• Please review our updated Terms and Rules here

Powertran Cortex

Glad to see you've got the 9938 mod article too, TMS9995! Your help has been invaluable to rescuing the system documentation for the Cortex. I had to stop adding to my earlier post to keep the 4-year old amused for a while. . .and I really understand that you were scanned out after the last batch of things you scanned in. That was a major contribution all by itself!

Have you tried Anadisk to get good images of the Forth and pascal MDEX disks? That one usually works on almost anything.

I can put anything else that trickles in up on WHTech too, as the data there is starting to show up on Cortex searches as well. My next task is to give an update to the folks at the Old Computer Museum.
 
Courtesy of TMS9995, there are now several MDEX disk images for the Cortex Emulator up on WHTech. Please check out the CORTEX_README file to see how to use them.
 
Hello all

I hardly ever use forums, so pardon me for being dumb.....

I take it PM is Personal Mail, if so how do you do it from this forum.

I don't have a problem with creating colour scans of everthing Cortex related so that it can go it the public domain and be of use to someone else.
The problem I am having at the moment is no access to an adobe writing software to create PDF's from my scans.
I did have adobe on my work laptop, but it seems to have turned into vapourware!

The only thing I can't scan without destroying it is the doorstop databook on the TMS9900 family, so I am not going to do that one.

I am truly amazed at the Cortex emulator, I had the same program running on the MK2 and the emulator at the same time.
Needless to say the MK2 was faster ;-). The emulator was running on an old lab PC the is a P3 clocked at 1GHz.
I would be interested in seeing seeing the source on how the Cortex emulation is done.

I have beed messing around with FPGA's for the last month and got to wondering about an FPGA replacement for the TMS9995 or even the whole Cortex (inc TMS9909 and the infamous 74LS2001). The are some IP cores avaliable for the TMS9995, but I dare not ask how much they cost for fear of going into a coma.

I am looking at building an updated Cortex on a single eurocard, using original TI chips but modern everthing else including a V9958 or the later V9959. The ancient DRAM would be replaced by a single and not so ancient larger SRAM. The "quaint, ancient, rare as hens teeth, expensive as ????? TMS2564's" will be replaced by a "nice, standard, common as muck, cheap as chips 27c512". The cassette interface can go as well..., LOLMA.

I have been scouring the internet for traces of a TMS9909 and some suitable diskdrive for it (BBC B ones should do the trick).
I have CDOS 1.2 on a 5.25 floppy (I don't know if it is any good).
That reminds me, would it be possible inthe emulator to map the Cortex drive to a physical PC drive?

I have been offered a TMX9909 by a former work colleague, he says it is a prototype TMS9909 but he can't gaurantee it will work in the Cortex or that the chip is in fact a worker.

Regards

Danny
 
Last edited:
It really depends on what format your scan saves them to, Cybernaut--but the OCR software I have will likely read whatever you get into an electronic format--I've even converted JPG pictures into documents with it--that was how I received the Cortex BASIC manual.

I sent you a PM about TMX9909s, so you should see a link under your name at the top left of the screen that will take you to your Private Messages. They will work, though.

The 9995 FPGA cores are expensive, especially when one considers that the chips themselves are still in the $40 to $80 range when they are available (and I have found a number of sources that still have them in quantity). I seem to remember that the cores were in the 3-5K range from the last time that one of my friends went looking.

I like the idea of a Cortex on a Eurocard. That would be a major interesting project. . .and if you are good at FPGA programming, you might want to take on the project of doing a CPLD for the 74LS2001, as we do have enough information now to do something on that front. I will get around to it eventually in any event, but my next year or two are getting pretty full project-wise. ;)

I'm not sure which application TMS9995 used to create his disk images, but one should be able to take them and restore them to physical disks with the appropriate program.
 
The are some IP cores avaliable for the TMS9995, but I dare not ask how much they cost for fear of going into a coma.

