No POST usually means you inserted it it the wrong direction.

I'm 99.99% sure it's plugged in the right way. I grab my POST test card, shove it in, and also plug in a speaker too, and I get this error code -

C7 C5.

The speaker also makes 10 horrible noises.

The mobo's BIOS is AMI, and so I look up the code above, and there's nothing there (in the POST test card booklet) for C7 C5 and AMI, so I look up the beep code info for 10 beeps and AMI, and it says -

"CMOS shutdown register read/write error"

I've tried clearing the CMOS (using the jumper method). Any suggestions to remedy this situation would be very gratefully received. 😀

Is this two beeps + 8 beeps or continuously 10 beeps?

Did you plug in the power supply for the CF card?
(Is the CF card small enough?)

The 10 beeps are all exactly the same length. These 10 beeps also have the same amount of "time space" between them. These 10 beeps all last for about 1.5 - 2.0 seconds each. After the 10th beep, there is a pause of about 15 seconds, then more of the same beeps are heard. It's at that point that I switch off, thinking that I am simply going to hear another 10 beeps again.

The power cable is plugged in to the CF IDE device. On the CF IDE device, there are 3 lights. 2 are lit, 1 is not lit. The 2 which are lit are for power and card detected. The 1 light which is not lit is for CF card activity. The lights appear to be functioning properly.

There's a very small chance that I have damaged the CF IDE device, I suppose. Perhaps I should put the CF IDE device back in to the slot 1 mobo, to retest it, just to make sure I haven't damaged it.

There's a couple of jumpers on the CF IDE device. 1 jumper is for master/slave & slave/master. I have left it on default - master/slave. The other jumper is for voltage. The options are +5v (default setting) and +3.3v. Curiously, the jumper setting is currently set to +3.3v. I tried the +5v setting, but the POST test card's LEDs seemed to indicate that there was a problem with 5v.

Only two jumpers? Usually there are THREE jumpers.

One of the jumpers chooses if power is taken through IDE bus (pin 20, is unused in most motherboards) or the external power. By default it drains power from the bus (at least in my adapter).

In my ZX Spectrum (fitted with a divIDE), connecting power through bus locks the computer (connecting the divIDE+adapter without CF boots OK).

There's definitely only jumpers J1 and J2. (Both of these jumpers have 3 pins.)

The booklet says this -

Model = CF-IDE40 V.E0
IDE pins = 40
IDE male/female = Female. (Damn, M/F would have been much more useful, then I could try using it with a IDE ribbon cable.)
Number of CF = double
VCC from pin20 = Yes
DMA support = Yes

Not sure what VCC from pin20 means, but reading your message, this may be relevant...?

I tried an Seagate 4GB Microdrive with a CF adapter and it was painfully slow (checked with SpeedSys) and resulted in many errors (after a few restarts Win98 refused to start because of corrupted files). If you succeed, maybe i'll try a real CF flash drive.

Vcc from pin 20 means that it gets +5v for the CF from pin 20 of IDE connector. In "standard" IDE connectors this pin is not used, in most female connectors it is a "key" pin (it is solid to avoid inserting in the bad direction).

My adapter (with one CF slot) has a standard floppy power connector, so you can get power from a PC power source or IDE bus.

Note that the electronic of the adapter get it's power from another place in IDE bus (not sure where), so there is a mod to get power for the CF from the electronic of the adapter (wiring from pin 15/IDE to PC power connector).

P.S.: +3.3v/+5v jumper selects CF power. Some CFs works with +3.3v, some other work better with +5v. If it matters, it is stated in the features of the CF card.

This is my IDE2CF adapter.

another nice choice for these retro machines are notebook hdds. anything made in this decade is fast enough and they are so quiet, easily hidden, etc. cheap too. the IDE ones just need a cheap adapter so you can power them and the sata ones just need the usual sata cables and a sata pci card. I had a Promise SATA150 card working fine once and it also has a IDE channel.

i also tend to dig up older ~80gig desktop drives.

anything that's recent that has fluid dynamic bearings. I refuse to hear whiny metal-on-metal ball bearing drives ever again.

I double-checked my CF IDE device was still working - on a socket 7 mobo. It works fine. So I dug out another 486 board to test it on (a G486VPA, link below), and it doesn't work.

http://stason.org/TULARC/pc/motherboards/D/DI … EV-0-REV-A.html

The BIOS HDD auto detection feature successfully sees the Transcend CF, but when it comes to booting up, the BIOS just hangs after displaying the familiar POST info table on the screen. The weird thing is that it always prints the letter j on the screen, after it displays the familiar info table. My POST test card shows FF 63, which I think is "thunderbirds are go! Get a move on! We're ready for some action!" Well, I got a bit carried away there.

Oh well. Looks like I'm gonna have to rule out using either CF IDE devices, and certainly any PCI-based IDE controller cards for use with 486 boards.

