Guru ramlib 80000004 when using Zorro cards. Hardware issue?

[jbenam]

Hi everyone,

I have this VERY long standing issue with my A4000D - I can't use Zorro cards which assign large memory spaces: for example, Mediator with PPC and Prometheus.

When using these, my A4000D gurus with 80000004 like clockwork. It's always 8000 0004. I am starting to suspect it's a hardware issue. I am testing with a 040, but I also have a 060 adapter handy.

Any hints where I should start looking at?

Few starter infos about my A4000D:
1) Motherboard recapped by me some years ago
2) A3640 recapped by me some months ago
3) PSU recapped by me around a year or so ago
4) Voltages are apparently perfect, when measuring with a multimeter
5) I have also tried with an ATX PSU and the same thing happens
6) The Voodoo on my Mediator works well and I can use Picasso and Warp3D
7) I have a socketed Buster 11. I have also tried with another Buster 11 and with the older Buster 9. Same issues
8) I have also tried without FastRAM - in that case instead of ramlib crashing, it's MasterControl
9) It's not an OS issue - I have installed, reinstalled and tried every combination possible of OS3.1 and OS3.9, along with both MMUlibs installed and not
10) Other than that, the A4000 is in splendid form and works without any issue

I think it has something to do with addressing space - AFAIR, both the Mediator and the Prometheus allocate 512MB in the Zorro III space.

This looks like a real head scratcher... Thanks for looking

[stevelord]

I sometimes get similar results. Not straight away but after a while and usually when doing large network transfers or anything involving lots of RAM. I have a ZZ9000 (256mb) and BigRAM Plus (256Mb).

[jbenam]

It seems it might have something to do with Zorro II / III mode?

Lots of people seem(ed) to have the same issues when using cards in Zorro III mode.

I had a Picasso II years ago and it worked flawlessly in Zorro II mode.

WinUAE also tells me in the board info section that the Mediator is Zorro II (is that correct?) while the Prometheus is Zorro III, which might explain why I can use the Voodoo but not the PPC?

[Hedeon]

WinUAE probably means the A1200 Mediator.

[jbenam]

Well, I have most certainly excluded power supply issues - I have tried 3 different ATX power supplies and I still get the same errors.

I have started checking the Zorro III address lines (since it happens when it allocates memory?) but it's all good with A24-A31. I thought of checking these lines first since they are used in Zorro III mode. I guess it's very-boring-multimeter time...

[jbenam]

Bump

I have reflowed the entire bottom left of the motherboard (everything was still kinda opaque by the leaky battery it had when I got this A4000) and now every solder point is nice and shiny.

I have resocketed Buster with a high quality PLCC socket. I have probed stuff randomly, checked a bunch of tracks on the motherboard going to the Zorro slot and I am still none the wiser.

I just don't know enough of the Zorro bus and how it works and what the Mediator / Prometheus is doing that makes the Amiga crash. I am sure it's an hardware issue, but what could be causing it? I would very much appreciate some pointers that could help me narrow my search.

Help please?

[trixster]

It sounds to me a little like it might be an issue with the logic chips on the mediator logic board - Wrangelr had a similar problem to you and it required rework be elbox to fix the problems. But you say you get the same problems using a Prometheus as well?

[jbenam]

Originally Posted by trixster

It sounds to me a little like it might be an issue with the logic chips on the mediator logic board - Wrangelr had a similar problem to you and it required rework be elbox to fix the problems. But you say you get the same problems using a Prometheus as well?

Yep, same exact issues with a Prometheus. BindDrivers gurus with 8000 0004 when booting the system.

I have read around that grelbfarlk fixed his motherboard replacing U701, U714 and U213.

Link: http://eab.abime.net/showpost.php?p=1281948&postcount=6

So I have done them same - programmed fresh GALs and see if something changed. Same error I didn't replace U213 since it verified fine in my programmer. The other two were PALs (which I can't verify with my TL866II) and as such were replaced by ATF16V8Cs. I think I will put back the original PALs since it's clear the new ones didn't make anything better.

