512k keyboard killed by a cable

[BEU]

I´m working with a 512k that was bought as defect. The 512k is now working but the keyboard is dead. Mouse is working. Noticed that the keyboard cable that was delivered with the 512k was not a keyboard cable but instead a telephone cable. se Picture



Larry Pina ( Macintosh Repair & Upgrade Secrets) says that this cable probably destroys the 8021 chip in the keyboard.
A functional KB from Mac plus (with correct cable) is working with the 512k. The defect KB+ correct cable: still dead.
As I have a working MAB 8021 chip from the mac plus keyboard my question is if its possible to program a new 8021 chip from a working Mac plus chip  ( Philips MAB 8021PB, R044, DSD 510, 341-0222B) or a working 128 k ( Philips MAB8021P, R044, DND427V3L, 341-0222B). All keyboards are M0110.  I have not found any ROM file to use with the programer.
 
Should I replace the 74LS123 in the KB??
Grateful for any help.

 

[trag]

I don't have anything useful to add, but wanted to say that I find this topic interesting as well. I looked into this several years ago, but don't remember what I concluded. I know I found one website that showed how to hack the contents out of an MC and it seems like it might have been the 8021, but I'm not sure. I need to hunt through my old bookmarks...

 

[trag]

Here's the link I was thinking of. It's for reading the contents of the 8048 or 8049. I'm not sure what relationship that has to the 8021, but it's what I had saved. Did the Plus extended keyboard use the 8048, maybe, while the original used the 8021? Anyway,might be some useful info there.

http://www.sbprojects.com/projects/8049spy/index.php

Looks like the 8021 is the same silicon as the 8048, but with fewer I/O pins brought out in the package. So, provided the necessary pins are available, the method listed at the above link should work for reading the contents.

 

[BEU]

Thank You trag for Your answer. The sbproject that You mentioned before is far to complicated for me. Previous I have managed to copy Hi and Lo ROM from Macplus and 512k to EPROM using a cheap ROM programmer, Stager G540. http://www.stg51.com/english/productshow.asp? It was easy to read the working ROM and to program a new EPROM.

Maybe thats impossible with the 8021?? Is there any risk for damage to the working 8021 if I try to read it in the programmer??

Both the working chips from 128k and Mac plus are 8021. No extended KB.

I found a previous thread where You discussed this problem but no obvious solution to this problem there.https://68kmla.org/forums/index.php?/topic/20051-mac-plus-keyboard-project-suggestion/.

According to Larry Pina this problem was not uncommon, but maybe You could buy a programmed 8021  nr 341-0222B from Apple at that time.

 

[Paralel]

You can't program it. It uses internal Mask ROM. However, newer versions of this particular "microcomputer" (that's what Phillips calls it...) have versions with no internal ROM that interface with an external EPROM. So, if you can read out the Mask ROM of another chip, you can program that to a compatible EPROM and use that as an external ROM for the "microcomputer".

That explains why you never found any ROM files. Since the ROM is Mask type it couldn't be "burned" like EPROM. However, there actually is a reason now to read out the ROM to file since an external EPROM can now be used.

Apparently someone realized that there would be times when one would need to update the ROM for whatever applications this chip was used for.

Unless your EPROM programmer says it can read the ROM out of this particular series of "microcomputer" trying to do so may indeed destroy the chip. I wouldn't try it unless you know for sure.

Apparently, for the MAB8021, if you apply 15V to the RESET pin, 15V to the PROG pin, and a TTL signal to the T1 pin, it will dump the internal Mask ROM to the Port 0 pin

 

[Paralel]

Also, unless my math is off, the max rate for the TTL signal is 116 kHz, but since the Mask ROM is only 1024 bytes, you probably don't need to transition signals that quickly.

 

[BEU]

Thanks Paralel for the information about the microcomputer. My programer can not program the 8021 so I will not try to read from the working 8021. Interesting if its possible to program to EPROM but I suppose thats rather complicated. 

I found some information regarding this problem in Macrumors.com with this picture

http://forums.macrumors.com/threads/macintosh-512ke-keyboard-issue.1891083/

Dont know if that is of any help.

I will try to replace the 74LS123 anyway and test again with the old 8021. The cable is reconnected so it is a correct macintosh keyboard cable now.

 

[Paralel]

Worst comes to worst, you should be able to use that 74 series to drive the T1 pin on the 8021 in order to get it to dump its ROM.

Also, I forgot to mention above, the TTL for the T1 pin needs to be positive pulse.

