This is an image I took with a digital oscilloscope, showing the D0 (data bit zero) signal plus some address bits on my Apple II+:
http://www.applefritter.com/node/20136
The logic high level looks really terrible. It is usually around 3.5 V, with occasional spikes at 5 V and occasional drops to 2.5 V. I checked on my IIe, which looks much better. Can anyone tell me if this looks OK, or should I assume something is very wrong?
.. that dies when the disk spins up? If it was the RAM, maybe you could throw the machine into an HGR mode and then spin the disk. Chances are good that if the ram (in the bank the hires is showing) goes bad, you'll see a bunch of hash suddenly appear as the disk starts. Then it's a question of "bad ram" vs "bad +12V ps". I wouldn't worry too much about the D0 trace, it's messy but it's all in spec. You might try looking at the +12V line when the disk starts ..
Yes, it is the faulty II+. It seems to work fine as long as I don't boot a disk. If I boot a disk, it always crashes somehow. Both disk and disk drive work smoothly in my Apple IIe.
So if I start with disk, it crashes with sth like
0005- A=52 X=DF Y=13 P=F4 S=FB
*
Thansk for the HGR suggestion. I just tried some combinations, with some interesting result.
1) Reset, then PR#6 - disk spins for a few seconds, then the ] prompt returns and disk continues spinning
2) Reset, HGR, then PR#6 - disk spins for a few seconds, then the ] prompt returns, and disk continues spinning
3) Reset, GR, then PR#6 - disk spins for a few seconds, then screen fills with inverted "É" characters and the ] prompt returns.
What could this mean?
I hooked up a logic analyzer to the address lines of the Apple II+ and got some strange results. From the error message above, it seems like the processor somehow crashes at address 0x0005. The processor references this address only once before it crashes, but only as a part of a read from memory into accumulator! The specific address that does this access is in Autostart at address F88C:
LDA (PCL, X)
And after that execution continues as normal at the next instruction.
The crash occurs a few 100 instructions later, so I couldn't do the full tracing.
EDIT: I added an image of the tracing here: http://www.applefritter.com/node/20154
Trigger point in the middle is at the reference to address 0x0005.
Just curious, but have you tried swapping out the RAM yet? Looks like the II is crashing when trying to do an indirect load from 0x05, so if the contents at 0x05 were incorrect it would lead to the wrong location giving bad data, right?
Have you tried any of the standard restore things such as reseating the chips, etc?
Vince
Thanks for the tip. I have already reseated and moved around pretty much all the IC's, including the RAM. I didn't see any consistent difference when doing so. Unfortunately I have no extra RAM chips to swap in.
As for the 0x0005 access, I just assumed that if the Apple crashes with such an error message, it means that it executed an instruction at that location. Surely the Apple can crash also from a LDA from that location, but how would it know which LDA instruction crashed it? Thing is I don't know what the error message means
just out of curiosity, do u have another apple II with socketed memory chips?
Hooray! After your post I decided to go through the RAM chips once more. This time I swapped chips between rows (I tried mostly within rows previously). I noticed a change in the crash error message for one swapped pair, and then tried swapping each chip with other chips. So finally I could point out one specific chip that seems to be faulty. All other chips seem OK. I really hope this is the last problem with the Apple.
Now I just have to try and get a replacement chip
If your memory is faulty, the particulars of what the ML monitor dumps won't necessarily mean much. Suppose your 6502 is chugging along happily fetching instructions from someplace in a RAM bank that has an intermittant failure. The next opcode that comes in could be anything, and the resultant effect on where the PC (program counter!) goes could be pretty random. By the time things grind to a halt in the monitor, the intermittant failure might be gone, and where the PC ends up is likely to be pretty random.
Anyway, I'd still recommend temporarily transplanting a different power supply into that machine. The 2's power supply is pretty easy to get in/out. Don't even take the old (current) one out; just unplug it from the MB while you plug a different one in.
The reason I'm betting on the PS is that there are lots of electrolytics in those switching supplies, which tend to dry out, especially after 20 or 30 years, and esp. in hot environments (like the inside of that unvented PS box!) - made all the more problematic by the fact that, in those PS's, several of the caps are right next to multiwatt power resistors.
Woops. Well, there was always that. Let me know if you get really stuck trying to find 4116s
Yes, that is a good idea, but I already tried it
Well there are at least 2 places to get them on the internet. Wow, I can't believe they are selling these for USD 0.50 a piece. Not that it really matters; the cost will be all shipping anyway
So, it was RAM? Glad to hear, I don't know why I didn't speak up sooner, sorry, I saw the post but was busy.
Vince