this is nice :3

did you ever consider having dedicated CF slots on the card itself?
I guess with the current layout you could only pur them on the back of the card.

Thanks!

Originally, I wanted to attach only IDE hard disks to it. But when I finished the interface and connected the drives and heard the noise of 4 spinning disks... I reconsidered my approach.
So now I am using some cheap eBay CF to IDE adapters that are mounted in 3.5" bays, just as ordinary hard disks.

4 layers?

Been searching for multi-i/o of this quality especially 4 layers at least.

Cheers,

very cool design! A little sad to see you went with AT spec card height, as this will interfere with fitment in the XT 286 and other similarly speced cases

Nice job pulling back the solder resist and beveling the corners on the edge connector. So many hobby designs have horrible edge connectors

4 layers?

Been searching for multi-i/o of this quality especially 4 layers at least and inductors and capacitors of two types (small and large) next to each power pins of "noisy" large intergrated IC, using more SMD saves your money through holes for pins for ICs, (costly).

This way stops the noise getting back into motherboard and interfere with sound card.

Otherwise, focus on heaviest ground plane so there is a strong current capacity to channel noise and power back into ground with and prevents voltage bounces. Also do not allow voltage plane tracks much as copper you can spare and not allow any necks in the tracks and try to use star topography.

High frequency (above 500KHz) signals should be 45 degree turns tracks if the circuit maker can allow, use curved paths. Also if the external i/o signal to the ports rely on dedicated return path, try to have isolated paths of equal length directly to the IC. Yes this is a requirement and is often ignored results in odd problems.

Cheers,

ALEKS wrote on 2020-06-07, 18:08:

Thanks!

Originally, I wanted to attach only IDE hard disks to it. But when I finished the interface and connected the drives and heard the noise of 4 spinning disks... I reconsidered my approach.
So now I am using some cheap eBay CF to IDE adapters that are mounted in 3.5" bays, just as ordinary hard disks.

gross! the noise of spinning rust is music

Yes, it's 4 layers, all traces manually routed. Inner planes are ground and power.
I know I could've reduced costs and board real-estate with SMD parts. But I went a no cost savings way on this design as it's something I wanted to do for a long time.
Back in the early 1990s if I only had something like this...

I've heard a lot of magnetic rusty music over time. Of them all, I enjoyed (or better said, loved) Seagate ST-351A|X. I still have a couple of them somewhere. That particular stepper motor sound has a place in my heart.

But for now I'm seeking for a bit of silence.

Looking great! Will this be avaiable as a kit?

I haven't thought about making this a kit, yet.
Sourcing some of the parts involve patience and eB*y digging. Take the parallel port controller for instance. Of course there are a couple of sellers in China that offer this chip but it takes a lot of time for it to arrive from China (about 1 month or so). Also I bought a fake floppy controller chip from the Internet which worked for couple of minutes then ... nothing. I paid $ 2.5 for it. Then I bought another one from a reputable seller. That one worked OK. So this also needs to be taken into consideration.

However, I still have a couple of PCBs left. I might think about selling the two remaining bare PCBs to recover some of the project expenses.

Another impediment would be programming of the EEPLD signalling controller for the two IDE interfaces. If you don't have an EEPLD programmer then it's impossible to get the firmware inside. That would mean that I have to program it and send the pre-programmed chip along with the PCB.

And then there's the metal bracket. I used one that I recovered from an old defective controller. I searched high and low for such bracket but all I could find within a reasonable price tag, is designed to fit PCI cards.

But anyway, I'll think about it.

Make it little longer by 1 inch or two and include inductors in your design. For RFI and power planes and few low ESR capacitors near the slot connector.

Cheers,

One more thing:

Is it actual fiberglass PCB material or the horrible material that i see all the time for some time that breaks easily? Cellphone uses these as well, GPU, LGA CPUs card, newer memory modules, etc. More of light tan and no fibers embedded.

