Heres a thought, is there a modern compatable chip with on chip large memory (the 256 colour 512 pixel mode types) and more advanced features? You could then possibly use the extra memory banked as an large bank switched memory with extra graphics upgrade daughter board for your TI?
@@Lantertronics is it fully compatable and simpler to program, or simpler to interface. If simpler to program, users could start with the old compatible modes, and work up exploring new features as they learn?
I have a prototype Hexbus video controller for the TI Compact Computer 40 that uses a TMP9118 and two 4416s. On the other hand, I have a TI-99/4 that uses a TMX9918 and eight 4116s. I wonder if the TI folks can get away with relaxing their 9918 with a 9118 and getting rid of their bad 4116s and jumping two 4416s in.
I'd start by replacing the 4116's since they're still available NOS from Jameco. If the 4116s hit the point that they're only available from dodgy eBay sellers, then then the hack in a 9118 route would have more appeal
With disclaimer, I could say it would seem to make sense to upload a video showing both a setup with the 4464 and the alternative 4416 _(on the TMS9118 but maybe also the TMS9918),_ and I say that because you are basing your rational partly on availability of chips _(not just fiddling with tristate)._ The tms9918 can use up to 192Kilobytes _(not talking kilobits here)_ video-RAM. Is that maximum the same for the tms9118? 8:46 Even though you mention the TMS9918, are you sure the TMS9118 can _(even though it does not have to)_ use the (256K) 4464 chips _(i.e. the 64K x 4bit)?_ If so, how many 4464 chips can be used to get the maximum RAM _(presumably the 192KBytes)_ for it? Just in theory. I'm thinking 12. But 8 would do, I assume _(even though for 32Kbytes, you'd only need two chips for your 8bit, I presume)._ My comment has no hate in it and I do no harm. I am not appalled or afraid, boasting or envying or complaining... Just saying. Psalms23: Giving thanks and praise to the Lord and peace and love.
Pretty low spec. What about something like used in the last max, Sega master system or the Sega Genisis. Plus, the advanced PC chips of the age were from NEC in Japan, the tech was used by IBM. There was a Motorola chip that is similar to the Amigas, that could work on 8 bit/16 bit/32 but bus. It was to be used on the Coco 3, but they went with a cheaper option. It seems to have disappeared at the time, as it is thought it was an Amiga technology based design through a transfer agreement, and there was litigation flying around about things, between Tramiel, Atari and Commodore. Maybe it is still out there. However, the rare Sinclair Spectrum Loki chipset went to Asia, as did the Microbee 68000 based design from memory (superior to the Amiga). There were a few other such Asian headed chips. The RCA/Philips digital disc system (forget the name) may have been one of them. You could even get a super cheap chip with coprocessor, an arm CPU chipset. But please, the older TI stuff was so bad and flashing.
The chip used in the Sega master system is based on 9918, and not generally available. The Genesis chip is also not generally available. The purpose of this project is to focus on technology that would have been available in the late 1980s and uses chips that are readily available.
@@Lantertronics So, the Sega chips were customs or exclusive. It's a wonder why no chip manufacturers made equivalents, as they were desparately needed by MSX? The ultimate simple chipset of this period, is the Acorn Riscos machines chipset unfortunately.
@@Lantertronics ok. I thought there was another source. We are fortunate to get the limited improvement we did anyway. 7800 chipset is out of the question. That was very nice, despite being underutilized. The Japanese had some very popular non MSX machines. The Chinese now do stuff. It could be done of those chips are still readily available and superior. Such a shambles the home computer market was. Little cutting edge after Commodore came in. Companies willing to do less, like the court computer 1 and 2, and Sinclairs. Tandy didn't even bother to ask a 4 colour pallet, 16 cor mode, or 512 pixel wide text mode, which would have made a world of difference to the cc1 and cc2 and mc10. The Sinclair's needed that bit more like that too, and 16 colour pallette. I wanted to like the Coco, but green on green two colour mode. The Nintendo chip. It was an outside job. If you can correct the poor colours (they do this to existing machines), then it's sort of ok. There is likely compatable chips on the market, maybe with ability to modify the pallet.
@@Waynesification I have had students work on the Atari 7800; they managed to port some source code for original 7800 games that was recovered from floppy disks found in a dumpster and get it to compile in a modern assembler:
Of course someone could implement all of this and much more in an FPGA, but that's not the point of the project. I'm trying to restricting myself to pre-90s technology. And more to the point... it might be "easier" for *someone else,* but I don't know Verilog or VHDL and have never deployed any code to an FPGA. It would take me much longer to learn those tools and to implement all of this than it was for me to figure out this memory issue and lay out and stuff and solder the PCB. And... even if I could do that, it wouldn't be as much fun. ;)
Great information for this new TI-99/4A user and fan.
Great video. I'm building another 65C02 computer that will hopefully use the TMS9918. I live near Chattanooga so it's great to hear from a "local".
Thanks!
I haven't gotten mine working yet. It think I may need to try different values for one of the oscillator capacitor on the TMS9918.
Heres a thought, is there a modern compatable chip with on chip large memory (the 256 colour 512 pixel mode types) and more advanced features? You could then possibly use the extra memory banked as an large bank switched memory with extra graphics upgrade daughter board for your TI?
