English Amiga Board


Go Back   English Amiga Board > Support > support.Hardware

 
 
Thread Tools
Old 01 August 2011, 02:04   #1
Photon
Moderator

Photon's Avatar
 
Join Date: Nov 2004
Location: Eksjö / Sweden
Posts: 5,079
Unstable ACA630

I will return it if I really have to, but it would be nice if I could get some ideas for what to have a look at before I send it.

After 6 months of no stability problems my ACA630-25 has gone unstable.

Symptom is that I sometimes get red, sometimes green screen directly at boot. Sometimes I get graphics crap after 1 or 5 or 40 minutes (chipmem options in acatune are off). Sometimes I can use it for 3 hour sessions. Every session ends with a freeze or a guru, or a freeze, then a guru. After a first crash, usually power-off, wait 5 seconds, power-on gives a green/red screen right away.

PSU, grounding is ok. 5V is ok with both PSUs I tried (34W brick and 120W PicoPSU). Right now only a screen, an A602 and a CF card in IDE are attached.

After lots of unplugging peripherals and testing, and making sure everything is perfect for the ACA, I've concluded that everything works perfectly with ACA off, and unstable with it on. I've come up with 4 reasons:

1. It gets too hot. (But why guru after 5-10 minutes sometimes then?)
2. Capacitors need changing (but why perfect with 68000+peripherals?)
3. Bad contact to 68000 (but it sits perfectly. clean pins again??)
4. Card has become unreliable (maybe after overheating.)

I hope someone in ACA owners club recognize the description and can help me! Or others who had unstable turbo cards in their A600 and found a cure for it.

HEEEELP............

Edit: I should add that there IS heat in there. Top side of the case is of course barely finger temperature, but just noticed underside of the case is hotter than any laptop PC I've used (somewhere between 45-50°C, I'd say), so I'd definitely think twice before putting this in my lap!!

Anyone else have such warm temperatures?

I will buy one of those "heat extractors" and put it to blow out of the floppy slit to see if the problem goes away.

Last edited by Photon; 01 August 2011 at 04:15.
Photon is offline  
Old 01 August 2011, 02:19   #2
FOL
PSPUAE DEV

FOL's Avatar
 
Join Date: Nov 2006
Location: Barry / UK
Age: 43
Posts: 5,902
Send a message via MSN to FOL
Quote:
Originally Posted by Photon View Post
I will return it if I really have to, but it would be nice if I could get some ideas for what to have a look at before I send it.

After 6 months of no stability problems my ACA630-25 has gone unstable.

Symptom is that I sometimes get red, sometimes green screen directly at boot. Sometimes I get graphics crap after 1 or 5 or 40 minutes (chipmem options in acatune are off). Sometimes I can use it for 3 hour sessions. Every session ends with a freeze or a guru, or a freeze, then a guru. After a first crash, usually power-off, wait 5 seconds, power-on gives a green/red screen right away.

PSU, grounding is ok. 5V is ok with both PSUs I tried (34W brick and 120W PicoPSU). Right now only a screen, an A602 and a CF card in IDE are attached.

After lots of unplugging peripherals and testing, and making sure everything is perfect for the ACA, I've concluded that everything works perfectly with ACA off, and unstable with it on. I've come up with 4 reasons:

1. It gets too hot. (But why guru after 5-10 minutes sometimes then?)
2. Capacitors need changing (but why perfect with 68000+peripherals?)
3. Bad contact to 68000 (but it sits perfectly. clean pins again??)
4. Card has become unreliable (maybe after overheating.)

I hope someone in ACA owners club recognize the description and can help me! Or other who had unstable turbo cards in their A600 and found a cure for it.

HEEEELP............
What happens if you have the ACA630 attach but have the disable jumper enabled.
If its a dodgy solder joint this will help show it.
FOL is offline  
Old 01 August 2011, 02:34   #3
Photon
Moderator

Photon's Avatar
 
Join Date: Nov 2004
Location: Eksjö / Sweden
Posts: 5,079
FOL, I've connected a switch to the 68000/68030 jumper. I've checked that I have a good connection between jumper legs and switch legs.

I guess the switch could have gone bad internally, so that's one idea to try at least - either replace or remove the switch.

