Stuart

Is there really on the 99000? Couldn't see any mention of it in the (preliminary) data manual. Have you found some details somewhere?

Looks like 10 Ohms. Which is consistent with the pictures you've posted - a 10 Ohm resistor has colour bands brown-black-black-gold, and those last three bands seem to correspond with what's left on the resistor.

But that creates a problem that I think someone raised in speccery's thread: some programs manipulate register values by reading/writing direct to where the workspace is stored in RAM (it's a perfectly valid programming technique). So the contents of a cache would lose sync with the register values in the workspace in RAM. Unless you had some sort of system to detect when this happens.

[Pedantic mode on!] It's actually a 3 MHz processor, same speed (but more efficient) than the TMS9900. The 12 MHz crystal is internally divided to produce a 4-phase, 3 MHz clock. [Pedantic mode off!]

No, you need to measure it out of circuit. Switch off the PEB, pull the resistor wires off those two pins you added, and measure across the resistor with it not connected to anything.

So to put names to faces, that Matt on the left, Cory on the right. Who is "Team Red" in the middle?

Yep, they grey lines are existing circuit, and the black lines the new stuff, with the connections being made either by piggybacking or by separate wires. There's one pin on one of the original ICs to cut as well and a wire soldered to it - look for the red 'X'.

The 16-bit 32K mod I have (http://www.stuartconner.me.uk/ti/ti.htm#32k_memory_expansion) is similar to that on the Mainbyte site but needs fewer chips, already incorporates the wait state defeat, and includes a switch (mount it on the back panel) to disable it so it doesn't affect any 32K on the sideport. The instructions aren't as comprehensive as the Mainbyte site though so you'll need to be experienced enough to be able to identify the ICs and connection points on the PCB.

There's one here! http://www.stuartconner.me.uk/ti/ti.htm#i2c_interface

Which design did you use for the upgrade? Something you put together yourself, or is it on the net somewhere?

I2C you can do through the joystick port - http://www.stuartconner.me.uk/ti/ti.htm#i2c_interface. The issue though is that pretty much every type of device will need a software driver to send/receive data specific to that device. Straightforward enough for the Arduino/Raspberry crowd where the population is in the millions, and the computers have got a gig or so of RAM.

Not sure if you've found it or not, and might be of interest - there are ports of LSX Unix and V6 Unix running on the TMS9995-based Mini-Cortex (http://www.stuartconner.me.uk/mini_cortex/mini_cortex.htm). Requires a little more hardware than the TI-99 provides unfortunately.

Cortex BASIC uses CPU RAM/ROM and needs 32K extended memory. Integer values in the range -32768 to +32767 are stored internally in an integer format, otherwise floating point format is used (you can't specifically define a variable as an integer). Cortex BASIC is **very roughly** 1.5 times the speed of XB. There's a little bit of benchmarking in this thread: http://atariage.com/forums/topic/245962-cortex-basic-in-80-column/?p=3376857. There's a bit about benchmarks using the 'top hat graph' here as well: http://atariage.com/forums/topic/215138-bitmap-mode/?p=3183359. All very non scientific as two BASICs can't run the exact same BASIC program ...

Really? I think you'll find that NOP just assembles to JMP $+2.

An excellent point! I done the disassembly quite a few years ago so memory is a bit vague. The disassembler I used runs under XB, so what I think I done was a partial dump of the MiniMem RAM, swapped to the XB cartridge and reloaded that dump at >2A00. So the addresses in the listing have an offset of >4600 from the actual RAM addresses in the MiniMem. So for example address >2B8E in the listing + offset >4600 = actual address >718E in the MiniMem.

Two files attached which may be of help/interest. "MiniMem LBLA Listing.doc" is a disassembly of the MiniMem with the LBLA loaded. Most of the code is the LBLA, but there is also some LINES and other stuff in there. It's a 'raw' listing - some data is interpreted as instructions, and this sometimes makes later instructions be misinterpreted, and so on. The first column in the listing is the memory address, the second column is the data at that address, and the third column is the disassembled instructions. The first and second columns are correct even where the disassembled instructions are out of sync. "MiniMem LBLA & Utilities Cartridge.a99" is a text file for my LBLA + TIBUG + Disassembler cartridge. It contains commented LBLA source, but has a number of changes from the original plus a bit of code shuffling so it can run from cartridge. All labels in the code in the form "Axxxx" relate to memory addresses in the first doc. The first bit of the LBLA, starting at >2B18, is workspaces and data. The main entry points are >2BA6 for "NEW" (label LBLANEW in my code) and >2BAC for "OLD" (label LBLAOLD in my code). There's then some shuffled code at A2BB0, but my code re-syncs with the listing at A2C3A. I've got various handwritten notes on how it works as well. Give me a shout if confused. MiniMem LBLA Listing.doc MiniMem LBLA & Utilities Cartridge.a99

Only if it is oxygen-free copper.

I would have thought it was the other way round - the shielding is there so it passed the stringent FCC tests for emissions coming *from* the TI. If you can use it and you don't get interference on your radio or TV, I don't think you'll have any problems.

Just a note that all my spare PCBs have now gone. I'm happy to order some more if a couple of people are interested, but they're likely to be more expensive, depending on whether or not those nice people at Sitopway include any spares in the order. Stuart.

There no top (component side) shield on the QI motherboard ...

Yep. If you go from handy +5V and 0V points on the motherboard to the pins on the connector on the end of your ribbon cable, which plugs into the cartridge. Might make no different at all but probably worth a go.

Might be worth trying to add some separate, thicker power and ground cables to supplement the connections you've already got through the ribbon cable. Should be easy enough to remove again if they make no difference.

Hmm. I reckon there's got to be a switch or sensor associated with that small board. In the first two photos above there are some silver pins coming out the opposite side to where the wires come in - can you see where they go to? With the photo of the removed motor, there are two bits of black plastic on the motor shaft. Are they both free to rotate on the shaft, or locked to the shaft? The piece nearest to the motor - is that completely round or does it have anything projecting from it? Is there anything set into the plastic, like a small magnet?

I think it has to be OLD "DSK1.CATALOG" - uppercase and with quotes.

Can you do a couple of close-up pics of that small board from various angles? What's on it? Do you have a multimeter available?