English Amiga Board


Go Back   English Amiga Board > Support > support.Hardware

 
 
Thread Tools
Old 08 April 2015, 12:01   #1
Futaura
 
Posts: n/a
Faulty A1200 motherboard help - chip ram suspected

I'm looking for some advice on my faulty A1200 1D4 motherboard (ESCOM). All the advice found on this forum has gotten me this far, so I figured it would be the best place to ask. As well as being my main development machine until a few years ago, it was also the very first A1200 to boot OS4 successfully (I know this because it was me who fixed the kernel to work with a 603e PPC). Although I've not used it much for the past few years, it is mainly used for WarpDT benchmarking, with my A1XE being my main development machine. For these reasons I'm keen to keep it alive if at all possible.

I have a custom made A1200 tower, with BPPC/BVision, which had worked perfectly for years until recently. One day it stopped booting and started crashing/rebooting with a blinking power LED right after the BVision boot logo appeared. To cut a long story short, after initially suspecting the BVision, I have now confirmed it is the A1200 motherboard which is the culprit. I stripped it down to just a HDD, keyboard and mouse attached, and it still crashes during boot. I've tried numerous power supplies and have recapped the board (all 14 capacitors), but the problem remains.

With a vanilla AGA WB3.1 boot, I can sometimes get Workbench up, but after opening various windows sometimes there is graphics corruption and sooner or later it crashes completely and the display completely corrupts and it either reboots or the display goes off altogether (no RGB output). Occasionally, I get a yellow screen on booting or a red software failure guru before the boot menu appears. And today I saw a green boot screen for the first time, hence...

My suspicion is one or more of the chip ram memory chips is faulty, or I guess it could be a chip associated with the memory bus (Alice?). Memory testers from Aminet tend to lead to an instant crash, although it seems the upper 1Mb tests ok. If I boot with no startup-sequence and a single partition, I can crash it by copying a 512Kb ROM image to the RAM Disk, or even by using addbuffers to add a large amount of buffers to the boot drive. I guess the memory could have being going bad over a long period of time, as normally with my BPPC/BVision fitted, not much chip ram gets used - it could be critical locations have now gone bad.

Any other ideas? Have I missed something?

I did wonder if it could be the ROMs, but seems unlikely I think. I've found that I can get replacement memory chips reasonably cheaply from China/eBay, so I am considering piggybacking the chips to see if I can least deduce which chip is faulty, although replacing the chips completely may be beyond my soldering capability. At least, I think piggybacking should work as part of fault diagnosis, right?

If all else fails, I guess will have to seriously consider selling my working BPPC (240MHz 603e, 68060, SCSI) and BVision .
 
Old 08 April 2015, 12:45   #2
vulture
Registered User
 
Join Date: Oct 2007
Location: Athens , Greece
Posts: 1,088
I'm not sure piggybacking would work in your case since, whatever chip is bad, is not completely dead and can still accept/send signals. I'm by no way an expert on this, so, please anybody who is, correct me if I'm wrong. Now, I know that's not what you've asked, but I think that finding and installing another motherboard wouldn't be very expensive and it'd definately be less work to do. In any case, good luck mate!
vulture is offline  
Old 08 April 2015, 13:23   #3
Futaura
 
Posts: n/a
Trouble is with getting another motherboard, it is probably going to be older than mine, so is probably going to develop or have a problem too. I figured it would be better to fix mine, especially as the history of another 20 year old secondhand board would be unknown. And it is a 1D4 board with all the timing fixes too, and it has always worked perfectly with my BPPC.
 
Old 08 April 2015, 15:05   #4
kipper2k
Registered User

 
Join Date: Sep 2006
Location: Thunder Bay, Canada
Posts: 3,838
Quote:
Originally Posted by Futaura View Post
I'm looking for some advice on my faulty A1200 1D4 motherboard (ESCOM). All the advice found on this forum has gotten me this far, so I figured it would be the best place to ask. As well as being my main development machine until a few years ago, it was also the very first A1200 to boot OS4 successfully (I know this because it was me who fixed the kernel to work with a 603e PPC). Although I've not used it much for the past few years, it is mainly used for WarpDT benchmarking, with my A1XE being my main development machine. For these reasons I'm keen to keep it alive if at all possible.

I have a custom made A1200 tower, with BPPC/BVision, which had worked perfectly for years until recently. One day it stopped booting and started crashing/rebooting with a blinking power LED right after the BVision boot logo appeared. To cut a long story short, after initially suspecting the BVision, I have now confirmed it is the A1200 motherboard which is the culprit. I stripped it down to just a HDD, keyboard and mouse attached, and it still crashes during boot. I've tried numerous power supplies and have recapped the board (all 14 capacitors), but the problem remains.

With a vanilla AGA WB3.1 boot, I can sometimes get Workbench up, but after opening various windows sometimes there is graphics corruption and sooner or later it crashes completely and the display completely corrupts and it either reboots or the display goes off altogether (no RGB output). Occasionally, I get a yellow screen on booting or a red software failure guru before the boot menu appears. And today I saw a green boot screen for the first time, hence...

