Apple IIe Card for Mac crashing during loop

[Elemenoh]

I have a couple of Apple IIe cards for Mac. One works fine. The other will crash after a few seconds running a simple basic loop like:

10 PRINT "Hello, Tinker Different"
20 GO TO 10
RUN

It's been recapped, reflowed and ultrasonically cleaned. Have you experienced this issue before? Any advice on how to narrow down the trouble?

 

[retr01]

That Apple IIe Card seems to overflow at the infinite loop and panics. Interesting.

Saw this on MG's Apple II Site:

LC //e Card - Problematic Software

Nearly all software that crashes on the Apple //e Card for Macintosh LC is likely because of access to $C029. This location is the NEWVIDEO register on the Apple IIgs, which is used for, among other things, turning on and off super hi-res graphics modes.​

​

TN.MISC.15 cautions against using this location without first identifying the machine. Ironically, the document is mistaken about the outcome of the IIc Plus and doesn't mention the //e Card.​

​

On the //e Card, accessing $C029 switches out the main firmware in favor of the active alternate firmware bank. Subsequent attempted use of common firmware routines will not go well.​

​

Other software may produce unexpected output on the screen depending on how it uses the disk or if it uses DGR graphics.​

As for the infinite loop, I wonder if it triggered something and caused a panic. Yet, the other Apple IIe Card works fine with the endless loop.

@Elemenoh, I have questions:

1. How long did the loop last before crashing?
2. Was it an immediate crash or a few seconds?
3. Was there any error message or indication of the Apple IIe Card crashing?

 

[retr01]

Paging @mg.man - is that something you are familar with?

 

[Elemenoh]

@retr01 the loop lasts 5-7 seconds before a crash. A longer string results in a faster crash.

It displays an error dialog “The application ‘unknown’ has unexpectedly quit, because an error of type 1 occurred.”

The Mac will keep working but if I relaunch the Apple IIe app the Mac will hard freeze. Rebooting puts it back into a usable state.

 

[Elemenoh]

Also the RAM has been swapped already with no change to the symptom.

 

[Elemenoh]

Another data point, I ran the ROM self-test by holding ⌘+⌥ during IIe startup. It gets through most of it, including 80 col for a while but eventually crashes just like the loop. Any insight based on where it crashes here?

Also, I'm using it in a Color Classic here, but it has the same behavior in LCs and another Color Classic.

 

[Elemenoh]

I swap tested parts from the known-good card. Swapping in the GEMINI from the good card to the 'bad' card resolved the problem. Start-up diags pass and loops work. I'm running some extended APPLE-CILLIN diags on it now.

There was some crap under one corner of the GEMINI. Just to be thorough I'm going to put the parts I pulled from the bad card onto the good card and see what happens.

 

[retr01]

Yep! As I suspected, some traces are causing those crashes under that Gemini ASIC. Now you can fix it.

 

[Elemenoh]

Yep! As I suspected, some traces are causing those crashes under that Gemini ASIC. Now you can fix it.

The traces are okay. There was just some crud in that area that needed cleaning. Will try the GEMINI from the bad card on the good card to see if that gunk was the problem or the chip itself is bad.

 

[Elemenoh]

I swapped in the Gemini from the bad card to the good card. The good card with the bad Gemini failed in the exact same way. So somehow the chip died.

 

[retr01]

I swapped in the Gemini from the bad card to the good card. The good card with the bad Gemini failed in the exact same way. So somehow the chip died.

Now the question is, how did that happen?

 

[retr01]

Mega II is the ASIC in the Apple IIGS to do the functions of an Apple II. The Mega II was designed by Apple circa 1986 with "Aragon Project 2562-0001" stamped on the die. However, it has been reported that the Apple IIGS made little use of it.

The word is that Apple made updates to the Mega II and renamed it Gemini, which was designed by Apple and manufactured by LSI Logic for the Apple IIe Card.

So, Gemini is essentially from the Mega II. I do not know what updates.

