Mine worked in an 815EP based board with 1400 Tualatin S .

I can't get my Asus P3B-F to finish post with that card installed .

See also Re: Corruption issue when using rloew's TRIM.EXE (TRIM utility for DOS) with FreeDOS FDISK 1.2.1/1.3.1 partitioned DISK

I used to have one of those and it worked with no issues in a couple of different Socket 7 motherboards; IIRC, I used it in a TXP4, a K6BV3+, and a P55T2P4 with K6-2+ and K6-3+ CPUs.

This controller was definitely still unstable after recapping. Before reflowing any of the parts on the board I washed it in the sink to clean away any oil and debris and to prepare the surface for soldering.

I tested it again after washing without having done any rework and now it works fine. I've been testing it a lot, lots of reboots, reinsertions, swapping slots, and installing Windows XP. The system is completely stable.

This isn't the first time that washing a part has fixed it for me. I once washed a dead motherboard and that's all it took to bring it back to life.

Maybe there was some tiny debris shorting one of the pins. It looked fine at a distance but it was super gross under a microscope.

Kahenraz wrote on 2021-10-02, 05:44:

Maybe there was some tiny debris shorting one of the pins.

This happens more often than we think, and sometimes it's on the inside of chip pins and difficult to spot.
Especially if the board in question is in your workspace, soldering and cleaning near it.

I wonder if this particular fault was due to a whiskers. A gentle scrub would be enough to knock them loose.

Could also be dried electrolyte that had leaked from those SMD caps doing something

Why did you make 3 different threads about this?

Different threads for different topics. One is repair, one is about speed sensitivity, and the other is compatibility with CompactFlash cards. It helps when filtering if someone is using with specific keywords and filtering by searching only by thread titles.

Except for the fact that the card is not speed sensitive, so you've potentially created a red herring. And the CF problem is likely caused by virtue of your adapter. 80 conductor detect is a simple pin sense, the real 80c cables have a missing conductor that your CF adapter probably has grounded

Kahenraz wrote on 2021-10-02, 05:44:

This controller was definitely still unstable after recapping. Before reflowing any of the parts on the board I washed it in the sink to clean away any oil and debris and to prepare the surface for soldering.
...
I tested it again after washing without having done any rework and now it works fine. I've been testing it a lot, lots of reboots, reinsertions, swapping slots, and installing Windows XP. The system is completely stable.

You're saying you got this working on a BX board? If so, which model please?

I'm going to re-cap (With Panasonic FT or FP SMDs - forget which one I have)... But first I'll remove the caps and give it a good wash with soap and a toothbrush and then blow off all the remaining water. We'll see if that helps.

I am not having a perfect experience with the Marvell adapter and BX southbridge combo...

I tested this on my ASUS P2B 440BX.

I use a mild dish soap and water with a brush. It's possible to knock off SMD components if you use too much pressure so scrub gently. I use a can of compressed air at the end not so much dry it buy to make sure that there is no soap stuck underneath any of the chips.

See here for my repair thread:

Promise UltraATA133 TX2 (PDC20269) repair

maxtherabbit wrote on 2021-10-03, 15:34:

80 conductor detect is a simple pin sense, the real 80c cables have a missing conductor that your CF adapter probably has grounded

It's the other way around. The pin is grounded on the 80 conductor cable.

Sorry, all I could remember for sure was it was a ground sense but forgot which way it went

Kahenraz wrote on 2021-10-07, 19:27:

I tested this on my ASUS P2B 440BX.
<snip>

Eureka, you are a genius sir. I was very encouraged by this, because my board is a P2B-S.

I removed one of the capacitors with an iron thus preserving its legs... you're right, I don't either like removing SMT radials this way, because it almost always lifts the pad, but I needed at least one intact for testing. It tests over 10 Ohms, so it's obviously no good, and it stands to reason that the rest would also test similarly, although I twisted them off so I can't test them.

I then washed the card and scrubbed it with a toothbrush and orange cleaner... I like to use the liquid orange cleaners on PCBs (often works wonders with RAM sticks)... Then I used my blower device to blow out and dry every last water molecule, especially from under the QFP.

One last thing, I was using a 40-conductor cable and I'm certian that that was also causing problems, as the Ultra133 BIOS was detecting the drive as UDMA-6. I switched to an 80-conductor cable.

The drive is an Intel SSD530 120GB with a Marvell SATA to IDE bridge. I also ordered a Sunplus adapter today and I might test that as well.

Then I proceeded to successfully install Windows 98, and everything seems to be working fine.

So thanks to you I now have an ideal setup. UDMA-6, Trim support, and now all that's left to be done is fiddle with drive alignment... I might tinker with that too. I don't think DOS likes partition offsets, but Rob Loew once posted at MSFN that you can set the FAT32 offset independent of the partition offset, so I might play a little with his RFORMAT utility. For now I just used Super Fdisk.

If anyone needs a direct replacement, see attached photo for the part number.

That's very good work. Thank you for the additional information. These cards are very good and I would hate to see then destroyed or thrown away. Once repaired they are excellent.

