I was one of 6 design engineers on the TMS9918/28/29. I defined the memory interface, helped define all the video processing, including how the sprites would work, and did all the sprite logic and circuit design. The 9918 ran at a 5.4xxxMHz clock rate (U.S. color Burst of 3.58xxxMHz times 3/2 and used a 10.8xxx Oscillator). It was designed in 5-micron Ratioless NMOS. It was a very high clock rate, for that level of technology which necessitated a heat sync.
Hi Everyone. This time Retro Relix reviews and tests the 1981 TI-99/4A computer from Texas Instruments. Afterwards, we play some games in BASIC, from ROM Cartridges and SD Card images. Enjoy...
You didn't play Hunt The Wumpus! :( It's not a computer I look back on fondly, the slow basic and ludicrous peripheral system (the good keyboard and voice synthesizer didn't make up for it). But it's good to see the same sort of modern addons for this computer as there are for others. I told my "parents" to get a C64 or Apple II so software could be "shared". Nobody else had a TI99. They never grasped the concept, only looked at the price tag.
one year for my birthday, My dad bought me a TI99-4A. hearing the sound of the data loading from the tape drive brought back some memories. I so miss that sound! I used to do a lot of coding in TI Basic, though I originally learned to code on an Apple II E in High school.
My first home micro, I adored this, and it was the world to me back in 1981 I would like to say. I searched and searched and found a UK group that sold tape produced games that came from the States. It served me well for many years. It was the start of a my love of computers, and ironically, also the spurred me on to my first degree, and my career as a Computer Engineer. Thanks for the video it really bought back that Christmas many years ago.
Someone else gave a good overview of the extensible architecture. The console ROMs also knew how to look for subprograms by name in the cartridge ROMs and the system memory from a loaded library (some revisions even knew how to find multiple cartridges! but the hardware was never made to support it). This was how cartridges like MiniMem could extend TI BASIC, or how Terminal Emulator II could add real text-to-speech support. It was a really advanced design, strongly reminiscent of minicomputers at the time, not home computers. And the TI-99/8 would have been nearly fully backwards compatible having a very similar architecture and capable of supporting nearly 16Mb of RAM! Sadly, there were too many compromises and problems (not all of then technical) which robbed this console of so so much of its potential.
Love these videos, must take some time to slot together for our pleasure! Well done, takes us all back to those old days of fun. Now everything is so clinical
I went to a science/computer summer camp the year these came out. They made the Commodore and Apple offerings look antique. The only real competition were the Atari computers. I tried to get my parents to buy one after camp but we could not find them for sale? Ended up with an Atari 800. Those were on sale everywhere.
I only heard of TI-99/4A from this video (well aware of the many calculators of course). I also started my quest on an Atari 800, back in 1982. Action! language was great for me (sort of Pascal/C mix of a language in a ROM cartridge).
16 bit but just a 8 bit can be used ,as it had just a 8bit bus ,good video I have 2 TI99/4a and have used one of them powers on the one I given away last year..
Yeah. The architecture on the TI-99 was overly complex, particularly the device read/write bottleneck; however it did have some “high speed” RAM - all 256 bytes ! 🤣
This is my dad's first computer! I think the TI-99/4A had the potential to be a great computer, but unfortunately, it was just a failure on arrival. I would like to see what Texas Instruments could've done had they sticked around just a little longer in the computer industry.
The quality of visuals and particularly the exposition and narration made this an immensely satisfying video to watch. As much as I enjoy waffling in the retro-computing sector-I am trying to relive the past after all-your results-oriented approach was refreshing. While this may have been the first video I've seen on the channel, it will not be the last. Well done indeed.
Nice video! Thanks! Because the games on Cassette tapes that you showed in the video are missing and needs to be preserved in the TI community, could be possibile to have a dump of them, please?
