SmartDiskII Interface (by BTB) Troubles with my DIY built card

If you use a 2732 eprom instead of a 2716 you should program your data in the upper half of the address range. Normally you can enter a start address where your programmer starts writing the data . If you enter  (800 hex) as start address everything will work as expected.

In post #2, 'driessens_nl" wrote:

" If you enter  (800 hex) as start address everything will work as expected. "

Uncle Bernie comments:

Assuming that the firmware EPROM is programmed correctly, the firmware will be there and it will work.

But the real problem with BTB's SmartDisk II and Disk II substitute cards is a timing issue with the other 2732 he uses as a substitute for the state machine PROM in the 'Woz Machine'. This timing is just too tight and can't be met with most 2732 EPROMs. Some later versions of the 2732 are fast enough, though, especially some later 27C32 made in CMOS using aggressive die area shrinks.

I wrote about this problem here on Applefritter and showed the timing calculations to prove my point, but can't find the thread anymore, which has been started by another victim whining about his BTB cards not working.

(I myself have two specimen of said BTB cards, which don't work, donated to me by an Applefritter member, so I have evidence other than plain theoretical calculations).

The fact that most 2732 - especially the early ones with large die area - are too slow for the 'Woz Machine' while some 2732 are fast enough may explain why building these contraptions is a hit-and-miss for most builders.

I could develop an easy fix for these two BTB cards which would be cheaper and easier than finding a fast enough 2732 (this is why I was donated these two specimen mentioned above) but I will do this only if there is enough interest from the Apple II aficionado crowd. I'd need to design a small PCB which goes into the place of the state machine 2732. There are three possible ICs which could go onto this PCB, replacing the 2732:

- a 256 x 8 bipolar PROM (expensive to get and needs professional programmer)

- one of these super fast CMOS EPROMs made by Cypress Semiconductors (expensive to get, but I have lots of them)

- a GAL22V10 (easy to get and easy to program --- many decades ago I put the whole Woz machine into one of them, and when I take the registers out, and specify combinatorial outputs, it can replace the PROM, but with a slightly different footprint)

These are the options we have to fix the issues BTB has created. Easy enough. But it will take maybe 2-3 hours of work to create the PCB. And then these must be ordered at OSHPARK and populated and put into the BTB cards. And then they must be tested using real Apple II. All this is a lot of work and there can be more bad surprises - I didn't do a design review yet of the rest of the circuitry on these cards but at a first glance there are hints he may have been struggling with making the soft switches work. Who knows what else may be wonky. It's generally not easy to design Apple II slot cards that work, due to the botched 6502 bus architecture - which was OK for the 1970s but turned into a problem when static memories got faster and faster. Ironically, the 6800 had the "VMA" output which, if put to use in the proper way, can avoid the issue with the crumbling addresses the 6502 bus architecture has (this happens at the end of PHI2, and the remedy is to have a magical gate that can look into the future - negative propagation delay - or a whole state machine to replace this magical gate with a real thing).

If you have a non working BTB card, comment here how far you want to go to fix it, and if you want to contribute your time and money to develop the fix. I want to keep my own time investment into this at a minimum.

- Uncle Bernie

I believe that BTB said NOT to change any of the solder jumpers on the board so I haven't but I read somewhere that sometimes you have to copy the file multiple times to fill the ROM. Is that what I need to do and if so how? (in blocks or all at once by cutting and pasting the data multiple times before burn).

 Timing issues aside - the easiest way to run a 2K 2716 binary on a 4K 2732 is to burn the file twice.  I think your XGECU burner software may have an option for this, or you could load the 2716 binary into the 2732 space twice.  driessens_nl explained tht you want it in the upper half of the space so doubling the binary achieves this without having to figure out where to start pasting in the code.