QuadRAM Quadboard, parallel interface

[deanimator]

Hi,
Does anyone have a schematic for the 256K Quadboard made by QuadRAM? I do seem to have some difficulties trying to get the parallel port working on mine - it does not show up in the system regardless of it being set to LPT1 or LPT2.
So far I was able to rule out address or resource conflicts as I am able to set up a Hercules card on LPT1 at base address 3BCh correctly - at the same time, the Quadboard won't work on LPT2, or when switched as LPT1 without the Hercules card.
Next to the RTC chip on said card, there's an 8255 PPI that I presume might also be used for parallel I/O, but I'm just guessing there, and without a schematic I get lost in there (it's a full size ISA). I've already tried to solder in a 40pin socket and try out a different 8255, to no avail - I'm guessing there's a fault somewhere in the TTL glue/address decoding logic. The rest of the card seems to work fine however, incl. the serial interface and the RAM on it.
Even though the card was originally equipped with a nickel timekeeping battery, it was desoldered in time, before any electrolyte leakage and subsequent corrosion of the traces could occur. (I did not test the RTC as I do not care much about that.)

It is the "early" Quadboard with 4 banks of 4164 DRAM, with its parallel port in an external bracket, connected to the card via a DIL-14 socket.

Thanks for your help.

Regards

 

[hornbetw]

The 8255 shouldn't be part of the LPT port. It wasn't used on the original PC LPT ports, so it would need different software in the BIOS to function as an LPT port. http://minuszerodegrees.net/quadram/quadram.htm has some info on a couple of Quadram boards.

 

[deanimator]

Yes, that was my second thought - the PPI is there to support the RTC chip.
It's that third board mentioned in the link above. The manual only mentions that the design of the parallel interface is "virtually identical to IBM's design" so I think they are referring to the "IBM Printer Adapter" ISA card.
They do omit a couple of signal lines in it, but they do mention some modifications that can be done to the serial interface - in words only, without a schematic.
To go along with the Quadboards there was a datadisk called "QuadMaster" which included a "diagnostics" BASIC script called QUADTEST.BAS. Unfortunately for me, the parallel port test just sends some characters to a connected printer.
Well, off to study the IBM Printer card schematic...

 

[modem7]

So far I was able to rule out address or resource conflicts as I am able to set up a Hercules card on LPT1 at base address 3BCh correctly - at the same time, the Quadboard won't work on LPT2, or when switched as LPT1 without the Hercules card.

Be aware that the terms 'LPT1' and 'LPT2' can be used in different contexts.

When card makers refer to LPT1 and LPT2, they are referring to hardware addresses. The norm is:
LPT1 = 378 (actually, the range 378-37A hex)
LPT2 = 278 (actually, the range 278-27A hex)

Right now, you are probably saying to yourself, "Hang on, but on my Hercules card, LPT1 is 3BC !"

And things get more confusing. DOS is dynamic. For example, what DOS refers to as LPT1 is simply the first port found, LPT2 is the second. If I only have one LPT port and it is at 278, DOS will assign that as 'LPT1', but if I run certain software, that software will refer to the same port as LPT2 !!!

it does not show up in the system regardless of it being set to LPT1 or LPT2.

What exactly is "system" ?

When the power-on self test (POST) executes, one of the things done is shown at [here]. So, you could use DEBUG as shown, to see what the POST found.

 

[deanimator]

What exactly is "system" ?

With a Hercules card and the Quadboard DIP switched to "LPT2" per manual, i.e. Page 19 Table 5 "Parallel Printer Port Address Selection" (to my same dismay, the folks at QuadRAM used "LPT" names instead of base addresses in hex in their manual), the BDA after POST fills the first parallel port base address stored at address 0:0408 equates to value of 03BCh, i.e. BC 03 in the RAM, and the second base address @ 0:040A equates to two zero bytes, followed by another four zeros since there are no tertiary and quarternary printer ports in the system. A "CheckIt" diagnostics program shows one and only LPT1 with a base address of 03BCh.

Now, if I take my Hercules card out, DIP switch the 5150 to expect a 80x25 CGA, put in my Twinhead CGA clone which has nil parallel interfaces in it, and DIP switch the Quadboard to use "LPT1" per manual, i.e. same Page 19 Table 5, the BDA after POST starting at address 0:0408 shows zilch and zeros all around; to be exact, 8 zero bytes, right until 0:0410, where the 2 equipment bytes follow. This time, the CheckIt v3 diagnostics program does not show any parallel ports in the system at all.

As for a third experiment, which turned out to be a little bit less sophisticated than the previous two, I've connected a homebrew Covox-style DAC to both ports, one by one, and turned up the volume. In both cases, with the amplification turned significantly up, I was able to hear the data bus chirping... but only in one case did the DAC make a loud "pop" into the speaker; and that was with the Hercules card, almost at the end of the POST, when the BIOS in a 5150 enumerates serial and parallel base addresses.

I might be wrong in my approach, and my second pair of eyes can be deceiving enough, but I'm starting to suspect I have an actual hardware problem with the Quadboard card... something I have suspected in my first post prior.

 

[deanimator]

Minor update: went on to desolder, socket and test some of the bus logic chips near the parallel out socket in an external tester - and I have also attempted to contact the successors of Quadram if there were any remote possibilities of old electical schematics surviving the test of time; so far, unsuccessfully. I'd just get a second parallel port card if only it weren't a 5150 with 5 ISA slots in total.