Ah, it says "Host adapter configuration error" - no matter if I use old 1.1 ECU files or the newest (2.12)..
The 2.11 BIOS doesn't rear its head at all.

There shouldn't be any correlation between the BIOS revision and the ECU. When I upgraded the BIOS on both of my cards from 1.x in two different systems where they were both already properly configured in the ECU, nothing changed with respect to EISA config.

If you were getting host adapter configuration error with the original ROM before the BIOS upgrade, something is wrong with your configuration.

Did you move the card to a different slot or something?

ltning wrote on 2024-01-11, 15:09:

- With original BIOS (1.1), the BIOS message shows, but complains about invalid EISA config and doesn't even try to detect devices
The board is configured using newest EISA .CFG/.OVL files. I'll try original BIOS with older .CFG/OVL to be sure, but..

This sounds like a software compatibility problem between the Adaptec BIOS, the mainboard BIOS and/or the configuration utility. I'm afraid I can't give you any hints on that one.

ltning wrote on 2024-01-11, 15:09:

- With new BIOS (2.11, unpatched), I get no BIOS messages from the board at all

On the other hand, that looks like a different issue. With the 284x controllers, we found that there are different hardware revisions that require different kind of BIOSes. The cards requiring a 1.x BIOS are called 284xVL, whereas the cards requiring a 2.x BIOS are called 284xA. It's possible that the same applies to the 2740/2742. In that case, the claim on my github page, that the 2740 (without A) is supported is wrong. I only actually tested it on an 2740W.

On ebay product images, I saw that the 2740 uses an Altera CPLD, whereas the 2742A uses an Adaptec ASIC (both cards also have the AIC-7770 main SCSI controller chip). This might change the way the BIOS has to initialize the card.

mkarcher wrote on 2024-01-13, 09:06:

On the other hand, that looks like a different issue. With the 284x controllers, we found that there are different hardware revisions that require different kind of BIOSes. The cards requiring a 1.x BIOS are called 284xVL, whereas the cards requiring a 2.x BIOS are called 284xA. It's possible that the same applies to the 2740/2742. In that case, the claim on my github page, that the 2740 (without A) is supported is wrong. I only actually tested it on an 2740W.

On ebay product images, I saw that the 2740 uses an Altera CPLD, whereas the 2742A uses an Adaptec ASIC (both cards also have the AIC-7770 main SCSI controller chip). This might change the way the BIOS has to initialize the card.

Both of my cards shipped with 1.1 and work fine with your BIOS

mkarcher wrote on 2024-01-13, 09:06:

ltning wrote on 2024-01-11, 15:09:

- With original BIOS (1.1), the BIOS message shows, but complains about invalid EISA config and doesn't even try to detect devices
The board is configured using newest EISA .CFG/.OVL files. I'll try original BIOS with older .CFG/OVL to be sure, but..

This sounds like a software compatibility problem between the Adaptec BIOS, the mainboard BIOS and/or the configuration utility. I'm afraid I can't give you any hints on that one.

Which config util should I be using? The one that I could find for the board (AIR 486ES) is just an old version of the generic AMI tool. Using newer versions of the AMI tool doesn`t change anything. That said, I have a few Compaq-branded cards (SCSI) that I'm trying to use, but those consistently give syntax errors in the .CFG files, according to the ECU. One might wonder if Compaq has deliberately made those files incompatible with ECUs not of their own making..? Any ideas?

mkarcher wrote on 2024-01-13, 09:06:

ltning wrote on 2024-01-11, 15:09:

- With new BIOS (2.11, unpatched), I get no BIOS messages from the board at all

On the other hand, that looks like a different issue. With the 284x controllers, we found that there are different hardware revisions that require different kind of BIOSes. The cards requiring a 1.x BIOS are called 284xVL, whereas the cards requiring a 2.x BIOS are called 284xA. It's possible that the same applies to the 2740/2742. In that case, the claim on my github page, that the 2740 (without A) is supported is wrong. I only actually tested it on an 2740W.

