PDA

View Full Version : A4000D weird IDE/hard drive problems need help asap



-Acid-
25th December 2014, 22:59
Hardware in use (it's an A4000D in a Power Tower)

A4000D rev B main board (recapped by Hikey a few weeks ago)
Power Tower conversion has been made
Cyberstorm Mk1 060 with 128mb ram
2Mb chip plus 16mb main board ram
3.1 roms
Buddha Flash
Buster 11 upgrade recently

Now to the problems.

First thing is it takes 30-35 seconds to get to the kickstart screen, it was like this when I put it all back together when I got the board back but it wasn't doing this before i had it recapped. At first I thought it was because I didn't connect the floppy drive but it's the same with it conencted.

At first the system was not recognising the 128mb fast mem on the processor but today when I switched it on the kickstart screen popped up in about 3 seconds flat and loading workbench from floppy showed all 144mb of fast ram.

At this point I plugged in my HxC and had a quick shot of Turrican 2 to make sure it worked, I turned off and disconnected the HxC and put the floppy drive back on and it was still booting fine in 3 seconds as it should.

Now the black magic starts up again. I took my big hard drive out of my A1200 to put in the A4000, it should work without modification as it is set up using CWB for 060, SFS formatted, IDEFix97 it is perfectly working when I put it back in my A1200 which has the same 3.1 roms, 060 lots of fast.

Yesterday when I first put the hard drive in the Amiga would not see it. 20 boots and the first two partitions were seen in the early boot menu but it would not boot from DH0. Again I got it to see the drive so in HDToolBox I had a look and it sees the partitions with the correct geometry. But anyway I saved changes to the drive and immediately three partitions showed up on the WB screen. I could quickly browse them to make sure my files were still there,DH0 was fine (it's the first 2Gb on the drive) but after DH1 bombed instantly (it's 60Gb roughly) as expected as IDEFix isn't loaded, on reboot the Amiga could not see the drive again. I checked the drive in my A1200 it's still fine all partitions and files are all there.

At this point I was getting annoyed so turned it off and went to bed. Then today I decided to have another look at it, I removed the hard drives and booted it for the first time. Kickstart booted in 3 seconds flat again looking for a floppy, so I powered it down connected the HxC and still booted normally.

Okay I thought hard drive time, decided to get my old hard drive out I keep as a backup plugged it in and powered up. 35 seconds to kickstart screen and cannot see hard drives again. I then tried the big drive again same problem, connected them both to Buddha and still no drives are seen.

Does this sound like the IDE controller is shot? and if it was would it affect the Buddha as well? if it's not what could be the problem and how much do you think it would be to get it fixed? (assuming it is fixable). I forgot to mention earlier, I bought the system on Amibay a few month ago and this is the first time I have tried a hard drive in it so I am not sure if the this problem existed before it was recapped or if recapping it has caused it (could going from Buster 9 to 11 have caused any problems in this area?).

dw0ntU5m3
26th December 2014, 00:22
Sounds like you have multiple issues going on. It doesn't seem all that likely that both HDD controllers are both bad. If you needed to repair the IDE on your A4000 motherboard the two little 16V8s at U901 and U902 would be cheap to replace or upgrade, at most $20 for the pair of chips and a little bit of labor, I can't imagine that costing more than a case of beer. You mention a couple things here, memory not showing up on your Cyberstorm, Buster 11 upgrade and intermittent detection of IDE. I would suggest trying to simplify a bit and make sure that when it's taking a long time to get to kickstart, you then boot a workbench disk either from floppy or your HxC and see if your RAM shows up. When this happens open up showconfig and see if your Buddha flash is there. Are you changing the Master/Slave mode on your drives? Are you using Cable Select? Going from Buster 9 to 11 in itself isn't going to cause any problems but if it's socketed it's not a bad idea for troubleshooting to remove Buster and try cleaning it with IPA and reseating it. And when any of this stuff is happening what's your IDE light doing? If it's on constantly that's bad. For the onboard IDE it should blink a second after startup briefly then go dark for a few seconds until it starts to boot, though I don't know what the normal pattern for the Buddha is.

-Acid-
26th December 2014, 12:45
Yes the memory is registered after it takes ages to boot, drives have been set as master on both but one is set to master with slave but didn't have a slave attached. Show boards indicates the Buddha is seen (id 4626) as well as the three memory nodes. I do not have the extraction tool for the buster so I won't be attempting that any time soon. As for the hard drive light I attached one drive today and it attempted to boot from it very quickly as normal, but it didn't have workbench installed so just went to the cli prompt. So I changed drive to the one with workbench on and it went straight back to the 30 second delay but it did attempt to boot from it, but it reset immediately to make idefix resident then boot so the big partitions can be accessed but after it reset to do this the amiga waited 30 seconds to get going again and then doesn't boot from it again. I then boot up an Install disk and HDToolBox can see the drive is there, a few reboots later and it did actually boot from it properly and get to the Workbench screen however as the icons were in black and white so something clearly didn't load correctly and as soon as I try to click on a hard disk icon or the ram disk icon the system crashes.