Dodgy solder joint where, exactly?


Does this behavior - perfect in DISABLE mode - rule some things out, if so which?
Photon is offline  
Old 01 August 2011, 07:59   #4
Jope
-
Jope's Avatar
 
Join Date: Jul 2003
Location: Helsinki / Finland
Age: 41
Posts: 8,698
The 68000s in A600s sometimes have cracked solder joints between their legs and the motherboard.
Jope is offline  
Old 01 August 2011, 08:33   #5
Predseda
Puttymoon inhabitant
Predseda's Avatar
 
Join Date: Mar 2007
Location: The City of Townsville
Age: 43
Posts: 6,602
Send a message via ICQ to Predseda
I have similar problems, I think I will send my A600 mobo to inspection to Amigakit on autumn. I am almost sure that the problem is with those cracks o processor's feet and the ACA is okay. I am unable to repair it on myself, though
Predseda is offline  
Old 01 August 2011, 09:03   #6
twiggy
\m/

twiggy's Avatar
 
Join Date: Nov 2008
Location: Devon, U.K.
Age: 32
Posts: 477
Out of those I only get the occasional boot colour and I used to get graphical glitches/freezes every so often, but I haven't for a while now (I think due to my bodge job cooling solution). But yes, mine used to get very hot too.
twiggy is offline  
Old 01 August 2011, 12:33   #7
FOL
PSPUAE DEV

FOL's Avatar
 
Join Date: Nov 2006
Location: Barry / UK
Age: 43
Posts: 5,902
Send a message via MSN to FOL
Quote:
Originally Posted by twiggy View Post
Out of those I only get the occasional boot colour and I used to get graphical glitches/freezes every so often, but I haven't for a while now (I think due to my bodge job cooling solution). But yes, mine used to get very hot too.
I dont think heat has anything to do with it.
Im over clocking mine and mine got very hot before I over clocked it.
Now removing the heatsink and not re-applying the heat compound correctly, may cause problems.

I even had it running for atleast 4 days straight (with case on) using a stock PSU. It was working perfect.

If it works fine in 68K mode, then I can only assume that there are no damaged solder joints on the CPU. Ofcourse without seeing it, we are only guessing. Can you take high res shots of each side of the 68k.
FOL is offline  
Old 01 August 2011, 14:09   #8
Photon
Moderator

Photon's Avatar
 
Join Date: Nov 2004
Location: Eksjö / Sweden
Posts: 5,079
Turned on the A600 just now and got an instant green screen. It had been off for 10-11 hours, so that green screen didn't come from heat, that's for sure!

Turned it off, waited 3 seconds, on. Green screen.

Turned it off, waited 3 seconds, on. Boots! Freezes after 4 minutes of working on it.

Turned it off, waited 3 seconds, on. Green screen. Let it reset 2 more times and on the 3rd reset it booted. Still working after 10 minutes now...

Edit: Still working after 20 minutes, I've decided to remove the A602.

I've never had green screens in 68000 mode, but it might be possible that cracks have appeared now, either in chip ram chips or CPU.

I'm not taking the bootscreen color extremely seriously though, other times it has been red, or guru'd. But I will let it cool down (barely above room temp on the underside right now), remove the A602, and start it in 68000 mode and use it for a few hours.

If I then get any problem at all, I will let you know. Either way I can use my tube of solder assist and re-heat the 68000 legs.


Edit 2: After 36 minutes of using it, I got something that sort of wants to look like a recoverable alert. (Yellow blinking on black a few times, then returns to what I was doing.) Except it's not a frame with text but rather a smaller block+crap. I ran my chipmem utility to see if any chipmem in the A602 was used (very long shot, was coding at the time and know no mem over $40000 was used) and discovered that the alert is triggered by AllocMem().

Now, this is overanalysing it of course. But it does seem to point to memory problems, or memory bus problems (such as bad contact).

Couldn't rule out heat, though. Underside is warmer than body temp after 36 mins of use...!

