But they're definately not reliable. On a turboR e.g., only the first 5 bits are returned, the rest is pulled up to 1.

Quote:

But they're definately not reliable. On a turboR e.g., only the first 5 bits are returned, the rest is pulled up to 1.

Maybe reliable is not the right term Every circuit diagram of memory mappers I have seen has the facility to let the cpu read the written contents of the mapper registers. In fact that makes mapper hardware so complicated.

What the standard for me is vaque about is what happens if there is more than one memory mapper available in a MSX. The I/O ports in the MSX are not coupled to slots, so if more than one mapper is available, which mapper returns what? Or is it the strongest mapper (electrical) which wins? It seems in a turboR the S1990 wins that fight and keeps the upper bits high.
And when written, are all mappers given the same contents for the I/O registers?

I think you are right. Although it does sound a bit strange that circuits would be designed to have two devices driving the bus during a read.

Anyways, I think having the memory mapped in slot 3-3 may be necessary in the HB-F1II, so I'm probably going to try that. I could post a wiring diagram when I'm done.

That would always be good: share the knowledge

OK, I now have 512KB in slot 3-3 ^_^ although it is using cartridge slot 1.

I'll describe how I built the cart but first something a bit funny. I mentioned before how it was only seeing 256KB instead of 512 and I didn't know why. Well, today after making the modifications to make the RAM appear in slot 3-3 instead of slot 1, I looked in the S1985 manual and found that MA18 (the highest address line) is only enabled if one of the keyboard inputs is low at power on. It's tied to a pull up resistor on the board. I thought about this for a moment and decided I was tired of doing work. So I put the machine back together thinking that 256KB is good enough for now and I'm probably lucky that the computer even works after all this playing around.

So I played games for a while and then I came to write this message. Then I thought "wait... keyboard input? pull up resistor? hmmm..."

I grabbed a nearby cardboard box and used it to press down all the keys on the MSX, and turned it on. Guess what? Now the full 512KB is enabled. Just gotta hold down some keys when I turn it on...

Thank you, HansO, for providing all the great documentation on your website. Without it, (that and the extra time off from work I had this month) I couldn't have done this successful (and educational!) project.

Here is how I built the cart:
Cut an ISA card to get a chunk of PCB with a 50-pin edge connector on it. Solder a 32-pin socket to a piece of prototyping board. Drill holes and use small nuts and bolts to attach the proto-board to the ISA chunk.

The 628512 pinout is like this:
pins 1-12, 23, 25-28, 30, 31 are address lines
pins 13-15, 17-21 are data lines
pin 16 is ground, pin 32 is Vcc
pin 22 is /CS
pin 24 is /OE
pin 29 is /WE

I wired up A0~A13, D0~D7, /OE to /RD, and /WE to /WR, as well as power and ground.

Then inside the HB-F1II, I had to solder wires to pins 3-7 of the S1985 to get the five highest address lines.

Then I desoldered pin 6 of the 7432 (this is the /CAS signal for the onboard RAM) and ran a wire from that pin to /CS on the 628512 socket. Lastly, I had to desolder pin 12 on the 7408 and run a wire from that pin to pin 10 on the 7432.

Wow, this is so cool! Thanks, now I have a new fun project with my two Sony MSX2 computers