Issue with TI-99 I can't figure out

[tandrews_ny]

I have been on a TI-99 repair spree. I have one where there is text (letters) where the color bar should be. This is on the top and bottom of the left color bar. Everything else looks good. I've tried replacing a few VRAMs and have had no luck. I've tried different GROMs too.

 

Anyone seen this before? I can get a picture if necessary but I'm currently at work.

[+9640News]

I've not seen this before, but I might be able to offer a thought to guide you on further diagnosis if nobody has other ideas.

 

Seems like there is a diagnostics module out there. Not sure how "smart" it is to do some testing with it.

The other idea, assuming you can get a clean Extended basic screen, is to write a character, then increment that character by 1 until you have scanned all the characters. Then, display color blocks and see if you get what you are expecting. This may help rule out whether you have a 9918 issue, or if there is some broken or shorted trace to either the VRAM or from the 9918 the source of the problem.

 

If you are writing a character pattern of "FFFFFFFF" and you see "FFxFFFFF" it may be able to guide you towards a particular trace at fault.

 

You could also test out sprites as well and move them over to various areas of the screen. Might be simpler to use something like TI-Artist with a mouse pointer. You could also save colors to the screen as well for testing.

 

Just an idea. Since you have replaced the VRAM, it is obviously a project you would like to solve, rather than adding it to the discard pile and pulling out another system.

 

Beery

[tandrews_ny]

I've not seen this before, but I might be able to offer a thought to guide you on further diagnosis if nobody has other ideas.

 

Seems like there is a diagnostics module out there. Not sure how "smart" it is to do some testing with it.

The other idea, assuming you can get a clean Extended basic screen, is to write a character, then increment that character by 1 until you have scanned all the characters. Then, display color blocks and see if you get what you are expecting. This may help rule out whether you have a 9918 issue, or if there is some broken or shorted trace to either the VRAM or from the 9918 the source of the problem.

 

If you are writing a character pattern of "FFFFFFFF" and you see "FFxFFFFF" it may be able to guide you towards a particular trace at fault.

 

You could also test out sprites as well and move them over to various areas of the screen. Might be simpler to use something like TI-Artist with a mouse pointer. You could also save colors to the screen as well for testing.

 

Just an idea. Since you have replaced the VRAM, it is obviously a project you would like to solve, rather than adding it to the discard pile and pulling out another system.

 

Beery

 

Sadly, putting a cart in results in just a mess. I haven't tested basic, but I can tell you the numbers don't render either. I can't determine if the issue is even RAM related. It did have one bad module which I replaced.

[+9640News]

With no carts in, see if you can get to a viewable screen. If you can, I would start to think dirty cartridge port or bad connector(s).

[tandrews_ny]

With no carts in, see if you can get to a viewable screen. If you can, I would start to think dirty cartridge port or bad connector(s).

will try tonight.

[+Lee Stewart]

It does , indeed, look like a VRAM issue. If it is only one of the VRAM chips, you simply missed replacing the right one(s). See end of post #1 of New to the group? TI-99/4A FAQ, Hardware and software resources READ FIRST and, paricularly, this VRAM troubleshooting page at ninerpedia.org.

 

...lee

[tandrews_ny]

It does , indeed, look like a VRAM issue. If it is only one of the VRAM chips, you simply missed replacing the right one(s). See end of post #1 of New to the group? TI-99/4A FAQ, Hardware and software resources READ FIRST and, paricularly, this VRAM troubleshooting page at ninerpedia.org.

 

...lee

https://www.ninerpedia.org/wiki/File:Vram_df.pngTHATS the exact issue my friend. I have bookmarked those pages for future reference. I wasn't sure what to search as sometimes...its not easy to articulate my thoughts. I'll be fixing this tonight!

Edited May 31, 2018 by tandrews_ny

[tandrews_ny]

Another one saved. This is two I've repaired. Thanks so much. Waiting on the ram I ordered to do a full replacement job on these.

Edited June 1, 2018 by tandrews_ny

[+Lee Stewart]

Another one saved. This is two I've repaired. Thanks so much. Waiting on the ram I ordered to do a full replacement job on these.

 

You are too much!

 

...lee

[Ed in SoDak]

I wish I woulda had this info back when I bodged an attempt at socketing and replacing the VDP RAM in one of my consoles. I never got it to work and the guy I sold it to as non-working also gave up! Originally, probably only one chip was bad, but my shotgun approach killed not only the perp, but all the witnesses!

-Ed

[tandrews_ny]

I use a desoldering gun and heatgun. No pulled traces yet. Something I was wondering...how do I replace the grom chips? Are they just eeprom ?

[+mizapf]

No, GROMs are TI-only. They can be described as streaming ROMs: You set the start address, and with every read access you get the next byte from its memory. I'm not aware of any similar circuit outside the TI world.

[tandrews_ny]

No, GROMs are TI-only. They can be described as streaming ROMs: You set the start address, and with every read access you get the next byte from its memory. I'm not aware of any similar circuit outside the TI world.

So if they are bad, no real replacement is available? I have one that just gives me a blue screen, nothing else.

