

[terminalcancerman]

One of the chips in that bank, the one to the immediate SW of C7 had verdigris at the package seam and is suspect. Thank you for your confirmation that these are potentially related.

 

[finkmac]

check continuity of all traces/legs in the video circuit... the redrawn schematics on github are good for that.
also consider just removing the affected ICs that are near the capacitors... cleaning under them and replacing with brand new parts.

 

[terminalcancerman]

check continuity of all traces/legs in the video circuit... the redrawn schematics on github are good for that.
also consider just removing the affected ICs that are near the capacitors... cleaning under them and replacing with brand new parts.

Ok, thanks so much

 

[obsolete]

Yeah, fink is right, that row of chips is the root cause of most video problems. I have also found that no vsync can be a PAL problem.

 

[terminalcancerman]

Yeah, fink is right, that row of chips is the root cause of most video problems. I have also found that no vsync can be a PAL problem.

Fortunately the multiplexers are jellybean parts so I’ll do a quick swap. Thanks for this tip. I’ve identified the Bolle Jedec files I would need, so I’ll grab a stack of GALs as long as I’m having parts shipped

 

[terminalcancerman]

Update. I replaced the video mux chips and now I’m getting a pretty simple simasimac pattern. I believe an analog board recap should fix this right up. Thanks for all the suggestions, everyone.

 

[obsolete]

No, that's not simasimac, but it's better in a way. This is a logic board problem, not an analog board problem. You have working video, just missing a vertical line. Probably a broken trace somewhere.

 

[joevt]

Update. I replaced the video mux chips and now I’m getting a pretty simple simasimac pattern. I believe an analog board recap should fix this right up. Thanks for all the suggestions, everyone.

Since every 8 pixels is broken, you're looking for an 8 bit data bus probably.

Looks like Data Bit 0 is broken. The right side of each byte is stuck black. The left side of each byte is Data Bit 7. I compared your image to an image taken from infinitemac.org with bad disk - assuming your Mac and the infinitemac draw the floppy in the same location - using the Preview.app selection tool to count the pixels to the floppy icon - also the top right corner shows the right most line of pixels is black instead of checkerboard in your image.

 

[terminalcancerman]

Since every 8 pixels is broken, you're looking for an 8 bit data bus probably.

Looks like Data Bit 0 is broken. The right side of each byte is stuck black. The left side of each byte is Data Bit 7. I compared your image to an image taken from infinitemac.org with bad disk - assuming your Mac and the infinitemac draw the floppy in the same location - using the Preview.app selection tool to count the pixels to the floppy icon - also the top right corner shows the right most line of pixels is black instead of checkerboard in your image.
View attachment 90828View attachment 90829View attachment 90830

Thank you for so much detail. I'll start by testing out the SDQ lines on the VRAM ICs and work from there. If it's bit 0 then it should be pin 2 on one of them but I'm not familiar with the schematic so it could potentially be any of them. And if only one bit is bad, then I should be able to get a pattern by scoping the lines and determine if an IC is bad.