Cheers,

Where did you get the fake FDC? You worried me a little with that since I ordered a nat semi PC8477BV-1 from a chinese seller on ebay

I have found one broken lik for PLD data:
http://www.alexandrugroza.ro/microelectronics … in/FIRMWARE.JED

BTW should we have some on-top thread for (community members) homemade HW? I think it would be pitty if such effort will fall deep under other threads and one need to dig it so better would have one thread with a short description of projects and links to related threads or authors web...

>maxtherabbit
Chinese sellers on ebay/ali/etc... It's always a bit lottery. You can get a NOS part or you can get a recycled part with reprinted marking or some fake part or part that didn't passed testig and should be trashed. There's no 100% way to avoid just you can try to choose the seller by number of sold parts and his rating instead the lowest price. But those Chinese are smart, they know the rating is important and they rather send you money back than let you give bad rating as it impacts their business more than a few $. I have quite a good luck with chips from ali, e.g. I replaced SuperIO on my MB where I damaged 1 line of LPT and it's working fine. I heard about some problems with power components htat are often below spec or remarked lower raring parts but I also had a luck there when I replaced IGBT in induction cooker. I tested the transistor before use and it matched params except higher VCEstat. I had bad luck with Li-Ion battery for smartphone that had ~1/3 of rated capacity I also saw photos of cells that had a smaller cell inside and some dummy stuffing 😀

Great job! I wish I had the skills to pull something like this off. How's the performance with your card?

RayeR wrote on 2020-06-08, 00:17:

... It's always a bit lottery

I know, I was just hoping to find out if it was the same seller that his bunk one came from of not

pentiumspeed, the substrate is 190Tg (glass transition temperature) FR408-HR with a dielectric permittivity of 3.69 at 1 GHz. The solder pads finish is electroless nickel immersion gold (ENIG). I'd say these PCBs are of very high quality.
I didn't use capacitors near the ISA supply rails because my (also DIY) system backplane has a lot of them next to each ISA slot. I'm attaching a picture for references. I agree though that some additional RFC beads on the supply rails, could not hurt at all and could improve high frequency filtering.

maxtherabbit, I don't remember where I bought the fake controller. I bought the parts during a year time span and only did the design this year. I think it was fake, otherwise why would it work for only a few minutes? I tried reading a floppy disk and it partially worked. Then I tried writting a floppy and it wrote the FAT12 OK but the files themselves were semi-garbage. Then I began encountering "controller not responding" errors. There is also a possibility that my chip was ESD-damaged and not fake. Who knows... One thing is for sure, the defective controller had a thinner font on the chip marking. The other one has a thicker font.

RayeR thank you for the broken link. I fixed it right away. It should be good now (I mismatched file case-sensitivity; apparently the server doesn't allow uppercase). Indeed, it's a lottery with vintage electronics parts. But mostly I had success in buying from eB*y.

chrismeyer6, I'm pretty happy with the performance. With 74ALS245 buffers I am getting a constant ~2.2 Mb/s transfer rate. The CPU is Pentium 233 MHz and the ISA bus runs at 8.3 MHz. I'll try to overclock it to 11 MHz (PCICLK/3) to see what happens. Here are the SPEEDSYS results.

That back plane is a work of art. I'd be happy with it's perfromance as well seriously great job.

ALEKS wrote on 2020-06-07, 15:29:

I'm not sure whether this project presents any interest but I thought about sharing my work with other retrocomputing fellows out there.

I think projects like these are awesome! 😁

chrismeyer6 wrote on 2020-06-08, 02:33:

Great job! I wish I had the skills to pull something like this off. How's the performance with your card?

^This

Lovely project! It would be very relaxing to assemble one of these. 😀 However sourcing the EOLed parts will be a majore pain for sure. And the EEPLD is of course a problem, because most people won't have a programmer. Since there are so many cheap multi I/O cards still out there I guess the market for this one is rather small, even though it has lots to offer. Thank you for sharing the project!