TI99/4A Motherboard Replacement (In Progress)

[Stuart]

29 minutes ago, rkrenicki said:

I am trying to find a datasheet for the non-A version of the TIM9904, but I have not yet found one...

Look at page 249 onwards of http://ftp.whtech.com/datasheets and manuals/Datasheets - TI/9900-FamilySystemsDesign-1stEdition/9900-FamilySystemsDesign-08-Product Data Book.pdf.

[+Ksarul]

35 minutes ago, rkrenicki said:

I am trying to find a datasheet for the non-A version of the TIM9904, but I have not yet found one...

Here you go. . .

TIM9904 datasheet.pdf

[fabrice montupet]

1 hour ago, rkrenicki said:

So, just to be 100% clear:   You are using a TIM9904/74LS362 with a 48MHz Crystal, 3.3uH Inductor and a 22pF capacitor, and that works for you?

 

I am trying to find a datasheet for the non-A version of the TIM9904, but I have not yet found one...

I have to find my notes for having all precisions of any values (it was in 2018) and I will back to you.

You can donwload another TIM9904 documentation on my website,  here: http://www.ti99.com/ti/data/documents/TIM9904.pdf

Edited April 2, 2021 by fabrice montupet

[+mizapf]

12 hours ago, fabrice montupet said:

I didn't made the connection with this (great) movie because some replicas have been differently translated in French!  The "Flux Capacitor" as been replaced by "Convecteur temporel" ("Time Convertor" en English).

 

Ah, you have a similar problem like we have in German! ?

 

This was a particularly weird thing in the German version. The "flux capacitor" was originally translated as "Fluxkondensator" (Kondensator is the correct electrical term for capacitor). However, the dubbing staff or the actors, don't know, misunderstood it and took "Fluxkompensator" which is now the term known from the movie. It does not really matter whether you compensate the flux or have a capacitance to charge ... as long as the Libyans don't find you.

 

Edit: At least that's what people say. Come to think of it, it may have even be done on purpose to improve lip syncing. "Kompensator" is closer to "capacitor" than "Kondensator"; the "nd" does not sync well with the "p", but the "mp" does. And about the technical term, who cares. They are doing an awful lot of effort to get things into sync, maybe this is why dubbing is called "Synchronisation" in German.

Edited April 3, 2021 by mizapf

[rkrenicki]

Alright, I think I can safely rule out a clock problem.   Both with a new 0.33uH inductor and the old-stock inductor, I get a solid 2.99964MHz clock on 4 phases, both normal and inverted.   I tried one of my other TIM9904's and also got the same clock exactly.

 

Phi-0

 

/Phi-1

 

 

I am going to socket the RAM and ROM on my QI console sometime this weekend, to that I can test those chips.   They are still big question marks to me right now.

Edited April 3, 2021 by rkrenicki

[Stuart]

Have you checked for activity on the ROM and RAM chip select pins?

[rkrenicki]

58 minutes ago, Stuart said:

Have you checked for activity on the ROM and RAM chip select pins?

Yes, /ROMEN, /MB, and /WR all have normal looking waveforms.   

 

I went through and probed the rest of the lines, and it seems that Non-buffered A3 and A4 may have an issue as they are staying mostly low.  I started pulling chips one at a time until I saw normal strobing on those lines, and it looks like the U610 ROM may be the issue here.

Edited April 3, 2021 by rkrenicki

[Tornadoboy]

Here's some ideas/pipe dreams from me for perhaps the enhanced version:

 

- As already mentioned, a reset button, internal 32k, speech synth, parallel port, VGA, etc. 

 

- Built in SD card interface, perhaps positioned so it can come out of the cartridge slot using a 3D printed filler plate/SD slot, which could also hold the reset button in a recessed spot to keep it from being pressed accidentally. Seems to me somebody has already designed an SD side car that's open source which you might be able to incorporate. If you need help with that I can design the filler plate and other 3D printed parts.

 

- Ability to mount a TIPI internally.

 

- Ditch the cassette port and pipe the VGA out through that spot since it's the right shape for that connector, if need be put in a mini-jack for cassette that uses the hole left from the original video out.  

 

- Simplified power input to maybe just 24 or 12 volts, that way we're not dependent on the original large wall-wart and can use common off-the-shelf replacements. Again maybe you could get rid of the cassette port altogether or at least leave an unpopulated spot for a jack if someone really wants it, put the power input with the original video out was, put a VGA port where the cassette port was then use the spot where the original power supply plugged in for other things such as the network jack for the TIPI.  

 

- EPROM for several permanent, built in ROMs for things people are always using like X-basic, Editor/Assembler, etc.   

 

- Soft power-on like modern PC power supplies. Not really that important, but it would get rid of the "bonk" when the power slams on with the normal switch and might be kinder to the hardware over the long term. 

 

- Ability to switch to an overclock mode. I don't know how practical that is for this chipset and overall design, but it would be cool to be able to kick it into turbo for particularly slow stuff, rather like you could with older PCs like a 486. 

 

- Built in ability to use a USB keyboard. That also might give one the ability to remove the original keyboard, put a 3D printed or CNC'd filler plate over it and use the space for other hardware? Or if you want to stick with the original keyboard make it so it can be mounted in a separate case and connected remotely either through simpler wiring or even a wireless connection.