Last edited by Photon; 01 August 2011 at 14:52.
Photon is offline  
Old 01 August 2011, 14:22   #9
Predseda
Puttymoon inhabitant
Predseda's Avatar
 
Join Date: Mar 2007
Location: The City of Townsville
Age: 43
Posts: 6,602
Send a message via ICQ to Predseda
Quote:
Originally Posted by Photon View Post
Turned on the A600 just now and got an instant green screen. It had been off for 10-11 hours, so that green screen didn't come from heat, that's for sure!

Turned it off, waited 3 seconds, on. Green screen.

Turned it off, waited 3 seconds, on. Boots! Freezes after 4 minutes of working on it.

Turned it off, waited 3 seconds, on. Green screen. Let it reset 2 more times and on the 3rd reset it booted. Still working after 10 minutes now...
I recognise that. My did the same on an Amiga party. Fortunatelly I had my reliable A1200 with myself as well.
Predseda is offline  
Old 01 August 2011, 14:59   #10
Photon
Moderator

Photon's Avatar
 
Join Date: Nov 2004
Location: Eksjö / Sweden
Posts: 5,079
Predseda, ah ok - saw you mentioned that. I'm sure you've had a look by now - did you see anything untoward? Does it work reliably now? What did you do to make it reliable?

Sorry for editing posted posts, but I figure test results pertaining to the same post should not be spread out into separate posts.

OK, I think we can rule out any problems in 68000 mode. Discovered this just before I was to turn it off and cool down for a while:



After the above use for 36 mins and getting the Recoverable Alerts, I Amiga-M'd back from CLI I got guru and a green screen. This kept going on. I decided to use this to track down the fault.


I did this just after getting the 68030 mode issues:

1. A power toggle with switch to 68000 mode and it boots fine.

2. A power toggle with switch to 68030 mode gives black screen, no boot.

3. Poweroff for 10 seconds the second time gives black screen, no boot. (68030 mode)

4. Poweroff for 30 seconds the third time gives boot! But just after booting comes the green screen...

5. Poweroff for > 1 min, removed A602, poweron, instant green screen.

6. Poweroff for 5 seconds, 68000 mode, boots fine. Has worked for 8 minutes now.

I can't see how this behavior is consistent with a mobo or 68000 problem, with or without microcracks or heat. It certainly was not cold inside this time, for the 68000 or the 68030.

I will approach the ACA630 problem as any A600 turbo card problem now. And assume that the weak point is the socket-on-smd-68000 and the heat. I will inspect the card. If after addressign those there is a problem I will reply here.

Last edited by TCD; 01 August 2011 at 15:11.
Photon is offline  
Old 01 August 2011, 16:21   #11
Photon
Moderator

Photon's Avatar
 
Join Date: Nov 2004
Location: Eksjö / Sweden
Posts: 5,079
OK, went shopping. With the A600 cooled down to room temp, 68000 mode boots fine as always. This should rule out microcracks on the mobo (?) leaving only the turbo card, since the CF/adapter works in 68000 mode.

After working for 5 minutes in 68000 mode, power off, 3 secs, power on in 68030 mode gives green screen. I interpret this as "68000 can talk to chipmem, but 68030 has trouble talking to chipmem".

I'll have a look at the turbo card, re-seat it, and add the heat extractor (mostly for equalizing the temp for testing).
Photon is offline  
Old 01 August 2011, 16:21   #12
FOL
PSPUAE DEV

FOL's Avatar
 
Join Date: Nov 2006
Location: Barry / UK
Age: 43
Posts: 5,902
Send a message via MSN to FOL
Another long shot. I have all types of wierd issues when the expansion edge connector is dirty.

I even had one A600 giving joystick faults, turned out to be dirty connection on A602/3 expansion slot / edge connector.

Quote:
Originally Posted by Photon View Post
OK, went shopping. With the A600 cooled down to room temp, 68000 mode boots fine as always. This should rule out microcracks on the mobo (?) leaving only the turbo card, since the CF/adapter works in 68000 mode.
When the ACA630 is on, it will be pushing against pins, if desoldered, the then the pins will be push away and not make contact.
With out the ACA630 on, the pins will more than likely rest in a position where they are making contact.

