• Please review our updated Terms and Rules here

Heads up: Compaticard I on eBay

The CCI is an odd beast, as it's not register-compatible with the 5170 convention or the CC IV. It's an early product. ImageDisk, for example, won't work with it correctly.

22Disk will, however, if you specify the "CC" flag in the DISKETTE.CFG file.

Personally, I'd steer clear.
 
raifield said:
Almost as bad as the $140 8-bit ISA 1.44MB floppy controllers.
Click to expand...


This seems like another thing that could be solved with a project like the XTIDE. 8 bit HD floppy controllers are pretty rare now and the eBay hoarders are asking a premium for them. I'm not, of course, volunteering as I have neither the time or the requisite technical skills to design such a thing. However, I wonder how hard it would be to harvest a chip, (say something by winbond) off of one of the ubiquitous cheapy Taiwanese 16 bit ISA combo cards (ide, fdd, serial and parallel)?

Does anybody still make an inexpensive floppy controller chip that you could interface to the PC/XT bus using TTL parts?
 
You'd have to buy NOS, but they're around. I'd recommend the National Semi 8473 or 8477. Both are available in PLCC, support up to 4 drives and require very little in the way of interface "glue". Both have sufficient drive capability that render external buffers unnecessary. I've got a tube of WD37C65 chips in PLCC, but they require 2 crystals (the National chips require only one) and external buffers.

SMSC also made a few comparable chips. I'd stay away from Intel, as they broke FM support a long time ago in the 82077AA-1 and succeeding chips. I suspect that lack of FM capability might be a deal-breaker for some.

Edit The part number for the 48 pin DIP version of the 8473 is DP8473AN; the number for the 44 pin PLCC is DP8473AV. The PLCC version supports only 2 drives, rather than 4. The 8477 part number is PC8477 and also includes support for 1Mbps (2.88M format) data rate and comes in a 68 pin PLCC. There's a seller on eBay offering a lot of 20 for about $40. He says that he's got more than 10 lots of 20 available.
 
Last edited:
Hi
We've used the WD37C65 on various N8VEM projects like the N8 and the SBC-188. Sergey used the FDC9266 on the Zeta and it works well too.

I've proposed making a PCI multi IO board before but that fell flat unfortunately. Not all projects are viable though like the ISA OPL2 board, etc.

An ISA FDC or even multi IO board would be fairly easy to do. One important ground rule though... all the IO goes out the top and not out through the ISA bracket. Getting the board and the ISA bracket lined up has proven to be difficult for past ISA projects. Getting ISA brackets to begin with is a real PITA as well.

The main issue with an ISA multi IO is most of the chips are SMT which will cause some hobbyists grief. It doesn't effect the design or bother me any but you can almost hear the howls of protest already if there was a big fat 100 pin QFP in the center of the PCB. I do have the part libraries and footprints for the WD37C66x controllers since I handbuilt them for the (also failed) P112 clone project.

If anyone is willing to organize and manage the project (like Hargle did on XT-IDE) I'd be glad to write up a schematic and PCB layout. Getting prototype PCBs would not be hard either. However I won't participate in full or partial assembly of boards or group buys of parts.

Thanks and have a nice day!

Andrew Lynch
 
lynchaj said:
Hi
We've used the WD37C65 on various N8VEM projects like the N8 and the SBC-188. Sergey used the FDC9266 on the Zeta and it works well too.

I've proposed making a PCI multi IO board before but that fell flat unfortunately. Not all projects are viable though like the ISA OPL2 board, etc.
Click to expand...

Tell ya what, Andrew. Do a board using one of the NSC chips--give me a bank of DIP switches (2 per drive), 4-8KB of ROM and I'll write the BIOS for it. If you want to do 4 drives, put two 34 pin headers on the board.

The good thing about the NSC chips is that you only need a single 24MHz crystal. I probably have 100 of those in my hell box.

PCI is a PITA. Most FDCs are rigged for ISA interface, right down to the DMA needs.

If it's more a matter of getting a floppy card onto a modern floppy-less sytem, we could look into using an MCU and one of the FDCs with a USB interface. There's some driver work there, but maybe not too bad.

It all depends, as you say, on demand.
 
If that would be done, i would get one. 5.25" usb with full read and write on my win 7 pc would relieve alot of issues..

Chuck(G) said:
Tell ya what, Andrew. Do a board using one of the NSC chips--give me a bank of DIP switches (2 per drive), 4-8KB of ROM and I'll write the BIOS for it. If you want to do 4 drives, put two 34 pin headers on the board.

The good thing about the NSC chips is that you only need a single 24MHz crystal. I probably have 100 of those in my hell box.