On ebay product images, I saw that the 2740 uses an Altera CPLD, whereas the 2742A uses an Adaptec ASIC (both cards also have the AIC-7770 main SCSI controller chip). This might change the way the BIOS has to initialize the card.

I think there's an Altera on mine, I'll have to double check when I'm back at that desk. Does anyone have another 1.x BIOS I could try? One thing I have not tried is to verify whether the BIOS chip even works right; I can try to flash the original BIOS on the EEPROM I'm trying to use, just to be certain.

Thanks for chiming in,
/Eirik

If you look at the CFG files for those compaq adapters you still see that are basically empty. All the meat is in the OVL file, which as you observed, only seems to work with the compaq branded ECU

ltning wrote on 2024-01-13, 15:41:

EEPROM I'm trying to use

Let me stop you right there, what EEPROM? Because there aren't many which are compatible with a 27C128

maxtherabbit wrote on 2024-01-13, 15:43:

If you look at the CFG files for those compaq adapters you still see that are basically empty. All the meat is in the OVL file, which as you observed, only seems to work with the compaq branded ECU

But can I use a compaq ECU on this board?

maxtherabbit wrote on 2024-01-13, 15:44:

ltning wrote on 2024-01-13, 15:41:

EEPROM I'm trying to use

Let me stop you right there, what EEPROM? Because there aren't many which are compatible with a 27C128

The original BIOS is a 27c256, actually ..

ltning wrote on 2024-01-13, 16:00:

maxtherabbit wrote on 2024-01-13, 15:44:

ltning wrote on 2024-01-13, 15:41:

EEPROM I'm trying to use

Let me stop you right there, what EEPROM? Because there aren't many which are compatible with a 27C128

The original BIOS is a 27c256, actually ..

I stand corrected, I was remembering (correctly) that the ROM image was only 16KB in size, but both of my cards did in fact ship with 27C256 EPROMs that just had the upper half blank.

Still, what EEPROM are you using in place of the 27C256? The only one I know of that is pin compatible is the Winbond 27E257...

ltning wrote on 2024-01-13, 16:00:

But can I use a compaq ECU on this board?

Likely not without patching it. It is perfectly correct that many Compaq OVL files only work with the Compaq ECU, but for some reason, but during startup of the Compaq ECU, it performs some low-level port I/O to trigger some Compaq specific functionality. These I/O instructions corrupt the EISA configuration on most other EISA mainboards. While there is a common way how the EISA configuration RAM (a battery-backed 8KB SRAM) is accessed on most generic EISA mainboards, Compaq chose a different implementation (an EEPROM) with a completely incompatible interface. I once did patch the Compaq ECU to remove a write to an I/O port in the 800..820 range (IIRC) so it worked on non-Compaq systems, but alas, I don't remember the details.

maxtherabbit wrote on 2024-01-13, 14:56:

Both of my cards shipped with 1.1 and work fine with your BIOS

Thanks for confirming!

Hi, I'm starting to build my 486 EISA build with this Adaptec AHA-2742-T controller. Re: Bought these (retro) hardware today
I'd like to patch it, but my board came with the v1.1 bios, so I can't do this.
I'd be appreciated if you will share a newer bios file with me.

Ok I can confirm my card also works with 2.x BIOS; it just doesn't like AT28C256. With AT29C256 it loads the BIOS just fine.

I still can get the card to actually *work*, but that seems to be due to a dad EISA config storage DALLS, which on this board is separate from the CMOS settings, it seems.

ltning wrote on 2024-01-14, 16:10:

Ok I can confirm my card also works with 2.x BIOS; it just doesn't like AT28C256. With AT29C256 it loads the BIOS just fine.

I still can get the card to actually *work*, but that seems to be due to a dad EISA config storage DALLS, which on this board is separate from the CMOS settings, it seems.

