Intellivision TutorVision found w/games

[AnnaWu]

You'll need the Tutorvision exec (aka wbexec); there is a tutorvision grom file but it is optional as the standard Intellivision grom file would work.

 

Rename the Tutorvision system files exec.bin and grom.bin and place in the jzintv\rom folder. Jzintv recognizes the tutorvision exec and makes the necessary changes when running the emulator. Using jzintv is no different than with Intellivision cartridge rom files. Here is an example command.

 

"path\to\jzintv.exe" -z3 %1

 

Where %1 represents the filepathname of the cartridge rom file and -z3 is for 1024x768 resolution. If you're using windows you can place the above command in a .bat file and then drag and drop cartridge rom files onto it.

 

Thank you so much for your great help, mr_me-.

 

Anna

[Jeffrey Bouchard]

I'll have to try these. Hopefully I can find them and use them soon.

[Walter Ives]

On 4/11/2017 at 11:03 PM, intvnut said:

Attached is a PDF version that's a little friendlier to read.

Here are some comments on the observations you made in that document.

8-bit memory.  Upper byte always reads as 0.

For 8-bit memory upper byte of bus is driven by STIC1a, lower byte by 2114's. That's why CSRAML- and CSRAMH- are separate.

UNK_1 and UNK_2 are signals I don’t have better names for.

They were supposed to be /BUSRQ and /INTRM. How about using /BUSRQX and /INTRMX, where the X stands for broken?

Quite why this circuit is necessary remains a mystery. I suspect the ASIC was intended to generate /BUSRQ and /INTRM, but generated them incorrectly.

Suspicion confirmed. Perfection is difficult, and it was expected that the STIC1a would need a revision. With the 1983 drop in sales there were plenty of original chipsets in inventory and the revision was deferred. When Valeski ran out of parts a few years later the economics were different and the original Toshiba design team had moved on, so he found it cheaper to use this fix than to rev the part.

One curiosity:  RO9580 is twice the necessary capacity for holding the Intellivision 1 EXEC.  What’s in the other 4K, or is it just disabled?

Hint: How big is the Intellivision I EXEC?

The STIC1A displays the 160th column, unlike the AY-3-8910-1.  This may have some implications for MOB-vs-border collision detection.

This was a deliberate change: APh just couldn't bear the thought of not displaying that last pixel. Damn trouble-causing perfectionists, they were. Mattel went along with it.

The STIC1A PSG implementation mostly appears to be an AY-3-8914 clone.  Importantly, there is zero evidence that it is an AY8930 clone.  The 8930 printed on the package is most likely a date code.

It's an AY-3-8914 clone, not an AY8930 clone. The STIC1A layout work was done in 1982/1983 and so long predates the AY8930. The 8930 printed on the package IS a date code.

The volume registers are 6 bits, like the original AY-3-8914.  Bits 5:4 behave unlike any other PSG observed to date for values 01b, 10b, however…

These discrepancies were noted and would have been fixed, but Morris cancelled all hardware development. Valeski decided to live with the discrepancies rather than pay for an iteration.

Because the jumper block involves GA9 and GA10, I speculate INTV Corp may have experimented with methods to correct the INTV88’s GRAM compatibility issues.

Cutting the traces connecting JP3/JP6 and JP4/JP5 disconnects GA9 and GA10 from the 6116 GRAM; inserting JP3 and JP4 grounds only the A9 and A10 pins of the 6116, reducing it to the Intellivision I/II size of 256 bytes (64 characters).

 

While this can be done manually on this board, such manual work is expensive and it's better done via a board layout change--including this geometry here makes implementing such a board layout change trivial. Furthermore, if you have two separate inventories of boards, one for World Book and one for Super Pro systems, you can in a pinch adapt one for the other if you have an inventory imbalance.

The STIC1A does not appear to provide a mechanism to disable the additional GRAM.

There WAS a mechanism for disabling the additional GRAM: tie the A9 and A10 inputs of the RAM chips to ground. STIC1A was always intended as a strictly cost-reduction measure. No new features were to be allowed, as programmers might take advantage of them and the incompatiblity not detected until cartridges were in the field. Any incompatibilities were bugs. This meant no additional GRAM was allowed, so no mechanism for enabling it was required. The only reason GA9 and GA10 were brought out of STIC1a at all was to address GROM.

 

It was expected that Intellivision systems would implement the 512 bytes of GRAM using whatever available devices were most cost effective. For the original Intellivision that meant two GTE 3539 256x8 static RAMs. By 1982 the economics had changed, so STIC1a based systems were expected to use two 2114 1024x8 static RAMs with their high order address line grounded, as was done on the Intellivision II and on Valeski's INTV 1987 board that used the CP1610A. The potential extra 512 bytes of GRAM already paid for would just sit there, tantalizingly out of reach. Of course, the additional GRAM would only have been useful in color stack mode, but in that mode it would let you bit map half of the screen.

 

Then along came World Book. The TutorVision system Valesky configured for them was conceptually a new private-labeled platform that did not have to be compatible with the existing Intellivision customer base. For this it made sense to upgrade GRAM to 2k x 8, potential incompatibility be damned. So that's what he ordered, then settled snug in his bed with visions of $ $ dancing in his head.

 

You may notice that the INTV 1988 board has no U13. U12 and U13 were originally a pair of 2114's with their A9 inputs grounded; they were combined into the 6116 at U12.

 

Valeski should have disabled the extra GRAM on those INTV 1988 boards he diverted for use in late-model Super Pro Intellivisions, but he didn't consider Worm Whomper compatibility worth the cost of reworking the boards to cut the two traces and install the two jumpers.

5.5kΩ (via two 11kΩ in parallel)

Parallel resistors were specified here because standard resistor value spacing is log scaled with 24 values per decade and rounded, whereas what's needed for resistive ladders is a power-of-two scale. The nearest standard value is 5.6k, about 2% too high. This is the lowest value resistor and so has the greatest effect, so you want its value to be the most accurate.

 

WJI