80-Col / 64k ROM Card Memory Corruption

18 posts / 0 new
Last post
Offline
Last seen: 3 months 3 weeks ago
Joined: Mar 12 2024 - 03:37
Posts: 7
80-Col / 64k ROM Card Memory Corruption

Morning, everyone! I've been lurking on these forums and elsewhere for a while taking in tips and tricks where I can to get my PAL Apple IIe up and running, but I've made an account as I've now run into a problem and I'd like a second opinion.

I bought a recreation 80-col/64k RAM card and it was going great for the first few days, but slowly more and more corruption would creep in and now it fails memory tests and text is illegible. I thought maybe the card had borked itself so I bought another one which just arrived yesterday and that one corrupted pretty much instantly as well; every text cell which pulls data from the RAM board displays "/" by default and text rendered in those fields in garbled, so you end up in a funny position of booting DOS 3.3 in 80-col mode and are greeted with:

(/L/O/A/D/I/N/G/ /I/N/T/E/G/E/R/ /I/N/T/O/ /L/A/N/G/U/A/G/E/ /C/A/R/D/)

To be safe I cut AX1 on the new card, but it made no difference (I have a B revision board).

My guess is some 74 series logic chip on the motherboard is on it's way out, but before I start disassembling the computer to de-solder chips I thought I'd get a second opinion first as well as a point in the right direction of what chip it might be on a PAL motherboard.

My IIe is unenhanced (boots with Apple ][ )

Thank you! :)

Offline
Last seen: 2 hours 13 min ago
Joined: Feb 27 2021 - 18:59
Posts: 551
not too many possible causes

The 74LS125 at location E1 is one possibility. Another is the MMU chip. Also check to see if the problem is the AUX socket itself.

 

Also spotted a VZ-1 and UB-Xa in the corner, but the real desirable synth to pair with a IIe is the K150-FS.

Offline
Last seen: 2 months 2 weeks ago
Joined: Jul 5 2018 - 09:44
Posts: 2587
robespierre wrote:The 74LS125
robespierre wrote:

The 74LS125 at location E1 is one possibility. Another is the MMU chip. Also check to see if the problem is the AUX socket itself.

 

Also spotted a VZ-1 and UB-Xa in the corner, but the real desirable synth to pair with a IIe is the K150-FS.

 

 

The only synth I own is a Casio SK-1, which is pretty much period correct for a //e.  I think I got mine around 1986 or so.

Offline
Last seen: 3 months 3 weeks ago
Joined: Mar 12 2024 - 03:37
Posts: 7
robespierre wrote:The 74LS125
robespierre wrote:

The 74LS125 at location E1 is one possibility. Another is the MMU chip. Also check to see if the problem is the AUX socket itself.

 

Also spotted a VZ-1 and UB-Xa in the corner, but the real desirable synth to pair with a IIe is the K150-FS.

 

I don't think it'll be the MMU as every other function of the computer is perfectly fine, from sound to graphics to card-slot read/write. I'll take a look at the LS chip as watching many 8-bit repair videos on YouTube has taught me, it's often the logic chips that go bad first! That was certainly the case with both disk units, and the corruption slowly creeping in over time with the first card makes me think a slow fail of said logic chip. Though are you sure it's E1? That's the CPU unless I'm reading the board layout very, very wrong. The only LS125 I can see is at position C12.

Sorry if these are all silly n00b questions and observations; I know nothing of 8- and 16-bit Apple Systems. All my 8-bit experience is with Commodore.

I own many, many synthesizers. Some may say too many.

Casio VZ1, Behringer UBXa, Sequential Prophet 5, Behringer RD8, Sequential Drumtraks (which is actually being sold), LinnDrum, Casio CZ1, Roland D-50 and a Kawai K4. A Behringer Odyssey and RD9 and Yamaha RX7 aren't pictured. :)

A song for your delectation: https://www.youtube.com/watch?v=fIb3roPBXxk

Offline
Last seen: 10 hours 39 min ago
Joined: Apr 26 2016 - 08:36
Posts: 707
(/L/O/A/D/I/N/G/ /I/N/T/E/G/E

(/L/O/A/D/I/N/G/ /I/N/T/E/G/E/R/ /I/N/T/O/ /L/A/N/G/U/A/G/E/ /C/A/R/D/)

 

 This has nothing to do with the memory expansion card in the AUX card slot.

