Need help with increasingly disturbing hardware problem

[murple]

rare_j wrote:
Sometimes old computers just get tired.
It's clearly a serious problem. Can you change your chip mem expansion or try running without it for a while?
After that, perhaps you should try a new motherboard.

Narrow it down bit by bit.

"Tired" isn't really a very helpful description, but he's on the right track as far as checking things. If you have an unknown problem, isolate it. Remove components that you can (which in a 500 means pretty much just add-ons like RAM, accelerators, and sidecar thingies) and see if the problem still comes up. See if any chips are unusually hot when its been running for a while. If you have a spare A500, maybe try swapping out some chips one at a time and letting it run until you get the crash.

If youre seeing a "deadbabe" error though... is that a normal Amiga error? If thats just an error from a certain virus, that seems kinda likely to me. Considering Amiga's have "guru meditations" though, it wouldnt surprise me if "deadbabe" is a real Amiga error.

[Zac67]

If it's not a virus, it may be
- defective/weak PSU: check voltages
- defective RAM: reduce/remove/replace RAM
- heat related: test system without covers

When I expanded my A3000 from 8 to 12 MB fast RAM (years ago) I began to get strange problems (corrupted graphics in ADPro, crashes, ...). I ran every RAM test I could find: nothing. Then I ran a test in part of the RAM while running applications: bang. Finally I tried running simultaneous tests of different parts of RAM: loads of errors. Removing the mainboard I finally found two address pins on a ZIP socket unsoldered.

Conclusion: There's something to be desired when it comes to RAM testing an Amiga.

I would concur with Zac67.

It looks to me as part of your RAM is misbehaving and you only start getting problems when your system's memory usage gets to a certain point.

[ChrisH]

I have AmigaForever, but I don't have a PC and don't ever plan on buying one (I bought AF years ago when I had a work notebook to use). WinUAE is OK, but still far from perfect. I'd still rather use a dying real Amiga than a flawed emulation (especially if it's running on Windows).

You really should try a modern version of WinUAE on a modern PC (say 2Ghz with 1GB ram on WinXP) - you honestly won't be able to tell that it isn't a real Amiga.  Well, except for the fact that it will feel like it's using a 800+MHz 68060, and that you can have a high-res 24-bit display, and 16-bit AHI sound...

It will even emulate the floppy disk drive noise, for full effect :-)

[motorollin]

eslapion wrote:
I would concur with Zac67.

It looks to me as part of your RAM is misbehaving and you only start getting problems when your system's memory usage gets to a certain point.

You can confirm this by filling up your RAM disk with files to use up all the RAM.

--
moto

[motorollin]

ChrisH wrote:

I have AmigaForever, but I don't have a PC and don't ever plan on buying one

You really should try a modern version of WinUAE on a modern PC

I know people mean well by suggesting "use UAE", but this is clearly not what the OP wants to do. His query is regarding faults using original Amiga hardware which he wishes to preserve. He obviously has no interest in UAE at this time. So suggesting using UAE is actually not very helpful, even if it was well intentioned.

I don't intend to offend anyone who wants to help, but it is a little annoying when every single hardware thread seems to have at least one person saying "use UAE".

--
moto

[Fixer]

motorollin wrote:
I know people mean well by suggesting "use UAE", but this is clearly not what the OP wants to do. His query is regarding faults using original Amiga hardware which he wishes to preserve. He obviously has no interest in UAE at this time. So suggesting using UAE is actually not very helpful, even if it was well intentioned.

I don't intend to offend anyone who wants to help, but it is a little annoying when every single hardware thread seems to have at least one person saying "use UAE".

--
moto

I know what you mean moto. I myself will always have a thing for the old hardware. It is fun memorable.

However the unfortunate reality is that a lot of the hardware is just dying, and only places like Amigakit are actually able to sell NOS 1200s. (I must commend them for that as well)

WinUAE is solid and has improved loads over the years, and is so quick to set up; I myself can't really blame people for recommending it.

You never know, maybe the Minimig will change all of this... ;-)

[freqmax]

Check the PSU and other parts for failing capacitors. They *dry* with age. Also consider that if you value your equipment. Consider to put it on an online UPS. And beware of any "modified sinus" stuff.

[motorollin]

@Fixer
I am not disputing the benefits of UAE. But there are those (myself included) for whom using the original Amiga hardware is part of the fun of the Amiga. UAE just isn't the same.

I'm sure most people are able to make the decision themselves whether to persevere with their original hardware or scrap it and move to UAE. The fact that somebody posts a message asking for help fixing their hardware suggests that they have chosen the former option, and aren't looking for information on UAE.

--
moto

[rkauer]

 Some ppl may feel tired of me saying always the same thing, but:

 CHANGE ALL ELECTROLITHICS FROM MOTHERBOARD AND PSU EVERY 3 OR 5 YEARS!:-o

 That prevents a lot hardware failures. And is cheaper than send the whole package to amiga center in France. :-)  

[murple]

motorollin wrote:
I am not disputing the benefits of UAE. But there are those (myself included) for whom using the original Amiga hardware is part of the fun of the Amiga. UAE just isn't the same.