@Ti99iuclt As requested, please find a cassette games upload update in the main description and a link here for your convenience. Enjoy! www.dropbox.com/scl/fo/edzr9ijb5vz44ng1qqkta/h?rlkey=vk7hfq76b7g71wtc9etzuqxlz&dl=0
TI BASIC is horribly slow because TI never intended for the Home Computer to have this powerful CPU. TI created a software language called GPL or Graphics Programming Language to use the capabilities of their 9918 VDP. They were working on a CPU designed to directly run GPL but the project was a failure. They had to get something together quickly so they did a bit of stripping down on their 9900 CPU. GPL machine code was a dead end but the console BASIC was written in GPL and there wasn't time before launch of the TI-99/4 (no A) to completely rewrite it for the 9900. So they wrote an interpreter. BASIC code gets interpreted to GPL then to 9900 code. Many of the cartridges produced by TI's programmers are also very slow running because they were written in GPL and have to go through the slow GPL to 9900 interpreter. IIRC some of those early TI cartridges are actually BASIC code in chips. TI Extended BASIC in a cartridge (and all 3rd party cartridge and disk BASICs) get interpreted from BASIC language to 9900 code and run a heck of a lot faster. Software written in assembly language and compiled to 9900 machine code is very fast on the TI-99/4A.
It's close but not exactly correct as I remember it as a designer of the TMS9918 and the TMS9995 that was to be used in the 99/2 and 99/8 (both of which were canceled before introduction. The Home Computer designers wanted a custom GPL CPU, but TI management nixed it. They were told to use the TMS9985, which will use the CPU core from the TMS9940 microcontroller (with embedded EPROM). The 9985 replaced the EPROM with 256 bytes of RAM. The Home Computer designer thought they would eventually get to design their own GPL CPU and thus keep with the GPL language and double interpretation (of Basic). The 9940 program was a disaster, and they never got it to work (due to commitments, a totally redesigned 9940A was eventually developed). The 9985 had 256 bytes of RAM on-boart with an 8-bit external memory interface. They split off the TMS9985 design when they were only part way through the 9940 design and copied most of the many bugs. The 9940/85 CPU was terribly designed. It has an 8-bit ALU and 16-bit registers, so any 16-bit operation took two CPU cycles. The 9940/85 was such a disaster that they considered using the 9980 (8-bit I/0 interface variant of the 9900) or the 9900. They went with the 9900 as it would be much faster. They more or less built a 9900 emulator of the 9985. The 9900 could access the 256 bytes of SRAM 16-bits wide, but then they built an 8-bit I/O interface with discrete TTL latches and logic. The plan was to cost-reduce with the 9985 when it arrived. The 9985 was such a mess that it was decided to design a whole new chip, first dubbed the 9985A and later named the TMS9995. I led the architecture of the 9995, which had a full 16-bit CPU (including the ALU) and 256 Bytes of SRAM with an 8-bit I/O. The 9995 running out of internal SRAM was about 3 or 4 times faster than the 9900.
Never seen those cassette games Spudz and Core! before as well as the some of the games shown in that pile of tapes, in fact they are not listed anywhere on the internet.
As requested, please find a cassette games upload update in the main description and a link here for your convenience. Enjoy! www.dropbox.com/scl/fo/edzr9ijb5vz44ng1qqkta/h?rlkey=vk7hfq76b7g71wtc9etzuqxlz&dl=0
@@RetroRelixRestorer TI made three versions of the VDP. NTSC with composite out, PAL with component out, PAL with composite out. Theoretically there's an NTSC version with component out but AFAIK one has never been found and the PAL version with composite out seems to have rarely been used. But now there' the FPGA F18A replacement with VGA out which works in place of the 9918A and its siblings.
I recall a mod at a user group meeting that seemed to double the speed. Anyone know anything about such a mod? The guy added a small board with some chips and jumper wires.
you want it complicated?? Yes !!! The memory is accessed via an I/O port (video processor) and the basic and other programs are executed via bytecode. On the other hand, the computer ran quickly. What he could do was observe if you had the memory expansion and the extedet basic. The processor only had the workspace register and the other registers were in the RAM. With the right architecture, multitasking would be possible.
