Hey everyone... I'm still working on that IIe that I bought late last year and have been slowly restoring. I wrote a message on the forums back in December or so and got pointed to Joe's for a new keyboard decoder chip and that fixed that problem right away. In the months that have followed, I have gotten some more gear and setup what, back in the day, would have been an awesome system, but a new set of problems have cropped up int he meantime. I suppose that I am stuck now, and could use a little advice.
I now have the IIe enhancement kit and the new decoder chip installed. On power-up I end up getting a screen of horizontal lines (solid while every other column in 40 columns), no beep, and a halt. Maybe it is acting like what one would expect when holding control-reset when powering on? I can resolve this situation by pressing OA-control-reset and then the machine boots and acts completely normal... a simple control-reset dumps into BASIC as on would expect.
Running regular 64k software, after about 30-40 minutes, I get crazy errors looking like RAM errors...screen glitches followed by bizzaro errors where I can't even get control with a cold restart or warm restart until I let the machine 'rest'. I suspect some component is warming up on the board causing the whole thing to go bad. I have checked the power rails and get good voltages, even when these failure conditions are present. With all the cards installed, it is probably a pretty good draw on the power supply, but I don't notice any overheating or anything janky with the power... maybe I am overlooking something, but I am pretty confident we are good there. Indeed, I have these problems whether I have a full array of cards installed (Aux, SSC, Disk II, Liron, RM Microdrive/IDE controller) or not.
It would seem that I can hasten these errors by attempting to run software using the original extended 64k or the Ramworks IIII card in the aux slot, i.e., AppleWorks or the like... I haven't been able to successfuly get into Appleworks. That being said, I also have not been able to completely troubleshoot with a clean install of ProDOS versions onto a 3.5 which I know has a version-appropriate match between AW and ProDos, in other words, I am not convinced that ProDOS 1.9 running on the IDE card isn't causing problems with old Appleworks 3.x, 4.0 etc.
On the initial startup, the OA-CA-Control-Reset diagnostics run fine. When the machine is "warm" there are always RAM errors, but I can't find a consistent pattern for them (as to what chips are failed). I went ahead and socketed the RAM and tried fresh chips and got the same behavior. This was a harrowing experience, as my soldering skills have deteriorated in the last 30 years, but I seemed to have done it correctly, as there are periods, no better or worse than before, when everything works. Thus, I don't think it is an actual problem with the RAM chips... I mean, it still could be, but with all of that work and no problems for a while after booting, it seems that I am playing a zero-sum game on that front.
I took a stab and tried a spare MMU in the system... that doesn't seem to make a difference. I don't have access to an extra IOU, unfortunately. I have also reverted back to the original (unenhanced CPU and ROMS and video ROM and even the old, busted keyboard decoder) and the behavior is the same.
Just tonight, for the first time (with the original 64k expansion card loaded, not the RW4), I did not crash during the mobo and RAM tests in the IIe Diagnostics 2.1... I did get an error in the that said "Main Logic Board RAM Test Failed" with an error number(?) R90400FD, and a request to refer to the Apple Service Technical Procedures. That book, of course, was useless, saying to "replace the system board". This was on a 'cold' machine, but it was accessing the 80 column card, so.. there is that little tidbit fitting into the puzzle again... not being able to address that aux memory without causing problems.
I do have access to an oscilloscope, but I have not done a lot of deep, systematic/diagnostic probing yet to figure out what is going on... I see clock pusles on the processor and ram lines that I'd expect during normal operations. I'll have to go back and check the power-on situation... I seem to remember that the clock was suspended there until the reset went through, but I didn't write down the specifics or the pinouts (sorry-- I'll try to followup on that!) I suppose for now I am looking for pointers and ideas on how to proceed in a more logical, diagnostic manner. I sense that there may be two separate problems... the booting issue (vertical lines) and then the warmed-up RAM failures. Any takers?
Thanks! -btd
In a situation like this one I would simply use my hot air gun on different parts of the motherboard in an effort to localize the problem to a particular chip. This universal strategy works on all types of electronics, old and new: https://www.youtube.com/watch?v=vQNk2g1opLM
In post #1, BTDonovan wrote:
"I suspect some component is warming up on the board causing the whole thing to go bad."
Uncle Bernie advises:
In this situtation (problem persists after RESET while machine still warm) and if it goes away when machine cooled down, your best option is the use of cold spray. Let machine run until the fault manifests, then use the thin tube that comes with the spray can to squirt a small amount on one suspect IC (I'd start with the bus drivers, then the DRAMs, then the ROMs, then working outwards from them according to the schematic).
