• Please review our updated Terms and Rules here

Commodore PET 4008 with strange issues?

It replaces the EDIT ROM (base address $E000).

The reason for this is that this device is generally in a socket - whereas the others had a habit of being soldered into the board.

The EDIT ROM takes care of the localisation for the keyboard and display etc. (in conjunction with a replacement character generator ROM) so was generally socketed.

Dave
I managed to find some 2716s in the back of the cupboard, but for the life of me can't find the pinout of the edit rom anywhere online (in my machine I believe it to be a mos 901447-29 based on the schematics and where UD8 is on my board, wouldn't be surprised if I'm mistaken) so do any of these sound familiar to you as known compatible?
Hitachi HN462716
TMS2716JL-45
I have one of the HN4262716s and two of the other. I don't yet have anything to burn them with.
Sorry to be so awkward, I did try to find the pinouts myself but no luck and you're objectively the most knowledgeable when it comes to the pet test, thanks again for sticking with it.
 
I should mention, I have no clue if they work, but I doubt they've come to any harm in an IC draw stuffed away in the back of a cupboard at work. Only one way to find out!
 
Do not use any TI TMS2716 parts. They use a Ebb of -5V.

Use can use any 2716 that is compatible with the standard Intel 2716 like the Hitachi, or use the TI 2516 EPROM or equiv.
Oddly you can us a 2732 part if you burn the data into the upper half of that 4K part (because pin 21 is A11 input on the 2732 and always held high in the PET). Or burn the 2K data into both halves of the 4K 2732.
 
Last edited:
Bow,
I just noticed said you don't have a programmer to program the 2716 or 2516 which takes a programming voltage of 25V. An old programmer is usually needed.
Send me your mailing address by the mail system in this forum and I will send you programmed 2716 or 2516 with the daver2 pettest code.

It will be gratis since it is for a PET. Actually I have a PETTEST4 laying around with a checksum of 75DD. I'll have to see if that is the latest version.
 
Bow,
I just noticed said you don't have a programmer to program the 2716 or 2516 which takes a programming voltage of 25V. An old programmer is usually needed.
Send me your mailing address by the mail system in this forum and I will send you programmed 2716 or 2516 with the daver2 pettest code.

It will be gratis since it is for a PET. Actually I have a PETTEST4 laying around with a checksum of 75DD. I'll have to see if that is the latest version.
Bless you sir that's a very kind offer, but I'm in the UK and shipping would be absolutely extortionate. Nonetheless, I greatly appreciate the gesture - that sincerely is good of you.

To be truthful, I think it's high time I invested in a programmer anyway, just in case any other ROMs fail in future repairs; I hope this isn't the last PET I'm lucky enough to work on, but hopefully I won't have to bug the forum quite as much going forward!

I'll use the Hitachi HN462716 I found with that in mind, and see if I can get a programmer that's compatible. Might take a while (at the very least I'll have to wait till I'm next paid), but I'll get back to you folks on the results when I do.
Cheers.
 
A GQ-4x programmer will do everything you need for the PET, these are readily available on ebay and a good price.

It is only when you shift to more vintage ROMs with additional power supply requirements like a 2708, that you have to shift to other programmers. A Bytesaver board works well for those, Martin Eberhard made a really convenient program for those. For more vintage parts I have had to move to a BP1400, and in one case an Eberhard programmer for the MM5204.
 
A GQ-4x programmer will do everything you need for the PET,
Hugo, knowing you, you would not have stated this unless you have used the GQ-4X with the 2716 and the 2532, the EPROMs needed for pin for pin replacement on the PET.

I did check their compatibility list and at least for their version 4 unit (GQ-4X4), they do list these parts. This seems like an amazing device, a USB interface, works on modern PCs and supports these old 25V Vpp parts! For about $100 USD.
But I would only buy from a reputable source not a far east vendor.
 
Hugo, knowing you, you would not have stated this unless you have used the GQ-4X with the 2716 and the 2532, the EPROMs needed for pin for pin replacement on the PET.

I did check their compatibility list and at least for their version 4 unit (GQ-4X4), they do list these parts. This seems like an amazing device, a USB interface, works on modern PCs and supports these old 25V Vpp parts! For about $100 USD.
But I would only buy from a reputable source not a far east vendor.
I was surprised what good value it was. The software disk installed effortlessly (on my HP computer running Windows XP), And it has worked without a hitch ever since. No problem with either 21 or 25V parts.