The TI-99/4 and TI-99/4A were 'future proof' out of the box due to how their memory mapped peripheral support worked. The OS in the console knows nothing at all about any hardware that's not built into it. It knows the cassette recorder, joysticks, VDP, memory, cartridge port, a keyboard for the joystick port, and that's it. As for the joystick port keyboard, I'm pretty sure that's what the MBX used. TI never did, though it's detailed in the technical documents. Other peripherals all contain firmware called a Device Service Routine. Attach the speech synthesizer, disk controller, 32K RAM, etc. and it announces its presence via an address in the peripheral area of memory, which gets scanned at power-on. A DSR may have no user usable commands, it just sits there and works, like the 32K card. Others like the disk controller seamlessly add new commands like OLD and SAVE. Remove the peripheral and all the commands the DSR adds are gone like they were never there. That complete lack of trying to cram support for every possible peripheral into the console OS/firmware has made the TI-99 a very upgradeable computer. Pretty much any hardware one can code a DSR for can be connected to one of these computers. That's what made possible so many 3rd party disk controllers, including some with MFM and SCSI hard drive support, RAM Disks, bigger than 32K RAM cards, realtime clocks, the ForTI MIDI card with four of the TI sound chips, and even a USB + SmartMedia card. That one was released just in time for the death of the SmartMedia card format. :( ISTR there has even been an IDE hard drive controller. Other 1980's microcomputers attempted to have support for every peripheral the company could think of built into the computer, and thus severely limited their capability, including many peripherals never getting made, or even prototyped. There were a lot of them with a slot or connector labeled "For future use" - a future that never came for them, while TI left theirs open to a future this little computer's engineers could not imagine.
Nice video! Just curious, for the video, is there a reason some type of standard (component, or component to HDMI) cable wouldn't work? It's great you were able to get video with the Extron box, but hat seems like a $$$ solution.
What I can gather is the YPbPr to HDMI uses RGB, whilst the Extron defaulted to YUVi which appears to be different as can be seen here learn.microsoft.com/en-us/windows/win32/medfound/about-yuv-video Happy to be corrected though 👍
Besides the normal RF-modulators that come with the TI. It depends on the TI-99/4A version (USA have 5-pin DIN plug and you can use a normal DIN to component yellow-red-white cable) and has quite good video output, for European version 6-DIN pins, need the composite YPbPr cable (never got it work with a TV), but used a GBS8200 board to get it to work (there are cheap ones from China). Better is to have the F18A-board (replace the video chip in the TI) for VGA output (high quality). And maybe on OSSC (upscaler) works with a TI-99/4A, but I always used it with a Myarc Geneve 9640 (very sharp output). On my TI99 VIDEOS channel you can see how sharp the output is with OSSC - th-cam.com/video/KJyjaapf0Ag/w-d-xo.html
spent hours playing parsec i was so good at it i could play all day lol those were the days when life was simple , now technology is all garbage and no fun.
I remember waiting weeks and weeks for the long-promised Arcturus (ZAXXON clone) cartridge. I finally called to see what was taking so long to be told it was cancelled and not coming out. Pretty good for a clone, but certainly no ZAXXON. Also waited for a Super Extended BASIC cartridge which never materialized. SIGH.
If you look below that "8306" date, you will see a "8218" so it seems it's even older than that by almost a year. However, it was not uncommon for older chips to be put in newer boards so it's hard to tell when the actual board was made/assembled.
It would be newer than the latest of the original chips. Also by original I mean there might not be a way to know unless one is clearly out of step with the others by a lot
You are not quite right about why TI basic is so slow. It's not because of accessing the 16K of video memory but because unlike other basics on other computers they all were interpreted languages meaning the computer had to take the basic command and convert it to machine code and then do the task. On the TI the basic language was double Interpreted, once from basic to the TI GPL language then from TI GPL to TI-9900 machine code. This caused the computer to do 33% more work than any other computer running basic. But the basic is a wonderful basic to use and more accurate with math than almost all other basics.
I have two more than complete systems (thousands of dollars worth of add-ons) that someone gave me a couple of decades ago. I haven't managed to find space to set one up yet, but I kind of hope I eventually do. The only one I ever played with was a store display. One guy I knew liked to type in a small program he'd written that shifted the keys' values over by one. Rebooting got rid of it, of course, but lots of fun. Still, I'm more of an Amiga guy. Damn, boy, you don't have to describe everything you're doing, that's what the video part is for. I despise idiocy, so you go on that lovely "do not recommend" list.
