• Please review our updated Terms and Rules here

4k EPROM recommendation for PET 4032 option socket

RobbyD

Member
Joined
Jul 25, 2022
Messages
31
Location
Washington State, USA
Hi folks,
I'm looking to burn a 4K ROM to put in one of the PET 4032 option sockets (i.e. $9000). Any recommendations on which EPROM(s) would be best to use for this nowadays?
Thanks!
 
Hi folks,
I'm looking to burn a 4K ROM to put in one of the PET 4032 option sockets (i.e. $9000). Any recommendations on which EPROM(s) would be best to use for this nowadays?
Thanks!
Hi,

I'm not 100% sure, but if it is like the UVeproms in earlier PET's the correct answer is the TMS2532JL.

(Not the TMS2532A, which has an awkward programming voltage that a lot of programmers do not support).

The IC's that will work are like these, just one example from the bay:

 
There may also be an EEPROM as well you can use. This may be better for development for you as it is quicker to erase. But I would have to have a hunt for the part number.

Just a thought... Are you using VICE to test your code in the first instance?

Also, what are you writing? It might be interesting for the rest of us...

Dave
 
Hugo - I just took a look at the source code for the DOS Wedge and it does in fact support Eprom at both $9000 and $A000. The binary for the ROM is not in the D64 image but in the ZIP file. You can download it here:


It's a 4K binary (but the code does not use up all 4k - it gets filled with FF so there's room to add even more functionality). It does a scan to see which version of BASIC is available. It will support Version 2 and 4 but not version 1. The code is based off the Wedge from the CBM 8050 test/demo disk. Just SYS to whichever location you installed the ROM and it will RTS back to BASIC and you'll be all set. I will definitely give this a try when I find an EPROM.

FYI - the GITHUB page is here: https://github.com/nils-eilers/CBM-DOS-Wedge

Dave - I'm not using VICE. I'm writing directly on the hardware using SUPERMON4. I suppose I'll get a proper assembler for it at some point but it's not needed at the moment. I'm writing random code to interface with the user port for various types of physical device control. Just kicking the tires at the moment but envision the PET to become a central hub for home automation, sensor monitoring, etc. I'm doing the initial dev work on the PET but I think the code will inevitably work on all 6502 systems from KIM 1 on up with some small hardware specific code logic running at initialization to determine the host system.

I just ordered a couple of the TMS2532JL EPROMS and will circle back to let you know if they work or not. Would love to find an equivalent EEPROM as well. I'll take a look at the data sheet and see if I can find an equivalent EEPROM.
 
I wrote the latest incarnation of the PETTESTER and I used VICE to test and develop that. I needed it to debug my non-working PET in the first place - and it grew from there!

I use the online assembler at asm80.com. Despite the name, it can assemble 6800 and 6502 code as well as 8080, 8085 and Z80 (and a few more besides).

A 2532 should be a direct replacement for a PET ROM. The 2732 pinout is different and, therefore, requires an adapter.

As Hugo has already stated, the 2532 and 2532A are both OK in the PET - but they do require different programming voltages - so you need to check what your EPROM programmer can handle.

It would appear that the EEPROMs skipped the 2832 variant... They went from 2816 to 2864. Of course, the 2816 is too small (2K) and the 2864 is a 28-pin device - so won't fit in the 24 pin socket (directly)!

It sounds an interesting project - and is what a lot of PETs ended up doing at the time. Not home automation - but in measurement and report production. I used ours at University for controlling an industrial robot for my final year project - but that was many, many years ago when the 8032 was brand new...

Dave
 
I think you're right about there not being an equivalent EEPROM. This website discusses doing a mod to the 28 pin variants to use in the TI-994A. Should also work for this purpose but I don't think I will go this route and just stick to EPROMs.


FYI - NTE makes an equivalent EPROM (NTE2532). Though it's a bit pricey.
 
I am going to try 68766C's. I have a tube of NOS here, they are 24pin and work in the vic20/64, just larger.
... and you can have two different images in the same chip, selectable with pin 21.

Another compatible chip despite being 28 pins is the versatile TI 2564 which can replace those non-JEDEC ROMs & EPROMS from 8 to 64Kb; the extra 4 pins are just used for programming & extra chip selects.

And of course there are adapters available to convert to JEDEC-standard EPROMs and EEPROMs; e.g.:
 
Few people seem to remember that there was also a Hitachi version of the 2532, for example HN462532.
Indeed; I *think* I've also seen a Hitachi version of a 24-pin 64K 2364-compatible EPROM, but I've never been able to find it again, so it's probably just another symptom of my failing memory ;-)

Philips also makes a 24-pin 64K EPROM (27HC641) but its pinout doesn't match either the 2564 types nor the JEDEC standard 2764! Go figure...
 