When I worked for TI a few years ago I wrote a "9900 family member" core in vhdl from scratch, I don't have access to it anymore but there is no real reason that a 9995 core couldn't be written if required, the hard part is verifying it, especially if you want to faithfully emulate the processor when doing strange things, for example:

IDT 'TEST'
*
RORG
LI R1,>0481
X R1
NOP
END

In a real 9995 you never reach the NOP and even NMI doesn't pull you out (a 990/10 needs powering down!!!) and is non-privileged code.
Anyone care to guess why NMI doesn't work?
 
When I worked for TI a few years ago I wrote a "9900 family member" core in vhdl from scratch, I don't have access to it anymore but there is no real reason that a 9995 core couldn't be written if required, the hard part is verifying it, especially if you want to faithfully emulate the processor when doing strange things, for example:

IDT 'TEST'
*
RORG
LI R1,>0481
X R1
NOP
END

In a real 9995 you never reach the NOP and even NMI doesn't pull you out (a 990/10 needs powering down!!!) and is non-privileged code.
Anyone care to guess why NMI doesn't work?

My guess is that for NMI (and all other interrupts (excluding Reset)), the interrupt is only acknowledged between instructions and that since the X instruction is executing itself, it never completes.

I tried it on my emulator and it bombs out with a Stack Overflow as the X instruction emulation just keeps calling itself :( I guess this is somewhat acceptable as you have clobbered the CPU at that stage.
 
spot on, nasty thing to do to a processor isn't it. A later member of the family made it such that if the target of an X instruction was another X it generated an illegal instruction interrupt to protect itself.
Interesting to hear what your simulator made of it.
 
I am truly amazed at the Cortex emulator, I had the same program running on the MK2 and the emulator at the same time.
Needless to say the MK2 was faster ;-). The emulator was running on an old lab PC the is a P3 clocked at 1GHz.
I would be interested in seeing seeing the source on how the Cortex emulation is done.

Thanks for the complement. I guess I could put the source up on WHTech if you want to have a laugh at my first C++/Direct X project! It started out in C as a DOS program. I then made it a Windows App using Borland C++ Builder, added Direct X and then converted it all to classes / C++. I'm quite proud of it really!


I have beed messing around with FPGA's for the last month and got to wondering about an FPGA replacement for the TMS9995 or even the whole Cortex (inc TMS9909 and the infamous 74LS2001). The are some IP cores avaliable for the TMS9995, but I dare not ask how much they cost for fear of going into a coma.

I have dreamt of putting the Cortex on a chip, but I think that's as far as it will go unless I win the lottery and have all the spare time in the World!

I am looking at building an updated Cortex on a single eurocard, using original TI chips but modern everthing else including a V9958 or the later V9959. The ancient DRAM would be replaced by a single and not so ancient larger SRAM. The "quaint, ancient, rare as hens teeth, expensive as ????? TMS2564's" will be replaced by a "nice, standard, common as muck, cheap as chips 27c512". The cassette interface can go as well..., LOLMA.

On both of mine I swapped out the 2564s for a 29F010 (or some FLASH equivalent) and replaced the TMS4500 & 4164s with a 628128 Static Ram (used with the Memory Mapper for 128K). You also get a slight speed increase as the TMS4500 no longer slows down the CPU during refresh (although removing the automatic waitstate generation on the TMS9995 is really the way to go, but that's another story...)


I have been scouring the internet for traces of a TMS9909 and some suitable diskdrive for it (BBC B ones should do the trick).
I have CDOS 1.2 on a 5.25 floppy (I don't know if it is any good).
That reminds me, would it be possible inthe emulator to map the Cortex drive to a physical PC drive?

I have been offered a TMX9909 by a former work colleague, he says it is a prototype TMS9909 but he can't gaurantee it will work in the Cortex or that the chip is in fact a worker.

I'm sure Ksarul has told you in his PM, but my understanding of TI at that time was the chips went TMP, TMX & TMS from prototype through production and I don't think the TMX9909 ever made past TMX :(

For your floppy drives, just use a couple of regular 3.5" PC drives. Cover the hole on the diskette so it thinks it's a 720K (ie non high density) and you're good to go.

Not sure about mapping a PC drive to the Cortex emulator. It's not like it had a hard drive on it to begin with. I have some utilities that can extract a file from a disk image and save it as a PC file. I suppose the other way would be to hijack the cassette save/load?
 
