PowerBook 100 Sad Mac 0000000E 0000FF00 Fix

[Androda]

Recently I picked up a PowerBook 100 for cheap because it lacked the CPU card. Fortune favored me in that the CPU card was available from a seller in Tunisia and not labeled for the 100 - they had it listed by part number.

Starting up the old machine resulted in the sad mac code:
0000000E
0000FF00

Sad mac code reference for the mac portable (since the 100 is basically just a mini portable): http://myoldmac.net/FAQ/SADerror-portable.htm

This means that the sad mac code is for a data bus failure, for data bus bits 8 through 15. I traced out the CPU card for awhile trying to find broken data lines, but all of them had a good connection to the plug (CPU card is socketed).

More tracing around revealed that the VCC pin on one of the PSRAM chips had a marginal connection. Light touch with the multimeter probe, no connectivity. Push even slightly harder, and it suddenly is ok.

Thus, I hot-aired all the PSRAM chips off the board and cleaned it all up (which was good, there was brown stuff underneath of unknown origin). After re-soldering the chips in place, my 100 no longer does the chimes of death and shows the floppy disk with question mark. Haven't tried booting yet, though I expect it's likely to work.

If your PowerBook 100 has even the slightest capacitor leakage from that one electrolytic on the CPU card, I suggest removing and cleaning all the PSRAM connections just to be sure.

edit: Yes, I also replaced all the other electrolytic capacitors.

 

[retr01]

Excellent fix, @Androda! How did the bench testing go?

 

[Androda]

Excellent fix, @Androda! How did the bench testing go?

The display is rather unhappy with the ceramic caps I used, very wobbly. Guess I'll switch to tantalum, as much as I dislike those. Or maybe I can find teeny tiny solid polymer electrolytics.

 

[retr01]

Aw! Back to the drawing board. I wonder if there is a chart that compares the capacitor types?

 

[Androda]

Ceramic capacitors can't always be used to replace electrolytics. Their big advantages are neither exploding (tantalum) or leaking (electrolytic) ever. But not all circuits like them.

 

[retr01]

Does it have to do with the conductivity? Do different types conduct slightly differently? That is a very small difference?

 

[Androda]

There are two major factors.

Ceramic capacitors have a dramatically lower "ESR" than most electrolytics or tantalums. This means that they react dramatically faster to voltage fluctuations, and not all circuits are ok with that. Maybe an electrolytic was used to provide 'bulk capacitance' but the ground plane it's connected to simply can't handle high frequency noise, for example.

Ceramic capacitors are also "inaccurate" in that the capacitance value changes depending on the voltage applied to it. So if you use a 10uF 16 volt rated capacitor and then apply 7 volts to it, the capacitance value will be less than 10uF - probably closer to 6uF. Does this always matter? No - because ceramics react so much faster, a lower capacitance rating can be ok. Some electrolytic to ceramic conversion guides even say you can often get away with half the total capacitance value when switching to ceramic.

So very low ESR combined with actual capacitance fluctuations can be a problem. If you need a guaranteed '10uF' in your circuit, better to use a solid polymer electrolytic for example. Their capacitance does not fluctuate based on applied voltage, and solid polymer caps don't leak. I avoid tantalum like the plague after a few experiences where replacing an electrolytic with a new tantalum ended up making the brand new tantalum capacitor explode. Not going to deal with that on my vintage stuff.

 

[retr01]

Thank you, @Androda, for the clear explanation. Some people still believe in avoiding electrolytic caps like the plague and go for tantalum because they do not want leaks. I now understand why that would not be good when I read your statement.

So, modern "high-quality" electrolytic capacitors are better than 30 years ago that they will not leak? Or they WILL leak the same some ten, twenty, or thirty years later?

 

[3lectr1c]

From what we're seeing in newer systems from the late 90s (which are now as old as SE/30s from the 80s were when they started to leak), they do still end up failing sometimes, but less often, and they seem to dry out and not leak. Certain types still do leak though, the original Xbox's CMOS capacitor being a great example.