Maybe use a romulator, for all the ram/roms or just to replace (or place) a rom image at the address you want?
Add the esp32 based stand alone programmer and you could change the rom using a web browser, no need to even open the pet.
 
Hi - I have been repairing a number of PETs recently and had this exact problem... You need 2532 EPROMS for all the sockets except the E000 Edit EPROM, which can use a 2716. The 2532 has three pins wired differently to the 2732, see attached datasheets... Not many programmers do a good job with the 2532 EPROMS either - I found the GQ-4x4 from here: https://www.gqelectronicsllc.com/ to be a good option - but even then, you have to find a good supply of EPROMS... I found many used ones to be unreliable...

So like many others, I ended up with an EPROM adapter board that takes up to a 27C512 (64KB) and allows selection of 1 of the 16 4K blocks to be addressed. This allows a couple of complete ROM sets to be programmed in, as well as diagnostic ROMs as well. I use the Original Commodore Diagnostics a lot, and added the switch to allow simulation of "removing" the original Commodore clip, to proceed to the second part of the diagnostics. The switch makes it really quick and easy.

If you are in the UK, I have some spare boards - I am looking at selling them next year, at some point - along with some other mod board I have made for PETs for a 32KB static ram, an MC3446 IEEE buffer replacement and a 74154 replacement, all of which make repairing the systems much easier...
 

Attachments

  • 2532 - HN462532 datasheet.pdf
    270 KB · Views: 1
  • 2732.pdf
    72 KB · Views: 2
  • IMG_4449 2022-12-21 08_56_45.jpg
    IMG_4449 2022-12-21 08_56_45.jpg
    1.1 MB · Views: 8
Your MC3446 buffer replacement sounds interesting. It is surprising how often we see the main fault on a PET fixed only to then find that the IEEE 488 port is also faulty, almost invariably due to failure of one or more of the MC3446 buffers.

As they are still just about available purists may wish to replace with like for like anyway but it would be good for there to be a drop in replacement available, if only for diagnostic purposes.
 
Your MC3446 buffer replacement sounds interesting. It is surprising how often we see the main fault on a PET fixed only to then find that the IEEE 488 port is also faulty, almost invariably due to failure of one or more of the MC3446 buffers.

As they are still just about available purists may wish to replace with like for like anyway but it would be good for there to be a drop in replacement available, if only for diagnostic purposes.
Hi there, thank you! - Yes, I have had a few failed ones (all 3 in one board!) and was getting the RAM replacement boards made - so I came up with a simple and cheap (if not so small!) version using 74LS07 buffers...

The info isn't on the radix.co.uk web site yet... still working on the IEEE->SD Card disk drive...

Oh the 74154 replacement uses a 74LS138 as the PET does not use the lower 8 4K select lines, except for the memory expansion port.. The next version of that will have jumpers to swap the F000 and 9000 sockets so the CBM diagnostics can be easily used in machines with soldered in ROMs...
 

Attachments

  • 20221021_142745737_iOS-2.jpg
    20221021_142745737_iOS-2.jpg
    2 MB · Views: 12
  • 20221021_153925728_iOS.jpg
    20221021_153925728_iOS.jpg
    708.6 KB · Views: 10
Your MC3446 buffer replacement sounds interesting. It is surprising how often we see the main fault on a PET fixed only to then find that the IEEE 488 port is also faulty, almost invariably due to failure of one or more of the MC3446 buffers.

As they are still just about available purists may wish to replace with like for like anyway but it would be good for there to be a drop in replacement available, if only for diagnostic purposes.
They are readily available
From Halbleiter.de in Germany. They have great new old stock MC 3446's and they are perfect.
But maybe like all of the other good things, they might not last like the song lyrics, all that glitters is destined to fade away again. So, maybe get them while you can!
 
Actually a small miss - quote, "shimmers", not glitters, the exact lyrics were:

'Cause I have found
All that shimmers in this world is sure to fade
Away again.

The song Shimmer, by the group Fuel.
 
Actually a small miss - quote, "shimmers", not glitters, the exact lyrics were:

'Cause I have found
All that shimmers in this world is sure to fade
Away again.

The song Shimmer, by the group Fuel.
No Hush lyrics? - lol - FYI, the Halbleiter.de site seems to be a blank site...
 
Back
Top