Ksarul

Though there are also a very small number of 4M SAMS test cards in circulation now. That number is not likely to climb all that much though, as the price for the memory chips it uses have doubled in the last year, making it uneconomical to purchase them at nearly $50 per chip--especially since one chip in three is defective on delivery. That alone would increase board costs by about $150, as a number of the chip sellers are carefully timing their shipments to arrive beyond the date where refunds are possible (and COVID delays aren't helping either). All future SAMS boards will be built to take advantage of them if present, but the user will have to find their own upgraded memory chips if they desire that change. I had enough chips to build my test boards (and the rest of the good chips I have will be installed on new boards to test full functionality of each board as I build it, before replacing them with a pair of standard 512K chips for delivery). I plan to have about a dozen of the upgrade-ready boards complete before Christmas. . .assuming life cooperates. I don't plan to take orders for them until the boards are ready though. . .more to follow when they are ready.

In the case of the SAMS card, it was always intended as a replacement for the 32K, as the card is swapping memory banks into and out of that 32K memory space. It NEEDS that space to work in a way that is useful to the programmer and the user, as only a 32K window of memory is visible at any one point in time. The TI just doesn't have a large enough memory map to do it any other way. It does not prevent the use of RAM Disks either, as it works fine with the Horizon RAM Disk (any flavor), the SNUG HRD 16, or the Australian Quest 200. The CorComp, Myarc, Atronic, and Foundation cards were other solutions intended to replace the 32K, but as the system can only have one 32K solution in it at a time, they aren't possible to use in combination with a SAMS card. One note, some later Horizon boards also have an option to fill the 32K space--and that option needs to be disabled with a SAMS board installed.

Same guy who is trying to sell a 99/4 engineering prototype for $2,500. . .after it failed to sell for $1,500.

Clips "might" work. I say that only from the standpoint of long-term stability for something that is going to take a lot of daily punishment from the typist using the keyboard (and just general cracking as the frame ages--something seen a lot in the cases Myarc used for most of their circuit cards (made in the same molds TI used for their later plastic card cases, which also have this problem)). That was why I went with the full-rail idea locking into a slot on the upper frame to provide that support over the entire length of the keyboard circuit card. All ideas are good though. Whatever works best in practice is what is important here.

it is a backup camera monitor, Bill. They work surprisingly well. . .

It has to be split, as the circuit card fits into a set of slots on either half of the frame to ensure stability. A single frame piece would probably require addition of screws which might also change thickness of the frame and possibly impact the ability to fit into the Tutor case. Of course, one could also make snap-in rail pieces that replace one or more of the slots in there now, as that would allow board insertion and maintain the stability too.

I still have to read it. . .

Well, if we follow the logic you used here, the total file size is still too large for any cartridge boards out there except for the Yellow boards and the boards Tursi designed for Dragon's Lair. It will exceed the 512K limit of the other options. The other issue is in the program's coding--it was designed to load from disk, so it is completely unaware of bank switching. You'd have to go into the program and change the load routines for each scene to load that scene from the banks where it is stored and then continue execution where it left off. On a program of this size, that type of surgery isn't exactly trivial. . .

Thank you for all of the fine work you do on these, @tmop69. Your work shows that compiled BASIC really fulfills its potential now--and the work of @senior_falcon to fine tune the compiler itself is just priceless. Many thanks to both of you!

They respun my 512K Red board, making all of the jumpers into links on the bottom of the board. The other components were changed to surface mount and also installed on the bottom side. Their version makes it very difficult for the normal user to change the settings, but it is otherwise functionally identical to one of the earlier revisions of my Red board (which is not open source--only the 128K Black boards and the 16K/64K Guidry boards were released that way).

