envagyok wrote on 2023-07-21, 12:26:

I buy a xtide card from internet, then i try it with ide-cf converter and cf card to my ibm ps2 xt
The BIOS of xtide show at boot, but not show the cf card.

Most CF adapters require external power, usually via a 4-pin connector from PSU. Some can be powered via pin 20, there an option for that on your XTIDE card, but I don't know if the adapter actually supports it.
Also, the jumper must be set correctly to what you have configured in your XTIDE ROM: See "jumper settings" at https://www.minuszerodegrees.net/xtide/rev_4/ … -%20general.htm

envagyok wrote on 2023-07-21, 12:26:

I think maybe wrong settings or wrong bios is the problem, so i try flash an xt version with xtidecfg software.
When i try flash, it hangs when start flashing, i wait an hour, but not start the bar, and time growing.

AFAIK the card switches must be set correctly to allow writing to the EEPROM chip. Your card seems to be set OK but was it also set this way when you attempted the programming?
Make sure you have ROM shadow in PC BIOS settings turned off - if in doubt disable them all for programming.

envagyok wrote on 2023-07-21, 12:26:

I try later put the 28c64 chip to one dlink isa network card, flashing with xtidecfg, but no result.

Won't work. Flashing must be done via XTIDE card, other cards support reading but not writing.
The good news is the XTIDE card should allow you to retry when all settings are correct, you can even program a blank EEPROM chip if you suspect this one is somehow faulty. Run the flashing program from a floppy, and not the HDD which is currently in use by XTIDE. This is important.

Now it's fixed itself.
I moved it to another xt card, the bios comes back on.
But unfortunately now it doesn't see any devices that I plug in.
Tried old HDDs, 3 cf cards with 3 ide adapters, nothing.
Tried to flash the bios again, the machine freezes the same as before.

Again, make sure your jumpers are correct. I'm pretty sure you should be using Hi-Speed mode setting, not Compatibility with 1.0.0 FW and rev1 boards, your's is (apparently modified) rev4. And that's just by looking at the photos.
Also you seem to be using a pretty old FW from 2013? Why, there has been some bug fixes since then.

Deunan wrote on 2023-07-21, 18:38:

Again, make sure your jumpers are correct. I'm pretty sure you should be using Hi-Speed mode setting, not Compatibility with 1.0.0 FW and rev1 boards, your's is (apparently modified) rev4. And that's just by looking at the photos.
Also you seem to be using a pretty old FW from 2013? Why, there has been some bug fixes since then.

I try hi-speed mode, and the latest r625 bios, and files, but no luck.

Then perhaps the issue is with the hardware. What did you buy, exactly? A kit or an assembled and tested device? Who supplied the chips?
I see you are using HC688 instead of HCT or ALS which might be a problem. The ROM seems to work but one of these 688 decodes ROM address and the other is for I/O ports for IDE. Could be one is just barely working and the other is not. At the very least try swapping them places but preferably get HCT or ALS chips as is required by ISA bus.

The second possible issue is the smaller DL series chips you have there, I'm not familiar with these but L suggest low power series and the circuit needs fast S chips here. This is a pretty important difference, the timing chain here is critical for correct IDE operation (there was even a change from rev3 to rev4 because of a small timing issue in this chain). So both '04 should be of faster series, S or ACT (HCT will probably be too slow), certainly not L.
The DS8205 is an old RFT brand 1 of 8 decoder, kinda like the '138 which the original design has, but I'm not sure about it's speed rating. S would suggest it's fast but if the replacement of the other chips doesn't help then also replace this one with HCT138.

First replace U6 (DL004D) with something much faster. Original design wants S04, but ALS or ACT should do as well. U4 should also be a faster chip but it's mostly doing delays so perhaps L is not a problem here. Still, I would buy another fast '04 for replacement because the original design used S04 here as well.
If replacing these doesn't help then replace U8 and U10 for HCT series 688. If still nothing try replacing DS8205 with HCT138. After that I'm out of ideas, you'd need to make sure the PCB is correctly assembled and start probing it with a scope or logic analyzer to figure it out.

EDIT: The DL032D should also be replaced, it's less critical but I doubt L series is fast enough. Seems like LS or HCT will do fine. This should be done along with replacing the '04 chips.

I buy it's as used from a seller from Germany.
He say it's work at him, make photo from him computer:

envagyok wrote on 2023-07-22, 18:54:

I buy it's as used from a seller from Germany.
He say it's work at him, make photo from him computer:

Ah, that would explain the old German chips. Well, his screenshots show it running on CGA, so possibly a slow 8088 system, which might be just slow enough for these chips to work.
You said you've tried a 386 and "ibm ps2 xt" which I assume is AT, not XT? Since there were no XT in PS2 series. So it's quite likely the chips are too slow now.

