True, Plus there’s the opensource descendant GNUStep and resulting distro etolieos.com or just run gnustep on top of FreeBSD.
I’ll hold you to that 
but do have fun camping too.
1 Like
Something like that would be very handy to get software onto vintage machines and make usable boot disks.
1 Like
Exactly, Of course a 9pin com port to RS232 converter comes in handy too
can’t forget the db9 to db25 pin converter either:
http://www.amazon.com/StarTech-com-10-Feet-Cross-Wired-Serial/dp/B00066HL50/ref=pd_sim_147_6?ie=UTF8&dpID=41kuafzkcNL&dpSrc=sims&preST=AC_UL160_SR160%2C160&refRID=1YJP8PCJNMVAVJNVJX09
Found an NEC 8401a-LS too. Not sure yet if it still powers up.
3 Likes
Nice fine. I bet we could fix it up if it needs a little work.
I found a very in tact TI99/4A I know they’re a dime a dozen on ebay but it might be fun to play with.
2 Likes
Status update:
If you guys haven’t seen the post on XM Core or reddit. The bbs is back online after aws had a maintence cycle. As for meeting up with me at the space that’s a little hard as of late since my transmission went out last week. I’ll be swinging by again once the part comes in.
Also, I did acquire a few things for the ICM which I’'ll be bring by with me at my next visit.
Still working on burning mac toast images on my new laptop, kind of hard to do much at the moment since it is the only machine I have that has a cd-burner and doesn’t support Intel VT-x or AMD-V extensions (ie no vmware/virtualbox) and since it runs windows 10 [its a gaming laptop] most of the older mac classic support software doesn’t work. No worries though, I do have qemu running and did get linux to work there so its just a matter of getting older mac toast images burnt correctly.
New machines that are currently on bid for the ICM:
- Timex Sinclair 1000 (pair of them with manual, ram upgrades and mp3 player to upload programs)They are nice little 8bits. Bring it in if you like. All vintage computers that work and can be on display are welcome at this point especially if they meet the museum’s main goal of teaching the history of the personal computer and internet from an hands on prespective.
1 Like
I believe the TI 99/4a. was an early 16bit computer in the era of 8 bit machines.
Yes it was, but it was a bastardized machine. TI purposefully crippled it to keep it from being used in business to replace their more expensive products.
It has the TMS9900 CPU, which is 16 bit, with 2 8-bit ROMs and 2 6810 SRAMs giving it a whopping 128 words of 16-bit RAM. The 16K of DRAM is sitting on the TMS9918A video chip which sits across an 8/16 bit bus conversion/arbiter IC. This architecture means the CPU crosses the 16 bit bus to the 8 bit bus and goes through the TMS9918A to get access to the RAM. Crazy!
Add to that the 16 pin serial GROMs they used for BASIC and for cartridges and you have yet another source of slowness.
To get speed out of it you need to get a PEB (expansion bay) and put RAM cards in it to get faster RAM on the CPU bus.
3 Likes
I have a pile of the TIs here… and 3 PEBs plus disk, cartridge, and cassette software if anyone is looking to find items for their collection.
I even still have a factory sealed voice synthesizer.
3 Likes
1 Like
oh!!! I would like just one! 
How much TI 99/4A history do you want to hear? 
I worked in TI’s Home Computer Division (HCD) in Lubbock, starting on December 30, 1979, as a co-op student. When I started at HC there was no 99/4A, only the 99/4. I worked a year in HC, returned to school at the beginning of 1981, then went back to HC for a 2nd co-op phase in the summer of 1981. When I returned the 99/4A had received a makeover, resulting in the 99/4A. Here were the major changes.
- Keyboard was changed from ‘chiclets’ style to slightly cramped traditional sculpted style
- VDP (Video Display Processor) was changed from TMS9918 to TMS9918.
- Monitor ROM was changed with the following enhancements
- Hook for user interrupt routine which was called after console ROM vertical sync interrupt processing (sprite movement, sound list processing) was done.
- Search for program headers in cartridges ROM, eliminating the need for a GROM chip (which only TI manufactured). The 99/4 monitor ROM only searched cartridge GROMs for program headers.
Because program headers were valid in cartridge ROMs, anyone who could put an EPROM on a PCB could make and market program cartridges. This caused TI to lose control of what cartridge was released as well as any revenue Ti would receive had the cartridge publisher had to include at least one TI GROM in the cartridge. I suspect that adding this capability was done at the engineering level, with no visibility by TI HC marketing and management as to the implications. However, the previous comment is speculation on my part.
The inclusion of a user vertical sync interrupt hook was is what enabled simultaneous game play and speech in the PARSEC game cartridge.
The difference between the 9918 and 9918A VDP was that the A-rev VDP had an additional graphics mode, graphics mode 2, that allowed full screen bit mapping but with a restriction that there could only be two colors (out of 16) in every 8 horizontal pixel group. This limitation was a practical limitation of how Graphics II mode was implemented - essentially the screen was divided into top, middle, and bottom sections. Each section an independent copy of a Graphics mode I screen. From an IC design standpoint, this mod was relatively easy to make and it didn’t require any extra bandwidth from DRAM - in fact it utilized it close to the theoretical maximum.
It has the TMS9900 CPU, which is 16 bit, with 2 8-bit ROMs and 2 6810 SRAMs giving it a whopping 128 words of 16-bit RAM.
Yes, there were only two memories on the 16-bit bus: 4Kx16 console monitor ROM and 128x16 scratch pad RAM. All of the other memories (expansion port RAM and cartridge RAM/ROM) or memory mapped devices (sound, GROMs, VDP memory) were on the back side of a 16 bit-to-8-bit bus converter.