Those games are a bit disappointing, the video-chip inside this computer is the same as in the MSX-1. So, 15 colours and transparent that happens to be just as black as black, and a resolution that was fine for the eighties. I saw better games on the MSX than on the Texas Instruments 99/4A. My impression is that whey did not even try to make something really good. Even the ZX Spectrum performed better and that computer did not even have a real video-chip!
That 16K is not system ram. It's graphics memory. There's only 256 bytes of system RAM. TI basic is slow because it is double interpreted, not because of the graphics memory.
12:19 I'm surprised their cable uses the old connector. It means you have to hack up the insides of the TI, and then the original power supply will fry the TI with AC if mistakenly connected. It's too bad they didn't just use a new connector with an adapter cable into the original one inside, avoiding any damage to original components.
@@Yeoman35 The modification involved removing the internal power supply and connecting the logic board power rails to the power connector on the back. That will certainly not accept AC.
Ah man, I owned a Texas for a very short period of time and Core was the only game I had. I have to admit it looks better than I remember it, would there be any chance of making this available to download please, this was the first ever home video game I ever played? Thanks in advance fella.
As requested, please find a cassette games upload update in the main description and a link here for your convenience. Enjoy! www.dropbox.com/scl/fo/edzr9ijb5vz44ng1qqkta/h?rlkey=vk7hfq76b7g71wtc9etzuqxlz&dl=0
Are the blue wires on the motherboard a mod, or did it come from TI like that? I almost got a TI99/4A. Back in the early 80s when I saw Toys 'R' Us advertising them for Christmas for $50. I ended up getting a C64 instead. I've found 2-3 TI99/4As in the trash, but haven't done anything with them. All of them were the all-plastic later models. These days, I don't even bother with my real C64 or Amigas.
So many errors in this video. First the 16 k is not system memory. It is video memory. There is only 256 bytes of system memory. Next when you opened it why did you lift it from the top instead of leaving it turned over and removing the bottom. The way you removed it could damage the bottom case. Next you removed the two screws on the heat sink which you don’t need to remove.
I can imagine the pain and copium of that thing's owners when they realized they've bought their "16-bit" micro without RAM, pitiful video hardware and overheating CPU.
9 minutes in and you've already translated the word Aluminium four times. Were you worried that Americans would be befuddled if you didn't do it every single time? Wait... What did that word mean again? If only I could remember!
WHAT?! The TI-99/4As were actually assembled/made in Italy? 😳😳😳 That's a first. Usually anything electronics is assembled/made in an Asian Country, especially back in the 70s and 80s.
Nice video. Technically, you have 128 words of System RAM (2x MC6810 SRAM chips) directly connected to the CPU. The 16K DRAM is slow not just because it's on the other side of the VDU chip, but because it's forcing the CPU to go across a 16 to 8 bit bus translator to the VDU to get data to/from those DRAM chips. Unfortunately, the Kernel ROMs have no code for supporting memory expansions, disk drives, serial ports, etc... even on the PEB. You must use a cartridge to enable the computer to see and use those expansions. Be very careful if you make your own video cables as there is power on one of the pins to power that external RF modulator.
I was one of 6 design engineers on the TMS9918/28/29. I defined the memory interface, helped define all the video processing, including how the sprites would work, and did all the sprite logic and circuit design. The 9918 ran at a 5.4xxxMHz clock rate (U.S. color Burst of 3.58xxxMHz times 3/2 and used a 10.8xxx Oscillator). It was designed in 5-micron Ratioless NMOS. It was a very high clock rate, for that level of technology which necessitated a heat sync.
Hi Everyone. This time Retro Relix reviews and tests the 1981 TI-99/4A computer from Texas Instruments. Afterwards, we play some games in BASIC, from ROM Cartridges and SD Card images. Enjoy...
You didn't play Hunt The Wumpus! :( It's not a computer I look back on fondly, the slow basic and ludicrous peripheral system (the good keyboard and voice synthesizer didn't make up for it). But it's good to see the same sort of modern addons for this computer as there are for others.