When the scsi bus is being scanned in HDToolBox it is taking ages when it normally blasts through the whole range in the blink of an eye. Maybe this is part of the problem?

dw0ntU5m3
26th December 2014, 14:11
Bad IDE cable? Have you ruled that out? What Filesystem are you using? The crash as soon as workbench loads is troubling but maybe that's just missing the 060 libs? The long delay might just be the 3.1 ROMS causing the spinup delay when there's no IDE present on the motherboard. And was this last round of testing with the Buddha or the motherboard IDE? The other thing is that when you move drives around HDToolbox complains a drive has changed and needs to be updated, and you don't do it can cause all sorts of problems.

-Acid-
27th December 2014, 03:24
So i've spent some hours on it again tonight and came up with some interesting facts.

I tried a different IDE cable and connected up the old empty drive (I first used WinUAE to put CWB 3.9 on it) and it booted quick as a flash however I noticed only the motherboard ram was in use. It seemed to be going fine until I ran Scout and clicked the system tab (which is when it would attempt to read the cpu and ram info etc I guess), it immediately crashed. The same thing happened esterday everytime I tried to run SysInfo to see what it was seeing. Everything described in this thread so far has been on the motherboard's IDE, 060 libs are fixed as I said before my other system has an 060 so everything software wise should be set.

Since I ran Scout and it crashed, the next few boots no hard drive is seen by the system, I left it a few minutes and tried again and it booted but as soon as I tried to open a disk icon it freezes. CPU ram is still missing.

So it varies between booting instantly and taking 30+ seconds, CPU ram sporadically is there but mostly isn't, ide cables have been checked and changed over, hard drives have been checked and work fine in WinUAE and my A1200. The Buddha flash card is always showing up in the early boot menu as working so the Zorro daughter board is connected properly and is working too.

Now tell me if I am going off track here, but could the CPU card and ram be causing this? I recently let go of my 030 card so cannot test it now to see if that is the fault, but everything was stable when I only had the 030 in it but I hadn't connected a hard drive then I was only testing with floppies.

Any ideas anyone?

dw0ntU5m3
27th December 2014, 03:47
First suggestion since you have a Cyberstorm MK-I is to try removing the RAM Module from the Cyberstorm and reseating it. I would troubleshoot this first, one good way of doing it is on your spare drive, create a new startup-sequence in S: that just contains:
Assign ENV: RAM:
Avail

You can just rename your standard startup-sequence to startup-sequenceGOOD and create a new file named startup-sequence.

That way you can very quickly boot and see if your RAM is showing up. Next step if reseating the Cyberstorm RAM module has no effect is to start removing SIMMS from the RAM Module, or just three of them leaving one, and change the jumper settings on the Cyberstorm if necessary.. When you can reproducibly boot with RAM on the Cyberstorm then proceed to check your IDE issue.

awmosoft
27th December 2014, 07:27
I had the same issue on my Amiga 4000. There were two problems with the Amiga 4000 : the updated scsi.device and a bad IDE cable.

When I removed these problems, then the problem goes away.

-Acid-
27th December 2014, 12:03
I had the same issue on my Amiga 4000. There were two problems with the Amiga 4000 : the updated scsi.device and a bad IDE cable.

When I removed these problems, then the problem goes away.
scsi.device has not been changed, even without a hard drive connected i get the problems so it can't be that and i have already eliminated the ide cable as the problem.

dw0ntU5m3
28th December 2014, 03:31
I really think you should troubleshoot the RAM issue first, partially detected, erroring out RAM is a big problem that can cause about any errors you can imagine to happen.

-Acid-
28th December 2014, 12:27
Will the cpu function normally without the ram module attached to it?

vibros
29th December 2014, 08:09
Hello,

the idefix command in your startup-sequence is needed only if you have a 4IDE card attached on the IDE port of the A4000D. If you have an Harddrive bigger of 4gb, patch the scsi.device, like already suggested. Then, in your situation, the A3630 was the best card to test your hardware properly.
IDE issue are related to U901 and U902 (but affect the whole system): you can take a look if this parts was reworked in the past.

Regards
Stefano

dw0ntU5m3
30th December 2014, 03:19
Will the cpu function normally without the ram module attached to it?
I have never used a Cyberstorm so I don't know for sure, but you could at least configure the jumpers for 0mb and remove the simms from it. The likelihood of it interfering at that point are pretty low I'd guess.

davideo
30th December 2014, 11:53
I then boot up an Install disk and HDToolBox can see the drive is there, a few reboots later and it did actually boot from it properly and get to the Workbench screen however as the icons were in black and white so something clearly didn't load correctly and as soon as I try to click on a hard disk icon or the ram disk icon the system crashes.

I've got an external CD drive connected to my SCSI chain. If for some reason it is turned off when I start my 4000D this is what happens to my system although it does work but the icons are Black and White.