When I was using it initially, I found that if it was set to a DS1225 Dallas NVram, it also programmed the Ramtron FMw08 FRAM. I notified the manufacturers and after awhile they added that to their supported device list, presumably they confirmed it worked too.

I used it to make a complete set of backup ROMs for the PET (which are in my Pet repair toolkit) as well as for many other ROM related projects, including the ROMs for the RM-65 video card.

And the 2716's for the AIM's Ramrobber:

 
A GQ-4x programmer will do everything you need for the PET, these are readily available on ebay and a good price.

It is only when you shift to more vintage ROMs with additional power supply requirements like a 2708, that you have to shift to other programmers. A Bytesaver board works well for those, Martin Eberhard made a really convenient program for those. For more vintage parts I have had to move to a BP1400, and in one case an Eberhard programmer for the MM5204.
Aha, funnily enough that's exactly the configuration I got as a "shot in the dark" second hand off a marketplace seller the other day. Should be in the post now, so I'll give it a go when it arrives. I did check its compatibility with my 27C16s and supposedly it is, according to the documentation - will check that first hand...
 
Does that little gadget have a boost converter to generate the +25V?
It must have, as it can program ROMs with either 21 or 25V voltages. It sometimes requests for its external power adapter to be connected, but that is just a 9V@200mA wallwart, most of the time it just runs from the USB power, so it must contain voltage converters to step up to 21 & 25V.
 
The unit arrived the other day and yes indeed, it seems to have a setting for the needed 25V. However, I foolishly assumed that the USB port was for data transmission!
I have most certainly not abandoned this thread, but seem now to be on the "side-quest" of finding a machine that has a working parallel port to connect it up to. I aim to do so by the end of next week, so I'll keep you all updated then.
 
The unit arrived the other day and yes indeed, it seems to have a setting for the needed 25V. However, I foolishly assumed that the USB port was for data transmission!
I have most certainly not abandoned this thread, but seem now to be on the "side-quest" of finding a machine that has a working parallel port to connect it up to. I aim to do so by the end of next week, so I'll keep you all updated then.


The USB is for data transmission, the data in the buffer on your screen, from the GQ-4x's software, is the data file that gets sent via the USB port to your device to be programmed, that you have put in the socket on the GQ-4x.

The only common programmer I know of that uses a parallel port, (rather than the GQ-4x's serial port) is the BP1400. I have one. But if you get one, make sure to get the original user manual and software installation disk with it.

Still, over the GQ-4x, the only thing so far that the BP1400 has helped me with, is programming some vintage OTP fuse style ROM's that were used on the Northstar Disk controller board and the Sol-20 keyboard.

These days it is much easier to find a computer with a USB port than with a genuine serial or parallel port. I use a vintage HP desktop that has all three ports, it was made in about 2001, and runs the now unsupported Windows XP, on a vintage Athlon 64 CPU. It has an over 20 year old mechanical disk drive, still working. As a precaution I cloned the entire machine, including all the hardware motherboards and card, and the disk drive, with cloning software (Paragon), so when it fails, which it will inevitably do, a have the clone ready to step in, with all loaded with my favorite programs, and I cloned the clone as well as a 3rd backup, including the VDU which I have two spares of. I don't want to be in the position that when the HDD fails I have to start from scratch (been there don't like it). So I am prepared. Interestingly the original HDD which was a "blue Caviar" type, has already outlasted many solid state drives three times over.

Most modern computers have trashed the ports that used parallel, serial and or USB cables and gone to wifi.
 
Last edited:
Can you post a photograph of EXACTLY what you have bought - including the I/O connections please.

If you bought a USB programmer - and it arrived with a serial or parallel port on it - you have been 'duped' and should return it...

Dave
 
Can you post a photograph of EXACTLY what you have bought - including the I/O connections please.

If you bought a USB programmer - and it arrived with a serial or parallel port on it - you have been 'duped' and should return it...

Dave
Aha, as a matter of fact I was about to do exactly that - unsurprisingly, had some issues with it!

I managed to find a windows XP machine with a parallel port (harder than you'd think in this day and age). I've installed the latest version of the willem software that I could find, hooked up the programmer and set the jumpers as instructed by the software with the correct chip specs input. It seemed perfectly capable of reading some of the old 27C16s I had found, but when I tried to program an empty one it failed miserably: neither did it write one bit correctly nor fail to write, it wrote something different to the buffer, darnit; the same occurred with the 2716 from hitachi as pictured. What's bizarre is that the erroneous data it has written seems to be the same or similar at a glance, and it sure doesn't match what was in the buffer before. IMG_20240507_103923.jpg I'm absolutely certain that it's something I'm doing wrong, but I'm unsure what at this stage. I'll try to blank them and start again, but in the meanwhile here's some photos of the process.

