Close

Drawing the Schematic

A project log for Recreating the Jupiter Ace

Recreation of the 'Fastest computer in the Universe' that runs FORTH.

cees-meijerCees Meijer 06/10/2019 at 18:079 Comments

As I'm nearly finished re-drawing the schematic by Bodo Wenzel and Isidro Nuñez Blanco De Arenas there are some issues as to where to draw the line between staying true to the original, and add modifications that will make it more useful. I love Grant Searles 100% replica of the PCB, but this will require all original parts which may be very hard to get. But I do want to stay close to the original, because otherwise you might as well just put  an Arduino board on it, and run an emulator.

Isidro's schematic already has some modifications, like using the 6116 RAM and a 27C64 EPROM which I decided to follow. Actually, after a suggestion from one of my followers I decided to go for a 27C512 and add some jumpers so it will be possible to select multiple images.

Next up is the decision to add a direct video output. The ASTEC RF modulator is not only rare, but it also serves no purpose so it makes sense to leave that out. After all, with every retro computer I get nowadays the first thing I do is add the video-out modification. I found a neat schematic for video out by '1024MAK' on the ZX forum:

After looking at the annotated schematic by 'nocash' I also noted some small additions

marked 'ACE4000'. /ROMCS is a contact on the edge-connector. Not sure what they are used for, but I decide to add them as well. After all, if not used, the resistors can just be replaced by jumpers, and the diodes left out. 

Discussions

Cees Meijer wrote 12/08/2020 at 09:27 point

@Curt J. Sampson  : Thanks for confirming it, and pointing me at your github. I have not done much with Commodore hardware yet, but now my Jupiter ACE is almost finished this might well be part of a future project...

  Are you sure? yes | no

Cees Meijer wrote 06/11/2019 at 06:49 point

@Ken Yap : Thanks for your input. I had not considered the crystal yet. As crystals are harder to get nowadays anyway it may be worth looking into that. The EPROM issue is also something that crossed my mind. Actually I thought the 27C64 was still widely available, but I just ran a quick search and it's marked 'out of stock' at almost every major supplier. So I was wrong. I will upgrade it to a 27256.

  Are you sure? yes | no

Ken Yap wrote 06/11/2019 at 06:33 point

There is another concern that occurred to me. Is the Ace dependent on crystal frequency? The 6.5MHz crystal may be difficult to source, it's not a commonly used frequency.

  Are you sure? yes | no

Cees Meijer wrote 06/11/2019 at 10:26 point

I've just checked, and both Mouser and DigiKey stock 6.5 Mhz crystals ( ECS-65-20-4X ), so that's not a problem.

  Are you sure? yes | no

Ken Yap wrote 06/11/2019 at 00:05 point

That /ROMCS line may be a way to select different images in a larger EPROM, one would have to look at the address line details.

Edit: Looking at the schematic, taking this line high forces the lines from the 138 decoder to be high which then disables the 2732s. However this turns on the AND gate in the 7411 which is marked For 2764. So it looks like a way of enabling the 2764.

  Are you sure? yes | no

Cees Meijer wrote 06/11/2019 at 07:01 point

I think so as well. It could be a way to disable the EPROM so you can connect one to the expansion port and the system will run from that.

  Are you sure? yes | no

Curt J. Sampson wrote 12/08/2020 at 02:48 point

Yes, that's exactly what it is. If on the expansion connector you ground that line or let it float  it will have no effect (because of the direction of the diodes the line cannot draw current), but if you pull it high it will override the '138 trying to pull either ROM select line low, and thus the internal ROMs will always have their select deasserted.

This is not only a great idea, but reasonably common. My National/Panasonic JR-200 also lets devices on the expansion bus  disable internal ROM as well, though only one of the two ROMs, and the Commodore 64 had a slightly more sophisticated scheme allowing either or both of the BASIC and KERNAL ROM to be disabled and the cartridge ROM to be placed in a couple of different locations. (I have some notes on that here, if anybody's curious: https://github.com/0cjs/sedoc/blob/master/8bit/cbm/address-decoding.md .)

  Are you sure? yes | no

Ken Yap wrote 06/10/2019 at 22:51 point

Maybe you could allow the use of the 27128, 27256 or 27512 with jumpers as they are all in 28 pin packages, in case the 2764 proves hard to obtain.

Actually depending on the state of the higher address lines, jumpers may not be necessary, you just have to burn the image into the correct half, quarter, or eighth of the larger EPROM.

  Are you sure? yes | no

Cees Meijer wrote 12/08/2020 at 09:24 point

We're nearly a year further, I recently completed the whole design and this is exactly what I did. I designed it for a 27C512, and added a jumper section. My ACE now works, but I have not tried with different ROM images yet.

  Are you sure? yes | no