muon wrote:

The hard disk is availabe/detected in the fdisk DOS command, but I do some partition, when the computer resets to proceed with the format of the new partitions , there is no partitions: the table of parttions is never saved ¿¿??

I have experienced such, too.
Often this can easily be solved by preformatting the drive using a more modern system. Easiest is using a Linux live CD only for creating the partitions and the DOS filesystem.

There is no way a 286/386/486 will detect the capacity of a SATA drive. This is more a capacity than an interface problem. Your best chance is to make use of the XT-IDE BIOS, can be done on a NIC but better on an actual card since your m/b doesn't have an IDE controller anyway.

konc wrote:

There is no way a 286/386/486 will detect the capacity of a SATA drive. This is more a capacity than an interface problem. Your best chance is to make use of the XT-IDE BIOS, can be done on a NIC but better on an actual card since your m/b doesn't have an IDE controller anyway.

I do not care that "some" capacity is lost. I don't like to use CF cards or other devices with NAND memory (SSD), due to their short longevity (W95, W98, WMe, MSDOS do not have TRIM). I have mounted the XTIDE BIOS in 3com card ( XTIDE ROM - 3COM 3C509B-C - W27E512-70 ) and it does not solve the problem.

retardware wrote:

muon wrote:

The hard disk is availabe/detected in the fdisk DOS command, but I do some partition, when the computer resets to proceed with the format of the new partitions , there is no partitions: the table of parttions is never saved ¿¿??

I have experienced such, too.
Often this can easily be solved by preformatting the drive using a more modern system. Easiest is using a Linux live CD only for creating the partitions and the DOS filesystem.

Could you say to me what linux distro you have used with more successful?

I did similar tests with SmartFdisk: I've create partitions in SATA disks with TB capacity ( small partitions (500 megas or less), 2GB or less) in a Pentium class machine with IDE controller integrated, but 286 machine recognizes only the hard disk: there is no partitions or any kind of struct in the hard disk.

I've tested "IBM/HITACHI Drive Feature Tool" too ( https://hddguru.com/software/2006.01.20-Hitac … e-Feature-Tool/ ). This tool allows to reduce the hard disk capacity from varios Teras to a few megas. Nothing change:

"The hard disk is availabe/detected in the fdisk DOS command, but I do some partition, when the computer resets to proceed with the format of the new partitions , there is no partitions: the table of parttions is never saved"

It's very frustring I can not use a 286/386 machines with HD and I have to load programs with floppy disk drive.

muon wrote:

I would like to use a mechanical Hard disk (no SSD) with my older motherboards (286, 386). As you know actually there is not av […]
Show full quote

I would like to use a mechanical Hard disk (no SSD) with my older motherboards (286, 386). As you know actually there
is not available mechanical IDE Hard disk. I thought in SATA HD with SATA to IDE adapter. It works properly in my Pentium / 486
motherboards with integrated IDE controllers, but if the motherboard has not IDE controller - and I have to plug a multi io+ ide ISA
adapter - doesn't works properly.

Hm.. Does the Multi I/O card work in the Pentium with the HDD attached to it (on-board stuff disabled) ?
If not, XTIDE Universal BIOS technically can also boot from the IDE connector of a sound card (Secondary IDE).

muon wrote:

I do not care that "some" capacity is lost. I don't like to use CF cards or other devices with NAND memory (SSD), due to their short longevity (W95, W98, WMe, MSDOS do not have TRIM). I have mounted the XTIDE BIOS in 3com card ( XTIDE ROM - 3COM 3C509B-C - W27E512-70 ) and it does not solve the problem.

Acknowledged. Though I have to say that I never lost a CF card to a DOS computer before.
They are special insofar as they support old IDE/ATA (and XT-IDE aka 8-bit IDE) natively and were intended to be useable by
professionals from the very beginning. While there are consumer grade cards on the market, the Compact Flash design in itself was prepared for embedded use.
All CF cards, even the oldest ones, have a simplistic form of Garbage Collection or wear-levelling built-in.
Unfortunatley, that "secret sauce" is not documented by the manufacturers, so there's no guarantee how long it lasts.
Anyway, there are industrial grade cards around that use SLC memory and are more 5v tolerant.