Although I didn't photograph it the very first capacitor I tried to remove slightly damaged the pad but did not break the trace. Every other capacitor I twisted off without issue and it was then very easy to remove them. It's good to see that you were able to verify that they had failed. All of the ones I removed tested bad as well but this was not reliable due to the damage upon removal.

The UDMA issue appears to be some kind of incompatibility with these IDE controllers. For example, I tried all of my CompactFlash cards on a Promise SATA150 TX2 Plus with a PDC20371 and it would only recognize as UDMA2. Whether any of these controllers work with a CompactFlash card at their rated UDMA speed is not guaranteed.

I did verify that the Promise SATA150 TX2 Plus does support UDMA6 with a real IDE hard drive so this is specifically an incompatibility with CompacfFlash.

To summarize, the Promise UltraATA133 TX2 (PDC20269) does support UDMA6 with some CompactFlash cards but fails with others. It does default to UDMA2 so at least it's not PIO speeds.

Promise stuff is dodgy to most of it belongs in the bin. Drivers as well youre lucky if you have a prpmise card that works on a board with a driver that isnt terrible ive had bad experiences more than omce with different setups also good experience

What would a good and reliable IDE controller be in your opinion?

Warlord wrote on 2021-10-08, 05:10:

Promise stuff is dodgy to most of it belongs in the bin. Drivers as well youre lucky if you have a prpmise card that works on a board with a driver that isnt terrible ive had bad experiences more than omce with different setups also good experience

meh all of mine have been reliable

Kahenraz wrote on 2021-10-08, 02:49:

That's very good work.

Thank you, you did a good job yourself... Which Nichicon series did you use there?

Although I didn't photograph it the very first capacitor I tried to remove slightly damaged the pad but did not break the trace.

Same here... The one on the left... But that's the one that I desoldered.

Warlord wrote on 2021-10-08, 05:10:

Promise stuff is dodgy to most of it belongs in the bin. Drivers as well youre lucky if you have a prpmise card that works on a board with a driver that isnt terrible ive had bad experiences more than omce with different setups also good experience

Their early stuff was not great, but I think the TX series were a collaboration with Marvell which was top notch. And this is the only card that will pass TRIM commands. I tried them all. ITE, SIL, VIA, ALI, you name it. So it's Intel and Marvell (but perhaps not anything older than ICH or ICH2 -- I had trouble with PIIX4E).... If ALI doesn't work, then chances are ULI (read ATI/AMD) probably doesn't either, though I haven't tried it on any SB4xx to SB9xx southbridge IDE port. Not sure about nVidia... But these systems are too new anyhow, the point is to have a system with ISA slots but also with a modern implementation of drive I/O.

It's also worth mentioning that the aforementioned Sunplus Sata to IDE adapter arrived today. It does not work with TRIM. The command runs, but it just hangs there and does nothing.

So Promise TX2 + Marvell bridge is the way to go.

mockingbird wrote on 2021-10-08, 20:01:

Thank you, you did a good job yourself... Which Nichicon series did you use there? […]
Show full quote

Kahenraz wrote on 2021-10-08, 02:49:

That's very good work.

Thank you, you did a good job yourself... Which Nichicon series did you use there?

Although I didn't photograph it the very first capacitor I tried to remove slightly damaged the pad but did not break the trace.

Same here... The one on the left... But that's the one that I desoldered.

Warlord wrote on 2021-10-08, 05:10:

Promise stuff is dodgy to most of it belongs in the bin. Drivers as well youre lucky if you have a prpmise card that works on a board with a driver that isnt terrible ive had bad experiences more than omce with different setups also good experience

Their early stuff was not great, but I think the TX series were a collaboration with Marvell which was top notch. And this is the only card that will pass TRIM commands. I tried them all. ITE, SIL, VIA, ALI, you name it. So it's Intel and Marvell (but perhaps not anything older than ICH or ICH2 -- I had trouble with PIIX4E).... If ALI doesn't work, then chances are ULI (read ATI/AMD) probably doesn't either, though I haven't tried it on any SB4xx to SB9xx southbridge IDE port. Not sure about nVidia... But these systems are too new anyhow, the point is to have a system with ISA slots but also with a modern implementation of drive I/O.

It's also worth mentioning that the aforementioned Sunplus Sata to IDE adapter arrived today. It does not work with TRIM. The command runs, but it just hangs there and does nothing.

So Promise TX2 + Marvell bridge is the way to go.

That mostly lines up with my own experience . JMicron IDE to SATA converters also pass TRIM (Marvell converters are compatible with more IDE controllers, though).

I'm about ready to give up on running this card with BX (Asus P2B-S)... Windows 98se installs, but I got a hard lock in Windows after a reboot post INF and Promise driver installation.

*Maybe* this card doesn't like Asus' early ACPI implementation, but I doubt that...

Looks like I'm going to switch over to my Tualatin killer platform a.k.a Abit KT7A + Athlon XP-M.

I lose two ISA slots but still retain one. But it's worth being able to use an MLC SSD and Trim.

Time to re-cap the thing and fiddle with it.