 

[Gorgonops]

There is a file called "mackbd.zip" listed in the MESS emulator repositories that *may* be the contents of that keyboard controller. Obviously I would investigate further before burning it into an OTP PROM or anything.

 

[BEU]

There is a file called "mackbd.zip" listed in the MESS emulator repositories that *may* be the contents of that keyboard controller. Obviously I would investigate further before burning it into an OTP PROM or anything.

Perfect. That ZIP file contains two BIN files: 341-0332-a.bin  and ip8021h_2173.bin. each file is 1 kb.

According to  https://github.com/MisterTea/MAMEHub/blob/master/Sources/Emulator/src/mess/machine/mackbd.c  that file is: "  Apple Macintosh original/512/Plus keyboard"

I will try to burn EPROM from those two and test on the KB.

Thanks for help. I will report after the weekend.

 

[Paralel]

Remember though, you will need to use a newer, compatible, replacement "microcomputer" that can address external EPROM. The external EPROM cannot be used with the old "microcomputer" as it has no way of addressing external EPROM. It only knows how to address its own internal Mask ROM.

 

[BEU]

Remember though, you will need to use a newer, compatible, replacement "microcomputer" that can address external EPROM. 

Ok perhaps to complicated for me.

I will try to find a 28 pin MCU EEPROM thats compatible with Intel 8021H.

 

[Gorgonops]

The Intel MCS-48 family has a dizzying array of members so I'm certain I can't say definitely what you want *doesn't* exist, but the only internal EPROM'ed member of it I know of that are at all common is the 8748/9. Also, as noted, there's the external-ROM 8035, but to use either of these you'll need some sort of daughterboard and I'm not sure if a program targeting the I/O of the smaller chip will directly work on these without modifications.

Assuming that schematic is correct, IE, it's just a key grid connected to GPIO pins on the MCU, it's possible the best solution if you can't find a drop-in MCS-48 that will work with the original code would be to adopt a small modern MCU like an AVR family chip into the socket and write your own scanning software. Here's someone who's built protocol adapters to use PS/2 keyboards on a Mac 128k-Plus jack using an STM32 MCU so sufficient documentation exists to pull this off. Modern 5v tolerant MCU's like the AVR could probably be dropped in there with a minimum of electrical fuss, it's just the software you need to worry about.

 

[Paralel]

The "microcomputer"s that can address external EPROM can do so without anything other than connecting the right lines. I read the manual for the internal ROMless version of this type of "microcomputer" and you just need to connect a couple of pins to the EPROM, apparently they did their best to make it really easy.

 

[Gorgonops]

"A couple of lines" is a little bit of an overstatement here. From what I can glean quickly the 8035L has a couple of options for addressing external memory but even the simplest uses some of the IO lines to implement a multiplexed address and data bus similar to the 8085, which means at very minimum you need an address latch as glue between it and a generic EPROM. (Not to mention a total count of about 24 wires between the latch, cpu, and eprom.) The 8035L is also a 40 pin package vs. 28 for the 8021 so, yeah... planning to "deadbug" the EPROM and the latch on top of the 8035 and hoping the other lines just sort of match up seems pretty optimistic. Thus I feel sort of confident in saying you'll need an adapter board to use different member of the MCS-48 family.

(If it's directly code compatible the internal EPROM 8748 is probably a better option than the 8035, but it's still not directly pin compatible unless the make a cut down one that's hard to find.)

 

[Gorgonops]

Minimal circuit for a 8035:

http://www.theresistornetwork.com/2013/07/programming-vintage-intel-mcs-48.html?m=1

 

[Paralel]

I was actually thinking of the MAB8401B as a sub-in for the MAB8021:

http://www.datasheetarchive.com/dl/Scans-005/Scans-00116775.pdf

You can actually just piggyback the EPROM right onto it. Or you can use the bond-out version MAB8401WP.

I don't understand why people keep talking Intel parts when he specifically quoted Philips MAB part numbers in his first post?

 

[trag]

I don't understand why people keep talking Intel parts when he specifically quoted Philips MAB part numbers in his first post?

The 8021 is a member of a MC family developed by Intel. The design was licensed to several other manufacturers, such as Phillips, but the source and relatives are Intel.

You see a similar thing with many parts. For example, I'm not sure who first came out with the 53c96 but you can find them from NEC and its descendants, as well as AMD and Zilog.

 

[techknight]

Dont quote me on this, but i "Think" that the mask 8021H is actually EPROM, but it doesnt have the window so its programmed once, and thats it.... 

I know there are some 8051s out there like that. 

Technically, with a bit of fuming nitric acid, you can open up a hole to the chip, and erase it with UV light and reprog like a regular windowed version.