IMG_20240507_103905.jpg
Above: the buffer after opening pettest v4 (if I recall correctly) binary version.

IMG_20240507_103855.jpg
Above: what it managed to write. Looks mostly ok, but it's missing the first value and it failed later on, not passing the comparison.

IMG_20240507_103914.jpg
Above: the settings used on the hitachi chip.
Below: the same as before, but for the 27C16. All of these EPROMS were blank beforehand. IMG_20240507_090009.jpgIMG_20240507_085814.jpg

If you think it best to return it and search for something superior I can try, but it may take some time to find one; programmers that work seem to be elusive (or maybe it's just me)...
Cheers

Edit: should mention that a 12V 4A supply was used in all cases. I highly doubt that 48 watts is too little to write an EPROM?
 
I'm sure I suggested the GQ-4x, which is linked via usb, you won't have any trouble at all. Also, it really does pay to buy something where the makers have bothered to fit it into a professional enclosure. Things on exposed pcb's are just begging to get shorted out by some debris on the bench.

I really like that Japanese HN462716, it has a beautiful gigantic die on a Gold base in it.
 
I'm sure I suggested the GQ-4x, which is linked via usb, you won't have any trouble at all. Also, it really does pay to buy something where the makers have bothered to fit it into a professional enclosure. Things on exposed pcb's are just begging to get shorted out by some debris on the bench.

I really like that Japanese HN462716, it has a beautiful gigantic die on a Gold base in it.
With retrospect you did indeed, but comparing its £100+ price tag I simply cannot afford to spend that much on hardware I'll only use occasionally. I'm not too bothered by the exposed PCB - if need's be I can fabricate or print an enclosure later down the line. Ultimately, I don't mind too much for lacking any features like USB transfer or a casing as long as I can write to EPROMs - if it comes to it, I might be able to invest in a GQ-4X later on, but money is tight so that isn't an option for me right now.

Edit: I must confess that I too adore the HN462716 - it really does feel era appropriate, I hope it works.
 
Last edited:
Just realised that the willem software is designed for 64-bit systems and mine is a 32. I'm going to try recompiling it for 32, but I don't necessarily think that's the main issue.
 
With retrospect you did indeed, but comparing its £100+ price tag I simply cannot afford to spend that much on hardware I'll only use occasionally. I'm not too bothered by the exposed PCB - if need's be I can fabricate or print an enclosure later down the line. Ultimately, I don't mind too much for lacking any features like USB transfer or a casing as long as I can write to EPROMs - if it comes to it, I might be able to invest in a GQ-4X later on, but money is tight so that isn't an option for me right now.

Edit: I must confess that I too adore the HN462716 - it really does feel era appropriate, I hope it works.
They seem to be a lot cheaper coming out of Canada, this seller even supplies it with an eraser which you should also have, but of course there is the shipping.

The GQ-4X's seem a little pricey (considering the exchange rate) on UK ebay.


I think vintage Uveproms with the large dies, probably, have excellent long term memory retention, because the cells in them store a larger amount of charge (compared to the small dies). So that long term charge shift, that can flip the bit will take much longer to do it. But it means the gate capacitances are higher, which slows the working speed of the device. But, for most applications it makes little difference, except perhaps high speed video ones where the access time needs to be fast. So for example a Character Generator Rom for video is better as a high speed type . Many of the 50 year old Uveproms with large dies, from the 70's era, still have intact data which is amazing. I'm not convinced yet the newer ones with the tiny dies will do as well in the long run, but they wouldn't have to try very hard to beat modern SSD drive units, many of which turn out to be hopeless. I have gone back to mechanical drives in some of my "modern" computers now. The SSD's lose data if the file is not accessed often enough and refreshed.

Interestingly the same "die shrinkage" happened with transistors. The original RCA 2N3055 had a die that is about twice the linear dimensions, 4 times the surface area, and nearly 16 times the volume of the later versions . This reduced the junction capacitances and sped them up to a much higher cutoff frequency. In some circuits that worked with the originals, the newer parts burst into RF oscillation.
 
Last edited:
Back
Top