I had a TI 99/4A for a couple years in middle school and learned how to program using the Basic language. The system came with a useful workbook that explained the Basic programming language well. The problem was storage with only a tape recorder type date storage which I didn't like cause my school had Apple computers with disk drives which operated quite a bit quicker than a tape recorder. A few years later, I sold it and bought a Commodore 64 which did the job a bit better with a floppy drive and some nicer features.
TI 99/4As were really popular in Argentina in the 80s. Schools and High Schools who had a Computer Lab had them. It was the first computer I was able to get my hands on, so I have fond memories of it. Great video Noel!
Noel's Retro Lab Yes. The Commodore 64 was also extremely popular, mainly because a local (licensed) version was produced: Drean Commodore 64. And in the Sinclair front, it was mostly Timex Sinclair 2068s. Then, the MSX explosion (with local producers too). Very peculiar. :D Edit: I forgot to mention, the thing in Argentina was the curious PAL-N TV system. So most computers needed the video section reworked. The local Spectrum manufacturer was Czerweny.
Right! I read about that PAL system in some of the forums. So the TI-99 available wasn't a clone or anything? I guess all they had to do was provide a slightly different RF modulator accessory than in Europe in this case.
Noel's Retro Lab The thing is Texas Instruments had a factory in the late 70s until 1983 in Argentina. So it looks like they made the Argentinian version there. It was an original TI, and most curiously, it was beige! Not the usual black and metal from the rest of the world. Edit: look! www.sitiosargentina.com.ar/imagenes-2008/ti-99-6.jpg
@@damianvila there was a black and metal TI99/4A made in Argentina too, and a beige one with the letter "Ñ" where the semicolon key is. Greetings from Buenos Aires Daniel
Clear explanation of what you are doing and why. i actually feel like i could actually fault find some of my own hardware in the future, thank you. Great video as always.
I loved my TI-99/4A It was amazing. I had several cartridges and a book with 99 games for the TI-99/4A basic! I had SO much fun with it. I even cataloged all of my G.I. Joes with the "Personal Records Keeping" program database.
@@NoelsRetroLab Good Question. I had only two cartridges: Demon Attack and Surgeon. My friends lean me Parsec and Soccer too. I never had PEB, so everything was tape. I did have Extended BASIC and I really "mastered" it. It was much more intuitive and powerful than PEEKs and POKEs of the other computers of that era. Unfortunately TI policies didn't help like you said. The (artificial) lack of running machine language without TI assembler cartridge was a mistake. Every game was interpreted BASIC. The good thing was my own games were even better than "commercial" BASIC games!
The TI-99/4A was my very first computer when I was 9. I will always have a love and fondness for it. It IS a bizarre computer. There is a lot of fascinating history behind it. Those bottlenecks were actually needed...although, it hurt them in the long run. Without all of those bizarre "features", we wouldn't have had the TI today. It would have been another "me too" computer or never even released. Those that had the TI back in the day either loved it or hated it. I was one of the former. And when you love it, you always love it. It's like that crazy uncle everyone has that's not quite right in the head but you love him anyway because he's so much fun. The TI has an amazing scene over on AtariAge as well..very active. I personally have 6 TI's with some on the original box. I also have the VGA mod (F18A) which makes the image AMAZING on my modern VGA monitor. I have the PEB with 1MiB of RAM, a floppy drive and serial/parallel. It's truly a great computer despite it's "flaws". FYI, the sound chip in it is the same as the ColecoVision. The Coleco also shares the same VDP. :-)
I agree, it's a very quirky computer! Very different from anything else out there. I'm looking forward to exploring it more once I get it up and running. I already have the FinalGrom99, but I might have to look into the F18A as well. Thanks for the tip!
@@NoelsRetroLab Yes, the FinalGROM99 is a must. Along with some sort of 32K RAM expansion. With those two, you can play almost every game ever made and run hundreds of software packages.
Thanks so much for making this interesting series of videos about the TI-99. Faced with the task of repairing one with corrupted video last night for a customer, I found your videos and they gave me a pretty solid understanding of what was initially a totally unfamiliar computer (I repair Atari 8-bits most of the time). Turned out to be the video RAM in my case, which I socketed and replaced as the first order of business, and the machine thereafter worked perfectly. I put sockets under the SRAM too, just for good measure. :)
Fantastic useful Video, I can't wait to see the part2! --- it is very clear and well done! I am happy that the power supply did his job too ;) thanks for the mention and support the TI-99 Italian User Club ;)
@@NoelsRetroLab Hi Noel, Ciro is your main TI99 man in Europe and has helped me with a few things TI99/4A related in the past. You'll find allsorts of useful stuff on his website and is very helpful. If you need a software resource for your TI99/4A then try the TI99/4A Italian User Club Website, the TI99/4A Gamebase and the TI99/4A forum over on Atariage.
Just discovered the channel today from a post you made on Adrian's channel. Really liking it. You do a great job on explaining the troubleshooting steps. Hopefully one day you and adrian can do a collab on something.
The TI99 was my first "real" computer my brother and I had. Before that my parents bought us a Bally Astrocade in 1977 (was supposed to be expandable with keyboard, etc but the company fell through). Eventually we got the Peripheral expansion box bought used from someone in the local computer club, and it was my first computer with a modem (300 baud). In 1986 or so, I gave in to peer pressure and bought a C128.
@@NoelsRetroLab We started out with the adventure cartridge that was the interpreter for the Scott Adams adventure International games that came on cassette tape. We also had Parsec, a space shooter and Tunnels of Doom, a simple dungeon crawler, on carts. Once we got the PEB we got some of the Infocom gamed on disk. Infidel and Deadline are two I specifically recall.
Hello, new subscriber here. TI 99/4A was the first computer I ever used, even if it was just for games. Parsec and Moon Mine are the ones I remember the most, mainly because they were utilising the speech synthesiser. If I remember correctly, sometimes when the cartridge was not inserted property, when switching the computer on it would not beep but produce a constant "c major" sound. Maybe this is what the probe shows.
Hello and welcome! It looks like if the TI 99/4A doesn't properly initialize, it doesn't shut down the sound chip on time and it makes that sound. So by not inserting it correctly you were probably getting that same effect (maybe the switch was detecting a cartridge and trying to read from it but there was no connection with the cartridge and hence no data--not sure, just guessing). But it's interesting you can get that effect even with a working TI. Cool!
That was an *excellent* explanation of the TI99/4A architecture -- I learned a lot about it. I actually haven't seen one since my brother had one back in the day. Clearly, not the simplest computer to troubleshoot. I am looking forward to Part 2.
This came up in my feed and so happy it did! My 1st computer was the 99/4A. Remember the days of typing in basic code from a book. I also wondered "what if" had TI managed to some how get a full 16bit system on the market at that early date. Looking forward to the next episode! Subbed.
Lovely computer. Very unique. Are you in Europe? If so reach out to the Italian TI club. If you're in the US, I think they go for really cheap over there. In other parts of the world, I have no idea 😃
Thanks! Although eventually you have to come up with a theory and test it out by replacing chips. I'm still not 100% sure it was the ROMs, but it was a good candidate I guess. We'll see soon!
A friend of mine had one of these as a kid. It's a very nice machine but it hampered by the double-interpreted BASIC and the inability to run machine code without a cartridge (either games or Extended BASIC, neither of which he had). This was 1983, and I soon persuaded him to get his dad to buy him a Spectrum - a lesser machine in many ways but able to run a ton of great games out the box. I bought it off him a few years later for next to nothing and acquired an extended BASIC cart, and found it a very useful computer. Sadly I sold it on myself rather than keeping it - they go for decent money in the UK now.
