Look at the backside of the Mitsumi card and see how many pins of that IDC40 connector are tied to ground. There were 3 custom CD-ROM protocols back in the day and (Mitsumi being one of them) all are only 8-bit. So if there's a lot of GND pins, it's a custom Mitsumi protocol card. Still usable but only with such drives.

feipoa wrote:

JP3, which has no label. There are jumper locations with headings of: I9, I3, I5, blank, I10, and I11. IRQ's? But what is blank for? On the Acculogic, the IRQ's selections are for 10, 11, 12, and 15. Seemed odd to me that the Mitsumi card would hae 3 and 5 available, but not 12 and 15.

There is a dip switch, SW1. The labels for each switch are A9, A8, A7, A6, A5, A4, A3, and A2. I assume these are addresses, like 1E8-1EF, but there is no table indicating which switch is for which address. And by default, the switch has A2-thru-A7 set to ON, and A8 & A9 set to OFF.

IRQ pins (JP3)
There is no gap - there are only 5 pairs of pins.
IRQ15 would be used for the primary hard disc controller
A machine with a CD-ROM would probably a have a mouse - so there goes IRQ12

I would assume that A2 to A9 represent the address on a one to one basis.

From what you've said the address is either:
98 7654 32xx
00 1111 11xx = base address 0x0FC
or
11 0000 00xx = base address 0x300

My money is on 0x300 being the correct address.

A look at the Mitsumi "mcdx" driver in the Linux 2.4 kernel shows that the expected base addresses for Mitsumi drives are 0x300+IRQ11 for the first and 0x304+IRQ5 for the second.

Given that JP3 is set to 11, and the DIP switch looks like its set to 0x300, this looks like further evidence that you have a Mitsumi interface rather than an IDE interface.

More info:
Looking at

http://computer-retro.de/Multi-IO.html

which shows pictures of
"Mitsumi CD-ROM Drive 16Bit I/F Card 74-1645A ISA CR623803"
and your card:
"Mitsumi CD-Rom Drive Controller 16Bit I/F Card 74-1881A"

The cards appear to be more or less identical

This link describes someone looking for info on the 74-1645A

http://discussions.virtualdr.com/showthread.p … 16-I-F-card-old

For me these bits of info more or less confirms that you have a non-standard Mitsumi CD-ROM interface card.

that mitsumi card is a for the old proprietary mitsumi interface, not IDE. It is from the same generation of non standard interfaces like the sony and panasonic one. It will work only with the earliest mitsumi cd rom drives, like FX001

So if there's a lot of GND pins, it's a custom Mitsumi protocol card. Still usable but only with such drives.

Yes, the whole top row is tied to ground. So Mitsumi drives only?

There is no gap - there are only 5 pairs of pins.

Yes, don't know how I missed that.

A machine with a CD-ROM would probably a have a mouse - so there goes IRQ12

I don't know. This card is from the early 90's and many people were using serial mice on IRQs 3/4 or even bus nice on IRQ 2/9. And audio on IRQ 5. Some systems had PS/2 mice on IRQ 12, but I don't think it was the majority. For my reasons for needing this card, IRQ 14/15 are unused as there isn't an onboard IDE controller.

OK, so you think the addresses are just a binary to HEX conversion? That is easy enough. But the DIP shows that up is ON, meaning '1'. My card is jumpered the same as in that photo, but IRQ is 10, DMA is 5.

I have some Mitsumi drives, but the slowest is 8x. I also have 12x and 16x. Will the 8x work on this card? Or will it fry the card/drive?
.
Given the age of the card, does NT4 and W95 have drivers built-in?

feipoa wrote:

I have some Mitsumi drives, but the slowest is 8x. I also have 12x and 16x. Will the 8x work on this card? Or will it fry the card/drive?

unlikely. IIRC the fastest proprietary mitsumi drives were 2x, maybe there were 4x but I don't think so. Easy way to tell is that IDE drives have Master/Slave jumpers, the proprietary drives do not have any jumpers, only a 40pin connector and a 4 pin audio connector.
.

feipoa wrote:

Given the age of the card, does NT4 and W95 have drivers built-in?

IIRC windows 95 should have drivers, NT4 maybe but I wouldn't bet on it.

feipoa wrote:

Yes, the whole top row is tied to ground. So Mitsumi drives only?
I have some Mitsumi drives, but the slowest is 8x. I also have 12x and 16x. Will the 8x work on this card? Or will it fry the card/drive?
Given the age of the card, does NT4 and W95 have drivers built-in?

Not just any Mitsumi, only specific older drives that used their custom protocol. I don't think there were any drives faster than 2x made for such 8-bit bus.
Don't connect IDE drives to that header, or custom Mitsumi CD-ROMs to IDE ports. Not safe. If in doubt just check if the drive in question also has the GND pins in the same spots.

As for drivers, I know there is one for Linux since I've used it (though I'm not sure how current distros handle ISA, if at all), and it is possible to get it working with MSCDEX once you find the correct SYS for that card. But NT or 95? I wouldn't count on it but I don't remember - I replaced my drive with an IDE one by the time I moved to 95.

OK, I won't be trying my Mitsumi 8x IDE drive on this proprietary Mitsumi ISA card. The only 2x and 4x drives I have are SCSI. I guess this card is getting thrown into the draw of expansion cards. Thanks for all the advice on this matter.

Deunan wrote:

feipoa wrote:

Yes, the whole top row is tied to ground. So Mitsumi drives only?
I have some Mitsumi drives, but the slowest is 8x. I also have 12x and 16x. Will the 8x work on this card? Or will it fry the card/drive?
Given the age of the card, does NT4 and W95 have drivers built-in?

Not just any Mitsumi, only specific older drives that used their custom protocol. I don't think there were any drives faster than 2x made for such 8-bit bus.

To complicate matters, some manuals for such drive/card combos wrongly referred to the Mitsumi interface as being an ATAPI/IDE interface. Needless to say, the drives don't work on a IDE/ATAPI interface...

Sold one of these drive/card combos back in the day (worked with interface card), only to get back a complaint "The drive doesn't work" because the buyer had tried hooking it up to a IDE controller.

feipoa wrote:

But the DIP shows that up is ON, meaning '1'.

That doesn't necessarily follow - it depends on how the electronics is implemented.

In all likelyhood the DIP switch is feeding multiple inputs on the chip, which have a resistor (inside or outside the chip) to ensure that an unconnected input is pulled to +5v, and hence gets
read as a logical "1".

When you close the switch, it connects that input to 0v.

So you have a situation where "ON" means the chip input is pulled to 0v, and 0v = logical 0.
When you leave the switch set to "OFF", the chip input is pulled to +5v (via a resistor), and +5v = logical 1.

You'll note that I posted:

From what you've said the address is either: 98 7654 32xx 00 1111 11xx = base address 0x0FC or 11 0000 00xx = base address 0x300 […]
Show full quote

From what you've said the address is either:
98 7654 32xx
00 1111 11xx = base address 0x0FC
or
11 0000 00xx = base address 0x300

0x0FC was the first answer, because the first assumption is that "ON" = 1
However when the address that yielded was obviously incorrect, I then inverted each bit yielding an answer which was worth investigating further.

If you want to learn more about the electronics side of this, look at the first two sections of https://www.elprocus.com/pull-up-and-pull-dow … h-applications/
(In case of TL;DR scroll down to the second paragraph "Pull Up Resistor" and the diagram shown there.)

/MD