Say if a typical integrated IDE port on a 486 can only cope with a max capacity of 8gb on any HDD. What are my options? I'm happy to give overlay software a chance, if it's reasonably reliable for FAT 16 DOS usage. I wonder if I can reduce the capacity of a notebook HDD using overlay software - that would give me the option of getting a new, reliable, and most importantly quiet HDD.

Did you use the same cylinder/head/sector count as on the board you installed it? Autodetection might choose the wrong translation.

h-a-l-9000 wrote:

Did you use the same cylinder/head/sector count as on the board you installed it? Autodetection might choose the wrong translation.

I tried all 3 of the auto detection selections - LBA, Normal & Large. However, I will re-insert the CF IDE device back in to the (working) socket 7 mobo, in order to write down the successful auto detection parameters. Then, I will re-insert the CF IDE device back in to this 486 board, and choose "user defined" HDD parameters and manually enter in the HDD parameters, instead of using the auto detection feature.

I'll try and do this soon. 😀

My Promise Ultra66 happily identifies big drives. At least 120GB. Not sure if it has the 137GB limit or not. The SATA150 identifies at least 640GB (biggest I've tried on it)

I'm surprised that PCI IDE cards are as much trouble for you as they seem to be. I only have experience with one PCI 486 mobo, as MSI 4144 that uses SiS 496/497. Both the Promise Ultra66 and Promise SATA150 TX2plus work in it...

I put either card in, its BIOS eventually shows up during POST and sets up the drive, and then it boots as 80h (primary IDE) just fine..... It didn't even take much in the way of BIOS voodoo tweaking. You do not use any BIOS HDD auto detection. Might want to put the BIOS primary master HDD to one of the preset choices. You also probably do not need to use any "PCI IDE Card" options in the BIOS cuz I think this is for cards w/o a onboard BIOS.

Funny thing is that 486s are so slow that they cannot hope to get too much from PCI. They only have as much memory bandwidth for everything as PCI theoretically has and all of that is also shared with cache, FSB, PCI, ISA, etc etc. And that's before considering that 486 PCI implementations always suck. Bus master DMA basically doesn't work. They may be bridged funky with VLB (usually are). 😀

h-a-l-9000 wrote:

Did you use the same cylinder/head/sector count as on the board you installed it? Autodetection might choose the wrong translation.

I tested the CF IDE device in a socket 7 board, just to write down the auto detect HDD drive parameter information. I then tested the CF IDE device in the socket 3 board again, but I had to power off / on the mobo many times, just to see the POST information on the monitor. There's a jumper on the CF IDE device that allows you to alternate between 3.3 and 5 volts. I tried both. Something is not right. Anyway, using the socket 3 board, I selected the BIOS auto detect HDD feature, and the information it displayed was the same as the socket 7 board auto detect HDD feature. So, I think the info is right, but the strange thing is that I'm sure the HDD auto detect information was different when I tested the CF IDE device on a slot 1 board. Anyway, to get to the point, the CF IDE device did not boot up on the socket 3 board. The screen just went black. I don't think this [using a CF IDE device] is a reliable solution for a socket 3 board - certainly not the ones I have tested it out on.

@swaaye: I reckon you've got a really good board there. Perhaps that is why your PCI IDE cards work in it. I bet that a lot of other 486 PCI boards out there are inferior to yours, and wouldn't work as well with PCI IDE cards. Like, for example, my entire range of 486 PCI boards! 🤣

I've tried my CF card on a few 486 boards, no faliures so far. My CF adapter has 3 jumpers though. This might also depend on the CF card itself as the drive intelligence resides in it (CF card might follow a later ATA version that is incompatible with the old BIOS).

h-a-l-9000 wrote:

I've tried my CF card on a few 486 boards, no faliures so far. My CF adapter has 3 jumpers though. This might also depend on the CF card itself as the drive intelligence resides in it (CF card might follow a later ATA version that is incompatible with the old BIOS).

That's very interesting. I want to get a different CF adapter (so that I can use a ribbon cable with it), and also a different CF card. Currently, I have a Transcend 133x 2GB card. It's probably "too good" for use with an old 486 mobo!

Can you recommend a "crap" CF card please? Thanks a lot! 😀

Update: I think my new CF IDE device is broken. I can't get my slot 1 board to POST with it in the integrated IDE port. As soon as I remove it, the board POSTs. I put the CF IDE device back in the IDE port, and the board won't POST.

This is interesting - if I remove the PCI graphics card, and put an AGP card in to this slot 1 board, the board POSTS OK (and the CF IDE adapter works!) But if I want the PCI graphics card to be in use, I must remove the CF IDE adapter, otherwise the board won't POST. Please note, that currently the graphics card "BIOS boot sequence" is set to PCI/AGP, so there's no problem with putting in a PCI graphics card.

Has this got something to do with voltages? (Sorry, bit vague!)

Might have to do with the missing jumper for pin20/external selection.