Right. Once I get it working I should get into why the default BASIC is so crippled, and maybe run a few comparison programs against other computers. But, all in all, it's a very interesting machine!
I did a head to head video of the TI99/4A vs the VIC-20. It was a close competition for sure, but the incredibly slow BASIC and really strange keyboard knocked a few points off. I do love my TI99/4A, the graphics can be lovely even though the games are a bit sluggish
I'll have to check it out. I've heard the BASIC is super slow, so I was planning on doing some benchmarks against other systems as well. What's the problem with the keyboard? It seems pretty good to me. Is it the tiny Enter key?
@@NoelsRetroLab The BASIC is about 1/10th the speed of a VIC-20 according to my benchmarks. Try editing anything with the keyboard - it makes almost no sense. Yes I am spoiled by growing up with the awesome keyboard and full screen editing of the VIC-20 and C-64, but just trying to edit or delete characters is a nightmare on the TI-994A.
@@10MARC But according to Texas Instruments at 2:00 the VIC-20 has a calculator-style keyboard. Oh for shame, TI, misrepresenting the competition like that!
Interesting video. Interesting computer CPU with its 15 bit address bus and the barely used 16 bit data bus. Just think that in the seventies, Texas instruments could have made a System on Chip with the CPU, VDP and Sound processor, hopefully all running at top speed and being much faster than the sum of the parts in the TI99.
Oh absolutely! It could have been a great machine. It has some similarities to the Sinclair QL that way, even though they're very different machines overall.
Very nice video as usual, thanks! One suggestion: the graphic showing the pinout and the oscilloscope probe on the pin you're probing is great but sometimes the probe graphic element covers the label of the pin that is being tested. May I recommend a more subtle graphic element such a red dot or a red arrow, maybe pointing from inside the chip so the pin label is not being covered? Just my 2p idea, not sure if everybody would agree! Thank you!!
Hi Tony, thanks! That's actually a good point. It crossed my mind before but never thought much about it. Maybe I just need to put the probe graphic smaller and from the inside of the chip. It doesn't reflect what you would be doing, but it would allow you to see the labels like you said. Or maybe I can try moving the labels INTO the chip, although that might not be always possible. I'll think about it. Thanks!
Hi Noel. This is a very interesting video. I think at this stage, I would be tempted to remove that suspect ROM and fit a socket. Program a 27 series EPROM and temporarily connect that using a breadboard and jumper wires.
True, that would have been a way to be able to continue without stalling. But I had to break up the video anyway, so hopefully they'll get here soon and I can continue. But if they're delayed for whatever reason, I like that idea to be able to move forward. Thanks!
I was one of the loons that thought with enough decent software and the TI99-8, or Myarc Geneve 9640, we would avoid the techno junkpile. Didn't quite work out that way but in another sense it did. Folks are still kinda interested in 45 year old tech. TI stuff is kind of inundating YouTub right now. Don't forget the real start of the show, the TMS9918A video processor. Damn!!! It lives.. Donn Granros author of Old Dark Caves and Legends ! and 2 (with assist from Ed Johnson) Who else was handling memory transfer in Scratchpad Ram? I'm guessing no one.
I recall part of the reason this ended up the way it did is that TI was unsuccessful at creating the 8 bit chip they designed the TI99 to use, so they crippled their minicomputer CPU and stuffed it in there.
@@NoelsRetroLab Pretty sure. It's been a while since I read up on it. I also have a TI99/4A here :) . It was the first computer my family had, well not THIS specific one. Mostly just played games on it and messed with the speech synthesizer. Man, those metal shields in these things are kinda ridiculous. And the patch wires? Jeezus. TI could've done much better.
@@NoelsRetroLab yes, their 8 bit CPU was not reliable and was not ready yet, so they used their functional 16 bit CPU instead. That ended up crippling the system in some ways.
In many ways it's a console architecture. Small amount of onboard RAM, most of your software comes from ROM cart. It seems like the "graphics ROM" setup is essentially a tile generator.
Looking at 11:56 - one can see the hole on the top of the TI plug - that hole lines up on a standard US dual-outlet's plate screw hole... it's intended so that you pull the plate retention screw, hold the plate in place, put the adapter in, and then screw through that hole on the adapter into the box... the wall-wart is heavy enough that a slightly worn socket might not hold it.
@@NoelsRetroLab We used to live in Portsmouth, so the ZX81 was a natural choice, then when we moved back to Malta (DAD ex-navy), the TI-99 was being sold for just 55 sterling or so... a no-brainer. Funnily, I moved to another great love of mine, the MSX... I also purchased a C64 later on, then an Amiga, and then PC. The MSX are under appreciated, but they were darned good... I sadly (stupidly, sold mine!)... and prices these days are silly...
What a synchronicity: Adrian Black has a similar problem with a 99/4a board right now! th-cam.com/video/LsL10nlMDwA/w-d-xo.html Edit: I now saw that you already noticed this and commented over there...
Yeah, I saw his part 1 and was very curious, but part 2 sounded very, very much like this board. Although from his enigmatic answer I gather it'll be something unexpected and not at all what I have happen here. We'll see!
Who the F is that 1 Person that keeps downvoting these videos on this channel? Somebody really must've lost control over his life. Get rid of your Apple-Products man...
Great Video and great Channel, looking forward to part two, I am mostly into spectrum I never heard of a TI-99/4A untill a year or two ago, I found one cheap at a car boot sale, it's a 6 pin pal version it came with a psu but no modulator :-( I was going to get a component cable from retro computer shack, but he stopped selling them saying that the signal wasnt compatible with modern component capable TV's is this true? I also heard there is a 6 pin cable to scart available, from one of the European countries, does anyone know where I might obtain either a uk pal modulator or a scart version? Thanks in advance, I'd like to fire it up some time and see if it works :-)
Thanks! I'm hoping that part 2 will be all about getting that video out, once I get past the ROMs. I don't think a plain cable to Scart will work because it outputs a weird component video. You need either composite or RGB. You could just get one of the components and use it as a monochrome display without much trouble. But to get real color, it gets more complicated. I also have this waiting (instagram.com/p/CCP7Kjxnzs6/) which is little board to output an RGB signal. But soldering that is going to be challenging! :-)
@@NoelsRetroLab Thanks for your reply, sorry I was not clear about the scart cable, what I meant was, one of the European countries I think it may have been france, came with some kind of modulator box that had Scart output :-) not just a straight cable from the din socket.
lol that leaflet really stretches the truth! It says the other systems don't have "animated graphics" (perhaps they meant hardware sprites), and that the Spectrum doesn't have music (OK the Spectrum only has beeps but I definitely heard music on it.) Also that the other systems don't have assembly language!!
I got an EOL 4A (built, IIRC, in Italy) cheap from Dixons in 1982. I never did score the Extended Basic, and it ended up mostly as a console for playing Scott Adams Games (there was a cart, and the games , like Pyramid of Doom , came on tapes.) The basic was limited, but it was good enough to give me solid As on the computing element of my engineering qualification I found the TI strangely fragile - the bloody thing fried twice and had to be shipped back to Italy for repair (Dixons handled that end - it had to have been the best travelled computer I had until the 00s when I was hot desking with a laptop across northern europe) It would colour cycle the screen while making various strange squawking and buzzing noises.
There is actually a mod to get around the 8/16-bit bottleneck. I saw it like 20 or 25 years ago. The computer basically ran at almost twice the speed. I recall a hole in the top of the case with a small additional circuit board sticking out and lots of jumper wires.
Oh wow! That must have been a serious mod. I'm going to have to look for it. Also, I imagine it would require all new software to use it. Still, super interesting.