The card you're plugging into the AUX slot is for auxiliary memory - another 64K over and above the base directly addressable memory on board.  Very few software titles take advantage of auxiliary memory.  A couple of game titles, ProTerm and Appleworks come to mind, but I'm having trouble coming up with any others off the top of my head.

 

The fault message you're getting is from the DOS 3.3 system master if it fails to store data in the upper 16K (the "Language card" in II+ speak).

The upper 16K resides on the motherboard - the IIe has 64K on board, unlike the II+ which had only 48K and relied on the Language Card for the remaining 16K.

 

Anyway...you probably have RAM issues on the motherboard, and if it's labelled "MT" then I'd give it a 99.9% chance that it's faulty and needs replacement.

 

Offline
Last seen: 2 hours 13 min ago
Joined: Feb 27 2021 - 18:59
Posts: 551
hack and slash

I thought the point was that the message appears with slashes only in 80 column mode.

In that mode, the even columns are displayed at $400-$7FF on the onboard memory, and odd columns at $400-$7FF in auxiliary memory. The 80-column monitor takes care of writing to either memory when a program displays text, but if the DOS 3.3 routines write their text output directly, it only appears in the even columns.

The reason for the slashes appearing, instead of spaces, is more the point, I thought. What puts slashes in the aux memory, or is it being written to because of a fault, and if so in which component?

Offline
Last seen: 10 hours 39 min ago
Joined: Apr 26 2016 - 08:36
Posts: 707
Ok, I see where you're going

Ok, I see where you're going with this - but the DOS 3.3 System Master doesn't invoke 80 columns, so there may be more than one issue at hand.

Offline
Last seen: 2 months 2 weeks ago
Joined: Jul 5 2018 - 09:44
Posts: 2587
baldrick wrote:Ok, I see
baldrick wrote:

Ok, I see where you're going with this - but the DOS 3.3 System Master doesn't invoke 80 columns, so there may be more than one issue at hand.

Yeah, it sounds almost like 80 columns is getting partially enabled, but memory on the AUX side isn't getting initialized correctly or something.  Maybe first the 74LS245 should be checked since I think that is uaully socketed and it is a readily availalble part.  Plus they tend to get beat on pretty hard so they fail fairly otten.

 

 

Offline
Last seen: 3 months 3 weeks ago
Joined: Mar 12 2024 - 03:37
Posts: 7
Ah! I forgot to clarify. I

Ah! I forgot to clarify. I did PR#6 after going into 80-Column mode. :)

That's a red herring. I forgot to switch back to 40-col before booking a disk.

Here's some pictures (and yes, the monitor needs opening up and readjusting; it got knocked around in transit and the yoke needs tweaking).

This is the card when you first call PR#3.

This is the when you call PR#6 to boot DOS 3.3

If you ESC+4 to go back to 40-col mode, some random corruption will remain in mainboard memory.

After calling "HOME" it works as if nothing happened and the ESC features of the card will remain active.

If I jump into Wozmon and write directly to screen memory starting at 0x0400 in mainboard memory then all characters appear as expected.

If I write a simple program to endlessly repeat "THE QUICK BROWN FOX JUMPS OVER THE LAZY DOG" in 80-col mode then you can see any characters in odd positions are never correct.

I wrote a short BASIC program which pokes A-Z into screen memory on the main board, and then A-Z on the card so you can see it in action (Warning, screen flicker!)

Everything in the main board RAM displays fine. AUX Card does not. There's also the case of uninitialised memory flashing up on screen before the program finished running.

When you switch between 80-col and 40-col mode with text already on screen and AUX data is copied back to main board RAM it copied back corrupted.