Yeah this is what I was alluding to, the 28C256 is not pin compatible with the 27C256

I've now attached a new battery to the Dallas DS1225 that seems to have been flaky; this hasn't made much of a difference. Although I'll say that I'm now able to have config changes to both the Adaptec and the SMC NIC in the machine "stick" through reboots warm and cold.
However, in a very un-scientific way, I changed more than one variable: I managed this using a (much) older ECU version; 2.01 instead of 3.40. So it's just as likely the older ECU is the reason I can now persist configs.

Not that it matters; the Adaptec still complains about invalid BIOS settings and will not install itself - no matter the BIOS version. I suppose I'll have to try on a different motherboard altogether.

maxtherabbit wrote on 2024-01-14, 17:06:

Yeah this is what I was alluding to, the 28C256 is not pin compatible with the 27C256

Now I know ;)

mkarcher wrote on 2024-01-13, 09:06:

On the other hand, that looks like a different issue. With the 284x controllers, we found that there are different hardware revisions that require different kind of BIOSes. The cards requiring a 1.x BIOS are called 284xVL, whereas the cards requiring a 2.x BIOS are called 284xA. It's possible that the same applies to the 2740/2742. In that case, the claim on my github page, that the 2740 (without A) is supported is wrong. I only actually tested it on an 2740W.

On ebay product images, I saw that the 2740 uses an Altera CPLD, whereas the 2742A uses an Adaptec ASIC (both cards also have the AIC-7770 main SCSI controller chip). This might change the way the BIOS has to initialize the card.

The 2740/42W has a different PCB with wide SCSI connections, unknown if the chips differ, but it certainly isn't the same as the 2740A.

So I gave up on the AIR 486ES board, and tested with an ECS mainboard .. And now things work as expected, across the board.
I repaired the DS1225 on the AIR board (used for EISA config storage), but despite being able to survive longer power outages, it changed nothing.
Using a much older - 2.01 - AMI ECU allowed me to persist settings that would otherwise not "stick". But despite having a complete and valid EISA config, the Adaptec BIOS complained about "Host adapter configuration error" and did not actually load the BIOS. Note that as long as the EISA config was invalid, the Adaptec BIOS didn't fire at all, so that's yet another failure scenario.

Anyway, on the ECS board, all is well.
For the record, this is a 274x with two 50pin SCSI ports and a floppy port, plus external SCSI in the back - but no wide-scsi. And it has the ALTERA chip and shipped with BIOS 1.1. Now running 2.10.

maxtherabbit wrote on 2024-01-01, 01:43:

Does anyone else have an AIC-7770 based card running in win95?

I'm getting a maximum of 5MB/s transfers in 95 on two totally different systems with different HDDs attached, in both cases I can get close to a full 10MB/s in pure DOS (speedsys) using the ASPI driver.

I have a system setup which is using the AHA-2842VL/2842A. It has an Am5x86 running at 160 MHz. I am using a Seagate 373307LW. I pulled it out for a quick test.

The SCSI BIOS splash screen indicates the BIOS is version 2.0.

From the DOS which comes with W95c (DOS 7.x)
Speedsys /HD returns 9337 KB/s (buffered read)
Coretest returns 9480 KB/s

Win95c
Coretest returns 8042 KB/s
RoadKil returns max 8.8 MB/s with a 512K block size. The breakdown was:
2K, 3.3 MB/s
8K, 4.9 MB/s
32K, 7.4 MB/s
512K, 8.8 MB/s

WinNT4
512K, 8.0 MB/s

The system is partitioned such that:
4 GB NTFS: NT4
4 GB FAT32: w95
60 GB FAT32: extra storage to be used by both w95 and nt4

I haven't been following this thread and didn't even know about it until now. So it sounds like mkarcher's BIOS fix would allow, for example, that w95 partition of mine to be 64 GB (in the above example), with the drawback being that the w95 read speeds would be reduced?

It would be nice if we had XT-IDE ROMs for SCSI VLB/EISA controllers.