My suspicion is one or more of the chip ram memory chips is faulty, or I guess it could be a chip associated with the memory bus (Alice?). Memory testers from Aminet tend to lead to an instant crash, although it seems the upper 1Mb tests ok. If I boot with no startup-sequence and a single partition, I can crash it by copying a 512Kb ROM image to the RAM Disk, or even by using addbuffers to add a large amount of buffers to the boot drive. I guess the memory could have being going bad over a long period of time, as normally with my BPPC/BVision fitted, not much chip ram gets used - it could be critical locations have now gone bad.

Any other ideas? Have I missed something?

I did wonder if it could be the ROMs, but seems unlikely I think. I've found that I can get replacement memory chips reasonably cheaply from China/eBay, so I am considering piggybacking the chips to see if I can least deduce which chip is faulty, although replacing the chips completely may be beyond my soldering capability. At least, I think piggybacking should work as part of fault diagnosis, right?

If all else fails, I guess will have to seriously consider selling my working BPPC (240MHz 603e, 68060, SCSI) and BVision .

did you try disconnecting HDD, and just have floppy connected (no keyboard) and boot a WB floppy with memcheck on it to see what that reports ?
kipper2k is offline  
Old 08 April 2015, 15:33   #5
Futaura
 
Posts: n/a
Quote:
Originally Posted by kipper2k View Post
did you try disconnecting HDD, and just have floppy connected (no keyboard) and boot a WB floppy with memcheck on it to see what that reports ?
Unfortunately not. The floppy drive died many years ago - it is one of those Panasonic PC drives that ESCOM used. I have the drive from my A600, but that won't work on the A1200 without reversing ESCOM's PC drive hack.

I did try without the keyboard attached, so just HDD, mouse and 4-way IDE interface attached. I've tried two HDDs and another IDE interface adapter. In my test set up, the motherboard only is powered by a standard A1200 PSU (tested with PC PSU too), whilst all other components like the HDD and floppy drive are powered from a separate PC power supply. For those reasons I'm sure it is not an external power issue, although could be some power issue on the mobo I guess.

The closest I have been able to try to your floppy suggestion is to boot with no s-s from the HDD after disabling all but one of the partitions and run memcheck from there, which leaves 1.75Mb free. I am quite surprised that a memory checker can cause such bad crashes with bad ram - it is only reading/writing ram and not trying to run code from ram after all. I might make up my own simple tool so I can test specific addresses.
 
Old 08 April 2015, 15:39   #6
kipper2k
Registered User

 
Join Date: Sep 2006
Location: Thunder Bay, Canada
Posts: 3,838
Quote:
Originally Posted by Futaura View Post
Unfortunately not. The floppy drive died many years ago - it is one of those Panasonic PC drives that ESCOM used. I have the drive from my A600, but that won't work on the A1200 without reversing ESCOM's PC drive hack.

I did try without the keyboard attached, so just HDD, mouse and 4-way IDE interface attached. I've tried two HDDs and another IDE interface adapter. In my test set up, the motherboard only is powered by a standard A1200 PSU (tested with PC PSU too), whilst all other components like the HDD and floppy drive are powered from a separate PC power supply. For those reasons I'm sure it is not an external power issue, although could be some power issue on the mobo I guess.

The closest I have been able to try to your floppy suggestion is to boot with no s-s from the HDD after disabling all but one of the partitions and run memcheck from there, which leaves 1.75Mb free. I am quite surprised that a memory checker can cause such bad crashes with bad ram - it is only reading/writing ram and not trying to run code from ram after all. I might make up my own simple tool so I can test specific addresses.
The crash when running a memcheck program could be the program waiting for a response. if you turn on the computer and just let it sit for a while do any chips get unusually hot ?
kipper2k is offline  
Old 08 April 2015, 17:09   #7
Futaura
 
Posts: n/a
Quote:
Originally Posted by kipper2k View Post
The crash when running a memcheck program could be the program waiting for a response. if you turn on the computer and just let it sit for a while do any chips get unusually hot ?
Not sure what the usual is, but all chips measured in the 30-40 degrees range (23 degrees room temp) after about one hour of playing the purple boot screen animation. The hottest being the Alice chip at 40 degrees. All four memory chips are among the coolest at 30 degrees.
 
Old 08 April 2015, 18:12   #8
kipper2k
Registered User

 
Join Date: Sep 2006
Location: Thunder Bay, Canada
Posts: 3,838
What about power supply, are you getting 5v at all IC's and HDD /FDD header ?

hook up an ohmmeter and check voltage as it boots to se if it is stable
kipper2k is offline  
Old 08 April 2015, 18:17   #9
Futaura
 
Posts: n/a
Quote:
Originally Posted by kipper2k View Post
What about power supply, are you getting 5v at all IC's and HDD /FDD header ?

hook up an ohmmeter and check voltage as it boots to se if it is stable
Is there a website that has the pinouts for all these?
 