Yeah, I didn't have any remotely similar EPROMs in hand. I think Adrian does because he does a lot of C64 stuff and has some 24-pin ROMs. I'm even going to have to do a small fixup in the EPROMs I ordered because I didn´t get the exact same ones it has.
The 99/4a was supposed to be a cheap, entry level computer.... it was going to use an 8-bit processor, but development on that processor failed.... so to complete and try to get something to market.... TI used the 16-bit processor they had on hand, and tried to nerf it as much as they could. They pretty much lost money on every unit sold.... there was no killer software app to make the thing take off like the apple ii and c64... the 6502 in those made them a lot more profitable. I still have my ti99/4a from 1982, it still worked last i hooked it up about 10 years ago.... i bought a few spares and want to re-cap it before i use it again.... 12 volt clocks are normal.... be careful with your probes!
You think they tried to nerf it on purpose, or just fit it into the 8-bit architecture they already had? It still seems like a fun system, but I could see how it was outcompeted by Commodore and other systems.
@@NoelsRetroLab i think they were trying to compete in the 8-bit home computer space.... but their 8-bit processor wasn't ready on time... they wanted to get the thing out to go against apple and commodore... they didn't make it a 16-bit "killer" machine (left that win for IBM) because no one would've paid the 16-bit killer machine price for a home use "toy" It was ahead of its time in some respects, and rushed into production. TI fell victim to the sunk cost fallacy and paid for it. They had their own chip fab, so they weren't going to go to intel or mos for an off the shelf cpu (like apple and commodore, well, commodore was mos....) Due to politics between mos and ti..... It's a shit show basically.....
Ti's marketing is the main reason the computer was not as successful as it could be. TI was a military contractor at the time, so the 99/a4 is really built well structurally. I would like to design a TI 99/4B which uses 16-bit chips to see what it would have been like.
I can totally see that. But even if the computer had been full 16-bit, I wonder if that would have made it stand out enough over the competition at the time. Unless there was something really different visually, I doubt it could have competed with cheaper home computers.
I'm just reading the TI 99/4A handbook and it states the t the computer has 16 K-bytes of RAM that can be used by the user and a system memory that contains TI-Basic... is it a lie?
The CPU access to TMS9918 VRAM is as you describe for the GROMs… You set an address, and then you can read/write sequentially, with the address auto incrementing. Also, I didn’t know the SN76489 sound generator was originally called TMS9919!
Oh I didn't know that about the VRAM access! I'm going to have to look at some assembly for TI-99 games to see what kind of convolutions they had to go through to create any kind of interesting graphic displays.
@@NoelsRetroLab It was complicated to access the GROMs and VRAM indirectly so the main ROMs implement a virtual machine called GPL (Graphics Programming Language) that makes them seem like normal memory and all programs, including the Basic interpreter, are written in that. As you might imagine most of the native power of the 9900 is thrown away due to this. And having the user's Basic code live in VRAM not only makes it extra slow but also keeps you from being able to use the higher resolution mode of the 9918A (which the 9918 in the original TI99/4 didn't have anyway).
@@jecelassumpcaojr890 Right. I did read about that. It's kind of crazy how they managed to make a potentially awesome architecture and then botched it up so badly.
@@TheLemminkainen Or you can use a voltage converter like I did if you already have a US power supply. Although the Meanwell is always going to be better (and avoid the coffee warmer plate completely) :-)
From what I was always told the T99 was bottle-necked to make sure it would not compete with the larger and way more expensive computer systems TI was selling. BTW The main bottleneck are the GROMs, reading from it is slow and code stored there has to be copied to the 255 byte "rollout memory" before it can be executed. Mine has 32K CPU RAM added and a 16Mhz XTAL for a 4Mhz effective clock speed. This gives a dramatic increase in performance, more than half that of a 8Mhz 68000, leaving all 8 bit home computers of the time miles behind. Note that the TMS9000 only has a 64K address space for memory, the way the processor is designed the address- and data-bus have to be the same width to avoid dramatic performance loss. The TMS9900 also needs fast CPU RAM to work well. This made larger computers based on multiple of these processors relatively expensive. The TMS9900 did find a home in process control though. In theory it can handle a whopping 8 x 64K GIPO lines in a separate address space. Just add more of the IO buffer chips as you go along.
Interesting! I didn't know about that theory. How did you expand your RAM? On the board itself? That'd be really interesting. It can't be through a cartridge because it's only connected to the 8 bit bus.
@@NoelsRetroLab The RAM expansion was done in the late 80s. I must have a article about it somewhere in the attic, if I can find it it will be in Dutch though. But please read on. First note: for the RAM expansion you need a MiniMEM cartridge (assembly cartridge) or similar to load code into the expanded RAM. The best place to add something to the 16bit bus is the system ROMs. Here you have easy access to all data and most address lines. For the 2 additional address lines just run wires to the CPU pins. The RAM upgrade was quite basic, just piggy-back four 8K SRAM to the system ROMs and run a few wires to the main board. However, these days it would be easier to cook-up a small PCB. Use and use wire-wrap IC sockets to give the PCB pins that fit into the ROM sockets, the ROM themselves are placed in those wire-wrap sockets. No need for a scary piggy-back soldering job. And while we're making a PCB anyways. 8K RAM is not that common anymore. Why not take a much bigger SRAM and use a latching TTL buffer IC for bank-switching. A 5V capable 512k x16 bit SRAM goes for about $15. The PCB would be quite simple to make. All that's needed is a RAM and 3 or 4 TTL ICs for the addressing and bank switching. I would be willing to help with such a project.
@@ezion67 That would be awesome! It would be quite an interesting project. I'm concerned about the cartridge though. Would you always have to insert it to load that code, then take it out and pop in a different cartridge of something you want to use? Although I suppose that with something like the FinalGROM99 it could all be in the same place.
@@NoelsRetroLab Hi again. You raised a interesting and not easy to answer point there. I needed some time to figure things out. And yea sorry, this is a long one. The 16bit RAM expansion was usually used with the assembly language cartridge, cause that was the only way to develop programs that run from the 16 bit RAM. Alternatively a Extended Basic program could be used to load those programs from tape or disk into the 16 bit RAM. A custom cartridge like FinalGROM99 only needs to have a loader program on it that transfers data from the cartridge to the 16bit RAM. However when are things ever that easy. For programs written to run from the 16bit RAM expansion there is no problem. With programs written to run from GROM it gets complicated. The short answer is NO, but maybe, you sort of can. The GROMs are a form of background memory. When a program runs from GROM a small part is transferred to the 255 byte scratchpad RAM, then the next part is swapped in etc, etc. The question is: Who regulates the swapping. I could not find a definite answer to this question, so the next part is "informed speculation" from my part. The TI99 has no memory controller that can generate interrupts to trigger the OS to do a memory swap. The easiest way is to let the programs themselves do the whole memory control thing. This means: 1) At the end of a block of code, you call a routine that loads the next block. 2) Whenever you do a jump outside the current block, replace the jump statement with a call that loads the appropriate block of code. My fear is that this is the solution TI went for. If the above assumption is correct. A play from GROM to play from RAM code translator is possible, but also a lot hardcore coding work. You could write a program that cuts out those "get new code from GROM" calls. But... this only will work if the addressing used is relative. You would need to change absolute addresses to match the actual place in memory. Doable, but a lot of work. Probably the best option is to use a 16bit RAM with programs specially written to make use of it. And if you really want a program written for GROM to go full throttle, then dive in and hack the machine code to make it work with the 16bit RAM.
It was my understanding that the memory split came about because the TMS9900 could only support SRAM... and static RAM is expensive. Hence the video controller does double duty driving the cheaper 16k of DRAM. At the time 16/8 splits were not uncommon. Most peripherals saw little speed benefit to make them cost effective, so 8 bit devices on 16 bit "internal" processors was pretty normal. Just look at the 8088, a cut down version of the 8086 that used a 8 bit bus despite being 16 bit internal with a 20 bit data path.