I told my "parents" to get a C64 or Apple II so software could be "shared". Nobody else had a TI99. They never grasped the concept, only looked at the price tag.
one year for my birthday, My dad bought me a TI99-4A. hearing the sound of the data loading from the tape drive brought back some memories. I so miss that sound! I used to do a lot of coding in TI Basic, though I originally learned to code on an Apple II E in High school.
My first home micro, I adored this, and it was the world to me back in 1981 I would like to say. I searched and searched and found a UK group that sold tape produced games that came from the States. It served me well for many years. It was the start of a my love of computers, and ironically, also the spurred me on to my first degree, and my career as a Computer Engineer. Thanks for the video it really bought back that Christmas many years ago.
You are very welcome 👍
Someone else gave a good overview of the extensible architecture. The console ROMs also knew how to look for subprograms by name in the cartridge ROMs and the system memory from a loaded library (some revisions even knew how to find multiple cartridges! but the hardware was never made to support it). This was how cartridges like MiniMem could extend TI BASIC, or how Terminal Emulator II could add real text-to-speech support. It was a really advanced design, strongly reminiscent of minicomputers at the time, not home computers. And the TI-99/8 would have been nearly fully backwards compatible having a very similar architecture and capable of supporting nearly 16Mb of RAM!
Sadly, there were too many compromises and problems (not all of then technical) which robbed this console of so so much of its potential.
Love these videos, must take some time to slot together for our pleasure! Well done, takes us all back to those old days of fun. Now everything is so clinical
I went to a science/computer summer camp the year these came out.
They made the Commodore and Apple offerings look antique. The only real competition were the Atari computers.
I tried to get my parents to buy one after camp but we could not find them for sale?
Ended up with an Atari 800. Those were on sale everywhere.
I only heard of TI-99/4A from this video (well aware of the many calculators of course).
I also started my quest on an Atari 800, back in 1982. Action! language was great for me (sort of Pascal/C mix of a language in a ROM cartridge).
Excellent video and work. Thanks for sharing
What a detailed video; thanks so much for this!
Thank you. I always try to capture the detail so others don’t make the same mistakes 🤔
I used to have one of those when I was a little boy.
16 bit but just a 8 bit can be used ,as it had just a 8bit bus ,good video I have 2 TI99/4a and have used one of them powers on the one I given away last year..
Yeah. The architecture on the TI-99 was overly complex, particularly the device read/write bottleneck; however it did have some “high speed” RAM - all 256 bytes ! 🤣
This is my dad's first computer! I think the TI-99/4A had the potential to be a great computer, but unfortunately, it was just a failure on arrival. I would like to see what Texas Instruments could've done had they sticked around just a little longer in the computer industry.
Another cracking video. The channel really deserves far more subs. Brilliant content.
The Delorean of 8 bit computers.
Ha ha ha, it certainly was - John DeLorean would have been proud 👍
Except that it's a 16 Bit machine, at least partially.
oh man i played parsec so much in my time!
I didn't know they made upgrades and new games, thanks for this video.
The quality of visuals and particularly the exposition and narration made this an immensely satisfying video to watch. As much as I enjoy waffling in the retro-computing sector-I am trying to relive the past after all-your results-oriented approach was refreshing.
While this may have been the first video I've seen on the channel, it will not be the last. Well done indeed.
In the 1970s 16-bit computers were like workstations.
Nice video! Thanks!
Because the games on Cassette tapes that you showed in the video are missing and needs to be preserved in the TI community, could be possibile to have a dump of them, please?
Yes, I’ll digitise them and provide a download link soon.
@Ti99iuclt As requested, please find a cassette games upload update in the main description and a link here for your convenience. Enjoy!
www.dropbox.com/scl/fo/edzr9ijb5vz44ng1qqkta/h?rlkey=vk7hfq76b7g71wtc9etzuqxlz&dl=0
@@RetroRelixRestorer thank you! I would like to exchange some messages with you in PM, Would that be possible, please?
TI BASIC is horribly slow because TI never intended for the Home Computer to have this powerful CPU. TI created a software language called GPL or Graphics Programming Language to use the capabilities of their 9918 VDP. They were working on a CPU designed to directly run GPL but the project was a failure. They had to get something together quickly so they did a bit of stripping down on their 9900 CPU.