PCI is a PITA. Most FDCs are rigged for ISA interface, right down to the DMA needs.

If it's more a matter of getting a floppy card onto a modern floppy-less sytem, we could look into using an MCU and one of the FDCs with a USB interface. There's some driver work there, but maybe not too bad.

It all depends, as you say, on demand.
Click to expand...
 
Hi,
Assuming we build the ISA board the above would be fairly simple. Basically it is a uC, Z80, or 8088 SBC core driving the ISA board circuitry. Pretty much like the SCSI to IDE/SD board is right now. However, lets approach this one step at a time and do the ISA board first. It still has some major challenges ahead.

The hard part is getting someone with leadership and management skills that can execute the project. That's the limiting factor not the technical doo-dads.

Thanks and have a nice day!

Andrew Lynch
 
Chuck(G) said:
Tell ya what, Andrew. Do a board using one of the NSC chips--give me a bank of DIP switches (2 per drive), 4-8KB of ROM and I'll write the BIOS for it. If you want to do 4 drives, put two 34 pin headers on the board.

The good thing about the NSC chips is that you only need a single 24MHz crystal. I probably have 100 of those in my hell box.

PCI is a PITA. Most FDCs are rigged for ISA interface, right down to the DMA needs.

If it's more a matter of getting a floppy card onto a modern floppy-less sytem, we could look into using an MCU and one of the FDCs with a USB interface. There's some driver work there, but maybe not too bad.

It all depends, as you say, on demand.
Click to expand...

Hi Chuck,
You are thinking the PC8477B PC-AT circuit from the datasheet right? There is not much to that circuit. Basically, a chip, a 74LS682, an oscillator, a couple of headers, some pullups, and some miscellaneous jumpers. They only discuss a couple of IO ports so we'd not even need a DIP switch for address selection. The DIP switches you mention are for the drive signals right? I just want to make sure I understand all your assumptions before I make something and we have to rework it.

I'll just rip the boot ROM circuit from XT-IDE since it is identical as far as I can tell. I am guessing with your background that you've got either a full blown FDC BIOS sitting on the shelf already or one that just needs minor tweaks. Thanks and have a nice day!

Andrew Lynch
 
lynchaj said:
Hi

The main issue with an ISA multi IO is most of the chips are SMT which will cause some hobbyists grief. It doesn't effect the design or bother me any but you can almost hear the howls of protest already if there was a big fat 100 pin QFP in the center of the PCB.
Andrew Lynch
Click to expand...

I don't know that we need to look at a full-blown multi-IO board as stand-alone 8-bit parallel and serial boards are pretty common, even modern ones with 16550 UARTs from SIIG for reasonable prices through the usual channels.

The problem that I am seeing is people are still looking for upgrade-ware that hasn't been manufactured for 20 or more years. A stand-alone 8 bit FDC card with bios support for HD floppy drives is really all they need (along with an XTIDE card) to move info in and out of a machine with an 8 bit bus. Additional features like those that the 8bit multi I/O cards of yore had would be nice, but I don't think they are really necessary and would probably add a level of complexity that some hobbyists are unable to handle.

A better choice than a PQFP would be if we could source a PLCC version of the controller chip and install a through-hole socket for it. Most folks on here probably have experience pulling a 286 in and out of its socket.

There ARE QFP through-hole sockets (I have a Zenith 386 SX with an AMP socket) but I'm not sure about the availability of them anymore.
 
dpatten said:
This seems like another thing that could be solved with a project like the XTIDE. 8 bit HD floppy controllers are pretty rare now and the eBay hoarders are asking a premium for them.
Click to expand...

I thought it was determined that you could use a cheap 16-bit ISA IDE all-in-one card and just simply disable the IDE portion. That would leave you with a ~$20 1.2/1.44 floppy controller, parallel, and serial card if you don't mind the 16-bit ISA fin hanging over the edge of the slot. (Note, I have not tried this but I'm sure I read about it either here or on another forum)
 
Hi
All of the multi-I/O chips I can recall are QFP. If there is a PLCC version of a multi-I/O chip that I've missed please let me know. Multi-I/O being a chip with 2 serial ports, a parallel port, an FDC, an IDE, and RTC devices on-board. I would love to be proven wrong though but I've spent a lot of time searching.

Yes there are QFP sockets but they tend to be highly specialized and expensive items (>$25 each) so they are impractical on hobbyist projects.

Thanks and have a nice day!

Andrew Lynch
 
Trixter said:
I thought it was determined that you could use a cheap 16-bit ISA IDE all-in-one card and just simply disable the IDE portion. That would leave you with a ~$20 1.2/1.44 floppy controller, parallel, and serial card if you don't mind the 16-bit ISA fin hanging over the edge of the slot. (Note, I have not tried this but I'm sure I read about it either here or on another forum)
Click to expand...