Yeah, all RAM was expensive, but SRAM even more so. The video controller manages the video RAM, but just for video display though. I believe it was extra slow to access it from the CPU, so I don't expect that many games used it for anything other than display data (I could be wrong though if they were really starved for RAM).
@@NoelsRetroLab No, the VDP's memory was used as system memory, where the video chip was literally handling the DRAM refresh and as a middle-man between the CPU and memory. Yes, it was painfully slow but where else was non-ROM software supposed to be stored? In the 256 bytes of SRAM? Time intensive loops were often loaded into the SRAM from RAM to be executed in smaller chunks, but on the whole RAM based software and variable space was loaded via the video chip. So if you typed a program into BASIC, guess where it's stored... on those chips on the other side of the video controller.
@@jasonknight1085 Whoa! I hadn't looked into that part of the architecture yet, but that explains a lot of why I was reading that BASIC was so slow. You're right of course, and that makes the architecture even more crazy!
For our "friend" which had a similar problem it was the CPU... Unfortunately replacing it ended up pretty tricky and he shorted one of the pins to 12V and damaged one of the ROMS... It did work in the end but he had to use an EPROM in an adapter (like he had tried earlier when he thought the ROMs might be to blame) and, of course, a replacement CPU (I was going to say "new" but I doubt any "new" ones can be found ;-) )...
Yeah, I just saw Adrian's third part on that repair. Very interesting. I'm hoping it's not the CPU, but thinking about it and seeing how those lines are always high, it sounds very possible. We'll see soon.
WOW - who ever made that chart SCREW Atari over big time. Judging by how the Atari 400 came with 16K standard (originally it was 8K), and a couple of other things, I would say this chart probably was made in the early 1980's. Now, let's set the record straight. I will note right here that they used the cheaper Atari 400 instead of the Atari 800, hmmmm, let's go with the lesser Atari, it makes it easier to bad mouth. I'll note both machines..... Standard keyboard - 400 Mylar, 800 Yes Basic built in - no, but they came with a basic cartridge so I'll change that to a "it came with basic" Speech capacity - yes, programmers were able to get POKEY to output speech Expandable memory - both could get up to 48K and later 128K Pascal - Atari had one of the best Pascals. I took Pascal in college and my professor saw the doc's for Atari Pascal and noted that it was the first home computer pascal that followed the standard that was suppose to be followed. Word processing - of course there was a word processor. Actually, one thing I wish someone could have done was to put most of the WP on a cartridge and have an 8K program driving it. Why? Because one could put a 40K WP on an Atari cartridge and it would only take up 8K of memory leaving 32 for the document. Compare that to Applesoft that used 40K ram for the program and had 8K for the document. Logo - Yes Atari had Atari Logo Electronic spreadsheet - Yes Atari had VisiCalc. DOH! Nice job LYING!
It also says that the VIC-20 has a calculator-style keyboard, when it really has exactly the same standard mechanical keyboard as the C64. And the VIC-20's number of rows and columns is reversed. There are other mistakes, as well.
Mmm... it is odd that it requires 16 volts AC between 1 and 2, and 8 volts AC between 2 and 4. That sounds more like needing a two-phase, 8 volts AC, with 1 and 2 connected to a dual-secondary transformer, with each secondary outputting 8 volts AC, and 4, as ground, connected to the middle point of the secondary...
It's a pity that it's not possible to add a RAM expansion to the fast 16-bit bus (the official RAM pack attaches to the slow 8-bit bus). That would've made it a much more competitive and interesting machine!
Right. I also wonder how exactly that RAM vs the RAM in the RAM pack is exposed to the CPU (meaning, do you have to use a port to access it, or is it mapped to certain addresses...). I'm not even clear in whether the SRAM itself can be used by programs or it's purely for the CPU's internal use as kind of registers. I need to research all of this more.
@@NoelsRetroLab I was reading the other day that there's a chip that links the 16-bit and 8-bit buses; it converts each 16-bit access from the CPU to a sequential pair of 8-bit accesses to ICs connected to the 8-bit bus. So (if I have it right), from the CPU's perspective everything looks like 16-bit memory, but most things (including the 32kB RAM expansion) are going to be slow because the CPU will be frozen until both 8-bit accesses are complete.
La verdad es que el retraso en el correo ha duplicado la espera para cualquier pedido, sobretodo desde USA. Es muy decepcionante que empiezas un proyecto y casi mejor te metes en otro porque has de esperar demasiado. Luego te pasa que cuando te llegan las piezas no te acuerdas tan bien de todas las historias que pueda tener lo que estabas arreglando.Yo actualmente estoy con 3-4 cosas.
@@NoelsRetroLab En el caso de reparar un ordenador es incluso más difícil, porque aunque sepas que un chip está mal y lo pidas, después puede ser otro también el que falla, y lo has de volver a pedir, lo que en tiempo se convierte en una pesadilla.
Just received a TI99/4A and found your video. Thanks to you I was able to test the power supply and troubleshoot the computer. Much Thanks !
I had a TI 99/4A for a couple years in middle school and learned how to program using the Basic language. The system came with a useful workbook that explained the Basic programming language well. The problem was storage with only a tape recorder type date storage which I didn't like cause my school had Apple computers with disk drives which operated quite a bit quicker than a tape recorder. A few years later, I sold it and bought a Commodore 64 which did the job a bit better with a floppy drive and some nicer features.
TI 99/4As were really popular in Argentina in the 80s. Schools and High Schools who had a Computer Lab had them. It was the first computer I was able to get my hands on, so I have fond memories of it. Great video Noel!
Oh wow, I had no idea! Very interesting.
Noel's Retro Lab Yes. The Commodore 64 was also extremely popular, mainly because a local (licensed) version was produced: Drean Commodore 64. And in the Sinclair front, it was mostly Timex Sinclair 2068s. Then, the MSX explosion (with local producers too). Very peculiar. :D
Edit: I forgot to mention, the thing in Argentina was the curious PAL-N TV system. So most computers needed the video section reworked. The local Spectrum manufacturer was Czerweny.
Right! I read about that PAL system in some of the forums. So the TI-99 available wasn't a clone or anything? I guess all they had to do was provide a slightly different RF modulator accessory than in Europe in this case.
Noel's Retro Lab The thing is Texas Instruments had a factory in the late 70s until 1983 in Argentina. So it looks like they made the Argentinian version there. It was an original TI, and most curiously, it was beige! Not the usual black and metal from the rest of the world.
Edit: look! www.sitiosargentina.com.ar/imagenes-2008/ti-99-6.jpg
@@damianvila there was a black and metal TI99/4A made in Argentina too, and a beige one with the letter "Ñ" where the semicolon key is.
Greetings from Buenos Aires
Daniel
Clear explanation of what you are doing and why. i actually feel like i could actually fault find some of my own hardware in the future, thank you. Great video as always.
Thank you! I love hearing that. My goal is for people to learn and feel they can start doing this themselves.
I loved my TI-99/4A It was amazing. I had several cartridges and a book with 99 games for the TI-99/4A basic! I had SO much fun with it. I even cataloged all of my G.I. Joes with the "Personal Records Keeping" program database.
I really hope you can fix that, so I can learn. The TI-99/4A was my childhood computer and I have two of those now.
So can I! :-) Since it was your childhood computer, what are some of the best games I should play on it once I get it working?