I have also reflowed Bridgette but nothing changed.

I have read around that some had luck changing the CPU card altogether, but sadly I haven't got a spare one to test my motherboard with.

[grelbfarlk]

My issue wasn't anything like your issue though. I could see autoconfig cards they just didn't work. Your Zorro-3 works, (though I'd question if Z3 DMA works). When you recapped everything did you replace the power socket on your motherboard? Or the six-pin PSU connector? Or attempt to clean the mating surface of the PSU connector on the motherboard and the pins on the PSU cable connector?
Your A3640 is that rev 3.1 at least?

[jbenam]

Originally Posted by grelbfarlk

My issue wasn't anything like your issue though. I could see autoconfig cards they just didn't work. Your Zorro-3 works, (though I'd question if Z3 DMA works). When you recapped everything did you replace the power socket on your motherboard? Or the six-pin PSU connector? Or attempt to clean the mating surface of the PSU connector on the motherboard and the pins on the PSU cable connector?
Your A3640 is that rev 3.1 at least?

Z3 DMA might actually be a very good point. Does the Voodoo3 in RTG mode use DMA? Because doing RTG stuff works well, but IIRC I start having problems as soon as I use VoodooMemOS. I will test this more asap. The Voodoo3 is okay and it has been tested in a PC.

Nope, still the original power socket - but I have soldered an ATX adapter on the bottom side of the PCB (straight on the contacts) for testing and the same happened (with three ATX power supplies) so we can exclude the power connector.

A3640 was rev 3.1 from the factory and I have recently upgraded it to 3.2 along with Speedgeek's state mod. The state mod didn't impact anything as I keep getting the same gurus as I did before I changed the GALs on the A3640.

[jbenam]

Bump

Re-socketed Buster with a good quality socket, measured a bunch of stuff but nothing came up. Still the same issues.

This is literally the last thing that prevents me from closing up my A4000D and enjoying it instead of having it strewn upon my bench. I have been stumped on this thing for almost two years now...

[DarrenHD]

Originally Posted by jbenam

Bump

Re-socketed Buster with a good quality socket, measured a bunch of stuff but nothing came up. Still the same issues.

This is literally the last thing that prevents me from closing up my A4000D and enjoying it instead of having it strewn upon my bench. I have been stumped on this thing for almost two years now...

I wish I had something else to suggest to you. I have three Mediators. One of them, a MK2 A4000 version, works perfectly except it needs an extra fan to be cooled. If not, I will eventually lose the RTG display.

The other two - a MK3 A4000 one and one for the A3000 (latest version, MK3 equivalent) work perfectly fine and I have no issues with them.

It seems it is some strange problem with your A4000 board as the Prometheus generates the same error. I once had a problem with my A4000T where Zorro 3 cards were not recognized. It turned out to be a bad solder connection on one of the Bridgette chip pins. (bad from the factory). Once that was fixed it worked perfect.

[NorthWay]

I have an A4000D that wont work with Z3 either. I see the Z2 and ID stuff, but what is supposed to happen in Z3 land just doesn't.

[jbenam]

Well, a slight update on this. I have finally desoldered, tested and resoldered every F245 and F646 on the left (RAM) side of the board, along with some of the smaller F-series. They are apparently taking care of the Zorro/Memory side of things. Everything tested correctly, so I cleaned up the board and resoldered everything, hoping it was a cold solder joint somewhere.

And well, no improvements at all.

I have also tested a bucketload of tracks and they seem all fine.

Any help? This thing has literally stopped me from enjoying my A4000D for years.

[trixster]

Have you tried a different set of kickstart roms?

I had some 80000004 issues during reboots that was solved by changing to new roms.

[Jpor]

Just reading through this. You are using either a Mediator or a Prometheus both 3rd party Daughterboards? What results do you have with an original 4000 Commodore Daughterboard? Ideally you need to go back to a basic A4000 setup and work your way from there. You have also tried cleaning up connections on the Daughterboard connections? Just in case you are not having firm pin connections. Tried Ami Test kit or DiagROM to test out things on the hardware side.
Like you I’m a stickler to getting a problem solved, but you may need to to start back and try things one at a time in a logical order

[jbenam]

Originally Posted by trixster

Have you tried a different set of kickstart roms?