That didn't work with my XT clone, a Laser/Vtech machine. I tried a half dozen different 16 bit cards and none were successful. They would work just fine with a 360 or 720 drive hooked up, but there was no way that I could configure driveparm to work with a 1.44 or 1.2 drive.

I finally snagged an 8 bit Longshine card with BIOS on Ebay last year for $20, but it seems like the supply of such things has dried up as that card is currently running about $140 on ePay
 
lynchaj said:
Hi Chuck,
You are thinking the PC8477B PC-AT circuit from the datasheet right? There is not much to that circuit. Basically, a chip, a 74LS682, an oscillator, a couple of headers, some pullups, and some miscellaneous jumpers. They only discuss a couple of IO ports so we'd not even need a DIP switch for address selection. The DIP switches you mention are for the drive signals right? I just want to make sure I understand all your assumptions before I make something and we have to rework it.

I'll just rip the boot ROM circuit from XT-IDE since it is identical as far as I can tell. I am guessing with your background that you've got either a full blown FDC BIOS sitting on the shelf already or one that just needs minor tweaks. Thanks and have a nice day!
Click to expand...

You've got it. The DIP switch is indeed for drive type. Since the FDC spans the range of ports between 3F0-3F7, you could map it into one of the unused gaps, such as 3F0, 3F1, 3F3 or 3F6. You could include a jumper for changing the decoding to 37x, but I don't intend (initially) on including a BIOS that will enable support of two controllers. There just isn't a need for a separate controller if you include two headers; one for drives 0 and 1 and the other for drives 2 and 3.

Otherwise, it's a pretty straight copy from the reference design (should be on the web).

And yes, I've got code kicking around...
 
Chuck(G) said:
You've got it. The DIP switch is indeed for drive type. Since the FDC spans the range of ports between 3F0-3F7, you could map it into one of the unused gaps, such as 3F0, 3F1, 3F3 or 3F6. You could include a jumper for changing the decoding to 37x, but I don't intend (initially) on including a BIOS that will enable support of two controllers. There just isn't a need for a separate controller if you include two headers; one for drives 0 and 1 and the other for drives 2 and 3.

Otherwise, it's a pretty straight copy from the reference design (should be on the web).

And yes, I've got code kicking around...
Click to expand...

OK, I think I have it now. You want an 8 bit read only port for the DIP switch so the CPU can determine the builder's pre-set drive type, right? That adds a 74LS32, a 74LS244, and an 8 position DIP switch plus a resistor SIP for pull ups. Not a problem. I am going to use address 3F0 for the DIP switch port to simplify decoding.

My plan is to include a jumper for 37x vs 3Fx IO decoding since it is only a 3 position jumper and adds a margin of flexibility. Use a jumper or wire it shut either way works fine.

Thanks and have a nice day!

Andrew Lynch

PS, the DIP switch pins are just connected to pull up high or GND, right? Do you want MFM or IDENT tied in to the pins also?
 
Last edited:
lynchaj said:
PS, the DIP switch pins are just connected to pull up high or GND, right? Do you want MFM or IDENT tied in to the pins also?
Click to expand...

Since we'll be operating in PC-AT mode, you need only pull IDENT high. You can leave MFM NC as that pin has an internal pullup. Since we're not creating a microchannel board, PS/2 mode is irrelevant.

The CC IV has some added logic (and external buffers) for support of 8" drives; for example, being able to assert reduced write current, but there external adapters for that (e.g. FDADAP), so that's not important. There's also logic to allow for several I/O addresses, selectable DMA and IRQ, but again, that's just icing on the cake and not strictly necessary. I seem to remember that the CC IV uses 3 20-pin GALs for its random logic. I'm guessing that you don't want to get into the GAL programming business... ;)

Thanks,
Chuck
 
lynchaj said:
Hi
All of the multi-I/O chips I can recall are QFP. If there is a PLCC version of a multi-I/O chip that I've missed please let me know.

Thanks and have a nice day!

Andrew Lynch
Click to expand...

No, I'm not aware of any such thing unfortunately, I mentioned it only because I thought that it would be easier, but I guess that's been looked into before...

What is the lead count on the QFPs? I know Aries manufactures some QFP to PLCC adapters (but I don't know their price or the lead density). It would add an extra step, increase the part count and require someone who is comfortable with surface mount soldering to make up a batch, but then the hobbyist could solder the adapted part directly to a thruhole board. I'm just rambling though.
 
You must log in or register to reply here.
Back
Top