So as I said, testing with ACA630 on in disabled mode can rule this out.
FOL is offline  
Old 01 August 2011, 17:23   #13
Photon
Moderator

Photon's Avatar
 
Join Date: Nov 2004
Location: Eksjö / Sweden
Posts: 5,079
Quote:
Originally Posted by FOL View Post
testing with ACA630 on in disabled mode can rule this out.
Yes, that's what I mean with "switched to 68030 mode" and "switched to 68000". DISABLE jumper on ACA open or closed. I have not yet unmounted the card from the 68000 chip.

Tests with 68030 on:

After just getting a red screen, I opened the case, and suddenly it could boot 7 times in a row! OK, maybe something is pushing on the ACA when lid is closed, thought I and screwed the case shut again. Nope, still worked. But after 30 seconds it guru'd. OK, open the case again. Green screen, next poweron black screen, next poweron it worked. After 30 seconds, guru.

OK, remove CF card and see if floppy disk animation appears. Nope, at least not after a few minutes. (BootSel was closed then, and no DF0 or DF1 connected.

Then I moved BootSel jumper to NoDisk, and still no CF card. Disk animation appears! I left it on for a few minutes, seemed to work.

Now I have only NoDisk jumpered, and it has worked for 10 minutes so far. I'm keeping the lid off and if I don't get a single guru in an hour or two, maybe I've found something.

It would be interesting to know if NODISK users have NO problems booting ACA630 when it's cool!
Photon is offline  
Old 01 August 2011, 18:21   #14
FOL
PSPUAE DEV

FOL's Avatar
 
Join Date: Nov 2006
Location: Barry / UK
Age: 43
Posts: 5,902
Send a message via MSN to FOL
Quote:
Originally Posted by Photon View Post
Yes, that's what I mean with "switched to 68030 mode" and "switched to 68000". DISABLE jumper on ACA open or closed. I have not yet unmounted the card from the 68000 chip.

Tests with 68030 on:

After just getting a red screen, I opened the case, and suddenly it could boot 7 times in a row! OK, maybe something is pushing on the ACA when lid is closed, thought I and screwed the case shut again. Nope, still worked. But after 30 seconds it guru'd. OK, open the case again. Green screen, next poweron black screen, next poweron it worked. After 30 seconds, guru.

OK, remove CF card and see if floppy disk animation appears. Nope, at least not after a few minutes. (BootSel was closed then, and no DF0 or DF1 connected.

Then I moved BootSel jumper to NoDisk, and still no CF card. Disk animation appears! I left it on for a few minutes, seemed to work.

Now I have only NoDisk jumpered, and it has worked for 10 minutes so far. I'm keeping the lid off and if I don't get a single guru in an hour or two, maybe I've found something.

It would be interesting to know if NODISK users have NO problems booting ACA630 when it's cool!
If you havent rmoved the card, I would try that first.
Re-seat it and try again.

Does'nt make sense having "no disk" jumper on and it works.
FOL is offline  
Old 01 August 2011, 19:38   #15
Photon
Moderator

Photon's Avatar
 
Join Date: Nov 2004
Location: Eksjö / Sweden
Posts: 5,079
Agree.

With lid off and NoDisk closed, it worked without a problem for 1.5 hours. I then power cycled 5 times to see if I could get an error. Nope. I then turned it off and went and made coffee to let it cool down and did the same again. No error. Moved the jumper back to BootSel and it has booted fine 3 times and been running for 10 minutes.

Sure is hard to make it fail when the lid is off, once in 20-25 tries if I look back at the last hours.

Conclusion: something is pushing or shorting the ACA630 when case is closed, or it gets too hot then. Had a look now as I closed the lid, couldn't see anything that could push or short the ACA630.

Leaving the lid on now for a while, but not screwed tight. To see if it's the push or the heat, and to see if it goes unstable.

Next step is to re-seat the ACA630 and make sure the heat goes down.
Photon is offline  
Old 01 August 2011, 20:36   #16
FOL
PSPUAE DEV

FOL's Avatar
 
Join Date: Nov 2006
Location: Barry / UK
Age: 43
Posts: 5,902
Send a message via MSN to FOL
Quote:
Originally Posted by Photon View Post
Agree.