This had substantial performance implications, as access a 16-bit word on the 16-bit bus only required 2 clock cycles while accessing the a 16-bit word on the 8-bit bus required 6 clock cycles (4 clocks were added to facilitate the bus conversion). Compounding this was the fact that all TMS9900 memory access were 16-bit (to save 9900 microcode space), thus forcing all accesses on the 8-bit bus to be in pairs and take 6 clocks.
The improved performance of the scratch pad RAM on the 16-bit bus is used in PARSEC to speed up the scrolling of the bitmapped. This is achieved by loading the inner data moving/scrolling loop from cartridge ROM into the scratch pad RAM. The overhead of moving the relative small chunk of code is more than recovered by the increased execution speed.
One area that the 99/4A had a real speed advantage over other home computers of the day was in floating point math computations. There was a full floating point package implemented in the 16-bit ROM, including all of the trig and transcendental functions. TI BASIC would kick everyone’s butt at pure floating point number crunching. Unfortunately there was no integer data type, so for loops had to do floating point arithmetic to compute and compare the loop variable 
One of the things that killed graphics performance was the fact that the VDP memory was not directly in the 9900 address space. Instead it was on the ‘back’ side of the VDP and was accessed through 4 registers that were mapped into the 9900 memory space as shown below. To read a byte of video memory three VDP accesses were needed: Write two address bytes+data r/w flag to the address counter, then read data register. The address counter was autoincrementing after a data access, so data at successive address could be accessed by multiple sequential reads of the read data register. Any need to read data at a non-sequential address or to write data would require the reloading of the address counter and r/w data access flag. Organizing the video memory map so that video data could be accessed sequentially during game play is one of the keys to good performance in VDP systems.
- read status - check sprite coincidence and vertical blanking status
- write address -
- read data
- write data
The GROMs (Graphic ROMs) were actually PMOS memories that were originally designed for use in calculators. GROMs had the same address/data port interface as the VDP. The benefit of this is that they could be packaged in a 16-pin DIP as only 11 control signals were needed. The downside is that no assembly language could be stored in a GROM. GROMs held either data (lots of it for PARSEC) or interpreted code called GPL (Graphics Programming Language). In some sense one could view GPL as byte code executing on a VM, where the VM was in the console ROMs.
- 8-bit data bus
- address/data port select
- R/W
- chip select
What appears on the surface to be a very short-sided design technically by TI, one of the top tech companies at that time (and still today), isn’t necessarily so. Here’s why.
The 99/4 was originally supposed to use a TMS9985 chip. This was to the TMS9900 what the 8088 was to the 8086 - a 16 bit machine internally but with an 8-bit bus to allow low cost systems to be built. With an 8-bit bus only half the number of RAM and ROM (and related buffer) chips would be needed. The 9985 was also to have 256 bytes of internal RAM at the same address (0x8300) as the 99/4 scratchpad RAM. But the chip wasn’t ready when the 99/4 design was nearing production. TI couldn’t afford to have cartridges and peripherals that all needed 2x the number of memory chips. This would also have forced a redesign of the cartridge and expansion ports with higher pin count connectors to accommodate the 16-bit data bus.
In order to move forward and meet the schedule for the 99/4, they added the 16 bit-to-8-bit bus convertor logic and RAM to make the 99/4 look like it had a 9985. This also maintained software compatibility with the 9985, so that when the 9985 became available the existing motherboard could be replaced with the 9985-based board that would achieve substantial cost reduction. The 9985 was never released.
All of the above happened before I set foot inside of TI, as the 99/4 was introduced in October of 1979. I was told an apocryphal story of why the 9985 was never released. Supposedly the 9985 designer was a savant, a genius, but a different sort of personality that didn’t fit into TI’s cultural mold. The designer’s manager tried to make him fit the mold and he walked out the door with the 9985 unfinished. When others tried to pick up his work they couldn’t make heads nor tails of it - apparently his way of thinking was different enough that what work he had done was unusable by conventional thinkers. Again, this is an apocryphal story. But from what I observed in my time at TI I’d bet more that it happened that way than that it didn’t…
I didn’t mean for this to be tome. But it’s barely scratched the surface of the 99/4(A), both from a technical perspective and all of the fun and crazy things that happened.
My second co-op phase was probably the most fun and enjoyable time I’ve ever had at a job. All the TTL parts I ever wanted for projects (they were expensive in 1981). Got to team with Jim Dramis to create PARSEC, ported the line-by-line assembler from the TM990/189 University board to the Mini-Memory cartridge, wrote the ‘Lines’ eye-candy demo to show how to use Graphics II mode, wrote the disk protection and graphics routines for the Control Data Plato interpreter cartridge (spent 2-3 weeks in La Jolla for that). To top it off, us game programmers would get reimbursed for money we spent at the arcade - it was research, right - and we may have slipped a few beers into the reimbursement request.
6 Likes
Maybe I can teach a class on ‘TMS9900 assembly language programming with the TI 99/4A’
5 Likes
In 1983 myself, TI programmers Jim Dramis (PARSEC, Munch Man, Car Wars) and Garth Dollahite (TI Invaders) joined with two ex-TI HC manager/marketer types to form Sofmachine. It didn’t work out, as TI exited the home computer market a year after we formed and our business/marketing types didn’t raise any money. But we did manage to get a contract with Atarisoft to implement 3 games - Pole Position, Jungle Hunt, and Vanguard - on the TI 99/4A. Here’s our first check from Atari for $35,000 - in 1983!
!
8 Likes
Please do!! It would be great to learn about these systems.
Feel free to post something up on the calendar and cross link here in this tread.