@@NoelsRetroLab Good Question. I had only two cartridges: Demon Attack and Surgeon. My friends lean me Parsec and Soccer too. I never had PEB, so everything was tape. I did have Extended BASIC and I really "mastered" it. It was much more intuitive and powerful than PEEKs and POKEs of the other computers of that era. Unfortunately TI policies didn't help like you said. The (artificial) lack of running machine language without TI assembler cartridge was a mistake. Every game was interpreted BASIC. The good thing was my own games were even better than "commercial" BASIC games!
I have one too!! Looking forward to part 2!
The TI-99/4A was my very first computer when I was 9. I will always have a love and fondness for it. It IS a bizarre computer. There is a lot of fascinating history behind it. Those bottlenecks were actually needed...although, it hurt them in the long run. Without all of those bizarre "features", we wouldn't have had the TI today. It would have been another "me too" computer or never even released. Those that had the TI back in the day either loved it or hated it. I was one of the former. And when you love it, you always love it. It's like that crazy uncle everyone has that's not quite right in the head but you love him anyway because he's so much fun. The TI has an amazing scene over on AtariAge as well..very active. I personally have 6 TI's with some on the original box. I also have the VGA mod (F18A) which makes the image AMAZING on my modern VGA monitor. I have the PEB with 1MiB of RAM, a floppy drive and serial/parallel. It's truly a great computer despite it's "flaws". FYI, the sound chip in it is the same as the ColecoVision. The Coleco also shares the same VDP. :-)
I agree, it's a very quirky computer! Very different from anything else out there. I'm looking forward to exploring it more once I get it up and running. I already have the FinalGrom99, but I might have to look into the F18A as well. Thanks for the tip!
@@NoelsRetroLab Yes, the FinalGROM99 is a must. Along with some sort of 32K RAM expansion. With those two, you can play almost every game ever made and run hundreds of software packages.
I cut my programming teeth on the TI 99/4A computer. That computer is responsible for my career today.
Thanks so much for making this interesting series of videos about the TI-99. Faced with the task of repairing one with corrupted video last night for a customer, I found your videos and they gave me a pretty solid understanding of what was initially a totally unfamiliar computer (I repair Atari 8-bits most of the time). Turned out to be the video RAM in my case, which I socketed and replaced as the first order of business, and the machine thereafter worked perfectly. I put sockets under the SRAM too, just for good measure. :)
Fantastic useful Video, I can't wait to see the part2! --- it is very clear and well done!
I am happy that the power supply did his job too ;) thanks for the mention and support the TI-99 Italian User Club ;)
Thanks again!
@@NoelsRetroLab Hi Noel, Ciro is your main TI99 man in Europe and has helped me with a few things TI99/4A related in the past. You'll find allsorts of useful stuff on his website and is very helpful. If you need a software resource for your TI99/4A then try the TI99/4A Italian User Club Website, the TI99/4A Gamebase and the TI99/4A forum over on Atariage.
My first home computer! Long live the TI-99/4A. 😀
Just discovered the channel today from a post you made on Adrian's channel. Really liking it. You do a great job on explaining the troubleshooting steps. Hopefully one day you and adrian can do a collab on something.
Thanks! Adrian is one of the biggest influences on me starting this channel, so that would be awesome.
The TI99 was my first "real" computer my brother and I had. Before that my parents bought us a Bally Astrocade in 1977 (was supposed to be expandable with keyboard, etc but the company fell through). Eventually we got the Peripheral expansion box bought used from someone in the local computer club, and it was my first computer with a modem (300 baud). In 1986 or so, I gave in to peer pressure and bought a C128.
Mine too! Back inn1983 my late father showed me them. They were stock piled I think Macy’s for $49. Wow. The games were good though.
Do you remember any particularly good games you'd recommend to me? I'd love to test them out once I get the machine working.
@@NoelsRetroLab We started out with the adventure cartridge that was the interpreter for the Scott Adams adventure International games that came on cassette tape. We also had Parsec, a space shooter and Tunnels of Doom, a simple dungeon crawler, on carts. Once we got the PEB we got some of the Infocom gamed on disk. Infidel and Deadline are two I specifically recall.
Hello, new subscriber here. TI 99/4A was the first computer I ever used, even if it was just for games. Parsec and Moon Mine are the ones I remember the most, mainly because they were utilising the speech synthesiser. If I remember correctly, sometimes when the cartridge was not inserted property, when switching the computer on it would not beep but produce a constant "c major" sound. Maybe this is what the probe shows.
Hello and welcome! It looks like if the TI 99/4A doesn't properly initialize, it doesn't shut down the sound chip on time and it makes that sound. So by not inserting it correctly you were probably getting that same effect (maybe the switch was detecting a cartridge and trying to read from it but there was no connection with the cartridge and hence no data--not sure, just guessing). But it's interesting you can get that effect even with a working TI. Cool!
That was an *excellent* explanation of the TI99/4A architecture -- I learned a lot about it. I actually haven't seen one since my brother had one back in the day. Clearly, not the simplest computer to troubleshoot. I am looking forward to Part 2.
Thanks! Glad you liked it. I'm looking forward to Part 2 as well :-)
Great troubleshooting video! Looking forward to part 2.
Thanks!
This came up in my feed and so happy it did! My 1st computer was the 99/4A. Remember the days of typing in basic code from a book. I also wondered "what if" had TI managed to some how get a full 16bit system on the market at that early date. Looking forward to the next episode! Subbed.
Thanks! Glad you enjoyed it. For me it's completely new, but it's a cool and different system. I'm enjoying it so far!
BTW, that power board has more than one function. Not only does it provide DC power, it also keeps your coffee warm on the the TI.
:-)
I used to put my M&Ms in the "tray" to soften the chocolate but the candy shell was still crunchy. Mmmm, yummy.
yep, that's the first computer we had in our home. I learned BASIC on this computer.
This was my first computer back in the day. Would love to find one again (that I could afford!).
Lovely computer. Very unique. Are you in Europe? If so reach out to the Italian TI club. If you're in the US, I think they go for really cheap over there. In other parts of the world, I have no idea 😃
eBay may have what you want. Read closely!
Carefully!
It's a bit messy in there. :) Can't wait for part 2.
Yes, it is! :-)
I like how some knowledge saves you having to swap out chips until it works. Saves a lot of work, soldering and potentially introducing new faults.
Thanks! Although eventually you have to come up with a theory and test it out by replacing chips. I'm still not 100% sure it was the ROMs, but it was a good candidate I guess. We'll see soon!
A friend of mine had one of these as a kid. It's a very nice machine but it hampered by the double-interpreted BASIC and the inability to run machine code without a cartridge (either games or Extended BASIC, neither of which he had). This was 1983, and I soon persuaded him to get his dad to buy him a Spectrum - a lesser machine in many ways but able to run a ton of great games out the box.
I bought it off him a few years later for next to nothing and acquired an extended BASIC cart, and found it a very useful computer. Sadly I sold it on myself rather than keeping it - they go for decent money in the UK now.
Right. Once I get it working I should get into why the default BASIC is so crippled, and maybe run a few comparison programs against other computers. But, all in all, it's a very interesting machine!
As always very methodical troubleshooting. Looking forward to part b.
Just bought a TI99/4A here in the UK, Noel.
I did a head to head video of the TI99/4A vs the VIC-20. It was a close competition for sure, but the incredibly slow BASIC and really strange keyboard knocked a few points off. I do love my TI99/4A, the graphics can be lovely even though the games are a bit sluggish
I'll have to check it out. I've heard the BASIC is super slow, so I was planning on doing some benchmarks against other systems as well. What's the problem with the keyboard? It seems pretty good to me. Is it the tiny Enter key?