On memory chips: you can start with less than 16 chips, you just have to follow the instalation order listed in the manual. The bottom layer has to be filled before starting to fill the top layer (which can also start with less than 16 chips). Depending on the jumper settings, you can use 128K OR 512K chips, but all installed chips must be of the same capacity. Note that when using the 512K chips, each layer has a maximum of 8MB. Two layers is then too much for the CFG program to segment into drives. You could use the split board option to make a 16M board into two 8M RAM Disks--this has not been tested yet with more than one layer of chips (giving two 4MB configurable RAM Disks), so you will be in uncharted territory if you try it. The HRD 4000 also decodes AMA/B/C, as I used the same circuitry on the 4000B. I'd have to do some research on earlier schematics to see when that circuitry first arrived. There are other minor differences between the 4000 and the 4000B. Several pull-up resistors were added, the DSR RAM chip must be a 32K chip (the 8K option is gone), and the chip select is no longer controlled by the seriously suspect linkage through the center LED. The board also has the options to select between the various battery types baked into it. I'd have to pore over the older schematics to answer your other questions on earlier board revisions, @mizapf

One other note: as the console has a QI motherboard, that means that two signal pins on the cartridge port are not connected, and this affects some cartridge software (only a few cartridges used either of those signals, but the FinalGROM board is one that will definitely have trouble because of it).

@AwkwardPotato, Those October '78 schematics are definitely for a Dimension 4, as they pretty much match both of my Dimension 4 board sets. Both sets were manufactured in late 1978 or January 1979, so the schematic timing (to include the changes to a 99/4-like board) through that fall and the winter months make a lot of sense.

Not at all. The JEDEC file from the reverse-engineered PAL matches the TI equation set you posted. I saw no differences, so his new equation set is good.

Dont forget that there were two different variants of the 99/4 being produced in the late 1979/early 1980 time period: the one most people have seen and the one with the internal speaker and volume control slide switch. With luck, one set of schematics will be found in there for each type.

This matches the new JEDEC equations for the 32K card.

I strongly suspect that the 99/3B documents were part of the "legendary" references that have shown up in the literature now and then indicating that some engineers presented management a Z-80 variant for the 99/4 to reduce cost, as the Z-80 was a lot cheaper than a TMS9900. The anecdotal references said they were shot down because it wasn't a TI part. I suspect this schematic set is the first solid proof that such a beast actually existed. . .

Sometimes, the intent of the post doesn't come through they way it was intended to. Electronic messaging is problematic that way. I usually read my posts a couple of times to try and make sure that it really says what was intended in a way that makes sense to the reader. I don't always succeed either. The earliest demonstration of an UberGROM was in 2012, using the first board revision that exists only in the hands of the three developers. We fiddled with it for a few years to get it really stable and to develop some comprehensive documentation for it. We met our design goals--and the huge number of TI cartridges that never existed in physical format (but which do as UberGROM images now, thanks to the efforts of @acadiel, @Gazoo, and others), were actually the main driver for this board on my part. I am eternally grateful to @Tursi for the work he did to develop the software and to @acadielfor the comprehensive testing he did along the way. Both became my good friends as the project progressed, and I am really glad to have been able to spend in-person time with both of them at various Chicage Faires. Your intent was definitely good--and knowing what has been done with it may inspire others to do their own thing with it too.

Turtles has also been getting pictures of things I've been digging out as part of a rearrangement down in my basement lair. . .every time a new thing comes to light, he temporarily pillages it for a photo session.

And the label on Rabbit Trail is one that we recreated to replace a really gnarly one that came with the cartridge. . .

I'll try and go through all of these tomorrow. Yes--there are some interesting differences that make sense to be opposite (I think TI was taking a signal that was considered to be a normal low and inverting it to get a high input)

I have plenty of my own disks, so getting it to a floppy isn't an issue--and since I build lots of cartridge boards, if the disk also has the cartridge binary on it, you can be sure I'll make a test cartridge to play with (and I'll definitely want the final version of the cartridge to go with the three or so sets I want to buy).

I will need at least three of these--and the price is more than acceptable.

The SST Expanded BASIC Compiler is exactly what it says it is--a compiler for a subset of Extended BASIC. The weird Sideport adapter looks like a nice one-off to rotate the keyboard foot to a vertical position, probably to make if fit on a narrower desk without hanging over the edge. I suspect the cable is a monitor cable. . .

I've been able to download both in Firefox--and they both open with Acrobat Reader.