Have you tried the card in a different, newer, machine?

lvader wrote:

it comes up with no drives detected, BIOS not installed.

This sounds like the Promise card's detection routine is running and just not detecting drives. Is that the case? If the Promise routine is coming up the problem is probably not the bus.

I've had Promise SATA150 TX2 and Promise Ultra66 cards working on a 486.

The card is also detected by Windows 95, I guess I'll load the 95 drivers and see if anything changes.

Can you post a pic of your card? Some variants did not have BIOS chips.

Did you disable your motherboard's onboard IDE ports? Perhaps there is some conflict in this regard. Try a different HDD and ensure it is jumper set to Master. I've noticed on some Promise Ultra100 cards that my HDD only worked on the middle position of the ATA100 cable. Try that. I'm not sure if this was a motherboard or Promise problem, but moving the HDD to the middle (or was it the end?) fixed the problem. I think it was the middle. This issue is evident on my 430TX board. Also, there is a max HDD size that the Promise will work with - probably less than 250 GB should be OK?

I can confirm that at least the Promise SATA TX2plus works on my 486 board.

Alternately, has anyone had success in getting DMA transfers from a SiS496/497 or Umc8881/8886 chipset? Some PCI sniffing program I was looking through claims the UMC chipset supports DMA for IDE.

I tried pretty much everything I can thing of. Windows 95 sees it but when I load the drivers it comes up with device disabled (code 22).

I've tried loads of motherboard BIOS combinations. Different drives, different cables, moved PCI slots, removed other cards.

You may need a Win95-specific driver. For troubleshooting purposes, you might want to try Win98SE and NT4. When installing NT4, be sure to hit F6 at the blue screen to install your Promise drivers.

lvader wrote:

I looked at the Promise web site and the thing that stands out is the minimum requirement is PCI 2.1 and my motherboard is 2.0.

For conventional PCI slots PCI 2.1 only added optional support for 66MHz PCI operation. (And some things for PCI-X slots which are n/a to this system.)
The card working in a PCI 2.0 slot is not and issue. I know it works. It will just run at 33MHz vs the optional 66MHz.
IIRC the backwards compatibility to PCI 2.0 is required by the PCI 2.1 spec so that should be true for any card, not just promise.
.
Assuming the card is not bad I'd guess the problem is related to IRQ assignments.
The card should take *one* IRQ for all drives on it. (Not an IRQ for each drive.)
The motherboard may be trying to use that IRQ for some other device thereby screwing things up.
.
And I agree with what another said - that the problem might be the master-slave jumpers on the drive. (Or which connector on the cable.)
I've had some models of drives that only worked on promise cards if set them to 'master' and others that only worked if I set them to 'cable select'.
.
Here this is if you haven't found it yet.
http://www.promise.com/media_bank/Download%20 … 0_Manual_En.pdf

Thought I'd resurrect this thread so I could add my experience.

I have a Promise Ultra 100 TX2 that I recently upgraded to the newest bios from the Promise website. Before I upgraded the bios on the controller, I had a very old firmware version. Had the same behavior with the old firmware too, which is why I upgraded it.

I have several socket 7 boards. When I put the Promise Ultra 100 TX2 in a Socket 7 board, it detects drives on bootup and installs its bios and works OK in Win98se, even with the onboard IDE enabled. Works with a quantum fireball ex, SD2IDE drive, CF adapter, and a hyperdisk DOM.

I have two Socket3 486 boards with PCI slots , FIC VIP (old Via chipset, no integrated IDE controller) & Gateway BAT4IP3 (i420EX chipset, integrated IDE controller). Both are using AM5x86 chips, no overclock. One has WB enabled, the other is WT.

When I put the Promise controller in either of the socket 3 boards, it does not detect any of the drives at boot, does not install its bios, and does not show any drives in Win98se. Same drives & cables & power supply as I used in the Socket 7 boards. Only trying one drive at a time, jumpered as master. Same behavior with another IDE controller enabled or not. Completely vanilla Win98se, no service packs, install using the same automated install script as the Socket 7 boards, using the same out of box VGA driver and the latest Promise IDE driver from the Promise Website.

