I've seen similar issues to this reported by other people, all of which were usually solved by cleaning and reflowing, but I've now tried so many times that I'm completely stumped and would appreciate expert advice on what to try next. I have a mostly-built SE/30 Reloaded, using this PCB, with all of the surface mounted components installed. All of the array logic is also installed. The only ICs I skipped were the Bourns filters. These chips were all working pulls from an SE/30 that worked reliably 3 months ago, but started developing some issues with trace rot from residual corrosion. I didn't want to deal with electrolyte wicking down vias or anything in the future, since it was not originally my machine, and I am frankly not cut out for 6-layer PCB repair.
Since I'm trying to get a minimally working system that walks the bus, or ideally blares some "no RAM" death chimes, I have not installed any connectors except for the speaker pins, the power/video socket, the PLCC ROM sockets, the CPU/socket and the battery holder. Unfortunately, all that seems to work are the reset circuit and the video PALs with a stable NEC Simasimac pattern. Nothing gets particularly hot and I can't find shorts to ground anywhere obvious. I've been reflowing things that don't look perfect, while checking pairs of adjacent pins for continuity without anything beeping at me. (Well, except 1, 6, 10 and 11 on UH7, but those are all ground!)
Somehow, the clocks are all over the place. I see something like 24Mhz from Y2 sometimes, but it's never anywhere near 31.3344. The 16Mhz clock generated from GLUE using this is absent. The scope shows the waveforms from Y2's pin shaking violently, too. When I put in a battery with the board off and probe Y1, I see the perfect, intended 32,768hz from the RTC. However, when I power the board up from DC, it's complete havoc. The reset circuit goes from 0V to 5V as it should, every chip connected and showing the same voltages, but the clocks from Y1/Y2/Y3 are all over the place, and the waveforms on the scope bounce up and down violently. Have I shorted something obvious and polluted my clock signals? I'm probing the legs of the oscillators/crystals themselves, so it really feels symptomatic of a short somewhere, but nothing's jumping out at me. Maybe there's some residue from flux/washing stuck between the legs of the CPU socket? I'm just unsure how it could have this sort of an effect. Especially since I have an old junk board with its CPU removed that does approximately the same thing when powered up.
The AB is recapped and reflowed and works great, the power supply is a modular Meanwell replacement with good +5/+12/-12 every time I probe it, and I've checked the voltages on everything thoroughly. The Sony chip putting RESET at +5V implies the voltages stabilize within what the hardware expects. I do not have a working SE/30 board to test with, but this Mac was originally an SE, and my SE board works fantastically in it with no transient issues. Something on the board has to be the X factor.
Since I'm trying to get a minimally working system that walks the bus, or ideally blares some "no RAM" death chimes, I have not installed any connectors except for the speaker pins, the power/video socket, the PLCC ROM sockets, the CPU/socket and the battery holder. Unfortunately, all that seems to work are the reset circuit and the video PALs with a stable NEC Simasimac pattern. Nothing gets particularly hot and I can't find shorts to ground anywhere obvious. I've been reflowing things that don't look perfect, while checking pairs of adjacent pins for continuity without anything beeping at me. (Well, except 1, 6, 10 and 11 on UH7, but those are all ground!)
Somehow, the clocks are all over the place. I see something like 24Mhz from Y2 sometimes, but it's never anywhere near 31.3344. The 16Mhz clock generated from GLUE using this is absent. The scope shows the waveforms from Y2's pin shaking violently, too. When I put in a battery with the board off and probe Y1, I see the perfect, intended 32,768hz from the RTC. However, when I power the board up from DC, it's complete havoc. The reset circuit goes from 0V to 5V as it should, every chip connected and showing the same voltages, but the clocks from Y1/Y2/Y3 are all over the place, and the waveforms on the scope bounce up and down violently. Have I shorted something obvious and polluted my clock signals? I'm probing the legs of the oscillators/crystals themselves, so it really feels symptomatic of a short somewhere, but nothing's jumping out at me. Maybe there's some residue from flux/washing stuck between the legs of the CPU socket? I'm just unsure how it could have this sort of an effect. Especially since I have an old junk board with its CPU removed that does approximately the same thing when powered up.
The AB is recapped and reflowed and works great, the power supply is a modular Meanwell replacement with good +5/+12/-12 every time I probe it, and I've checked the voltages on everything thoroughly. The Sony chip putting RESET at +5V implies the voltages stabilize within what the hardware expects. I do not have a working SE/30 board to test with, but this Mac was originally an SE, and my SE board works fantastically in it with no transient issues. Something on the board has to be the X factor.