 

- Maybe a smaller, more reliable connector in the expansion port that perhaps has more abilities than the original and would be easier to get female connectors for. It would rule out using an original PEB (unless an adapter can be made... replacement PEB interface card that doesn't need the "fire hose?"), but people could design newer, better things to replace it with. 

Edited April 9, 2021 by Tornadoboy
I'm an illiterate bastard :p

[+dhe]

Any news rkrenicki?

[Tursi]

On 4/3/2021 at 4:02 AM, mizapf said:

Edit: At least that's what people say. Come to think of it, it may have even be done on purpose to improve lip syncing. "Kompensator" is closer to "capacitor" than "Kondensator"; the "nd" does not sync well with the "p", but the "mp" does. And about the technical term, who cares. They are doing an awful lot of effort to get things into sync, maybe this is why dubbing is called "Synchronisation" in German.

That's kind of fascinating, actually! I didn't know either the French or German terms before

 

[rkrenicki]

I just received another Non-QI TI99 in the mail.  It is pretty beat up, but it does function.   I have not had a chance to open it up and socket the ROMs yet.

 

I may actually go though and socket everything and use this as a testing reference while I am at it.

[+dhe]

If I remembering correctly, you can burn your own ROM's with 2532, also another way to test the content storage.

Edited April 11, 2021 by dhe
Fixed EPROM type/number per update from Ksarul

[+Ksarul]

3 hours ago, dhe said:

If I remembering correctly, you can burn your own ROM's with 2432, also another way to test the content storage.

That should be 2532s, not 2432s. Note most of these are the "A" variant which can be programmed at 21V, as some of the non-"A" variants need 25V, even though a lot of them seem to be 21V-tolerant for programming. Most current programmers won't program either variant, as they seem to have dropped programming voltages greater than 18V a few years ago. That said, look at the device lists to see if you still have the capability to program them.

[atrax27407]

TMS2532As are spec'ed at 21V. TMS2532s are nominally 25V chips. However, they program  very nicely at 21V (and by extension the Motorola and Hitachi chips) with a proper adapter on the TL866 when you select a TMS2732A as a template.

[+dhe]

Atrax, do you know if anyone make a board and adapter to burn 2532's and the newer pro version. I got in two late and couldn't find a TL866. I think mine is the newer TL866II.

[atrax27407]

The TL866II plus will NOT do TMS2532s. The max voltage on it and the T56 is 18V which is insufficient to program the chip. It requires at least 21V using an adapter on the TL866CS/A..

[+Ksarul]

Most of the modern programmers that will do 2532s are well out of the usual hobbyist price range, generally in the $500 to $800 range.

[atrax27407]

It looks like even the counterfeit TL866A/CS market has dried up. I could only find one on eBay.

[rkrenicki]

So, it looks like the ROMs are not the problem.   I have not probed around since trying both sets of ROMs in both boards, so I have no theories at this exact moment.   I have been working nights, so I have not had much time to work on the PCB.   

 

In the meantime, I have been continuing to clean up the schematic, as well as validating it to the Bunyard manual. 

So far I only found one missing connection on a bypass capacitor, and a power filter inductor that had both sides tied to +12V for some reason.   

I've also corrected ALL of the power delivery notation on the schematics, as they were incomplete.

 

So.. no real progress, but after this week, I will have my days back and I can actually do some more diagnostics.

In the meantime, I have ordered some new MCM6810s and 4116 RAM chips from Jameco, as well as some proper 64 pin 2.54mm pitch sockets (instead of the hacked up 40pin and 24pin socket arrangement I am using today).  Also, on the next prototype PCB, I am not going to populate all of the keyboard/joystick, cassette, and IO parts, until I can get it to boot.   While I could pull all of the parts off, it really isnt worth the time to do so.

Edited April 14, 2021 by rkrenicki

[rkrenicki]

After two weeks of night work, and beautiful weekends that meant that I ended up working outside.. I finally found some time to sit down with this again.   I am satisfied that the schematics match.  I did clean up some notations to make it a little easier to read, but no actual connection changes.

 

I did build one of the 8bit Museum's Retro Chip Testers a few days ago.  Turns out that the 20x4 LCD module that I had does not fully work, so I just ordered a replacement before I can go through and test most of the chips.

 

As for the PCB itself, I did fix a few things on the silkscreen, some of the mounting holes were off by a fraction of a millimeter, and some power related fixes that were noted above.   I will send this latest version off for manufacturing before I try again.  

[Lumpy]

Any update on your progress? This is a very interesting project and I am hoping to purchase a PCB if/when you work out the bugs and hopefully offer the boards for sale.

 

[rkrenicki]

There has been little progress as work is ramping up now that we are approaching the end of the pandemic here.   We also had a death in the family, so my workshop is now overrun with items from their estate, making it difficult to get anything done in there right now.

I need to use some vacation time here soon, so hopefully I can get a day or two in the near future to dedicate to this project.

[Lumpy]

Family always takes priority and vacation is important for mental health so take care of those things. As I said, I am very interested in this project but i can certainly wait until you are ready.

[Tornadoboy]

On 6/14/2021 at 11:05 PM, Lumpy said:

Family always takes priority and vacation is important for mental health so take care of those things. As I said, I am very interested in this project but i can certainly wait until you are ready.

Ditto, worry about the important things right now, the rest can wait for as long as necessary. 

[fjlt]

hola estoy interesado, en cuanto pueda salir a la venta la placa o los gerbers de kickad quiero una. Gracias