@@NoelsRetroLab The BASIC is about 1/10th the speed of a VIC-20 according to my benchmarks. Try editing anything with the keyboard - it makes almost no sense. Yes I am spoiled by growing up with the awesome keyboard and full screen editing of the VIC-20 and C-64, but just trying to edit or delete characters is a nightmare on the TI-994A.
@@10MARC But according to Texas Instruments at 2:00 the VIC-20 has a calculator-style keyboard. Oh for shame, TI, misrepresenting the competition like that!
Interesting video. Interesting computer CPU with its 15 bit address bus and the barely used 16 bit data bus. Just think that in the seventies, Texas instruments could have made a System on Chip with the CPU, VDP and Sound processor, hopefully all running at top speed and being much faster than the sum of the parts in the TI99.
Oh absolutely! It could have been a great machine. It has some similarities to the Sinclair QL that way, even though they're very different machines overall.
The Italian TI-99 club... This world rocks, there something for everyone, like, literally everyone ^-^
Very nice video as usual, thanks! One suggestion: the graphic showing the pinout and the oscilloscope probe on the pin you're probing is great but sometimes the probe graphic element covers the label of the pin that is being tested. May I recommend a more subtle graphic element such a red dot or a red arrow, maybe pointing from inside the chip so the pin label is not being covered? Just my 2p idea, not sure if everybody would agree! Thank you!!
Hi Tony, thanks! That's actually a good point. It crossed my mind before but never thought much about it. Maybe I just need to put the probe graphic smaller and from the inside of the chip. It doesn't reflect what you would be doing, but it would allow you to see the labels like you said. Or maybe I can try moving the labels INTO the chip, although that might not be always possible. I'll think about it. Thanks!
Hi Noel. This is a very interesting video. I think at this stage, I would be tempted to remove that suspect ROM and fit a socket. Program a 27 series EPROM and temporarily connect that using a breadboard and jumper wires.
True, that would have been a way to be able to continue without stalling. But I had to break up the video anyway, so hopefully they'll get here soon and I can continue. But if they're delayed for whatever reason, I like that idea to be able to move forward. Thanks!
I was one of the loons that thought with enough decent software and the TI99-8, or Myarc Geneve 9640, we would avoid the techno junkpile. Didn't quite work out that way but in another sense it did. Folks are still kinda interested in 45 year old tech. TI stuff is kind of inundating YouTub right now. Don't forget the real start of the show, the TMS9918A video processor. Damn!!! It lives.. Donn Granros author of Old Dark Caves and Legends ! and 2 (with assist from Ed Johnson) Who else was handling memory transfer in Scratchpad Ram? I'm guessing no one.
I recall part of the reason this ended up the way it did is that TI was unsuccessful at creating the 8 bit chip they designed the TI99 to use, so they crippled their minicomputer CPU and stuffed it in there.
Was it an 8-bit CPU they were trying to design? That would make a lot of sense.
@@NoelsRetroLab Pretty sure. It's been a while since I read up on it. I also have a TI99/4A here :) . It was the first computer my family had, well not THIS specific one. Mostly just played games on it and messed with the speech synthesizer. Man, those metal shields in these things are kinda ridiculous. And the patch wires? Jeezus. TI could've done much better.
@rezargamer Right? I read somewhere that you can actually replace the onboard GROMs with GROMs from a cartridge.
@@NoelsRetroLab yes, their 8 bit CPU was not reliable and was not ready yet, so they used their functional 16 bit CPU instead. That ended up crippling the system in some ways.
In many ways it's a console architecture. Small amount of onboard RAM, most of your software comes from ROM cart. It seems like the "graphics ROM" setup is essentially a tile generator.
This was recommended to me today and I'm glad it was! Great video, you very clearly describe what you're doing and why. Subscribed!
Thank you! Glad you like the approach.
Excelente VIDEO!
Your videos are great! Both the narrative and the editing are awesome.
Thank you very much! I really appreciate that!
Very good vídeo with a lot of interesting information.
Congratulations for your great job.
Thanks! Glad you enjoyed it!
Looking at 11:56 - one can see the hole on the top of the TI plug - that hole lines up on a standard US dual-outlet's plate screw hole... it's intended so that you pull the plate retention screw, hold the plate in place, put the adapter in, and then screw through that hole on the adapter into the box... the wall-wart is heavy enough that a slightly worn socket might not hold it.
I have one I need to repair. Thanks for the instructional video.
You're welcome. Good luck with it!
So damn weird. I've been looking at buying a TI-99/4A.... and THIS video pops up in my subs feed. I need to start making some tinfoil hats.
The first computer I got after the ZX81.... I still have mine and use it from time to time...
Interesting. Most people wiht a ZX81 move on to the ZX Spectrum, not the TI99. Did you like it back then? What did you move to afterwards? A C64?
@@NoelsRetroLab We used to live in Portsmouth, so the ZX81 was a natural choice, then when we moved back to Malta (DAD ex-navy), the TI-99 was being sold for just 55 sterling or so... a no-brainer.
Funnily, I moved to another great love of mine, the MSX... I also purchased a C64 later on, then an Amiga, and then PC.
The MSX are under appreciated, but they were darned good... I sadly (stupidly, sold mine!)... and prices these days are silly...
Your start-up screen looks like Tandy/Radio Shack! Thanks!
What a synchronicity: Adrian Black has a similar problem with a 99/4a board right now!
th-cam.com/video/LsL10nlMDwA/w-d-xo.html
Edit: I now saw that you already noticed this and commented over there...
Yeah, I saw his part 1 and was very curious, but part 2 sounded very, very much like this board. Although from his enigmatic answer I gather it'll be something unexpected and not at all what I have happen here. We'll see!
Who the F is that 1 Person that keeps downvoting these videos on this channel?
Somebody really must've lost control over his life. Get rid of your Apple-Products man...
Haha, you're not a true TH-cam channel without a few haters :-)
Great Video and great Channel, looking forward to part two, I am mostly into spectrum I never heard of a TI-99/4A untill a year or two ago, I found one cheap at a car boot sale, it's a 6 pin pal version it came with a psu but no modulator :-( I was going to get a component cable from retro computer shack, but he stopped selling them saying that the signal wasnt compatible with modern component capable TV's is this true? I also heard there is a 6 pin cable to scart available, from one of the European countries, does anyone know where I might obtain either a uk pal modulator or a scart version? Thanks in advance, I'd like to fire it up some time and see if it works :-)
Thanks! I'm hoping that part 2 will be all about getting that video out, once I get past the ROMs. I don't think a plain cable to Scart will work because it outputs a weird component video. You need either composite or RGB. You could just get one of the components and use it as a monochrome display without much trouble. But to get real color, it gets more complicated. I also have this waiting (instagram.com/p/CCP7Kjxnzs6/) which is little board to output an RGB signal. But soldering that is going to be challenging! :-)
@@NoelsRetroLab Thanks for your reply, sorry I was not clear about the scart cable, what I meant was, one of the European countries I think it may have been france, came with some kind of modulator box that had Scart output :-) not just a straight cable from the din socket.
Thanks for the video =)
My pleasure! 😃👍
Great channel
Thank you so much!
lol that leaflet really stretches the truth! It says the other systems don't have "animated graphics" (perhaps they meant hardware sprites), and that the Spectrum doesn't have music (OK the Spectrum only has beeps but I definitely heard music on it.) Also that the other systems don't have assembly language!!
the psu connector is the valuable part
I got an EOL 4A (built, IIRC, in Italy) cheap from Dixons in 1982. I never did score the Extended Basic, and it ended up mostly as a console for playing Scott Adams Games (there was a cart, and the games , like Pyramid of Doom , came on tapes.)