One more thing, this XTIDE is now set to address 0x300, which is the default setting but make sure it's not in conflict with some other card - like a sound card, or MIDI adapter. Note that if you change that, the ROM configuration must be updated as well - I take it you have now figured out how to flash it since you tried a newer version? As explained on the page I've linked above you have to set (in the configuration program) "XTIDE rev2 or modded rev1" as the device type, port 0x300 (or different if you modify the "I/O addr" switches), have the 2 jumper on top set to hi-speed setting.

If everything looks to be OK in the config then I would buy the '04 and '32 chips first (these will be cheap, S might be hard to find but ALS or ACT will work as well) and replace them. All the chips are socketed so you are not going to have any trouble with this. Make sure you try with HDDs first and only when something is detected move to CF cards. To make sure this is not CF card or adapter issue.

I try it my ibm ps2 model 30, what is intel 8088.
I try with my 386 40mhz mb, and try my turbo xt motherboard without turbo.

Ok, then i search the chips to change it.

envagyok wrote on 2023-07-23, 10:24:

I try it my ibm ps2 model 30, what is intel 8088.

Hm, I always thought that there were no 8088 in the PS/2 series. And I looked at some pages and they all say Model 30 had 8MHz 8086. The 8086 has a 16-bit data bus, unlike 8088, so system timings will be a little different one way or another.
In any case, I'm out of ideas so changing those few chips is a good way to at least rule them out. As I've said these are not expensive chips.

Yes, Model 30 is weird. I'd should never had have existed in first place. 😔
The 80286 was the very baseline level here, with the ability to run real OSes.
Those 808x toy "processors" are fine for PS/1 series, at best.
By the late 80s, IBM should have had installed an NEC V30, at the very least, to provide rudimentary 286 instruction set compatibility with the rest of the PS/2 line.
Not sure what IBM was thinking when making the Model 30.
The best thing about it is the VGA monitor, I assume.

Edit: I'm talking about the low-end Model 30 only.
There also was Model 30 286, with a 286. That one seems to be okay.
It makes sense in so far, as that it fits into the PS/2 line.

"Not sure what IBM was thinking when making the Model 30."

I was purchasing computers for our business back then. We did not end up going with the PS/2. But I think the "XT" models like the Model 30 were made available for those who did not want to invest in Micro Channel and only wanted machines to run their current software and maybe re-use some of their ISA cards. Just conjecture though.

Seaken

seaken64 wrote on 2023-07-23, 18:40:

"Not sure what IBM was thinking when making the Model 30."

I was purchasing computers for our business back then. We did not end up going with the PS/2. But I think the "XT" models like the Model 30 were made available for those who did not want to invest in Micro Channel and only wanted machines to run their current software and maybe re-use some of their ISA cards. Just conjecture though.

Seaken

Hi. 🙂 What I meant to say is that the Model 30 has no relationship to the PS/2 line.
It doesn't have the PS/2 BIOS, the PS/2 bus, nor the 80286 processor.
It doesn't belong in the PS/2 line. If it was called the PS/0 Model 30, everything had been fine.
It's like with STD ( Star Trek Discovery). It's not bad per se as a show (on its own), but it equally ruins the franchise.
Or, let's take the Battlestar Galactica remake from the 2000s.. Same story.

Edit: Also, the relevance of the V20/30 CPUs shouldn't be forgotten.
Turbo XTs had them for a reason. These processors do address decoding in hardware and have an updated instruction set.
Past circa 1985, there was no place for 8086/8088 anymore.
Just like the 8080 had been replaced by the Z80. No one in his right mind woukd have had made a new CP/M computer on basis of an 8080 processor.
If it had to be intel, an 8085 was minimum.

Same goes for x86, I think. An 80186 was an alternative to the NECs and 8086/8088.
Despite the IBM PC incompatible internal hardware, it could be operated by external support chips similar to 8086.
It also had interrupt masking/trapping abilities, like the 80286 had, so it could emulate an FPU in software (x87 TSRs).

It probably had something to do with IBM wanting to only sell "PS/2" at that moment in time, but still wanting to offer a lesser machine for their existing customers who didn't want to convert to MCA.

We ended up staying with the IBM XT until we ended up switching to the Pentium with Micron. Then eventually Dell. We left IBM behind. But we were small and really didn't need PS/2 power machines back then.

Seaken