A4000T won't boot with A4000T ROMs, but boots with A4000D ROMs

[8bitbubsy]

I got an Escom A4000T recently and I did a full recap of it, and everything looks perfect, it has no cap leakage left (just a minor leak on one via near a Zorro slot). It runs an A3640 040 CPU card, also recapped and known to be working good. Stupid as I am I never tested it before the recap, but many caps had started to leak, and I got told it was unstable.

The problem is, if I boot the system with the stock A4000T 3.1 ROMs, it only gets to a gray screen and stay there forever. It passes the tests then just freezes. I've tried to wait for 20 minutes, nothing happens. I've checked the continuity from the SCSI/floppy board to the motherboard pins, and they all have a good connection... Also looked all over the board for bad solder connections, but there are none! Looked for shorts from my soldering too, can't see any and no chips are overheating.
If I put in a pair of A4000D ROMs it boots right up from the IDE connector and everything works nicely. I also tried the A4000T ROMs in my A4000D and the system boots right up after a while, so I know it's not the ROMs... a bad ROM checksum would lead to a red screen anyways.

Any ideas? Everything is correctly connected, the dip switches on the SCSI board are default, and so on. I've tried to boot with and without an SCSI drive, with and without the SCSI termination, with and without an IDE drive, with and without a floppy drive. Also resoldered some chips on the SCSI board, and some chips related to the SCSI circuitry on the A4000T mobo. I suspect something is busted in the SCSI circuitry, but I don't know...
Oh, also worth mentioning is that if I connect an IDE device w/ the A4000T ROMs, it takes four seconds then the machine crashes after two power LED blinks, then reboots and gets stuck in the grey screen. Weird, because I know the IDE stuff works fine with the A4000D ROMs. Chip RAM and Fast RAM was tested with MBRTest-2, no problems there.

[SpeedGeek]

I think you have correctly identified the SCSI circuit as the source of your problems. Check the solder connections and PCB tracks near the NCR SCSI chip as well as P651 and P652.

You might want to check the SCSI components of the disk module as well but if you still can't find the problem, it's probably easier to replace P651 & P652 than the NCR SCSI chip as the next step (also you could upgrade the original versions to the PIO2 versions).

Good luck!

[8bitbubsy]

First off, P651 and P652 work fine because the machine boots from IDE with the A4000D ROMs. I traced _HD_STAT and _PCB_SCSI (from P651) to verify that they indeed go into U602 (SCSI ID logic chip) and U600 (SCSI controller).

Secondly, the only crucial component for operation on the SCSI module is _+5V_SCSI, which I have verified is the correct ~4.3v (goes through a fuse and a diode, so it loses around 0.7v). I checked this voltage on the A4091 SCSI card too, and it's indeed ~4.3v, for the SCSI terminator stuff.

The logic chips on the SCSI module are for the floppy drive connector.

Man, this is getting frustrating! I also tried to swap U610 and U611 (DS2107S) from an A4091, but the problem still persist.
I should try to look at the resistors and stuff connected to U600, before eventually trying to desolder the whole component (gee...). Could be a dead U602, but not sure ...

[dannyp1]

I bought a 4000T maybe 13 years ago that had what sounds like close to the same problem you have. It booted with the old roms that came with it but when I bought and installed 4000T roms it wouldn't boot. I sent the board away to an expert and the problem was the motherboards scsi chip. The person working on it obtained the last new chip that Software Hut had and it fixed the problem. I'm thinking the computer came to me with 3.0 roms and said that in the EBay posting. That should have tipped me off that something was up but I didn't realize at the time that 4000T's never came with those roms.

[8bitbubsy]

Yeah, it's most likely the SCSI controller that is bad... It doesn't get hot, which is common when a chip has failed, but chips can fail in so many ways...

[8bitbubsy]

So I desoldered the main SCSI chip (U600), and now the system boots with the A4000T original kickstarts, so the chip was indeed defective. I use an A4091 instead, which is basically the same controller.
The A4000T is now up and running, had to replace the old leaking caps, used high-quality Panasonic HD SMD ones (5000hrs @ 105C) and Nichicon PW(M) low ESR series for the through-holes and the PSU.

[amigasith]

Aah - very nice I love those repair stories

[kneehighspy]

me too, enjoy people digging in and getting their feet wet to solve their problems. so the 4000T came with onboard scsi along with onboard ide and the 4000D only had onboard ide?

[Jope]

Yes.

[jbenam]

Damn, I'm jealous now - I need to source an A4000T

[Locutus]

Better have a big wallet.

[8bitbubsy]

I got the A4000T for free (for shipping costs) because the caps had leaked, but it lacked the front plastic panel, so I had to get that from eBay for a rather insane €63.. The A4000T also came with no CPU card.
And then I had to pay for replacement caps, and VAT/taxes on the machine and the panel, but all in all it was still worth it! Waiting for some 5.25" rails and a 5.25" floppy drive bay in the mail so that I can fit the DVD-ROM and another Chinon FZ-357A that I had laying, as DF1.

[jbenam]

Originally Posted by 8bitbubsy

I got the A4000T for free (for shipping costs)

Weeeell congrats on making me EVEN more jealous!

Good catch, man! Enjoy it, it's a beautiful machine

[Davka]

These restored "free" Amigas are the best

[A3000T]

Do you have MBRTest-2 software?
My disk went bad and I need the files.
a3000t@yahoo.com

[Kin Hell]

Iirc....Un-DMS this file for MBRTest-2 software.

http://amiga.resource.cx/install/M1230XA-101.dms

From here:

http://amiga.resource.cx/exp/mbx1230