I think I can rule out main system RAM being fully working. Everything boots and runs as expected (I was even playing some oregon trail not long ago).

Offline
Last seen: 3 months 3 weeks ago
Joined: Mar 12 2024 - 03:37
Posts: 7
softwarejanitor wrote:Yeah,
softwarejanitor wrote:

Yeah, it sounds almost like 80 columns is getting partially enabled, but memory on the AUX side isn't getting initialized correctly or something.  Maybe first the 74LS245 should be check

Socketed is what I like to hear :))

 

Online
Last seen: 1 min 46 sec ago
Joined: Mar 2 2024 - 12:19
Posts: 3
.

Looks like the lower 4 bits of aux memory are stuck on. In 80-column mode the even columns will only display "/","?","O","_","o", and the cursor ($2F, $3F, $4F, $5F, $6F, $7F)

You said you have two memory cards. Do both cards exhibit the same behavior?

Offline
Last seen: 3 months 3 weeks ago
Joined: Mar 12 2024 - 03:37
Posts: 7
des1 wrote:Looks like the
des1 wrote:

Looks like the lower 4 bits of aux memory are stuck on. In 80-column mode the even columns will only display "/","?","O","_","o", and the cursor ($2F, $3F, $4F, $5F, $6F, $7F)

You said you have two memory cards. Do both cards exhibit the same behavior?

Yeah I looked up an ASCII table for the IIe after posting all those pictures and you're right; the lower nibble is stuck on.

Both cards don't display the same behaviour because I'm fairly certain the first card I have has simultaniously failed rather spectacularly. Corruption creeps in and the AUX memory is still incorrect, so I think a RAM chip on the card, a logic chip on the card, or both, have up and died.

These are the cards I have. The card displaying backslashes (and thus the one I assume works properly) is the one with socketed chips.

The "good" card is from here: https://www.laser.com/product_info.php/item/Apple_IIe_E80-Column-Apple_IIe_Extended_80-Column_Text_Card__64kB_Memory_Expansion/cPath/288_292/products_id/1989

The totally borked card was random eBay sluge.

Offline
Last seen: 3 months 3 weeks ago
Joined: Mar 12 2024 - 03:37
Posts: 7
So on my lunchbreak I've been

So on my lunchbreak I've been looking at a picture of my motherboard and datasheets for various chips. I think the two possible options are the Demultiplexer (SN74LS138N) and the previously mentioned Bus Tranceiver (SN74LS245N).

Both are socketed and both are dirt cheap and available brand new from Texus Instruments, so I'm going to buy a couple units of both.

The reason why I think it might be the LS138 is because that chip handles slot select, and sometimes when I tried to call PR#3 the disk drives would jump to life as if I had called PR#6. I thought it was a feature given I hadn't booted DOS (aborted with CTRL+Reset to immediately bring up BASIC), but now that I think about it, it's strange behaviour as I've not handed control over to that slot.

Offline
Last seen: 2 months 2 weeks ago
Joined: Jul 5 2018 - 09:44
Posts: 2587
benanderson89 wrote:So on my
benanderson89 wrote:

So on my lunchbreak I've been looking at a picture of my motherboard and datasheets for various chips. I think the two possible options are the Demultiplexer (SN74LS138N) and the previously mentioned Bus Tranceiver (SN74LS245N).

Both are socketed and both are dirt cheap and available brand new from Texus Instruments, so I'm going to buy a couple units of both.

The reason why I think

 

I think you are probably on the right track.  It will be very interesting to see what the problem turns out to be.  Chances are the bad card can be repaired fairly easily by replacing the bad chips.  Sadly it looks like that one may not be socketed, but it's not too hard to de-solder chips if they are bad.  I just clip off all the pins at the top then they can be easily removed one at a time.  Removing chips and saving them is much harder if you don't have a purpose build desoldering station with the right attachments.  Of course don't solder in the new chips, solder in a socket so next time if necessary the repair is easy.