Agreed... telling someone to use UAE is kind of like telling someone of the benefits of a shiny new machine-made violin and how its easier than trying to use some beat up old Stradivarius.

It may be in my imagination, but Amigas have a feel you dont get from emulators. The video display has a warmth you don't get from a PC display emulating an Amiga. An emulated Amiga just doesn't seem to have quite the same look, sound and feel.

Also, for some of us, the challenge of keeping old machines working is a big part of the fun. With an emulated Amiga, you're just using a PC with disposable parts that can be replaced at your choice of several hundred local McPC shops. Some people enjoy maintaining and using vintage cars. We enjoy maintaining vintage computers.

If you just want to run some old programs on something thats not really an Amiga, UAE is great. UAE doesn't do a damn thing for me though.

[ChrisH]

@motorollin
I wasn't suggesting UAE, I was merly pointing out that his comment on UAE was waaaay out of date.

And sorry if every other thread gets a comment on UAE, but I don't usually read this site, so I had no idea.


Personally speaking, I've been using Amithlon & then WinUAE for over 6 years (since my BlizzPPC+BVision was unstable by design), and have never looked back.  I find it quite difficult to understand why someone would still want to use incredibly slow & unreliable hardware to do modern stuff like browsing the internet, playing/encoding MP3s, etc.

Any way, sorry for diverting the thread.

[da9000]

Zac67 wrote:
When I expanded my A3000 from 8 to 12 MB fast RAM (years ago) I began to get strange problems (corrupted graphics in ADPro, crashes, ...). I ran every RAM test I could find: nothing. Then I ran a test in part of the RAM while running applications: bang. Finally I tried running simultaneous tests of different parts of RAM: loads of errors. Removing the mainboard I finally found two address pins on a ZIP socket unsoldered.

Interesting, Zac. What's your take on why the errors would show up during concurrent tests? Maybe increased bandwidth caused more heat and thus caused the pins to dialate and thus affect contact?

[Zac67]

No - the missing address bits caused some memory locations to be blended together: e.g. you write $00000000 to location X and $FFFFFFFF to location Y, and after that you'd read $00800000 from X, because one bit appears shared between the two addresses. The chip has the address bit disconnected and thus can't distinguish between both memory locations.

None of the available test could show this, they just seem to write a pattern to a whole block, read it back and compare the results. When the very same pattern is used on the whole block, you can't see a problem - you'd have to use different patterns to locate non-uniqueness problems.

I knew the problem was in the last 4 MB block (Amy worked fine before), so I used 4 instances of the same tool to test 1 MB each - running them simultaneously immediately showed that memory addresses were not unique and a bit more testing (plus a tiny C routine) revealed there actually were several different address patterns that overwrote each other. Before disassembling the machine (quite a feat on the 3k) I had it narrowed down to a single chip and the exact pins that must have been the problem. I expected damaged trace, bent pins but found - the contacts cleanly suspended in the middle of the hole without any solder.  :-P

If you want to try this: boot the system w/o startup-sequence and shell run the RAM test tool for, say 1 MB each. A bad contact on an address bit will show up very quickly; maybe there's also a better tool around today, I can't remember what I used back then.

Errors in the RAM chips may be very hard to find - on a PC I've had memtest86 running for nearly a day before a single bit error showed up that had repeated crashed the machine. If it's in a vital area, it'll cause a crash quickly, but if it's intermittant, the test software may have difficulties locating it.

PS: Just checked on the Viper: the RAM is soldered to the board?? That'll make it very hard to swap for testing and also very hard to find a replacement...

[da9000]

Zac67 wrote:
No - the missing address bits caused some memory locations to be blended together: e.g. you write $00000000 to location X and $FFFFFFFF to location Y, and after that you'd read $00800000 from X, because one bit appears shared between the two addresses. The chip has the address bit disconnected and thus can't distinguish between both memory locations.

Aha! I get it! Yes, that makes a lot of sense. I guess it would not be very hard to make a memory tester do "zoned writes", so that each zone has different patterns, and then proceed to read each memory zone and see if it has the original pattern or not. This way only a single instance is needed. Now I wonder if memtest86 does this... (I've not seen something to this effect)

Zac67 wrote:
disassembling the machine (quite a feat on the 3k) I had it narrowed down to a single chip and the exact pins that must have been the problem. I expected damaged trace, bent pins but found - the contacts cleanly suspended in the middle of the hole without any solder.  :-P

Hahaha! That's impressive "divide and conquer" problem targeting :-)

Zac67 wrote:
If it's in a vital area, it'll cause a crash quickly, but if it's intermittant, the test software may have difficulties locating it.

When it's in a vital area, is when I call someone "lucky" ;-)

Zac67 wrote:
PS: Just checked on the Viper: the RAM is soldered to the board?? That'll make it very hard to swap for testing and also very hard to find a replacement...

I didn't check the Viper in too much detail, but if it's surface-mount RAM, the trick that I use, which has worked successfully with DIP DRAMs (like those in A500s) is to "piggy-back" a known working DRAM chip on top of the surface-mounted ones and see if that affects the tests. It can be a bit tricky since all the pins have to touch and stay that way until the end of the tests, but it sure beats desoldering ;-)