With lid off and NoDisk closed, it worked without a problem for 1.5 hours. I then power cycled 5 times to see if I could get an error. Nope. I then turned it off and went and made coffee to let it cool down and did the same again. No error. Moved the jumper back to BootSel and it has booted fine 3 times and been running for 10 minutes.

Sure is hard to make it fail when the lid is off, once in 20-25 tries if I look back at the last hours.

Conclusion: something is pushing or shorting the ACA630 when case is closed, or it gets too hot then. Had a look now as I closed the lid, couldn't see anything that could push or short the ACA630.

Leaving the lid on now for a while, but not screwed tight. To see if it's the push or the heat, and to see if it goes unstable.

Next step is to re-seat the ACA630 and make sure the heat goes down.
Nothing should push against card, unless you have alot of extra's installed?
FOL is offline  
Old 01 August 2011, 22:42   #17
Photon
Moderator

Photon's Avatar
 
Join Date: Nov 2004
Location: Eksjö / Sweden
Posts: 5,079
Nothing did, of course. I just didn't remember from memory how it looked inside. I must have something like 5-17mm air above the cooling shield of the ACA630, depending on if it's below the keyboard or back at the rear.

OK, cleaned the 68000+ACA socket and re-seated the card. I mounted the heat extractor "instead of a floppy drive" to the top half of the case, air intake down. (No use sucking in air from outside through the top case grille, eh ) I mounted a small fan on the ACA cooling shield, to take the heat off it and out into the air so the heat extractor can blow it out. After 10 minutes you can feel slightly warmer than room temp air coming out of the floppy drive, so it seems to be doing something at least.

I've closed the case and shut the trapdoor, to make sure the test is for real. Two short sessions so far, no problem yet. I will now code for at least 4 hours under sharp conditions.


(Is there some kind of "burn-in" or test program to run for the ACA630, FOL? Perhaps you can recommend the memtest program you use for it at least?)
Photon is offline  
Old 01 August 2011, 23:29   #18
Predseda
Puttymoon inhabitant
Predseda's Avatar
 
Join Date: Mar 2007
Location: The City of Townsville
Age: 43
Posts: 6,602
Send a message via ICQ to Predseda
Well I had those problems somewhere in february. I havent turned on my A600 since. Until last week - I had to test my newly repaired monitor, so I connected it to my Amiga 630 and let it loop in Mr. Nutz demo for two and half hours. It didn't hang up, which it had did after about some 40 mins in february. I know, it doesn't mean anything, but in my opinion it points to some mechanical problem. I think it "got fixed" after I moved my Amiga from a party back to my office in my car, so something inside has now better contact or so. My suspicion is still microcracks on CPU. OR - it really can be caused by a heat.
Predseda is offline  
Old 01 August 2011, 23:56   #19
Photon
Moderator

Photon's Avatar
 
Join Date: Nov 2004
Location: Eksjö / Sweden
Posts: 5,079
After 1.25 hours it's perfectly acceptable to have my A600/030 in my lap

So it should mean I have really lowered the temperature buildup inside. And still working, haven't even rebooted yet. Fingers crossed...
Photon is offline  
Old 02 August 2011, 00:34   #20
FOL
PSPUAE DEV

FOL's Avatar
 
Join Date: Nov 2006
Location: Barry / UK
Age: 43
Posts: 5,902
Send a message via MSN to FOL
Quote:
Originally Posted by Photon View Post
After 1.25 hours it's perfectly acceptable to have my A600/030 in my lap

So it should mean I have really lowered the temperature buildup inside. And still working, haven't even rebooted yet. Fingers crossed...
You done too much, you cant find an issue like that.
I would have re-seated and then tested. If you were happy, then you could have added extra cooling.

Be nice to know exactly what it was.
FOL 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
Wanted: Aca630 werdy MarketPlace 1 31 January 2012 09:22
Unstable A1200 Composite DDNI support.Hardware 8 22 September 2009 01:29
Unstable Miggy NovaCoder support.Hardware 11 08 June 2008 15:36
Hostages... unstable Kristian95 project.WHDLoad 2 10 May 2008 18:01
Save state very unstable?! Anubis support.WinUAE 7 24 April 2007 12:33

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 10:29.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, vBulletin Solutions Inc.
Page generated in 0.14132 seconds with 14 queries