I'm not sure which application TMS9995 used to create his disk images, but one should be able to take them and restore them to physical disks with the appropriate program.

I use the CDOS Disk I/O routine to read one sector at a time and send it via the TMS9902 RS232 port to a PC that simply writes it to a file.
 
Last edited:
I've been reading & downloading the documentation, don't worry. But time is one thing I'm ridiculously short of at the moment. Plus my 'scope did "the big firework" a couple of weeks back and I ended up spending quite some time repairing that. Computery stuff (actually ALL fun stuff :( ) is very much on a back burner till I can get the loft conversion (and heating) fully functional. That should be within the next 6 months or we all freeze and I go loopy. and I WILL be adding my questions to the list!
 
I've been reading & downloading the documentation, don't worry. But time is one thing I'm ridiculously short of at the moment. Plus my 'scope did "the big firework" a couple of weeks back and I ended up spending quite some time repairing that. Computery stuff (actually ALL fun stuff :( ) is very much on a back burner till I can get the loft conversion (and heating) fully functional. That should be within the next 6 months or we all freeze and I go loopy. and I WILL be adding my questions to the list!

Cool. Good luck with everything. Just wanted to tease as you were instrumental on getting me to dive in at the deep end! Glad i did though :)
 
Thanks for the complement. I guess I could put the source up on WHTech if you want to have a laugh at my first C++/Direct X project! It started out in C as a DOS program. I then made it a Windows App using Borland C++ Builder, added Direct X and then converted it all to classes / C++. I'm quite proud of it really!

I wouldn't laugh, I am not that good a programmer that I would even know where to begin writing an emulator for a computer that is over 27 years old.

I have dreamt of putting the Cortex on a chip, but I think that's as far as it will go unless I win the lottery and have all the spare time in the World!

Its been done for the Amiga, Minimeg. Maybe the same hardware could be pressed into service for the FPGA Cortex.

On both of mine I swapped out the 2564s for a 29F010 (or some FLASH equivalent) and replaced the TMS4500 & 4164s with a 628128 Static Ram (used with the Memory Mapper for 128K).

Good minds think alike :D

You also get a slight speed increase as the TMS4500 no longer slows down the CPU during refresh (although removing the automatic waitstate generation on the TMS9995 is really the way to go, but that's another story...)

You learn something new everyday. It's not really new to me, I just forgot that I knew it LOL

I'm sure Ksarul has told you in his PM, but my understanding of TI at that time was the chips went TMP, TMX & TMS from prototype through production and I don't think the TMX9909 ever made past TMX :(

No wonder the bessed things are as rare as the proverbial rocking horse doodoo!

For your floppy drives, just use a couple of regular 3.5" PC drives. Cover the hole on the diskette so it thinks it's a 720K (ie non high density) and you're good to go.

I was thinking about this, but I wasn't to sure that it would work. But then again I have the "chicken and egg syndrome" the only copy I have of CDOS is on a 40T SS SD 5.25" floppy and you need a drive to load it in.

Maybe I could reverse the direction of the your imaging technique, and get the Cortex to read the WHTech CDOS image from the PC via its RS232 port. BTW, my copy of CDOS is actually 1.11 serial No 53 (1984), and not 1.2 as I stated previously.

Not sure about mapping a PC drive to the Cortex emulator. It's not like it had a hard drive on it to begin with. I have some utilities that can extract a file from a disk image and save it as a PC file. I suppose the other way would be to hijack the cassette save/load?

I was thinking more along the lines of mapping a PC floppy or USB stick to one of the floppy drives on the emulator. I wasn't actually thinking about the PC's HDD or full on "map network drive" type of thing. Just simply mapping A: to CD0 :D job done, but I wouldn't where to start with programming that cos it sounds easier said than done.
 
Last edited:
Tmx9909

Tmx9909

Originally Posted by tms9995:
"I'm sure Ksarul has told you in his PM, but my understanding of TI at that time was the chips went TMP, TMX & TMS from prototype through production and I don't think the TMX9909 ever made past TMX"

Here's the chap to ask: <http://www.linkedin.com/in/longley>.

