Addressing problem with 16C2552 dual UART for RC2014

[Steve Clarke]

Hi All,

I've built Etched Pixels EPDSER 16C2552 dual UART board and I'm having a very bizarre addressing issue which I can't seem to work out.

As I understand it putting the address jumper at pin 5 to pin 6 on J2 should put the first address of the UART at 0xA0, with the second port appearing at 0xA8. However, according to ROMWBW's hardware inventory this seems to put the base address at 0x20, making it think there's a PPIDE card present. In this position survey.com within CP/M shows 0x20 through 0x2F. Other pins pairings have equally unexpected addresses. None of which correspond with the documentation (80,90,A0 etc). I can provide a list if required.

I had concerns about some of the the parts I was using. Therefore I have used 2 different 74HCT1238 parts (ST M74HCT138B1 and TI SN74HCT138NE4) and I had trouble sourcing the 1N517 diode so used a 1N5817 I had a few spares of but I think that should be fine?

I've also checked for all the usual mechanical issues and with chips removed found no shorts or bad connections. I traced every pin according to the PDF and they are all as expected.

Any help or suggestions would be greatly appreciated.

Cheers,

S.

[Alan Cox]

> As I understand it putting the address jumper at pin 5 to pin 6 on J2 should put the first address of the UART at 0xA0, with the second port appearing at 0xA8. However, according to ROMWBW's hardware inventory this seems to put the base address at 0x20, making it think there's a PPIDE card present. In this position survey.com within CP/M shows 0x20 through 0x2F. Other pins pairings have equally unexpected addresses. None of which correspond with the documentation (80,90,A0 etc). I can provide a list if required.

The A7 line is connected to E3 (the must be 1 enable) on the decoder
and the decoder drives the chip enable. If it's appearing at 0x20 not
0xA0 (or at both) that suggests to me you've got a lack of continuity
between A7 and pin 6 of the HCT138 (bad joint, pin not in socket, bad
backplane joint etc).

Check the resistance between U2 pin 6 and the CPU card slot A7 line
and see if it is properly connected or floating high, thus letting the
board enable.

>
> I had concerns about some of the the parts I was using. Therefore I have used 2 different 74HCT1238 parts (ST M74HCT138B1 and TI SN74HCT138NE4) and I had trouble sourcing the 1N517 diode so used a 1N5817 I had a few spares of but I think that should be fine?

Shouldn't matter which HCT138 you use. The diode is only used for
interrupts so your only symptom with an unsuitable diode would be the
IRQ not working I think

> I've also checked for all the usual mechanical issues and with chips removed found no shorts or bad connections. I traced every pin according to the PDF and they are all as expected.

In which case I would definitely check between the CPU slot and U2 pin
6. Clearly the CPU end of it is properly connected but the other might
not be ?

Alan

[Steve Clarke]

After an accidental private conversation with Alan this has been resolved.

The pin out of J2 for the address allocation was not as expected. Confusing matters, ports A and B on the board are reversed with RomWBW using port B as the primary console.

Many thanks to Alan for his designs and his assistance.