| Attachment | Size |
|---|---|
 AppleIIScreenGarbage.jpg | 25.39 KB |
| AppleIIScreenGarbage2.jpg | 243.07 KB |
Hello,
I have an Apple //e and after a rebbot - most of the time - I have garbage on screen.
Sometimes with a blinking cursor and sometimes only garbage.
I try it 10x and perhaps it could be ok once.
Best
Erik
Could it be a bad RAM chip?
If you hold down the closed-apple key when you turn it on, it might be able to get into diagnostic mode where it will perform a RAM test.
Also, it's a long shot, push down all of the socketed chips, because sometimes if the board is manhandled, some of them can pop out a little. Touch the power supply with one hand while you do this, with the power supply turned off and plugged into the wall.
I was thinking the same thing. Flaky RAM can cause that kind of error.
Hi,
thanks! But the RAM seems to be ok. I already tested it and the message was: SYSTEM OK.
All Chips in a sockets were removed and put again in. Same issue.
Best
Erik
When you say "reboot" are you turning off the power? If so, you need to leave the computer off for 30 seconds before powering on again (or get a ROMXe). If that is the issue, you can also hold down the open-apple key when powering on to force a cold boot.
The //e built in self test is not really that thorough as far as RAM test goes. I'd try Apple Dealer Diagnostics or XPS Diagnostics //e or something similar which has a thorough RAM test.
Is there a good cheap tester for the RAM chips? For my II Plus I used https://kosciuskomedia.com/product/4116-dram-tester-w-lcd/ (which does say you can modify a 4164 for testing, but I have not tried). It's sold out at the moment though it looks like.
I've seen 4116 testers on eBay recently, but nothing for other common DRAMs like 4164, 4416, 4464 or 41256 chips. Something that could handle all of those and maybe a few others would be really awesome to have for anyone who is a retrocomputing enthusiast or works on vintage hardware.
The other issue is that all but the earliest //e motherboards come with the DRAM soldered down, not socketed. So repairs involve first de-soldering the old chip, then soldering in a socket. I generally just snip off all the pins close to the package on the old chip then pull each one out one at a time with just a touch of heat on the back side with the iron. A little solder wick to mop up excess solder and it's an easy-peasy fix. If you have access to a nice rework station which has vacuum solder sucker and the attachments for heating all the pins at once it is even easier of course, but I've never had one of those at home. I did at one time work at a place that had one, and it was sure nice to use.
It is certainly not necessary to heat all the pins at once, and even vacuum rework tools are nice, but not necessary, since it is only a 2-layer board. Using a quality solder sucker (such as an Edsyn Soldapullt) and proper technique you can remove solder from all of the holes of DIP ICs.
Oh, I never said it was necessary to have a pro rework station... just awfully nice. Like really, really nice. What you say is true. Even a half ass cheapo solder sucker can do an adequate job if your technique is good, but as always, the better tools make a tough job possible and an average job easy.
Yes I turned off the power. If I wait let say 5 seconds an hold down the open-apple-key when powering up, it seems to be ok. Thanks a lot for this hint. But is it normal this behavior???
Best
Erik
This behavior falls into the category of "uncommon but not rare."
It occurs because the ROM tries to distinguish between a warm-start and a cold-start by comparing the contents of memory address
$03F3against the contents of memory address$03F4, and testing for a specific relationship that's stored in those bytes when the computer has been initialized: an exclusive-OR against the bit pattern10100101(A5).Occasionally, one particular computer's RAM chips exhibit a natural affinity toward matching that bit pattern...just by chance...which causes the ROM to misinterpret the pattern as an indication that the computer does not need to be initialized at cold start.
If your Apple //e often exhibits this issue, a simple solution is to shuffle the RAM chips so that their natural affinity trends away from these bit patterns. For instance, swap the RAM chip at row 6 with the chip at row 7 and that will effectively flip the affinity in the lowest two bits. (ie: if the chips tended to toward matching
10100101, they will now trend toward10100110) Swap arbitrary pairs of chips until the problem is resolved to your satisfaction.Apple IIe RAM chip positions.JPG
Incredibly helpful and useful post, thank you!
Just some clarification here. The issue is not that a RAM chip has an affinity toward a certain bit pattern. It's that many old RAM chips can retain some of their contents for an unbelievable amount of time after the power has been removed. Thus moving the chips around WILL NOT solve this issue. If the page 3 data retains its values while other parts of the RAM lose their content, then the next power on will look like a warmstart but will have corrupt memory. Even the Apple documentation warns that you should only power back on after at least 30 seconds.
While Woz was very clever in designing many parts of the Apple II - and loved to accomplish things in software instead of hardware - this is one area where I think he let us down. The original Apple II's did not even have a true power on reset circuit. This was corrected at first with the Disk II controller card and then also incorporated into later revisions of the motherboard. Even the keyboard controller and latch received power on reset circuits as the motherboards improved. Any of these circuits could also have been used to discriminate between a cold and warm start. But that's easy to say 40+ years later.
If the issue really bothers you - and you could benefit from its other features - you might look into the ROMXe product. Along with its ability to offer multiple ROM images and fonts as well as a real-time clock, it will also guarantee a cold boot at every power on.
Wow, that seems to be great but a little bit expensive to solve this problem. Hmmm...
Best
Erik
Fair enough. Yes this would not be cost effective just to solve this issue.
But consider that a $40 clock + $45 MultiROM board would still not get you all of the features of the ROMXe plus the extra fonts. And you don't even have to give up an expansion slot! Seen this way, the consistent cold booting just comes for free.
A new pricing structure is also coming soon that will substantially drop the price of the ROMXx modules if you don't need the extra fonts ;)
I'll have to get my RomX back out. I had to remove it when troubleshooting my II Plus. Now that I've fixed the language card which was the issue with RAM corruption on that system, perhaps I'll try the RomX again.
When the MSFT Language card had a bad 74LS* chip and bad RAM IC, I ran into very similar issues as described here with inconsistent booting, especially if you don't leave the system off long enough. The information here is very helpful for me if I need to troubleshoot this again. I also have a 4116 DRAM tester on hand.
https://www.ebay.com/itm/254648489833
https://www.ebay.ca/itm/124803897060?hash=item1d0ee456e4:g:NLQAAOSwl0Ng51ed
I was the first person to notice this problem. The cause is definitely the RAM.
Since then, I've come across a specific RAM chip that takes care of this problem. You'll be surprised how well these chips work.
Get the HM4864P-3. It's a RAM chip by Hitachi.
When I want RAM chips, I don't buy them separately. Instead, I get a 64k 80 column card that is socketed.
On eBay, such a card available. It's item number 142942741346
This looks like a classic RAM persistence issue.
Try booting (or powering up) holding the open-apple key down.
It should boot normally.
If it does then you have a RAM persistence issue.
Sometimes this clears itself up with normal regular use.
long periods of powered-off state are not firendly to these machines. They need to be used regularly.
Otherwise you'l need to replace the RAM.