My guess is the root issue happens during boot up when the Promise controller fails to detect the drives.

What part of the 486 systems could be affecting the Promise Ultra 100 TX2 drive detection?

I guess I could try an older 486 CPU instead of the AM5x86 chips, but I hate to mess with all of the voltage stuff again.

douglar wrote on 2020-03-02, 16:46:

What part of the 486 systems could be affecting the Promise Ultra 100 TX2 drive detection?

Maybe the BIOS of the card is compiled for Pentium and uses opcodes the 486 simply does not understand, causing the drive detection to fail. Could be tested easily by installing a Pentium Overdrive, or maybe a Cyrix 5x86, which should support the whole P5 instruction set as well.

I’ve noticed that these promise cards work on some 486 systems and not others. Perhaps it is a PCI. Compliance issue. Even some socket 4 pentiums will exhibit your symptoms and not others.

Chiming in to echo similar experiences.

I have an ultra66, fasttrak 100 and ultra133. All three exhibit varying issues on my 486 similar to what Douglar described. In my P90, the ultra66 works marvelously with older drives (up to a 60gb WD HDD) and optical drives in W98SE, however it does not see my CF IDE adapter. The fasttrak sees no devices, and the ultra133 has various issues, intermittently not seeing drives, and it does not like soft reboots. I am satisfied with the ultra66 and have an ultra100 coming to see if it fares as well.

I installed an Ultra100 TX2 and it is as well behaved as the Ultra66 (and gave me ~3% improvement in Winstone 94 which is my primary measurement) whereas the Ultra133 TX2 remains glitchy. At least for a Pentium I suspect that will be the sweet spot.

Is there an easy way to find BIOS revision history for these controllers?

I have had problems with this controller too.
What I have found out is that it makes trouble when the IRQ is not assigned properly (by mainboard BIOS).
You may use SpeedSys 4.78 to find out the IRQ assignment.
Perhaps it will work in another PCI slot.

However, I'm running it in a 486 motherboard with UMC chipset.

The Ultra100 TX2 works fine in my UMC 8881-based MB-8433UUD system, however I cannot soft reset.

derSammler wrote on 2020-03-02, 16:52:

Maybe the BIOS of the card is compiled for Pentium and uses opcodes the 486 simply does not understand, causing the drive detection to fail. Could be tested easily by installing a Pentium Overdrive, or maybe a Cyrix 5x86, which should support the whole P5 instruction set as well.

I tried a Pentium Overdrive CPU and it didn't make a difference. Still no go.

I have to assume that the issue could be with the old BIOS, but it's probably with the PCI implementation on the 420ex chipset & VIA 82C505 chipsets. (Both are PCI 2.0)

Most of the non-working cards don't even show up on the Plug-N-Play identification or if I search for non-plug-n-play hardware in windows.
The Maxtor 150 is a little different. It shows up on the list of PCI cards during the plug and play arbitration phase and accepts drivers under windows 98se, but windows locks up if I try to access the drive.

1Controller / Adaptec(SII) Promise 66   Promise 100   Maxtor 150   VT6410
2Motherboard  ---------------------------------------------------------------  
3  Via82C505     No          No          No            No            Yes
4  420EX         No          No          No            No            Yes
5  430TX         No          Yes         Yes           Yes           Yes
6  440BX         Yes         Yes         Yes           Yes           Yes

I'll update if I get more cards or newer BIOS

In my experience, for some boards that do not work with the Promise Ultra66-133, may work with the SATA150 TX2plus's IDE port.

feipoa wrote on 2020-03-12, 09:57:

The Ultra100 TX2 works fine in my UMC 8881-based MB-8433UUD system, however I cannot soft reset.

I have the same soft reset issue and would love to know what it takes to fix it!

I am using a Promise SATA150 TX2plus and TX4 (4 SATA 150 ports) and both work on a Asus 486 AP4 Motherboard (Intel Aries 420EX chipset). Connected with a 256GB SSD, boots to DOS using UDMA5 mode using it's own bios extension . Benchmarks give 63MB/s in DOS with no drivers loaded. Very good for 486 PCI throughput.