New member saying hi

Porchy · 20:31, 30 November 13

Hi!

I've very recently got my first Amstrad ever. I was a C64 user when I was young and that was pretty much me.
After that I was mainly consoles of some sort until the PC entered into my life.
I've been collecting and repairing arcade hardware for some years now but have always had a collection of retro computers.
Fast forward to about 3 weeks ago when I was asked if I could dump the PAL chip from an Amstrad 6128. I've dumped or reversed hundreds of PAL chips in the past with few issues, this one however was something very different and it got me thinking so I got myself a 6128 to do some testing on and I've fell in love with the hardware.

Nice to be here and hope to get involved with this a lot more.

Jon

redbox · 20:39, 30 November 13

Interesting article on your website.

Why were you trying to dump the chip?

Gryzor · 20:55, 30 November 13

Welcome mate! What do you know, there's still unknown stuff on our CPC

(hat-tipping to redbox for the mention)

Porchy · 20:58, 30 November 13

Quote from: redbox on 20:39, 30 November 13
Interesting article on your website.

Why were you trying to dump the chip?

Someone asked me if I could for a project they are working on.
I've made a lot of progress since that post on my site. I've got all three of the latches done but have found something else that has slowed me down a bit. It's almost done and seems to work (albeit with limited testing) but my goal is to output exactly like the original chip.

Porchy · 21:00, 30 November 13

Quote from: Gryzor on 20:55, 30 November 13
Welcome mate! What do you know, there's still unknown stuff on our CPC (hat-tipping to redbox for the mention)

Thanks. I actually thought all this was known, am I wrong?
The operation of this chip is so interesting (at least to me) that I will be writing up my finding at some point.

Thanks for the welcome

TotO · 21:41, 30 November 13

I have asked to Porchy if it was possible to dump the CPC 6128 HAL/PAL chip, to know exactly what it does.
I'm sure that will help more to understand our computer bank switching, put informations on the wiki and then copy this chip to repair bad CPC and done new one.

You already done a great work.
So, welcome here!

Porchy · 22:03, 30 November 13

Quote from: TotO on 21:41, 30 November 13
I have asked to Porchy if it was possible to dump the CPC 6128 HAL/PAL chip, to know exactly what it does.
I'm sure that will help more to understand our computer bank switching, put informations on the wiki and then copy this chip to repair bad CPC and done new one.

You already done a great work.
So, welcome here!

The recreation seems to work ok for the things I've tried so far but I can see from logic analysis that they aren't identical. The CAS lines are using a flip-flop which I can implement in test logic but can't get it to fit in with the multiple equations I have. Like the D type latch on the X lines, I've never seen a PAL chip use used in this way unless it was a registered device.

TotO · 22:15, 30 November 13

Yes, it's a really strange use.
The first time that I have contacted you, I hopped that will be an easy task as you know.

I hope that some peoples may help by giving ideas. May be it will be more easy when you will use your own CPC.

TFM · 06:34, 01 December 13

Welcome Porchy! Enjoy the forum and the wiki.

Bryce · 10:58, 01 December 13

Hi Porchy,
wlecome and nice work. I haven't tried to dump the PAL up to now, but I'd be really interested in the results so that a modern replacement could be available and possibly a simple internal 6128 compatible upgrade for the 464.

Bryce.

Porchy · 12:34, 01 December 13

well as of this morning I think I may have cracked this thing.
Ive programmed a replacement GAL16V8 device and everything seems to be working well and looks identical in simulations.
GAL devices, although obsolete are still plentiful, cheap and best of all reprogrammable

TotO · 13:25, 01 December 13

Great work!!!

Has I said to Gryzor, there is no CPC HAL/PAL entry on the wiki.
May be, it's time to create one for peoples accustomed with that.

gerald · 13:57, 01 December 13

Quote from: Porchy on 12:34, 01 December 13
well as of this morning I think I may have cracked this thing.
Ive programmed a replacement GAL16V8 device and everything seems to be working well and looks identical in simulations.
GAL devices, although obsolete are still plentiful, cheap and best of all reprogrammable

Nice, we now can replace a defective PAL (But I never had one

)
BTW, can the equation you use on the GAL be programmed on an original HAL/PAL16L8 ?

Also the use of a 16L8 instead of 16R8 is one more proof that Amstrad tried to get the most for the lowest price !

Bryce · 14:04, 01 December 13

Any chance you could upload the current file you've created, then a few others (me for example) could test it on other real CPCs.

Bryce.

Porchy · 22:20, 01 December 13

Quote from: gerald on 13:57, 01 December 13
Nice, we now can replace a defective PAL (But I never had one )
BTW, can the equation you use on the GAL be programmed on an original HAL/PAL16L8 ?

Also the use of a 16L8 instead of 16R8 is one more proof that Amstrad tried to get the most for the lowest price !