Also, SATA controllers not necessarily support CHS or ECHS anymore (they do almost all expect LBA).
Depending on the type of SATA-IDE Converter, that may causes trouble.
Some converters may only convert the format (serial <-> parallel), leaving the data stream the way it is.
XTIDE Universal BIOS supports LBA, as does a Pentium BIOS. A 286/386 BIOS maybe doesn't even support ATA/IDE,
just the underlaying WD100x command set from the days of ST-506/ST-412 HDDs.

Edit: If you prefer mechanical HDDs, perhaps you can find a few IBM Microdrives for cheap.
They fit in full-size CompactFlash slots and were produced in smaller capacity (2GB and below, for example(.
Perhaps they work for you. I have little experience with them, though. Not sure how long they last by comparison.

Hi, I just experimented with XTIDE on my 386 with some HDDs and SSD. The results varies widely. As IDE controller I use quite common 16-bit ISA card with Acer chip, one IDE, one FDC, COM, LPT... XTIDE BIOS is placed in 8kB EPROM at 16-bit eth. card. I disabled all disk in AMI main BIOS.

HDD Seagate Medalist 10GB - detected OK, but before boot it do 2 power off/on cycles like drive reset? Why? Also when write data from some pmode dos program it starts power cycling. It doesn't happen with AMI main BIOS while can use only part of capacity.

HDD Seagate Barracuda 7200.7 40GB - detectes OK, no power cycling bu I hear heads move like after power on. Writting ok.

SSD Transcend IDE44-pin 32GB - not detected - empty device name string, it doesn' work neither with AMI main BIOS.

SSD Samsung mSATA in IDE44 mSATA adapter 32GB - even worse, cause freeze at BIOS POST screen while some characters on screen are missing or messed up. Looks like corruption on system bus when more devices talks at one time?

CF card Transcend 16GB in homemade CF-IDE adapter - detected OK, works as expected.

Note to SSDs - I tested them in Pentium Pro MB with PIIX3 and Samsung mSATA works fine while Transcend freeze at IDE detection.

Any idea what is the root cause of SSDs incompatability? I think that all devices have to support very basic ATA identify command in PIO mode so ut should at least show device name during detection. Or did ATA spec deprecated PIO mode at all on latest devices?

RayeR wrote on 2022-11-29, 01:28:

IDE controller I use quite common 16-bit ISA card with Acer chip, one IDE, one FDC, COM, LPT
...failures...
even worse, cause freeze at BIOS POST screen while some characters on screen are missing or messed up. Looks like corruption on system bus when more devices talks at one time?

picture of the controller? Re: Best ISA super i/o chipset? tldr is it one of those barebones with no 245 buffering of data lines?

RayeR wrote on 2022-11-29, 01:28:

Any idea what is the root cause of SSDs incompatability? I think that all devices have to support very basic ATA identify command in PIO mode so ut should at least show device name during detection. Or did ATA spec deprecated PIO mode at all on latest devices?

no idea, but would be great for someone (other than me:P) to investigate with logic analyzer
inspiration https://www.shackspace.de/2011/04/27/open-ses … -logic-sniffer/

It's quite a common controller, any low-cost noodle like there. Chipset is Acer M5105 + some buffers and PALs. I'll take some better photo later...

no buffering on upper 8 bits https://retronn.de/imports/hwgal/hw_acer_multiio.html

But is it real problem on cca 30cm cable with a single drive? There are buffers on MB-ISA and IDE sides...

It cant be helping.
Your IDE controller on this card is 2 chips, Acer rebranded PAL and one 245 buffer - exactly that kind of thin low cost noodle 😀 All of the rest is logically separate with Acer M5105 providing 2 Serials, LPT, FDC, and Gameport.
Here is similar one, except with two buffers https://www.vogonswiki.com/index.php/Elitegroup_C190
Here is one with provision for one, but manufacturer was too cheap for it https://retrocmp.de/ctrl/testfdc/acer-m5105_02.jpg

Good news is your card also has provision for second 245 buffer, here a populated one

Im going by pcb substrate and soldermask color. Its possible your card has this buffer and Im just barking the wrong tree.