GPL machine code was a dead end but the console BASIC was written in GPL and there wasn't time before launch of the TI-99/4 (no A) to completely rewrite it for the 9900. So they wrote an interpreter. BASIC code gets interpreted to GPL then to 9900 code. Many of the cartridges produced by TI's programmers are also very slow running because they were written in GPL and have to go through the slow GPL to 9900 interpreter. IIRC some of those early TI cartridges are actually BASIC code in chips.
TI Extended BASIC in a cartridge (and all 3rd party cartridge and disk BASICs) get interpreted from BASIC language to 9900 code and run a heck of a lot faster.
Software written in assembly language and compiled to 9900 machine code is very fast on the TI-99/4A.
There are now compilers available, then these programs run quite fast (there is someone on TI99 atariage) who converted many Basic programs already
It's close but not exactly correct as I remember it as a designer of the TMS9918 and the TMS9995 that was to be used in the 99/2 and 99/8 (both of which were canceled before introduction.
The Home Computer designers wanted a custom GPL CPU, but TI management nixed it. They were told to use the TMS9985, which will use the CPU core from the TMS9940 microcontroller (with embedded EPROM). The 9985 replaced the EPROM with 256 bytes of RAM. The Home Computer designer thought they would eventually get to design their own GPL CPU and thus keep with the GPL language and double interpretation (of Basic).
The 9940 program was a disaster, and they never got it to work (due to commitments, a totally redesigned 9940A was eventually developed). The 9985 had 256 bytes of RAM on-boart with an 8-bit external memory interface. They split off the TMS9985 design when they were only part way through the 9940 design and copied most of the many bugs. The 9940/85 CPU was terribly designed. It has an 8-bit ALU and 16-bit registers, so any 16-bit operation took two CPU cycles.
The 9940/85 was such a disaster that they considered using the 9980 (8-bit I/0 interface variant of the 9900) or the 9900. They went with the 9900 as it would be much faster. They more or less built a 9900 emulator of the 9985. The 9900 could access the 256 bytes of SRAM 16-bits wide, but then they built an 8-bit I/O interface with discrete TTL latches and logic. The plan was to cost-reduce with the 9985 when it arrived.
The 9985 was such a mess that it was decided to design a whole new chip, first dubbed the 9985A and later named the TMS9995. I led the architecture of the 9995, which had a full 16-bit CPU (including the ALU) and 256 Bytes of SRAM with an 8-bit I/O. The 9995 running out of internal SRAM was about 3 or 4 times faster than the 9900.
Never seen those cassette games Spudz and Core! before as well as the some of the games shown in that pile of tapes, in fact they are not listed anywhere on the internet.
Maybe I should save as a WAV file with a link?
@@RetroRelixRestorer Wow, if you could do that it would be awesome!
@@_.OX._ Leave it with me 👍
As requested, please find a cassette games upload update in the main description and a link here for your convenience. Enjoy!
www.dropbox.com/scl/fo/edzr9ijb5vz44ng1qqkta/h?rlkey=vk7hfq76b7g71wtc9etzuqxlz&dl=0
Doesn't your European version of the TI-99/4A have the TMS9929A video chip? Thanks for the video, too!
You know what, I think you’re probably right; my source was for the US version. Didn’t want to remove that heat sink given its workload. 👍
@@RetroRelixRestorer TI made three versions of the VDP. NTSC with composite out, PAL with component out, PAL with composite out. Theoretically there's an NTSC version with component out but AFAIK one has never been found and the PAL version with composite out seems to have rarely been used.
But now there' the FPGA F18A replacement with VGA out which works in place of the 9918A and its siblings.
I recall a mod at a user group meeting that seemed to double the speed. Anyone know anything about such a mod? The guy added a small board with some chips and jumper wires.
you want it complicated?? Yes !!!
The memory is accessed via an I/O port (video processor) and the basic and other programs are executed via bytecode.
On the other hand, the computer ran quickly. What he could do was observe if you had the memory expansion and the extedet basic.
The processor only had the workspace register and the other registers were in the RAM. With the right architecture, multitasking would be possible.