After one IC has been squirted cold, wait a few seconds, then do the power on reset sequence. If the problem went away, you have found the culprit. Otherwise, progress to the next IC. Keep the machine powered on to keep it warm.
Once a culprit is found, repeat the test until you are certain it's the one.
This process is tedious and takes a lot of time, but believe me, even if you know exactly how the circuit works, looking for thermal faults with an oscilloscope will take longer and most likely will get you nowhere.
Use of a heat gun as advised by CVT in post #2 can also help, but it is less precise, and unless it's a thermostatically controlled heat gun where you can set the air temperature, you risk to damage the machine you work on.
The cold spray, when the thin tube is used, is much more precise if you squirt with the proper technique. Precise location limited to one IC is important. I had cases where sloppy squirting hinted at a culprit but the real offender was in the row above it.
- Uncle Bernie
P.S.: don't waste expensive cold spray before you have monitored the power supplies until the fault manifests. It could be the warm up of the power supply causing out-of-spec or excessively ripply supply voltage rails. To see if the output filters are OK, the oscilloscope is best. A simple multimeter can't catch this.
Also a much cheaper compressed gas duster (like Ultra Duster™) can be used as freeze spray by simply holding it upside down.
For the general approach, it's always best to start simple and work out. By simple I mean remove all cards just power, motherboard, keyboard, and video out.Power is always your prime suspect. Have you verified the power supply is working well? A working supply does not mean a good supply. You have a scope, so use it, monitor the voltages as it runs, my prime suspect will be 5V since it's used everywhere. 12V is primarily used for driving drive motors, -5 and -12 are worth monitoring but the lower voltages are drive by big brother voltage -5 by -12, +5 by +12 so if one is bad both can be bad unless caps in the C19+ values are bad. With a 4ch scope watch all 4 voltages at the board input (the caps next to the connector are good, hope you know the + side of a cap with a negative voltage is ground.
If you picked up some cheap RAM from China, I wouldn't trust it. I may be good, but it may be bad I've seen too many rejects getting out when not coming from the good sources. Bad chips come from good sources too, but those are less common.
With a good supply, start with the 48K RAM check, this check everything just past zero page through 48K, note <space><enter> means the keys not typed out, and I'm writing this from memory so it may be slightly off... someone will correct if I've got something wrong.Boot and ctrl-reset to get to the basic prompt and call -151 to enter monitor:
] CALL -151* C050 C053 C054 C057 265:FF N 266<265:bFFFEM 266<265:BFFEV 265:0 N 266<265:bFFFEM 266<265:BFFEV 34:14<SPACE><ENTER>This test will fill then verify the lower 48K, the screen when filled to 0 will be black and FF will be white, it loops forver and if errors are encountered it will print the address at the bottom of the screen. Since heat may be an issue could add hot air to the system to help, but with just the mobo connected heat will be low. If it runs for a while wihtout errors, they start adding cards back in one by one starting with the one that
If heat is a thing for you you can try this over and over, but you do need to type it out every time unless you have a cassette recorder! When building I'd start with the bigger power draws (ie disk drive) and keep monitoring power levels. If you have the disk working the RAM check 4.1 would be useful too because that can do continuious sweep on regular and extended RAM as the system warms up. Keep building out until you find a failure, then remove the last device and try adding a different card. Was it the card or power issue? Swapping cards with purpose could help pinpoint an suspect.
All that said, I wouldn't be surprised if you have a part that gets flaky when hot, but it's also possible that you have a bad solder joint or trace that has bad contact and over time there's enough resistance to cause a part to heat up. A laser thermometer could be helpful for scanning the board for hotspots. (although this would be last in my list of things to check)BTW I hate that this is not a WYSWYG editor! Formatting blow, cuz I can't figure it out!
Hey everyone... thanks for your awesome comments and sorry for the delay at getting back to you all.
I finally got some canned air and started working through the IIe, and wouldn't you know it... after going into the 'unstable state' (bizzaro crashing after a while doing some sin/cos loops) , when I chill the IOU and reset, everything all comes back perfectly until warming up again.
So... now I am looking for an IOU, it would seem. Anyone got one handy? I could offer an extra, working MMU in exchange?
I still haven't gotten the boot thing going yet, but let's see if getting a 'fresh' IOU in there gets things working otherwise.