 

 

S.Elliott's picture
Offline
Last seen: 4 days 15 hours ago
Joined: Jun 23 2022 - 16:26
Posts: 211
just a cautionary note...
softwarejanitor wrote:
 
Sadly it looks like that one may not be socketed, but it's not too hard to de-solder chips if they are bad.  I just clip off all the pins at the top then they can be easily removed one at a time.

That's an good de-soldering method for beginners, but be sure your cutters cleanly cut through the leads without stressing the motherboard or bending the leads too much!

I've repaired an Apple //c that had three broken traces on the motherboard around a TTL that the owner had previously replaced.  He told me he had done it by cutting the leads one-at-a-time with a pair of angle-cutters, aka: "dikes".  I didn't actually see it happen, but apparently his cutters put enough stress on the board to break a few traces.

(It was easy to bodge a few wires around the breaks, but he sure wished it hadn't cracked the motherboard in the first place!)

S.Elliott's picture
Offline
Last seen: 4 days 15 hours ago
Joined: Jun 23 2022 - 16:26
Posts: 211
just a cautionary note...
softwarejanitor wrote:
 

Of course don't solder in the new chips, solder in a socket so next time if necessary the repair is easy.

 

QFT: "quoted for truth"

 

Offline
Last seen: 3 months 3 weeks ago
Joined: Mar 12 2024 - 03:37
Posts: 7
Good news! Sort of.I replaced

Good news! Sort of.

I replaced the LS245 bus tranceiver and the LS138 demultiplexer. No change when replcaing the 245 but it's nice to get a brand new chip in there regardless. The LS138, however, did actually improve things. When I start the IIe, CTRL+Reset out of the boot sequence (no DOS is loaded) and call PR#3 from BASIC, it immediately jumps into 80 column mode every time instead of running full steam through all the slots to hit slot six to start the boot sequence from disk.

So whilst the LS138 was a problem, it wasn't the problem. It's nice to know I have reliable slot selection now.

I tried the broken 80-col/64k card first (soldered chips), and when I called PR#3 for a brief moment it actually behaved in an identical fashion to the working card (socketed chips) where all AUX memory on screen were backslashes. It then corrupted again so, yeah, that card is cooked and I'm 99% certain it's the LS logic chip in the middle of it.

I wrote a quick machine language program in the mini-assembler to call the AUXMOVE subroutine (the apple technical manual states it's address $C312, it's not, it's $C311, so that was a lot of wasted time wondering why it was crashing).

I inject $CF into screen memory starting at $0400. I call the routine. I then change the byte at $8012 from $38 to $18 (SEC to CLC) so the AUXMOVE routine runs in reverse. Lo and behold, the lower nibble is off, pulling $C0 from the card.

It's baffling, since when in 80-col mode the opposite is true, the lower nibble is stuck on resulting in the backslashes.

I'll continue to ponder it. Makes me wonder if a flip-flop somewhere isn't flopping it's flips properly.

Offline
Last seen: 2 months 2 weeks ago
Joined: Jul 5 2018 - 09:44
Posts: 2587
S.Elliott wrote
S.Elliott wrote:
softwarejanitor wrote:
 
Sadly it looks like that one may not be socketed, but it's not too hard to de-solder chips if they are bad.  I just clip off all the pins at the top then they can be easily removed one at a time.

That's an good de-soldering method for beginners, but be sure your

 

Oh yes.  Cutting off the pins needs to be done carefully of course.  I use a small fine pair of cutters with a sharp point and not a big full sized side cutter.  Having the right rool for the job is best, but at least having the not-completely-wrong tool is often good enough.

 

I bought a real desoldering station and a hot air rework station recently but I haven't had opportunity to use them much yet.  They are things I really shuold have invested in years ago.  I made due with little hand held soldering irons for years until I got a decent soldering station and I've been amazed at how much more pleasant it has been.  I kinda got spoiled a few years ago when I worked at a place that had a lab with a nice soldering/desoldering station I could use.  I didn't know what I was missing before that I guess.

 

 

Log in or register to post comments