I found another spare ISA IDE controller. There's a difference that Samsung mSATA SSD doesn't hang the POST but it's not detected anyway, like the Transcend IDE. So it might be some issue at electrical level but there's another problem.

I also tried to put SSDs in one old 486 machine with similar ISA controller but it has HDD autodetection in AMI BIOS. Both SSDs was some way detected with low cylinder number (BIOS doesn't support 16k cyls, Heads=15, Sectors=63, size cca 120MB. The H, S values matches the datasheet of Transcend SSD, see https://docs.rs-online.com/1787/0900766b816b6ff8.pdf
There's noted CHS: 16383,15, 63. When XTIDE is put in it doesn't detect any SSD. In both cases with/without XTIDE it doesn't boot from SSD, it behaves like not connected and trying to boot floppy so neither MBR was loaded from SSD. Also there's a note in datasheet that the SSD supports all PIO0-4 modes and there's ATA command table where it supports device identify in PIO mode so I don't see problem here. BTW the Pentium Pro with intel PIIX3, where at least Samsung mSATA SSD works, supports only up to MWDMA 16,6MB/s not even UDMA33. I still don't understand how would go ATA identify differently on platter HDD, CF and SSD, it should be the same. Definitely need to see waveforms what's going on, maybe some timing issue...

rasz_pl wrote on 2022-11-30, 13:05:

Good news is your card also has provision for second 245 buffer, here a populated one

Oh, you're right. Mine has 33R resistors there. So I will suck off the solder and put a buffer there if it changes something...
Here's better photo before the mod.

I digged another 386 MB. It's highly inregrated
with VGA ET4000 and onboard IDE, FDC, a ports bu only 386SX 20MHz. It has some VLSI chipset. I don't have a manual, theres no visible partnumber, just BIOS prints a string NetMate. Onboard BIOS support HDD autodetection and it detected that 10GB seagate as cca 1.8GB with partially usable primary partition. When I pushed XTIDE BIOS there it works via LBA with full 10GB. But it still makes 2 power off/on cycles during detection.

Then I tried both SSDs there and surprisingly the Transcend IDE works! So it proved to support necessary PIO modes and CHS commands. It works with XTIDE (full capacity) and onboard BIOS too (limited capacity).

The Samsung mSATA SSD in IDE adaptor is detected by onboard BIOS too with limited capacity but it dispalays Error init drive error then and no boot. Contrary the XTIDE doidn't detect it at all. It seem like this mSATA doesn't know CHS but only LBA. The Device Identify command doesn't needs params so it works always and so onboard BIOS show at least some capacity but fails to read sectors later. Question remains why XTIDE cannot handle it when it supports LBA. I didn't go through sources but maybe it tries communicate via CHS first and give up when failed? Does it have some debug mode where could I see what ATA commands are send to drive?

Michal Necasek did a deep dive into ATA standard. He has a series of articles on IDE, starting with https://www.os2museum.com/wp/historical-ata-standard-drafts/ . He did look into history of IDENTIFY DRIVE command itself https://www.os2museum.com/wp/whence-identify-drive/

Hm, according to his name seems to be a man from my country but I didn't know him. I also have a bunch of PDF with ATA specs and I readed some parts as I programmed my SMB tool that can also send ATA commands to drive. I don't remember there would be something weird about identify cmd. I still have unanswered question about ATA TRIM command that would need consult with some expert but that's a different story...

Im making a 386SX machine, and today i connected a 80GB SATA HDD without trouble. I use EZdisk

You had a good luck with your SATA-IDE adaptor.

An alternative, scsi.

I was able to get 128gb working in windows 95 osr2 using a adaptec isa controller and a 15k rpm scsi 80 pin drive on adapter on my 486 class system running a POD

Here is another good mechanical option that should work as noted before. I use this all the time.

Microdrive

It runs windows ME on k6 class systems with a some slowness