I had some 80000004 issues during reboots that was solved by changing to new roms.

Tried with 3.1.4 and 3.2 and I get the same issues

Originally Posted by Jpor

Just reading through this. You are using either a Mediator or a Prometheus both 3rd party Daughterboards? What results do you have with an original 4000 Commodore Daughterboard? Ideally you need to go back to a basic A4000 setup and work your way from there. You have also tried cleaning up connections on the Daughterboard connections? Just in case you are not having firm pin connections. Tried Ami Test kit or DiagROM to test out things on the hardware side.
Like you I’m a stickler to getting a problem solved, but you may need to to start back and try things one at a time in a logical order

The Prometheus goes in the original A4000D Bridgeboard, but I have the same issues with the Mediator (which can also act as a Zorro busboard, if you use it without the bridge).

I have a very barebone A4000D setup (just the motherboard and a IDE-to-SD adapter) and everything has been cleaned with the best solvents around (I don't spare expenses when it comes to lab equipment I use on rare old stuff like the A4000).

DiagROM doesn't show any issues and neither it does Amiga Test Kit.

Other than this "little" Zorro issue, the A4000D is perfect and as stable as it can get.

As an interesting side-note, I get the same issues with VoodooMemOS set to Yes. It seems like my Amiga can't access memory mapped in Zorro III space.

EDIT: Actually I can use VoodooMemOS but only if I have 8MB of RAM. With 4MB I start getting all kinds of issues (8000 0004 errors everywhere on WB loading). Weird?
EDIT: Okay, it works with 8MB because it doesn't start using the Zorro III memory until way later - it does boot like that since the "basic" boot load order will fit entirely in the MB ram, which won't exhibit the issue.
EDIT: Amiga Test Kit tests the Zorro III RAM correctly, but as soon as I try to execute something from it, I get plenty of 8000 0004 errors.

[trixster]

Can you list exactly which zorro cards cause the problem, and also which PCI cards you tried too.

If you're using a voodoo3 and a ppc pci board (or PMC with pci adapter), have you used mmu-configuration to cache-inhibit the memory ranges of the pci cards?

Do you have access to something like a ZorRam or a Bigram? If so, do those large memory expansion zorro 3 cards work ok?

[jbenam]

Originally Posted by trixster

Can you list exactly which zorro cards cause the problem, and also which PCI cards you tried too.

If you're using a voodoo3 and a ppc pci board (or PMC with pci adapter), have you used mmu-configuration to cache-inhibit the memory ranges of the pci cards?

Do you have access to something like a ZorRam or a Bigram? If so, do those large memory expansion zorro 3 cards work ok?

Both my Mediator and Prometheus exhibit the same symptoms.

I am just testing with the Voodoo 3, right now. If I just use RTG, everything is fine. If I use VoodooMemOS, and I actually try to use it as Fast (executing code from it) the ramlib error happens.

I first became aware of this issue when testing the Sonnet library with my Killer M1. It allocates RAM correctly, then probably tries to execute something from it and it fails. Still, I can reproduce the same exact issue with my Prometheus (it fails when it loads the Prometheus driver) and I *think* also with my A2090. Sadly I don't have an A2091 otherwise I could put some RAM on it and see if this issue also happens in Zorro II mode.

I have tried CacheInhibit'ing everything to no avail - I have also tried disabling CPU caches entirely just to be sure.

Sadly I don't have access to a BigRAM/ZorRAM, but I suspect I would see exactly the same issues with them.

[DarrenHD]

Have you looked at the Bridgette chip? Inspect it's connections and ensure it's ok. I had an issue with Zorro 3 and it was due to a bad solder joint on Bridgette.

Also it wouldn't hurt to swap out Fat Gary. The gary chip surprisingly has it's fingers in lots of system functions.