The TI-99/4 and TI-99/4A were 'future proof' out of the box due to how their memory mapped peripheral support worked. The OS in the console knows nothing at all about any hardware that's not built into it. It knows the cassette recorder, joysticks, VDP, memory, cartridge port, a keyboard for the joystick port, and that's it. As for the joystick port keyboard, I'm pretty sure that's what the MBX used. TI never did, though it's detailed in the technical documents.
Other peripherals all contain firmware called a Device Service Routine. Attach the speech synthesizer, disk controller, 32K RAM, etc. and it announces its presence via an address in the peripheral area of memory, which gets scanned at power-on. A DSR may have no user usable commands, it just sits there and works, like the 32K card. Others like the disk controller seamlessly add new commands like OLD and SAVE. Remove the peripheral and all the commands the DSR adds are gone like they were never there.
That complete lack of trying to cram support for every possible peripheral into the console OS/firmware has made the TI-99 a very upgradeable computer. Pretty much any hardware one can code a DSR for can be connected to one of these computers. That's what made possible so many 3rd party disk controllers, including some with MFM and SCSI hard drive support, RAM Disks, bigger than 32K RAM cards, realtime clocks, the ForTI MIDI card with four of the TI sound chips, and even a USB + SmartMedia card. That one was released just in time for the death of the SmartMedia card format. :( ISTR there has even been an IDE hard drive controller.
Other 1980's microcomputers attempted to have support for every peripheral the company could think of built into the computer, and thus severely limited their capability, including many peripherals never getting made, or even prototyped. There were a lot of them with a slot or connector labeled "For future use" - a future that never came for them, while TI left theirs open to a future this little computer's engineers could not imagine.
Thanks for adding this detail 👍
Nice video! Just curious, for the video, is there a reason some type of standard (component, or component to HDMI) cable wouldn't work? It's great you were able to get video with the Extron box, but hat seems like a $$$ solution.
What I can gather is the YPbPr to HDMI uses RGB, whilst the Extron defaulted to YUVi which appears to be different as can be seen here learn.microsoft.com/en-us/windows/win32/medfound/about-yuv-video
Happy to be corrected though 👍
Besides the normal RF-modulators that come with the TI. It depends on the TI-99/4A version (USA have 5-pin DIN plug and you can use a normal DIN to component yellow-red-white cable) and has quite good video output, for European version 6-DIN pins, need the composite YPbPr cable (never got it work with a TV), but used a GBS8200 board to get it to work (there are cheap ones from China). Better is to have the F18A-board (replace the video chip in the TI) for VGA output (high quality). And maybe on OSSC (upscaler) works with a TI-99/4A, but I always used it with a Myarc Geneve 9640 (very sharp output). On my TI99 VIDEOS channel you can see how sharp the output is with OSSC - th-cam.com/video/KJyjaapf0Ag/w-d-xo.html
spent hours playing parsec i was so good at it i could play all day lol those were the days when life was simple , now technology is all garbage and no fun.
I remember waiting weeks and weeks for the long-promised Arcturus (ZAXXON clone) cartridge. I finally called to see what was taking so long to be told it was cancelled and not coming out. Pretty good for a clone, but certainly no ZAXXON. Also waited for a Super Extended BASIC cartridge which never materialized. SIGH.
If you look below that "8306" date, you will see a "8218" so it seems it's even older than that by almost a year. However, it was not uncommon for older chips to be put in newer boards so it's hard to tell when the actual board was made/assembled.
It would be newer than the latest of the original chips.
Also by original I mean there might not be a way to know unless one is clearly out of step with the others by a lot
Woooaah i didn’t know euro ti99 systems did output component,i tout it only did output composite video.
You are not quite right about why TI basic is so slow. It's not because of accessing the 16K of video memory but because unlike other basics on other computers they all were interpreted languages meaning the computer had to take the basic command and convert it to machine code and then do the task. On the TI the basic language was double Interpreted, once from basic to the TI GPL language then from TI GPL to TI-9900 machine code. This caused the computer to do 33% more work than any other computer running basic. But the basic is a wonderful basic to use and more accurate with math than almost all other basics.
