My Apple /// isn't working properly as of today :-( It has been sitting idle for a couple of months now (since I last used it).
I've tried booting a couple of diskettes but the only one that worked (so far) was the Confidence Program (other diskettes loads some tracks and the freeze/hangs).
If I select/execute Machine Status test (1st option on the menu) it cycles through the disk drives and then loads what I assume is the sound sample ("I'm OK - Machine status normal" or whatever it says) from the diskette and after that I only hear a tiny click from the speaker (hardly noticable) instead of the sound sample and the program/computer freezes/hangs before the interrupt system test (no further tests is checked/done). If I reboot and select the Memory Test instead it immediately freeze/hangs when the RAM configuration map is displayed (no "animated" or moving Apple cursor). The Video Tests starts and displays the 1st image of the Apple /// bitmap but when pressing RETURN it doesn't advance further with the next test.
So far I've done the "easy" part and reseated the socketed chips both on the mainboard and the 5V memory board.
It's a 256K version of the Apple ///.
I haven't measured any voltages but I assume that if I come this far the voltages are there?
BTW, when I looked at the mainboard I noticed a resistor that seem a little burnt on the right side of the rightmost cardslot (the rightmost of the 2 big green resistors next to the 2 transistors).
Any suggestions on how diagnose/troubleshoot this further is very much appreciated - thanks!
First, I would absolutely replace that resistor. if you cut it so that you leave as much of the leads as possible, you can bodge one on top of the existing leads with miniml fuss.
There are two memory cards, 5V and 12V. Remove each, reseat everything on them, and reinsert them. All software crashing, though, to me says bad RAM, power issue, connection issue, bad drive.
You didn't specify if the SOS splash appears on the disks that crash.
When you run the RAM test, if select the wrong memory size, the test will fail or crash. Be sure to set the test for the exact memory map configuration of your ///.
I would likewise check the drives using the DIsk // checker in Apple II mode, to rule out issues with them. Does it boot ][ software?
Further, I advise buying a AC mains circuit checker. Test the outlets to see if you have a wiring error on the outlet that you are using. I recently learnt ow fussy some 12V stuff can be with something as simple as an open ground. If the outlet passes, then I would look at the 12DC circuits first, particularly the 12V RAM card.
Look for blown caps on every PCB, including all daughter cardd and slot cards. Further, install a fan off of 12VCD.
Beyond this, you are looking most likely at bad RAM, a failed RAM card (check the caps on those), a failed mainboard cap,a failed PSU, or similar.
I had typed a longer and more verbose reply, but my browser ate it. I have flowcharts for the ///. but they will boil down to 'replace component X', and not cover low-level repairs.
Thank you for your reply!
Well, out of the disks I tried Confidence is the only one that actually boots. The other ones like Apple II Emulation or other SOS diskettes does not load all the way and start (they load some tracks from the diskette and then the computer freezes/hangs). There are no read retries, everything sounds and acts as normal with the floppy drive.
I'll measure the resistor (when I know what the value is supposed to be) but in the long-term, for sure I'll replace it. I will probably measure the ESR of the capacitors as well although I need my ESR-meter back from my friend first.
I don't think I can select memory size in Confidence, it has automatically detected the size to be 256KB (which is correct in this case).
BTW, the +12V I assume is for the motor in the floppy drive and possibly the sound amplifier section but are there other parts that need +12V (and there is -12V as well, right?). +5V I understand is to drive the TTL chips etc. and that I apparently have (although it might be low). Is there a potentiometer in the power supply to beef up the +5V a bit if needed?
I'll keep you posted and if anyone has some more specific things to check - please let me know!
As far as I recall, unlike Mac sweep boards, the voltages are fixed and regulated inside the PSU shell. You have +5, -5, +12, -12 typically, plus earth. Try checking your RAM boards. I'm pretty sure that resistor is identical to the one next to it, but you can check the schematics.
If you have the /// manual, I believe there is a PCB schematic in it. if not, there are a few on-line. Do the resistor, then check its path, see what VCC lines go to it. Look for bloated caps or leaking caps; and very carefully inspect your RAM boards. Do you have a second Disk ///?
To me again, I suspect RAM first, voltage second, Disk /// last. Don't worry about the Confidence Disk. Use the internal test built into the monitor: Hold Control, Open Apple, and then tap Reset.