The basic was limited, but it was good enough to give me solid As on the computing element of my engineering qualification
I found the TI strangely fragile - the bloody thing fried twice and had to be shipped back to Italy for repair (Dixons handled that end - it had to have been the best travelled computer I had until the 00s when I was hot desking with a laptop across northern europe) It would colour cycle the screen while making various strange squawking and buzzing noises.
There is actually a mod to get around the 8/16-bit bottleneck. I saw it like 20 or 25 years ago. The computer basically ran at almost twice the speed. I recall a hole in the top of the case with a small additional circuit board sticking out and lots of jumper wires.
Oh wow! That must have been a serious mod. I'm going to have to look for it. Also, I imagine it would require all new software to use it. Still, super interesting.
The CPU was based on a mini they were selling according to Wikipedia. en.m.wikipedia.org/wiki/TI-990
Worked on TMS9900 systems as a young software engineer. It was a mean machine then...
Just watched the 3 part repair in Adrians Basement of a TI 99/4a and he had the same problem with the roms but used an adaptor.
Yeah, I didn't have any remotely similar EPROMs in hand. I think Adrian does because he does a lot of C64 stuff and has some 24-pin ROMs. I'm even going to have to do a small fixup in the EPROMs I ordered because I didn´t get the exact same ones it has.
The 99/4a was supposed to be a cheap, entry level computer.... it was going to use an 8-bit processor, but development on that processor failed.... so to complete and try to get something to market.... TI used the 16-bit processor they had on hand, and tried to nerf it as much as they could.
They pretty much lost money on every unit sold.... there was no killer software app to make the thing take off like the apple ii and c64... the 6502 in those made them a lot more profitable.
I still have my ti99/4a from 1982, it still worked last i hooked it up about 10 years ago.... i bought a few spares and want to re-cap it before i use it again....
12 volt clocks are normal.... be careful with your probes!
You think they tried to nerf it on purpose, or just fit it into the 8-bit architecture they already had? It still seems like a fun system, but I could see how it was outcompeted by Commodore and other systems.
@@NoelsRetroLab i think they were trying to compete in the 8-bit home computer space.... but their 8-bit processor wasn't ready on time... they wanted to get the thing out to go against apple and commodore... they didn't make it a 16-bit "killer" machine (left that win for IBM) because no one would've paid the 16-bit killer machine price for a home use "toy"
It was ahead of its time in some respects, and rushed into production.
TI fell victim to the sunk cost fallacy and paid for it.
They had their own chip fab, so they weren't going to go to intel or mos for an off the shelf cpu (like apple and commodore, well, commodore was mos....)
Due to politics between mos and ti.....
It's a shit show basically.....
and what was the idea about the RAM? 256bytes were enough for some cartridge based game machine, but this looks like a "real" computer...
Ti's marketing is the main reason the computer was not as successful as it could be. TI was a military contractor at the time, so the 99/a4 is really built well structurally. I would like to design a TI 99/4B which uses 16-bit chips to see what it would have been like.
I can totally see that. But even if the computer had been full 16-bit, I wonder if that would have made it stand out enough over the competition at the time. Unless there was something really different visually, I doubt it could have competed with cheaper home computers.
@@NoelsRetroLab I agree. I just want to see what could have been.
Has anyone heard of Typo 2 for the 99/4A? I picked it up over the weekend and I'm having trouble finding any info on it
It's a rather uncommon game released by Romox. It was released around 1983.
I'm just reading the TI 99/4A handbook and it states the t the computer has 16 K-bytes of RAM that can be used by the user and a system memory that contains TI-Basic... is it a lie?
The CPU access to TMS9918 VRAM is as you describe for the GROMs… You set an address, and then you can read/write sequentially, with the address auto incrementing. Also, I didn’t know the SN76489 sound generator was originally called TMS9919!
Oh I didn't know that about the VRAM access! I'm going to have to look at some assembly for TI-99 games to see what kind of convolutions they had to go through to create any kind of interesting graphic displays.
@@NoelsRetroLab It was complicated to access the GROMs and VRAM indirectly so the main ROMs implement a virtual machine called GPL (Graphics Programming Language) that makes them seem like normal memory and all programs, including the Basic interpreter, are written in that. As you might imagine most of the native power of the 9900 is thrown away due to this. And having the user's Basic code live in VRAM not only makes it extra slow but also keeps you from being able to use the higher resolution mode of the 9918A (which the 9918 in the original TI99/4 didn't have anyway).
@@jecelassumpcaojr890 Right. I did read about that. It's kind of crazy how they managed to make a potentially awesome architecture and then botched it up so badly.
GP25A13A-R1B is external model from Meanwell which is perfect for TI-99
Oh cool, I didn't know about that external model. Thanks!
@@NoelsRetroLab i am looking NTSC model cos it outputs CVBS as standard. I have to use external PSU here in europe :)
@@TheLemminkainen Or you can use a voltage converter like I did if you already have a US power supply. Although the Meanwell is always going to be better (and avoid the coffee warmer plate completely) :-)
Did he get is running?
Hi when is part 2 coming?
I'm recording it now, but I won't know for sure until it's working. Stay tuned... :-)
From what I was always told the T99 was bottle-necked to make sure it would not compete with the larger and way more expensive computer systems TI was selling. BTW The main bottleneck are the GROMs, reading from it is slow and code stored there has to be copied to the 255 byte "rollout memory" before it can be executed. Mine has 32K CPU RAM added and a 16Mhz XTAL for a 4Mhz effective clock speed. This gives a dramatic increase in performance, more than half that of a 8Mhz 68000, leaving all 8 bit home computers of the time miles behind.
Note that the TMS9000 only has a 64K address space for memory, the way the processor is designed the address- and data-bus have to be the same width to avoid dramatic performance loss. The TMS9900 also needs fast CPU RAM to work well. This made larger computers based on multiple of these processors relatively expensive. The TMS9900 did find a home in process control though. In theory it can handle a whopping 8 x 64K GIPO lines in a separate address space. Just add more of the IO buffer chips as you go along.
Interesting! I didn't know about that theory. How did you expand your RAM? On the board itself? That'd be really interesting. It can't be through a cartridge because it's only connected to the 8 bit bus.
@@NoelsRetroLab The RAM expansion was done in the late 80s. I must have a article about it somewhere in the attic, if I can find it it will be in Dutch though. But please read on.
First note: for the RAM expansion you need a MiniMEM cartridge (assembly cartridge) or similar to load code into the expanded RAM.
The best place to add something to the 16bit bus is the system ROMs. Here you have easy access to all data and most address lines. For the 2 additional address lines just run wires to the CPU pins.
The RAM upgrade was quite basic, just piggy-back four 8K SRAM to the system ROMs and run a few wires to the main board.
However, these days it would be easier to cook-up a small PCB. Use and use wire-wrap IC sockets to give the PCB pins that fit into the ROM sockets, the ROM themselves are placed in those wire-wrap sockets. No need for a scary piggy-back soldering job.
And while we're making a PCB anyways. 8K RAM is not that common anymore. Why not take a much bigger SRAM and use a latching TTL buffer IC for bank-switching.
A 5V capable 512k x16 bit SRAM goes for about $15.
The PCB would be quite simple to make. All that's needed is a RAM and 3 or 4 TTL ICs for the addressing and bank switching. I would be willing to help with such a project.
@@ezion67 That would be awesome! It would be quite an interesting project. I'm concerned about the cartridge though. Would you always have to insert it to load that code, then take it out and pop in a different cartridge of something you want to use? Although I suppose that with something like the FinalGROM99 it could all be in the same place.
@@NoelsRetroLab Hi again. You raised a interesting and not easy to answer point there. I needed some time to figure things out. And yea sorry, this is a long one.