Old 08 April 2015, 18:21   #10
kipper2k
Registered User

 
Join Date: Sep 2006
Location: Thunder Bay, Canada
Posts: 3,838
Quote:
Originally Posted by Futaura View Post
Is there a website that has the pinouts for all these?
do a google search for each chip.. example "Amiga alice pinout"

gives you....
http://aminet.net/package/docs/hard/pinouts

when the amiga fails to boot, what colour is last screen seen ?
kipper2k is offline  
Old 08 April 2015, 20:37   #11
Futaura
 
Posts: n/a
Quote:
Originally Posted by kipper2k View Post
do a google search for each chip.. example "Amiga alice pinout"

gives you....
http://aminet.net/package/docs/hard/pinouts

when the amiga fails to boot, what colour is last screen seen ?
Most of the time it starts to boot fine, but crashes after the boot colour errors have passed with a grey colour. I've seen the green screen error code once, but it is usually yellow on the few occasions it doesn't start to boot. The only consistent thing is that it crashes during chip ram accesses - it is harder to crash with the BPPC attached as obviously most stuff is then using fast ram from the BPPC.

I've now tested the 5v pins on the Alice, Denise, Bungie, Gayle, CPU, CIA x 2, IDE header, floppy power and the 3 pins on each of the 4 memory chips, and the power seems pretty stable to me.
 
Old 08 April 2015, 20:46   #12
supaduper
Registered User
supaduper's Avatar
 
Join Date: Oct 2012
Location: Birmingham UK
Posts: 886
I did have a similar problem to you but on a A600 and it was indeed one of the RAM chips , there were two chips and basically I got a working one and swapped it with one chip and it was still the same so I got another one and replace the second RAM chip and Bingo it worked, so perhaps it may be a RAM problem.
supaduper is offline  
Old 09 April 2015, 13:49   #13
Hewitson
Registered User
Hewitson's Avatar
 
Join Date: Feb 2007
Location: Melbourne, Australia
Age: 37
Posts: 3,406
Quote:
Originally Posted by kipper2k View Post
hook up an ohmmeter and check voltage as it boots to se if it is stable
You'd probably want to use a voltmeter for that
Hewitson is offline  
Old 09 April 2015, 16:14   #14
kipper2k
Registered User

 
Join Date: Sep 2006
Location: Thunder Bay, Canada
Posts: 3,838
Quote:
Originally Posted by Hewitson View Post
You'd probably want to use a voltmeter for that

lol.. i think all ohmmeters do voltage and current etc
kipper2k is offline  
Old 06 May 2015, 13:04   #15
Futaura
 
Posts: n/a
Ok, so I am now certain that the chip ram is faulty. I made a simple test tool so I could more accurately narrow down the affected memory area... An entire 512K block is dead, from 0x80000-0xfffff. Any attempt from Exec to start allocating memory within this block causes havoc because it corrupts the memory list - the avail command gets a 0100000c recoverable alert.

If I preallocate this 512K block in s:startup-sequence so that nothing else can use it, the system boots and works fine. I haven't tried with my BPPC and everything hooked up yet, but I expect it should work also especially as most stuff will be using fast ram instead then.

So, I guess I am correct in assuming just one of the four memory chips is faulty? Does anybody know which chips correspond to each the four 512K blocks of addressable chip ram? Obviously, I would prefer just to replace the faulty one .
 
Old 10 November 2019, 18:28   #16
Futaura
Registered User

Futaura's Avatar
 
Join Date: Aug 2018
Location: United Kingdom
Posts: 21
As a matter of closure, it was the chipram that had gone faulty, as suspected. I found exact replacements for the 4 Samsung chips online, ordered them and AmigaKit replaced them for me in 2016. Unfortunately, the board instantly developed another fault the same day I got the fixed board back, so it went back to AmigaKit where it sat for 3 years without any progress. I gave up hoping they could fix it and arranged for somebody else to look at it and they fixed it last month!

So, my mobo is fixed and BPPC+BVision still work. System booted fine.

Unfortunately, my SCSI HDD then shut off after an hour uptime and now won't spin up, so the project continues. I guess the drive is dead, so looks like I'll be investing in SCSI2SD soon.
Futaura is offline  
 


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools

Similar Threads
Thread Thread Starter Forum Replies Last Post
Amiga 1200 Suspected DAC - not faulty?!? GadgetUK support.Hardware 13 07 August 2013 17:38
A1200 chip RAM components whitebird support.Hardware 0 24 April 2012 01:12
WTB: Any faulty/damaged A1200 RAM/Accelerator Wasagi MarketPlace 0 20 November 2010 20:11
A1200 with 1meg chip ram - WTF? DC33 support.Hardware 19 08 July 2007 18:00
Diagnose faulty chip ram IC's chaoticjelly support.Hardware 1 15 February 2007 23:44

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +2. The time now is 01:59.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2019, vBulletin Solutions Inc.
Page generated in 0.08183 seconds with 15 queries