Oops, my obsession with computers from my childhood strikes again!

[rubber_jonnie]

Apparently there is an excellent description in "Understanding the Apple IIe by Jim Sather, so I'm going to have (another) look there, but I also read some info on other forums (Don't berate me for visiting other forums!!!) that says some of the soft switching, like PR#3 to jump into 80 column mode is handled by the MMU, so I think that will complicate matters.

I've attached that PDF and just renamed to TXT for uploading purposes.

But we know that this part of the chip is still working correctly so we don’t have to replicate that. The aim is to do a minimal replication of the functionality which has failed.

Just to confirm, without the 80 column card the other peripherals also stop working after the MMU warms up?

If this is the case then it’s most likely the transceiver direction control that’s broken. In this case a simple 16V8 can be used. It will need address lines 13, 14 and 15 for address decode, R/W and /IOSELECT for CPU access and Phi0 plus DMA for card DMA access. You probably don’t even need Phi0 as that’s up to the peripheral timing.

Hey Stephen, just quickly, without the 80 column card, the machine works fine all the time and passes any diags you throw at it.

With an 80 column card fitted the base machine functions as normal, however if you switch to 80 column mode and run a simple test for about 10-15 minutes, the screen becomes corrupt and I believe it's the RAM on the card being corrupted. Hitting the MMU with freeze spray at this point makes the problem go away until it warms up again.

When the problem occurs if I reboot and load up the config program for the 80 column/RAM card, the card cannot be found unless I allow a cool down period or use freeze spray.

If I run a game that doesn't require more than the base system RAM it works fine, even with the RAM card fitted, however if I run a game that requires the RAM on the card, the game initially boots but eventually crashes once it's warmed up.

Hope that makes sense.

[rubber_jonnie]

@rubber_jonnie

Good question i will have to look may have been a 16v8 d part


Tom

Thanks.

[stephen_usher]

This does sound as if it’s the slot transceiver direction control becoming unreliable.

[stephen_usher]

Having watched the latest Adrian's Digital Basement it looks like the 80 column card isn't anything really special when it comes to the bus. Depending upon the type it may need extra signals from the motherboard though which aren't available from slot 3.

I'm now wondering if it's purely a timing thing and something on the 80 column card is going marginal so that it still works in the other machine but in this one the MMU has become too fast or too slow once it warms up.

[rubber_jonnie]

Having watched the latest Adrian's Digital Basement it looks like the 80 column card isn't anything really special when it comes to the bus. Depending upon the type it may need extra signals from the motherboard though which aren't available from slot 3.

I'm now wondering if it's purely a timing thing and something on the 80 column card is going marginal so that it still works in the other machine but in this one the MMU has become too fast or too slow once it warms up.

I just watched that video too, and it made me wonder what would happen if I installed an 80 column card in the No 3 expansion slot?

The difference with the IIe over the II is that the IIe just needs extra RAM to enable the built in 80 column mode plugged into the AUX slot, but the II needs extra hardware to do it. I get the problem with 2 different cards.

It also made me wonder if a normal RAM card would have the same effect when plugged into slot 3, or if the same issue would occur. I guess my question is "is it an AUX slot issue only or would the same problem occur in slot 3?"

I guess if it is the MMU, as it seems it is because of the freeze spray test, then it won't matter if it's slot 3 or the aux slot.

Since the

As it goes, I'm not that bothered about 80 column mode, but I do want the extra RAM, so it may be possible to just have a RAM card in another of the full expansion slots, but the cards I have only work in the AUX slot, so it'd mean buying another card that works in a normal expansion slot.

As a side note I left the IIe running a burn in test for all of yesterday with the AUX slot empty, and it was solid as a rock, so I am wondering "Why didn't this machine have an 80 column/RAM card when I bought it?"

The FDD controller was hand built which suggests it was an enthusiasts machine and I'd possibly expect an 80 col card, so has it had this issue since I bought it and I've never forced it to fail until now? Or has it become more marginal the more I've used it?

I think I need to do more testing and see if I can probe the signals that would affect the RAM contents on the AUX board.

Off to look at block diagrams and schematics again..

[rubber_jonnie]

This does sound as if it’s the slot transceiver direction control becoming unreliable.