The 16bit RAM expansion was usually used with the assembly language cartridge, cause that was the only way to develop programs that run from the 16 bit RAM. Alternatively a Extended Basic program could be used to load those programs from tape or disk into the 16 bit RAM. A custom cartridge like FinalGROM99 only needs to have a loader program on it that transfers data from the cartridge to the 16bit RAM.
However when are things ever that easy.
For programs written to run from the 16bit RAM expansion there is no problem. With programs written to run from GROM it gets complicated. The short answer is NO, but maybe, you sort of can.
The GROMs are a form of background memory. When a program runs from GROM a small part is transferred to the 255 byte scratchpad RAM, then the next part is swapped in etc, etc. The question is: Who regulates the swapping.
I could not find a definite answer to this question, so the next part is "informed speculation" from my part.
The TI99 has no memory controller that can generate interrupts to trigger the OS to do a memory swap. The easiest way is to let the programs themselves do the whole memory control thing. This means:
1) At the end of a block of code, you call a routine that loads the next block.
2) Whenever you do a jump outside the current block, replace the jump statement with a call that loads the appropriate block of code.
My fear is that this is the solution TI went for.
If the above assumption is correct.
A play from GROM to play from RAM code translator is possible, but also a lot hardcore coding work.
You could write a program that cuts out those "get new code from GROM" calls. But... this only will work if the addressing used is relative. You would need to change absolute addresses to match the actual place in memory. Doable, but a lot of work.
Probably the best option is to use a 16bit RAM with programs specially written to make use of it. And if you really want a program written for GROM to go full throttle, then dive in and hack the machine code to make it work with the 16bit RAM.
It was my understanding that the memory split came about because the TMS9900 could only support SRAM... and static RAM is expensive. Hence the video controller does double duty driving the cheaper 16k of DRAM.
At the time 16/8 splits were not uncommon. Most peripherals saw little speed benefit to make them cost effective, so 8 bit devices on 16 bit "internal" processors was pretty normal. Just look at the 8088, a cut down version of the 8086 that used a 8 bit bus despite being 16 bit internal with a 20 bit data path.
Yeah, all RAM was expensive, but SRAM even more so. The video controller manages the video RAM, but just for video display though. I believe it was extra slow to access it from the CPU, so I don't expect that many games used it for anything other than display data (I could be wrong though if they were really starved for RAM).
@@NoelsRetroLab No, the VDP's memory was used as system memory, where the video chip was literally handling the DRAM refresh and as a middle-man between the CPU and memory. Yes, it was painfully slow but where else was non-ROM software supposed to be stored? In the 256 bytes of SRAM?
Time intensive loops were often loaded into the SRAM from RAM to be executed in smaller chunks, but on the whole RAM based software and variable space was loaded via the video chip.
So if you typed a program into BASIC, guess where it's stored... on those chips on the other side of the video controller.
@@jasonknight1085 Whoa! I hadn't looked into that part of the architecture yet, but that explains a lot of why I was reading that BASIC was so slow. You're right of course, and that makes the architecture even more crazy!
I need to get clip meter leads like that.
For our "friend" which had a similar problem it was the CPU... Unfortunately replacing it ended up pretty tricky and he shorted one of the pins to 12V and damaged one of the ROMS... It did work in the end but he had to use an EPROM in an adapter (like he had tried earlier when he thought the ROMs might be to blame) and, of course, a replacement CPU (I was going to say "new" but I doubt any "new" ones can be found ;-) )...
Yeah, I just saw Adrian's third part on that repair. Very interesting. I'm hoping it's not the CPU, but thinking about it and seeing how those lines are always high, it sounds very possible. We'll see soon.
WOW - who ever made that chart SCREW Atari over big time. Judging by how the Atari 400 came with 16K standard (originally it was 8K), and a couple of other things, I would say this chart probably was made in the early 1980's. Now, let's set the record straight. I will note right here that they used the cheaper Atari 400 instead of the Atari 800, hmmmm, let's go with the lesser Atari, it makes it easier to bad mouth. I'll note both machines.....
Standard keyboard - 400 Mylar, 800 Yes
Basic built in - no, but they came with a basic cartridge so I'll change that to a "it came with basic"
Speech capacity - yes, programmers were able to get POKEY to output speech
Expandable memory - both could get up to 48K and later 128K
Pascal - Atari had one of the best Pascals. I took Pascal in college and my professor saw the doc's for Atari Pascal and noted that it was the first home computer pascal that followed the standard that was suppose to be followed.
Word processing - of course there was a word processor. Actually, one thing I wish someone could have done was to put most of the WP on a cartridge and have an 8K program driving it. Why? Because one could put a 40K WP on an Atari cartridge and it would only take up 8K of memory leaving 32 for the document. Compare that to Applesoft that used 40K ram for the program and had 8K for the document.
Logo - Yes Atari had Atari Logo
Electronic spreadsheet - Yes Atari had VisiCalc. DOH!
Nice job LYING!
Haha, seriously! I guess that's what most advertising tables do, but that one was particularly bad. Hilariously so actually 🤣
It also says that the VIC-20 has a calculator-style keyboard, when it really has exactly the same standard mechanical keyboard as the C64. And the VIC-20's number of rows and columns is reversed. There are other mistakes, as well.
I learnt to code on this machine. Nowadays I'm a .net architect 😅
Mmm... it is odd that it requires 16 volts AC between 1 and 2, and 8 volts AC between 2 and 4. That sounds more like needing a two-phase, 8 volts AC, with 1 and 2 connected to a dual-secondary transformer, with each secondary outputting 8 volts AC, and 4, as ground, connected to the middle point of the secondary...
It's a pity that it's not possible to add a RAM expansion to the fast 16-bit bus (the official RAM pack attaches to the slow 8-bit bus). That would've made it a much more competitive and interesting machine!
Right. I also wonder how exactly that RAM vs the RAM in the RAM pack is exposed to the CPU (meaning, do you have to use a port to access it, or is it mapped to certain addresses...). I'm not even clear in whether the SRAM itself can be used by programs or it's purely for the CPU's internal use as kind of registers. I need to research all of this more.
@@NoelsRetroLab I was reading the other day that there's a chip that links the 16-bit and 8-bit buses; it converts each 16-bit access from the CPU to a sequential pair of 8-bit accesses to ICs connected to the 8-bit bus. So (if I have it right), from the CPU's perspective everything looks like 16-bit memory, but most things (including the 32kB RAM expansion) are going to be slow because the CPU will be frozen until both 8-bit accesses are complete.
A short but good history of the TI99/4A: spectrum.ieee.org/tech-history/heroic-failures/the-texas-instruments-994-worlds-first-16bit-computer
Oh awesome! I'll definitely check it out tonight. Thanks!
TI logic:
Release a system with crappy keyboard
Update said system but only add 256 bytes of ram
Commodore Logic
Innovate and keep innovating
La verdad es que el retraso en el correo ha duplicado la espera para cualquier pedido, sobretodo desde USA. Es muy decepcionante que empiezas un proyecto y casi mejor te metes en otro porque has de esperar demasiado. Luego te pasa que cuando te llegan las piezas no te acuerdas tan bien de todas las historias que pueda tener lo que estabas arreglando.Yo actualmente estoy con 3-4 cosas.
Pues sí, así es :-( Por eso tengo como 3 o 4 proyectos empezados. A ver si los consigo acabar pronto.
@@NoelsRetroLab En el caso de reparar un ordenador es incluso más difícil, porque aunque sepas que un chip está mal y lo pidas, después puede ser otro también el que falla, y lo has de volver a pedir, lo que en tiempo se convierte en una pesadilla.
did it ever "run"? :D
Change the power supply, lose the coffee cup warmer!
Sorry, can't stand the lip smacking / mouth clicking