Without relevant data, I am unsure why a Gemini ASIC chip would fail unless it were one of the tiny percentages of defects.

 

[robin-fo]

G

Without relevant data, I am unsure why a Gemini ASIC chip would fail unless it were one of the tiny percentage of defects.

Well, ESD could harm a chip during improper handling and the effects may appear a long time after out of the blue. Was this card ever exposed to a harmful electric field, maybe wrapped in plastic foil or so?

 

[retr01]

Well, ESD could harm a chip during improper handling and the effects may appear a long time after out of the blue. Was this card ever exposed to a harmful electric field, maybe wrapped in plastic foil or so?

That is a possibility. This is a fascinating article, "What is ESD and How Does it Affect My PCB Design?"

Suppose that the die of the Gemini ASIC on one of @Elemenoh's Apple IIe Cards got hit by ESD. A trace or two on the die could be damaged. This is an example of trace damage on the die after ESD.

In more severe cases, the IC would show something melted into the plastic on top of the IC.

 

[retr01]

The United States National Aeronautics and Space Administration (NASA) has an excellent write-up by NASA Electronics Parts and Packaging Program (NEPP) about Microcircuit ESD Damage.

Here are two images showing ESD damage to a C2 MOS capacitor. What appears to be a slight pinhole on the top image shows to be substantial damage to the trace when examined at 4300X.​

​

 

[retr01]

This is an excellent opportunity to photograph this Gemini ASIC's die.

We should reverse engineer the Gemini to bring Apple II compatibility to more vintage Macs that can do 24-bit or even 32-bit. It would be neat to do that on all compact and modular vintage Macs like the SE/30, the SE, the IIci, the Quardas, etc.

A new thread on reverse engineering the Apple IIe Card to 32-bit compatibility, SE PDS, SE/30 PDS, NuBus, etc.

@Ron's Computer Videos started a discussion about making the Apple IIe Option Card software 32bit compatible.

Thoughts, @Kai Robinson? @Zane Kaminski?

AppleLogic has resources on many of Apple's ICs, including ASICs such as the Mega II, ADB Keyboard Controller, or the ADB GLU but not the Gemini yet. A Bit about Apple's ASICs is an article at AppleLogic giving an overview.

 

[tms9900]

Could also just be that the chip died of old age. Hope it's not too inappropriate to ask this, but @Elemenoh would you have any interest in selling the card with the bad chip on it for parts?

 

[retr01]

We need to take a picture of the die of the Gemini and compare it to the Mega II, as new cards are required to run Apple II programs on the SE/30, SE, IIci, etc., without emulation. Also can make it 32-bit compatible.

There are other custom chips on the Apple IIe Card that we need to look at to reverse engineer as well.

 

[Elemenoh]

I’ll give the dead chip to someone who can/will decap it and scan the die if that’s helpful. Otherwise will be keeping the bad card for parts.

 

[tms9900]

We need to take a picture of the die of the Gemini and compare it to the Mega II, as new cards are required to run Apple II programs on the SE/30, SE, IIci, etc., without emulation. Also can make it 32-bit compatible.

There are other custom chips on the Apple IIe Card that we need to look at to reverse engineer as well.

My 2¢: I feel like you'd have an easier time reverse engineering the card if you assumed the 65C02/IWM-connected side of the Gemini was largely the same as a normal IIe, and then connected a logic analyzer to the LC PDS side to see how the LC pokes and prods at the IIe hardware. Decapping the ASIC, be it by heating+softening or dissolving the package, and photographing it, would be relatively straightforward compared to actually making sense of the die photos. The people with the technical expertise to make heads or tails of the die are few and far between, and I think it's fair to say they wouldn't be content with the reward being "knowing they did their best to bring something amazing to the Macintosh community" (and just to be clear, I say this as someone who absolutely could not decipher those photos).

TL;DR Many ICs have been decapped and photographed in the broader vintage electronics scene, but the people reverse engineering the ICs seem to find more efficient ways of replicating their functionality than going off die photos (besides the more straightforward cases like photos of mask ROMs).