2:54 that isn't the system RAM but the video RAM, 3:45 that is the system RAM
I have two more than complete systems (thousands of dollars worth of add-ons) that someone gave me a couple of decades ago. I haven't managed to find space to set one up yet, but I kind of hope I eventually do. The only one I ever played with was a store display. One guy I knew liked to type in a small program he'd written that shifted the keys' values over by one. Rebooting got rid of it, of course, but lots of fun. Still, I'm more of an Amiga guy.
Damn, boy, you don't have to describe everything you're doing, that's what the video part is for. I despise idiocy, so you go on that lovely "do not recommend" list.
Those games are a bit disappointing, the video-chip inside this computer is the same as in the MSX-1. So, 15 colours and transparent that happens to be just as black as black, and a resolution that was fine for the eighties. I saw better games on the MSX than on the Texas Instruments 99/4A. My impression is that whey did not even try to make something really good. Even the ZX Spectrum performed better and that computer did not even have a real video-chip!
That 16K is not system ram. It's graphics memory. There's only 256 bytes of system RAM. TI basic is slow because it is double interpreted, not because of the graphics memory.
12:19 I'm surprised their cable uses the old connector. It means you have to hack up the insides of the TI, and then the original power supply will fry the TI with AC if mistakenly connected. It's too bad they didn't just use a new connector with an adapter cable into the original one inside, avoiding any damage to original components.
It is quite likely that the power supply will accept DC or AC as input.
@@Yeoman35 The modification involved removing the internal power supply and connecting the logic board power rails to the power connector on the back. That will certainly not accept AC.
I got one.
Ah man, I owned a Texas for a very short period of time and Core was the only game I had. I have to admit it looks better than I remember it, would there be any chance of making this available to download please, this was the first ever home video game I ever played? Thanks in advance fella.
Hi, yes - I’ll upload and send you a link soon.
Bless you sir, many thanks!
As requested, please find a cassette games upload update in the main description and a link here for your convenience. Enjoy!
www.dropbox.com/scl/fo/edzr9ijb5vz44ng1qqkta/h?rlkey=vk7hfq76b7g71wtc9etzuqxlz&dl=0
You're a gent and a scholar and as I thank you I'll be sure to mention this video and your channel on mine.
Can I ask you sir, how did you manage to get these wav files loaded, was it using a phone or something into the original hardware?
Are the blue wires on the motherboard a mod, or did it come from TI like that?
I almost got a TI99/4A. Back in the early 80s when I saw Toys 'R' Us advertising them for Christmas for $50. I ended up getting a C64 instead.
I've found 2-3 TI99/4As in the trash, but haven't done anything with them. All of them were the all-plastic later models. These days, I don't even bother with my real C64 or Amigas.
From what I can see, the majority of motherboards had the blue “bodge wires” probably to fix a trace design issue.
16bit memory map killed it instantly.
So many errors in this video. First the 16 k is not system memory. It is video memory. There is only 256 bytes of system memory. Next when you opened it why did you lift it from the top instead of leaving it turned over and removing the bottom. The way you removed it could damage the bottom case. Next you removed the two screws on the heat sink which you don’t need to remove.
I can imagine the pain and copium of that thing's owners when they realized they've bought their "16-bit" micro without RAM, pitiful video hardware and overheating CPU.
9 minutes in and you've already translated the word Aluminium four times. Were you worried that Americans would be befuddled if you didn't do it every single time? Wait... What did that word mean again? If only I could remember!
WHAT?! The TI-99/4As were actually assembled/made in Italy? 😳😳😳
That's a first. Usually anything electronics is assembled/made in an Asian Country, especially back in the 70s and 80s.
Nice video. Technically, you have 128 words of System RAM (2x MC6810 SRAM chips) directly connected to the CPU. The 16K DRAM is slow not just because it's on the other side of the VDU chip, but because it's forcing the CPU to go across a 16 to 8 bit bus translator to the VDU to get data to/from those DRAM chips. Unfortunately, the Kernel ROMs have no code for supporting memory expansions, disk drives, serial ports, etc... even on the PEB. You must use a cartridge to enable the computer to see and use those expansions.
Be very careful if you make your own video cables as there is power on one of the pins to power that external RF modulator.