Enter: F6E6G, then press RETURN. Let it run for about half an hour. You want to see only a grid of dots. If anything else appears, it marks a fault in the RAM boards.
If you any other error messages, please post them. Aside from RAM test, this detects faults on the mainboard itself. Here is a matrix that represents the RAM board locations based on the grid. According to my reference, it is this, but I think that two of those rows should be S2 through A17:
B9 B8 B7 B6 B5 B4 B3 B2
B17 B16 B15 B14 B13 B12 B11 B10
B9 B8 B7 B6 B5 B4 B3 B2
B17 B16 B15 B14 B13 B12 B11 B10
C17 C16 C15 C14 C13 C12 C11 C10
D9 D8 D7 D6 D5 D4 D3 D2
D17 D16 D15 D14 D13 D12 D11 D10
C9 C8 C7 C6 C5 C4 C3 C2
Again, anything other than a dot is a bad RAM chip. If this doesn't work, I will check the Appe Service Source docs for a 256K system test. I know this one works on the 128K 12V system.
Check if you have a 12B or a 5V board. Tell us the p/n.
There are some 5V boards on ebay now:
Also an internal Disk ///
and a PSU
if you plan on maintaining this system, I'd suggest stocking up on spares.
There is also a mainboard, which IMO is overpriced:
To me, this looks as if someone disassembled a /// system and is selling each component. I truly detest that.
We used to have a /// in the lab, and when it was retired, I took it home. It was a slick system with the //e emulation card, a profile crd, and some de bosrds, plus an RGB monitor. Sadly, someone stole it from me when I was in the ospital for my heart. A decade ago, they weren't worth that much, so I should have replaced it then, but I really only ever used it in //e mode, so aside from looking odd and having heat issues and limited slots, my //e and //gs systems can do everything that I need.
I've measured the power supply (without load) and these are the values:
+12V = 11.98V
+5V = 5.09V
-5V = 4.023V
-12V = 10.32V
I've also measured the +5V on a TTL chip and that was 4.87V I think.
I ran the inbuilt memory tester and it indicate faults but on the other hand I have a 256K 5V memory board so can that be trusted and if so, how would the grid pattern be interpreted? If not, is there a memory test in ROM for 256K or is this just something that was in the initial 128K design and never updated afterwards on shipped 256K systems?
I don't have any Apple /// manuals. I've found some documentation but the troubleshooting guide is like 3-4 pages only so if the Apple Service Source docs covers tests for 256K systems that would be useful. BTW, is this information available online?
This is what I get when running the inbuilt RAM test:
I've lifted every socketed chip on the mainboard and the 5V 256K board and re-insterted as well (to no avail).
What happens if you swap those chips out to other locations? Be sure to discharge yourself (ESD) before fiddling with the RAM chips!
Those 1s or Is indicate faulty RAM, but that test was designed for the 12V 128K card. if it gives errors in the new locations, then that is a good indication that the RAM is bad, and what you can try to do, is to pull the extra 128K off; then fit chips into a 128K configuration until the test shows all of them good. At that point, test to see if normal software boots again. Be sure to store RAM chips in an antistatic bag, or rail; and if you hae any antistatic foam for ICs, press them into it to protect their legs.
The main thing here, is to 'strip' te system down to 128K, and to get known good RAM in every spot, so that you can check if that is the only culprit. If that makes the system operable, then swap out the known good bank with the unknown bank, and test them. This should identify the exact bad RAM chips and you can order replacements for them. IDR how sensitive the /// is for memory speed. I think that 15ns RAM is typical on them.
We can discuss that once you do the above procedure. I wish that I saved some /// RAM cards. I have some for the Lisa, but nothing for the /// other than my original system manual and some SOS catalogue disks from WAP.
I'd advise buying the spare PSU and RAM card while they exist. If I still had a /// system, I would have bought all of those components and tucked them away.
IDK if your measurements are bad, but you should not be seeing voltage loss to that degree. Rebuilding the PSU is probably in order.
Will do some work outside in the garden now and get some fresh air (staying isolated because of the C19 virus is not a good thing!) and further tests will be done later today.
I'll do as you suggest, strip the memory down to 128K and see what happens after that.