***
New Product Definition Engineer
Texas Instruments - Houston
(Semiconductors industry)

1976 — 1979 (3 years)

Architect of the TMS9909 Floppy Disk Controller - precursor to modern DSP's, and TMS9911 DMA Controller. Transferred 9909 to design team at TI Bedford, England and evangelized TI processors in Italy. Wrote articles on TMS9940 16-bit microcomputer published internationally, and wrote datasheets and most original content of "The TMS9900 Family System Design Guide". Worked on SPICE models of microprocessors and a high-speed bus architecture.
***
 
I'd seen some posts from him before. He indicated that they were in production for ten years or so in that thread, but everything else I've seen cut that to about five. . .the most recent ones I've seen came from Malaysia and the Philippines though, so they may have been in use longer on that side of the ocean.
 
Powertran Cortex

I stumbled across this forum the other day after watching a TV program on early PCs.

I'm assuming that most people who still have a Cortex probably have all the documentation that came with it but just in case anyone needs additional info, I've got the following documentation that I could try scanning:

Cortex users manual
The 9900 family data book
The TMS9918A/TMS9928A/TMS9929A Video display processor data manual.
The TMS9995 16 bit microcomputer data manual.
Pages entitled "Reprint of article published in Electronics Today International (with corrections and amendments necessary to be consistant with kit)"
Including :
Block diagram for the complete Cortex.
Circuit diagram of the CPU and DMAC circuitry.
Circuit diagram of the memory section.
Circuit diagram of the video display circuitry.
Circuit diagram of the keyboard and control port.
Circuit diagram of the RS232 and cassette interfaces.
Circuit diagram of the floppy disk controller section.
Component overlay for main board..
Circuit diagram of the E-bus interface.
Circuit diagram of the keyboard.
Circuit diagram of the power supply.
Component overlay for the power supply.
The keyboard overlay.
Circuit diagram of the parrallel I/O board.

Parallel I/O project for the Cortex - ETI September 1985.

Cortex users group news letter # 3

Fig-forth users manual / memory map.

My Powertran Cortex has been tucked away in my loft for the past 20 years and I've always wanted to show my kids my early attempts at writing games for the PC.
The only problem I've got is that at least one of my eproms is faulty and I've never been able to get another copy.
I don't know if you can still find TMS2564 eproms but if necessary I would be willing to sacrifice my fig-forth eproms if somebody was able to reload it with the missing data for me.

Once again, If anyone needs a few pages of the above info scanned and e-mailed, just let me know.
 
Last edited:
Hi Andy,

Welcome to our growing little group! I think we now have about 9 known Cortex computers between us which is great.

Have you seen what's on ftp://whtech.com/Powertran Cortex/ ? We have quite a lot of info, so please take a look and see what you can add. I believe most of the TI data you mention is also pretty easy to come by.

I think what is really interesting to us is the Forth stuff! I don't know of any copies of this so it would be awesome to copy the ROMs and documentation. Do you have access to an EPROM programmer so you can copy the ROMs to a binary file? If you do, you can also download the CORTEX ROM image from the above link and blow yourself a new copy to get you machine going again in its native environment. The 2564s are kind of rare but it's pretty simple to replace one or all with something fairly modern. Not sure if this is something you are comfortable with? Ksarul is quite good at locating the obsolete ICs.

If not, I could make a copy of the ROMs and program the CORTEX Basic back in. We are all pretty dedicated to the preservation of this machine so I'm sure something can be sorted out.

Thanks again for checking in!
 
I don't know if you can still find TMS2564 eproms but if necessary I would be willing to sacrifice my fig-forth eproms if somebody was able to reload it with the missing data for me.

If you can't find 2564s You can probably make an adapter for 2764s which are easy to find, with a turned pin socket & some mod wire. the pinout's similar.

if you put your location on, we also know who's up the road and can post you the eproms more easily.
 
They have the SN74LS612N memory mappers in too item number 180273705894.

I've just rooted out my programmer, apparently it will program 2564s!
 
Last edited:
Back
Top