Edited June 1, 2018 by tandrews_ny

[+mizapf]

Only if you find the same GROM from another console.

[tandrews_ny]

surprised no one has made a replacement part. seems like it wouldn't be too difficult?

[+Lee Stewart]

surprised no one has made a replacement part. seems like it wouldn't be too difficult?

 

The functionality has been replaced by the UberGROM and the FinalGROM, but, alas, pin-compatible solutions to actual TI GROMs have not yet been implemented as far as I know. I do not know how difficult it might be to actually duplicate a GROM. I do know that the auto-incrementing functionality eats into the ROM space, leaving only 6 KiB available per GROM. I do not know that the auto-incrementing electronics takes the full 2 KiB that is unavailable for GPL code.

 

There are also other solutions to replacing the console GROMs—some old (HSGPL) and some new (others will need to jump in here ).

 

...lee

[+Ksarul]

There was a hardware replacement for the console GROMs, the O.P.A. S.O.B. It was a plug in replacement for the console GROMs using a pair of standard memory chips and a PAL to do the necessary register magic. There is a picture of one on a recent thread. . .unfortunately, the PAL JEDEC code used here isn't publicly available to make more of them and the original vendor no longer makes them either.

[Tursi]

The UberGROM code can be run on AVRs that are almost as small as the real GROM, and I started doing such a port at one point, but the AVR power pins are placed such that they can't be made pin-compatible. It's not impossible, something could be made with small programmable logic on a DIP-sized package (probably, or perhaps there's another microcontroller with a compatible pinout), but there hasn't been much demand.

[Opry99er]

That sounds pretty exciting, actually.

[iKarith]

The UberGROM code can be run on AVRs that are almost as small as the real GROM, and I started doing such a port at one point, but the AVR power pins are placed such that they can't be made pin-compatible. It's not impossible, something could be made with small programmable logic on a DIP-sized package (probably, or perhaps there's another microcontroller with a compatible pinout), but there hasn't been much demand.

 

Which chip do you normally use for that? GROMs hold 64k do they not? That'd probably require an ATmega1284P. Might a PIC18F27K40 fit? I'm sure it would if the GROMs were DIP24, but they're skinny DIP16s IIRC. Option #2 is much more ambitious: Some single chip that can replace both GROMs at once. That's still got the problem of the through-hole pins for the GROM sockets, unless you had something crazy involving a board with pins to fit both sockets on the bottom and pins to fit a one-sided board on the top with a QFP on it. That'd give you space, but it sounds janky.

 

Also, you'd like as not different socket adapters for the original and QI boards. Likewise one for the /4. Messy.

[RickyDean]

There was a hardware replacement for the console GROMs, the O.P.A. S.O.B. It was a plug in replacement for the console GROMs using a pair of standard memory chips and a PAL to do the necessary register magic. There is a picture of one on a recent thread. . .unfortunately, the PAL JEDEC code used here isn't publicly available to make more of them and the original vendor no longer makes them either.

I tried reading the two atmel AT22V10's and I didn't seems to be getting anything usable, though I was able to read the eprom and download it.

[+Ksarul]

Each of the original GROMs occupies an 8Kx8 space, and the console has three of them. Note also that each of the original GROMs only uses 6K of the memory space it occupies. The remaining 2K can be used with other devices (like the UberGROM), so it is not necessarily permanently lost space.

[Tursi]

Which chip do you normally use for that? GROMs hold 64k do they not? That'd probably require an ATmega1284P. Might a PIC18F27K40 fit?

I'm sure it would if the GROMs were DIP24, but they're skinny DIP16s IIRC. Option #2 is much more ambitious: Some single chip that can replace both

GROMs at once. That's still got the problem of the through-hole pins for the GROM sockets, unless you had something crazy involving a board with pins

to fit both sockets on the bottom and pins to fit a one-sided board on the top with a QFP on it. That'd give you space, but it sounds janky.

 

Also, you'd like as not different socket adapters for the original and QI boards. Likewise one for the /4. Messy.

As noted, real GROMs are 6k (in an 8k memory space). The 'UberGROM' as released does in fact use an ATMega1284, and 120k of that is usable for GROM space.

 

As for the ambitious, not as weird as you think. The console has just one GROM bus - so the built-in GROMs and the cartridge GROMs are equivalent* in terms of what they are wired to (*mostly). You can replace all built-in GROMs with a single external one - see my Youtube MPD experiments for using an ATMega to do just that. (Also proven by legacy hardware like the GRAM Kracker).

 

I've been considering similar to what you suggest for the MPD (if I ever get to it), but I'm not quite sure on packaging yet.

 

IIRC the QI and the original do at least keep the same layout for the 3 console GROMs.. so depending on surrounding clearance you could probably have a single compatible board. (I tried to verify, but my archive picture is blurry on that section ).

[iKarith]

My understanding is that a QI board is GROMs and a single jumper wire away from being 100% compatible with the standard board, so I'm very interested in this kind of thing. If you've got 120k free out on a 1284, you don't really need a 1284 for internal use unless you wanted to put an extended BASIC in the console. Which is not a bad idea actually. Hmm…