I also found the Apple /// Service Reference Manual (730(!) pages) which my printer is having "fun" with right now! :-)
I will rebuild the power supply soon as well (it's an ASTEC AA11190), hope that I can find a list of the capacitors needed alt. if someone sells a "cap kit" for it? Otherwise I'll just have to have a look at the PSU and document it.
ReActiveMicro sells a universal PSU kit:
Yeah - success! :-)
I swapped the memory chips from the 2nd row to the 1st row and tested 128K successfully and machine appears to function normally. I then populated the 2nd row with the swapped chips from the 1st row and then it failed memory test in Confidence Program. Still machine appears to be working normally (I guess it's common that most programs fit in the first 128K) so I can probably wrap this up for now and look for some memory chip(s) and also find a new resistor and replace the burnt one.
I can also confirm that the inbuilt RAM test (via the monitor @ $F6E6) only support/sees the 1st 128K as that test passes now with the faulty RAM still installed in the 2nd row - can be good to know that it is confirmed now if someone else have a similar issue. Anybody know if Apple released newer ROMs that supports a 256K memory test?
Of course I will run different test programs available but most likely everything is fine now apart from 1-2 bad RAM chip(s).
Many thanks to Timelord for your suggestions and ideas - great support! ;-)
Thats a good find with the memory test. There is no later/updated version of the monitor rom released by Apple, so this is quite a trap for 256k versions, which i expect most out there to be.
When SOS loads in the Interpreter (application), it always loads this into the highest available bank. So for a 256k machine, this is into the untested bank which gets you into your situation.
There is an internal memo I was reading again that was talked about on the drop -iii-inches podcast. It highlights the problems with the memory test, it seems it was never fixed. It is in this collection of memos:
A good source of information about Apple /// is located at apple3.org.
The burnt component is *not* a resistor, but an inductor (see L3 marking on the motherboard). According to the schematics, it's a 10 µH.
Right, L3 is an inductor and it's only purpose is to filter the 5V output voltage on the 15pin RGB video connector (J15) of the Apple III. I guess the most likely way to burn this inductor is by plugging something else into the SUBD-15 connector, which isn't an Apple III RGB monitor - something like an IBM PC joystick plug (which would fit mechanically).
I don't think many monitors would require the 5V output at the RGB connector. It was probably meant to allow the extension of some graphics converters, which might require power. I would leave the burnt inductor alone - unless you're actually having issues with a monitor or converter no longer working.
Very good to hear. Your next step is to put the known back 128K in the upper bank, and to isolate the bad chips. That, or just buy anoter complete set of 128K chips. If you ar very careful about static discharge conditions, and feel competent to do it, I would swap both banks, and to check the locations of every bad RAM chip; then order replacements for them.
You can leave it at 128K until new chips arrive. Are the chips 4464s?
I hear you, will leave it alone as you suggest as this isn't anything that I will use ever (most likely) and if I would then I can replace the inductor.
The RAMs are HM4864-3 (Hitachi & Apple logo on it).
I already put it together again with the faulty chips installed. Will have to dig up some RAM chips first and replace. Would be great if my EPROM programmer could test these RAM chips but I do not have it in front of me right now so I don't know, but that would be the easiest way to isolated the faulty ones quickly.
You know that there are 64kb DRAMs with 7bit and 8bit refresh cycles? I don't know the internals of the Apple III, but dealing with DRAMs this is the first important detail to know.
Means: if the Apple III hardware uses 8bit refresh you can use even 7bit refresh DRAMs.
The fast and easy way to isolate the faulty RAM is to install only that 128K bank into the board, and have the in-built monitor RAM test tell you which chips are bad. Each of those positions corresponds to a specific chip!
Or, just, you know, but two lots of the RAM from eBay...
Note: This does not constitute any form of advocation on my part, for eBay, for Paypal, or for the specific seller or auction. I provide the following purely for the purpose of notification and information.
These Hitachi HM4864 need a 7bit refresh (128 cycles). They are also good for a DRAM controller with 8bit refresh.
Time for an update!
Waited for ages for RAM chips which finally arrived a few days ago and problem/fault was isolated to a single chip which I replaced with one of the ones I bought and now Confidence Program passes without issues (several passes on the memory test), so it seems that the Apple /// lives again!
Thank you (everyone) for suggestions etc. in troubleshooting this problem! :-)