There's the Yamaha V9938 and V9958. I have some V9958s., but I wanted to start with the simpler 9918 chip.
@@Lantertronics is it fully compatable and simpler to program, or simpler to interface. If simpler to program, users could start with the old compatible modes, and work up exploring new features as they learn?
Any ideas what the 4164 chip variants you see on aliexpres mean? Like -15NL, -20NL etc
No idea...
Likely speeds.
It's usually the access time, in nano seconds.
I have a prototype Hexbus video controller for the TI Compact Computer 40 that uses a TMP9118 and two 4416s. On the other hand, I have a TI-99/4 that uses a TMX9918 and eight 4116s. I wonder if the TI folks can get away with relaxing their 9918 with a 9118 and getting rid of their bad 4116s and jumping two 4416s in.
Wow, I've never even heard of these TMP and TMX variants. And google is coming up short too. Wild!
It's feasible, I think would take a good deal of hacking and wire jumping on the PCB though. It would make be nervous to try it.
I'd start by replacing the 4116's since they're still available NOS from Jameco. If the 4116s hit the point that they're only available from dodgy eBay sellers, then then the hack in a 9118 route would have more appeal
With disclaimer, I could say it would seem to make sense to upload a video showing both a setup with the 4464 and the alternative 4416 _(on the TMS9118 but maybe also the TMS9918),_ and I say that because you are basing your rational partly on availability of chips _(not just fiddling with tristate)._
The tms9918 can use up to 192Kilobytes _(not talking kilobits here)_ video-RAM. Is that maximum the same for the tms9118?
8:46 Even though you mention the TMS9918, are you sure the TMS9118 can _(even though it does not have to)_ use the (256K) 4464 chips _(i.e. the 64K x 4bit)?_ If so, how many 4464 chips can be used to get the maximum RAM _(presumably the 192KBytes)_ for it? Just in theory.
I'm thinking 12. But 8 would do, I assume _(even though for 32Kbytes, you'd only need two chips for your 8bit, I presume)._
My comment has no hate in it and I do no harm. I am not appalled or afraid, boasting or envying or complaining... Just saying. Psalms23: Giving thanks and praise to the Lord and peace and love.
And the TMS9918 vs TMS9918A vs the TMX9918 and the TMP9918 ;)
It never ends. ;)
Pretty low spec. What about something like used in the last max, Sega master system or the Sega Genisis. Plus, the advanced PC chips of the age were from NEC in Japan, the tech was used by IBM.
There was a Motorola chip that is similar to the Amigas, that could work on 8 bit/16 bit/32 but bus. It was to be used on the Coco 3, but they went with a cheaper option. It seems to have disappeared at the time, as it is thought it was an Amiga technology based design through a transfer agreement, and there was litigation flying around about things, between Tramiel, Atari and Commodore. Maybe it is still out there. However, the rare Sinclair Spectrum Loki chipset went to Asia, as did the Microbee 68000 based design from memory (superior to the Amiga). There were a few other such Asian headed chips. The RCA/Philips digital disc system (forget the name) may have been one of them. You could even get a super cheap chip with coprocessor, an arm CPU chipset. But please, the older TI stuff was so bad and flashing.
The chip used in the Sega master system is based on 9918, and not generally available. The Genesis chip is also not generally available.
The purpose of this project is to focus on technology that would have been available in the late 1980s and uses chips that are readily available.
@@Lantertronics So, the Sega chips were customs or exclusive. It's a wonder why no chip manufacturers made equivalents, as they were desparately needed by MSX?
The ultimate simple chipset of this period, is the Acorn Riscos machines chipset unfortunately.
@@Waynesification The MSX chips were either the TI or Yamaha chips.
@@Lantertronics ok. I thought there was another source. We are fortunate to get the limited improvement we did anyway. 7800 chipset is out of the question. That was very nice, despite being underutilized.
The Japanese had some very popular non MSX machines. The Chinese now do stuff. It could be done of those chips are still readily available and superior. Such a shambles the home computer market was. Little cutting edge after Commodore came in. Companies willing to do less, like the court computer 1 and 2, and Sinclairs. Tandy didn't even bother to ask a 4 colour pallet, 16 cor mode, or 512 pixel wide text mode, which would have made a world of difference to the cc1 and cc2 and mc10. The Sinclair's needed that bit more like that too, and 16 colour pallette. I wanted to like the Coco, but green on green two colour mode.
The Nintendo chip. It was an outside job. If you can correct the poor colours (they do this to existing machines), then it's sort of ok. There is likely compatable chips on the market, maybe with ability to modify the pallet.
@@Waynesification I have had students work on the Atari 7800; they managed to port some source code for original 7800 games that was recovered from floppy disks found in a dumpster and get it to compile in a modern assembler:
Frankly at that point, it might be easier to just implement the whole shebang in a small, low cost fpga.
Of course someone could implement all of this and much more in an FPGA, but that's not the point of the project. I'm trying to restricting myself to pre-90s technology.
And more to the point... it might be "easier" for *someone else,* but I don't know Verilog or VHDL and have never deployed any code to an FPGA. It would take me much longer to learn those tools and to implement all of this than it was for me to figure out this memory issue and lay out and stuff and solder the PCB.
And... even if I could do that, it wouldn't be as much fun. ;)