The equations would need to be compiled for the PAL device which the more modern compilers don't support but I could do it if need be. Finding a blank PAL chip is the bigger issue.

arnoldemu · 10:29, 02 December 13

Quote from: Porchy on 12:34, 01 December 13
well as of this morning I think I may have cracked this thing.
Ive programmed a replacement GAL16V8 device and everything seems to be working well and looks identical in simulations.
GAL devices, although obsolete are still plentiful, cheap and best of all reprogrammable

btw I received your e-mail through cpctech, but I didn't have time to reply to it yet.
I don't know how the /CPU signal functions. It may just be that the cpu is reading/writing to the bus and not the gate-array.
So the gate-array knows to open the memory correctly - they share access to it.

I read in a later post that you've solved all the problems

arnoldemu · 10:31, 02 December 13

Quote from: Porchy on 22:20, 01 December 13

The equations would need to be compiled for the PAL device which the more modern compilers don't support but I could do it if need be. Finding a blank PAL chip is the bigger issue.

It would be good if you could. Please make sure you put your name next to it so you have credit for it where it's due.
I am pleased you made use of my truth tables that I worked out (but never tested

).

gerald · 11:21, 02 December 13

Quote from: Porchy on 22:20, 01 December 13
The equations would need to be compiled for the PAL device which the more modern compilers don't support but I could do it if need be. Finding a blank PAL chip is the bigger issue.

The purpose of this question was indeed to check that the equations can be compiled for the original PAL model. Just to confirm we are close if not identical to the original design.

Quote from: arnoldemu on 10:29, 02 December 13
I don't know how the /CPU signal functions. It may just be that the cpu is reading/writing to the bus and not the gate-array.
So the gate-array knows to open the memory correctly - they share access to it.

Yes, /CPU is indicating whether the RAM is allocated to the GA/CTRC (when high) or the CPU (when low).

Porchy · 14:36, 02 December 13

Quote from: gerald on 11:21, 02 December 13
The purpose of this question was indeed to check that the equations can be compiled for the original PAL model. Just to confirm we are close if not identical to the original design.
Yes, /CPU is indicating whether the RAM is allocated to the GA/CTRC (when high) or the CPU (when low).

The GAL devices are a superset of PAL devices so its just a matter of specifying which device within the equations file. In my case I compiled for a GAL16V8 which is a direct drop in replacement for all 20 pin PAL devices. There is a small problem with compiling for PAL chips as the windows based programs don't support them so will need to use old DOS programs. Its not a big deal but as PAL chips are write once devices getting hold of a blank one is getting harder and only really expensive or older programmers support writing them.

Ill continue to test what I've got and it works well for all 128k games I've tried so far.

Munchausen · 15:17, 02 December 13

Hello and welcome!

Great work reversing the PAL, this is great news for preservation of CPCs

Have you had any luck with registered devices (16R4)? The multiface (http://www.cpcwiki.eu/index.php/Multiface_II), a memory dumper for the CPC, has one that has not been reversed, and it would be great to reverse it so that it can be reproduced as it's becoming quite rare.

Have fun with your 6128!

Porchy · 15:42, 02 December 13

Quote from: Munchausen on 15:17, 02 December 13
Hello and welcome!

Great work reversing the PAL, this is great news for preservation of CPCs

Have you had any luck with registered devices (16R4)? The multiface (http://www.cpcwiki.eu/index.php/Multiface_II), a memory dumper for the CPC, has one that has not been reversed, and it would be great to reverse it so that it can be reproduced as it's becoming quite rare.

Have fun with your 6128!

Registered devices can be reversed. Work is underway to make this easier but its not ready yet.
Depending on how the registered device in used depends on how easy or difficult it will be to manually recreate.

Bryce · 15:51, 02 December 13

The MultiFace uses 2x PAL16L8A and 1x PAL16R4A, so it will be a pain in the arse to reverse engineer as you will probably have to simulate all three at the same time.

Bryce.

Munchausen · 16:37, 02 December 13

Yeah, but both the 16L8 were already were already reversed (by Giant). I emailed him and he said he was just stuck with the 16R4. The jedec files for them are available on the page I linked.

I know Charles McDonald has managed to reverse engineer some registered PALs automatically, or at least dump them (http://cgfm2.emuviews.com/) but I don't have any contact details for him.

EDIT: I see Charles has added a contact email address to his web page, so maybe it would be possible to contact him there.

Bryce · 16:41, 02 December 13

Cool, I didn't know that the other two are already reversed

Bryce.

Porchy · 16:58, 02 December 13

Quote from: Munchausen on 16:37, 02 December 13
Yeah, but both the 16L8 were already were already reversed (by Giant). I emailed him and he said he was just stuck with the 16R4. The jedec files for them are available on the page I linked.

I know Charles McDonald has managed to reverse engineer some registered PALs automatically, or at least dump them (http://cgfm2.emuviews.com/) but I don't have any contact details for him.

EDIT: I see Charles has added a contact email address to his web page, so maybe it would be possible to contact him there.

I know Charles well and am working with him on this. I dump almost all of my PAL chips using the hardware he designed. As a general rule the 16L8 chips are usually simple to reverse, it literally takes a few seconds but every now and then something crazy happens like with this 6128 one.

News:

New member saying hi