Well, there are a few things I can say for sure:

1. The problem occurs with 2 totally different AUX slot cards. All they do is add RAM.
2. The base machine itself is rock steady.
3. The problem seems to be the MMU as when I used the freeze spray the corruption went away immediately and returned once the MMU warmed up again.
4. The problem is corruption in RAM locations on the AUX card, as every other column is affected. The corruption immediately clears when freeze spray is used on the MMU.
5. Once the issue occurs, whichever card I use, the machine won't go into 80 column mode using PR#3 and with the 8MB card, loading the config disk I get an error message stating the card isn't installed.
6. With the standard 64K card, once the issue occurs, PR#3 causes the machine to hang as if you'd executed the command without a card in the AUX slot.

Theres a lot to take in, but it does all seems to centre around the MMU. What's annoying is not having another machine to compare against.

[stephen_usher]

Oh, OK. So it's not the slot interface at all but the aux memory interface. Is the MMU doing the 80 column stuff as well or is it another chip and the MMU is doing the memory paging? If the latter then potentially a circuit could be made up to replicate this easily enough.

[rubber_jonnie]

Oh, OK. So it's not the slot interface at all but the aux memory interface. Is the MMU doing the 80 column stuff as well or is it another chip and the MMU is doing the memory paging? If the latter then potentially a circuit could be made up to replicate this easily enough.

The RAM card in the AUX slot simply adds an extra 64KB of RAM and a little 74LS logic, but from what I can see the MMU deals with the RAM in the AUX slot with RA0-RA7 address lines (The aux slot is not like a full expansion slot).

MMU:

Screenshot 2023-04-16 at 19.51.52.png

AUX slot:

Screenshot 2023-04-16 at 19.50.48.png

I started trying to look this afternoon but got distracted building mu 2nd v2 Analog, and then ended up going off to do something else.

I'm not sure if I mentioned, but whilst it works fine with nothing in the AUX slot, if any RAM card is in there, it can still fail even if you don't put it in 80 column mode. I found that out doing burn in tests and it occurs when the machine doesn't see the extra RAM after about 10-15 minutes.

[stephen_usher]

So, it looks like the only signal from the MMU which is vital is the /CASEN line. The /EN80 will just switch between 80 and 40 column mode.

Given the the AUX slot memory is accessed via the main memory bus which doesn't go though the MMU this would suggest that there's not a connection problem there but there could be on the video bus.

Having an 80 column card in addition to the AUX would probably cause a bus conflict and memory fighting each other as well as the system already has the 80 column card built in.

By the way, if you put the 80 column card in a slot other than 3 you have to use a PR# command which matches the slot number to enable it.

[rubber_jonnie]

So, it looks like the only signal from the MMU which is vital is the /CASEN line. The /EN80 will just switch between 80 and 40 column mode.

Given the the AUX slot memory is accessed via the main memory bus which doesn't go though the MMU this would suggest that there's not a connection problem there but there could be on the video bus.

Having an 80 column card in addition to the AUX would probably cause a bus conflict and memory fighting each other as well as the system already has the 80 column card built in.

By the way, if you put the 80 column card in a slot other than 3 you have to use a PR# command which matches the slot number to enable it.

Ok perhaps I need to think about how I've described this. It isn't an 80 column issue, that is just one way I can replicate the problem. The problem is that the memory in the AUX slot becomes corrupted and then the card becomes inaccessible, as though there was no card installed.

This manifests as on screen corruption in alternate columns of the 80 column mode, or if running a game that requires more than 64K, the game will hang and on reboot, if you issue the command PR#3, it will hang, and never go into 80 column mode.

If I use my GW RAM2E II card, which is configurable between 64KB and 8MB, the same issue occurs, but when I try to run the config utility, it tells me the card isn't installed.

The AUX slot isn't really connected the same way as the expansion slots are.

Screenshot 2023-04-17 at 09.24.32.png

What I need to figure out is why is the RAM disappearing, and for now it seems to be the MMU based on the freeze tests.

Reading this:

Screenshot 2023-04-17 at 10.09.50.png

I'm wondering if it's something to do with the C07X line on pin 6 of the AUX slot that appears to control the bank select. If that had a problem, then the bank could vanish. I at least need to check it, its connected to UA12, a 558 timer.

It could also certainly be CASEN since that is coming from the MMU, so I'll have a look at that too.

I also see that it is possible to get an AUX port diag card, but lord know if they are available in any form now, and at what price.