Micro ATX - DoG

[DoG]

I have been gathering material for this for some time now and will hopefully start building before this year is over. I have a few STfm so I have to decide which one is going to be the candidate as a donor for this project. I have also made some small modification boards that I will be trying out and share in this thread later on hopefully.

More info can be found here about the project: http://ataripcb.pl/

ATX small.jpg (122.66 KiB) Viewed 2091 times

ATX_ram_small.jpg (27.62 KiB) Viewed 2091 times

[DoG]

This little adapter board made if you want to use a original Mega ST(e)/TT keyboard instead of the Eiffel. Replaces the PS/2 outlets and the PIC16F876 with a RJ11 instead.

Micro_ATX_Original_Keyboard.png (41.23 KiB) Viewed 2086 times

Made two board for the same job. So I will not be using the +12V/-12V components as well as the 1488/1489 for the RS232/serial communication on the main board. I will be using a MAX238 as on the H5 Phoenix since that runs on 5V. Adapter board will be sitting in the 1488/1489 sockets on main board. These was done and ordered yesterday so will take some time to arrive.

MAX238SMT_smol.png (53.39 KiB) Viewed 2086 times

MAX238DIP_smol.png (56.59 KiB) Viewed 2086 times

[DoG]

Did a bit of soldering during Christmas holidays. So all the birdseed is soldered on. I seem to have solder the socket for IC27 backwards when I look at the photos. Updated the PCB for the RAM module since no resistors was present on the address signals, WE or RAS0 and RAS1. The two CAS1 signals had 68Ω on the ram module and on the main board (under MMU) so removed the ones on the module.

I want to change the pullups on the main board to something stronger. Like 4.7K on the data bus and 3.3K on address bus (as the PAK3 manual recommend). However the resistor arrays on the mainboard is a bit of mess. RN3 and RN4 is data bus and RN6 and RN7 is address bus (most of them) so pretty easy to change. RN8 has the A1-A4 but also other signals. A21-A23 is 0805 resistors located under the GAL for TOS v2.06 decoding on the board. So have to do something about that. RESET and DTACK is however part of other resistor arrays on the board and they are currently 4.7K. They should be 1K on the schematics I have checked. Probably add 10K pullups on the data bus between the WD1772 and DMA as well.

2023-01-07 16.56.09.jpg (722.56 KiB) Viewed 1921 times

RAM module v2 uATX small.png (105.58 KiB) Viewed 1921 times

[tOriman]

Hi,

I'm watching your progress on this project. I think people can benefit a lot from your experience. Author also.

Regards
tOri

[DoG]

I fitted all of the small ICs. Changed RN3 and RN4 to 4.7K (data bus). Changed RN6, RN7, R44, R45, R46 and pin 6-9 on RN8 to 3.3K (address bus). RN8 pin 3-5 is still 10K (FC0-FC2) on the bottom of board. DTACK (pin 2 of RN9) and RESET (pin 5 of RN5) is changed to 1K on the bottom. Also changed VPA (pin 5 of RN9) to 2.2K on the bottom. The rest of the pullups is 4.7K. I also programmed the two EPROMs, the two GALS and the PIC although they are not fitted on the photo below. Soldered in the two plcc sockets for GLUE and MMU. I also tried the fit on some of the adapter boards in previous posts.

uATX without Atari chips_small.jpg (186.95 KiB) Viewed 1783 times

pullups_small.jpg (56.09 KiB) Viewed 1783 times

A21-A23 pullups_small.jpg (26.75 KiB) Viewed 1783 times

[HigashiJun]

Such a nice purple PCB...

[LarryL]

I fitted all of the small ICs. Changed RN3 and RN4 to 4.7K (data bus). Changed RN6, RN7, R44, R45, R46 and pin 6-9 on RN8 to 3.3K (address bus). RN8 pin 3-5 is still 10K (FC0-FC2) on the bottom of board. DTACK (pin 2 of RN9) and RESET (pin 5 of RN5) is changed to 1K on the bottom. Also changed VPA (pin 5 of RN9) to 2.2K on the bottom. The rest of the pullups is 4.7K.

Sweet!
I also like the purple board!

I had similar plans re the different pull-ups around the 68k
Designed a adapter PCB to separate address-, data-bus and other control signals
Any special reason, why FC0-2 should remain with 10k?
I checked different schematics and found them to be 4k7 usually…
But not 100% sure

I struggled a lot with DMA issues (read/write errors for floppy and ACSI), but now found a stable mix of IMP and Atari Chips (MMU&Glue is IMP, shifter is IMP with diode mod, DMA is Atari). Needed also the 10k pull-ups at DMA data bus (res net on bottom of PCB) and the 1k pull-ups at Floppy-Controller (SMD resistors under PCB)

Regards
Michael

[DoG]

I had similar plans re the different pull-ups around the 68k
Designed a adapter PCB to separate address-, data-bus and other control signals
Any special reason, why FC0-2 should remain with 10k?
I checked different schematics and found them to be 4k7 usually…

Good question. No special reason. Probably doesn't matter that much. FC0-2 on ST(f,m), Mega 1-4 and STe schematics that I checked never seemed to be changed by Atari over the years (10K on all of them). And I have not read anything about changing those pullups for any reason. IP0-2 did however have 4.7K on the schematics I looked at.

Smart to do a new board with pullups. Will look neater.

I struggled a lot with DMA issues (read/write errors for floppy and ACSI), but now found a stable mix of IMP and Atari Chips (MMU&Glue is IMP, shifter is IMP with diode mod, DMA is Atari). Needed also the 10k pull-ups at DMA data bus (res net on bottom of PCB) and the 1k pull-ups at Floppy-Controller (SMD resistors under PCB)

Yes, I figured that those pullups are needed. IMP seems happier with 4.5V on VCC. I have some boards with that mod done from factory.

[LarryL]

Probably doesn't matter that much. FC0-2 on ST(f,m), Mega 1-4 and STe schematics that I checked never seemed to be changed by Atari over the years (10K on all of them). And I have not read anything about changing those pullups for any reason. IP0-2 did however have 4.7K on the schematics I looked at.

Maybe I have mixed up FC0-2 and IP0-2
Will check again and probably have to add that to my adapter board
Thanks