derkom's attempt

[derkom]

I only suggested a few pins though :)

Assuming you are not using the blitter, make sure BR,BG,BGACK are high all the time.

Yeah, I pulled the blitter, just to remove the variable (and did set the jumpers right). This is turning into two steps forward, eight steps back, because it's not even attempting to boot most of the time now. :roll:

Going to have to do more close inspection and probably some more reflowing, but that is unlikely to happen tonight. Neighbours' fireworks are making too much noise to concentrate. :lol:

[derkom]

Okay, after (re-re-)reflowing all the PLCCs, the SILs, and the MFP, the system was still crashing after it warms up, so I started cold this morning and poked around at various components. Discovered that I can make the system crash by pushing on GLUE. I manipulated the springs in the socket a bit to get a firmer grip on the IC, and also cleaned up both the socket and IC with contact cleaner, and despite the chip feeling considerably more solid going into the socket, it's still crashing both spontaneously and when I press on GLUE.

I'm wondering if I should just replace the whole GLUE socket. And hoping that I don't have a crack in the board instead. :|

[exxos]

Are you using the 3M sockets ?

[derkom]

Are you using the 3M sockets ?

I am. Specifically went for 3M based on your recommendation.

This maybe just got a bit more interesting, because I decided to have a look at my solders under the microscope, and while they all look fine, I noticed that there was a whole lot more gooey flux residue there than I realised. (In one of my earlier reflow sessions, I used a paste flux that's a bit messy, although I used a liquid flux more recently.) I drenched it all in IPA and cleaned it up, and at least so far, it's not crashing, either on its own or when I poke at the GLUE. I wonder if the residue was getting just conductive enough to cause problems.

Time to leave GEMBench running for a while. I'll report back later.

[PaulJ]

@derkom I don't think flux is an issue unless you have some small wire clippings stuck in it. At worst it looks bad. Also are you using leaded or lead free solder?

[derkom]

@derkom I don't think flux is an issue unless you have some small wire clippings stuck in it. At worst it looks bad. Also are you using leaded or lead free solder?

That is my understanding of flux as well. Perhaps some long term damage from particular varieties, but not likely to cause the problems I'm seeing.

I am using lead-free solder.

So, GEMBench ran for about an hour, which is quite a bit longer than it's run in the past, and I was thinking the problem was solved, but then it crashed with four bombs. Hitting the reset button gave the following screen (weirdness in the middle is part of the GEMBench dialog box test animation, frozen):

IMG_20200107_165340.jpg

And power cycling at that point gives a white screen with black border. It'll now have to sit powered off for long enough to cool down before I can repeat the cycle. Very confusing.

[PaulJ]

@derkom , pretty obnoxious problem. You can try the heat tool on the glue when its cool and running and some freeze spray when its cold to regain normal operation. Unfortunately a bad solder joint or cracked trace can also exhibit this behavior with heading and cooling. Good luck.

[DoG]

Do ram test cycles with yaart maybe? It wasn't plummer flux you used because that can be corrosive on pcb?

[derkom]

Do ram test cycles with yaart maybe? It wasn't plummer flux you used because that can be corrosive on pcb?

Same issue with YAART. It'll run for a while, then crash. Starting cold, it'll get through a number of successful cycles before crashing, so the RAM is fine. (And besides, I've used both Falcon RAM and a SIMM.)

No, it was electronics flux, just a messy one. :) The board looks perfectly fine when it's cleaned up, just takes some effort.

[exxos]

Is the flux lead free compatable ?