[SOLVED] The most ridiculous cooling problem

Discuss CT60/CT63, CTPCI, SuperVidel and EtherNAT hardware here.

Moderators: Mug UK, moondog/.tSCc., [ProToS], lp, Moderator Team

mikro
Atari God
Atari God
Posts: 1284
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

[SOLVED] The most ridiculous cooling problem

Postby mikro » Sat Apr 08, 2017 8:59 am

Eventually, this is going to end up in Willy's mailbox but who knows, maybe someone will have a clever idea at least to explain what's going on?

When I got my CT60e, since the day zero I could observe a really disturbing problem: Falcon/CT60e is happily humming, everything works (TBH, I don't remember one single bus error so far, knock-knock), I felt like an old sport with a decade of CT60 experience, nice! And then it comes, out of nowhere -- everything freezes. No error, no crash, no action from my side. Just the whole system is frozen in ice. Can't hear the key click anymore, can't CTRL+ALT+DEL, nothing. Just hardware reset, works for a while and then the same. Doesn't matter whether I'm watching a demo, booting FreeMiNT or staring at empty desktop. Can be clean boot, full featured setup, whatever. It's completely software agnostic. It behaves same with Willy's 1.03c, 2.01 and even the original 1.03c.

Checked everything I could think of, even ran the tests in Diagnostic cartridge, verified NVRAM, IDE, floppy, RGB/VGA, different (stronger) PicoPSU, everything. No problem found, 030 mode works like a breeze.

I think you agree this sounds awfully bad. But since I am the old sport ;), I didn't give up. After hours of testing I noticed one thing -- the freezes happen much less often during evenings. Even more interestingly, one night I noticed they don't happen at all (same as it didn't happen during the first night when I was testing the CT60e -- I could watch demos for hours). You, living in a country where nearly nobody cares to use bricks ;), already have a hunch -- it's Autumn in Australia, nights are getting colder and colder! Houston, have a cooling problem.

So you think "that was easy, just use a fan / replace existing fan / check your thermal glue". Trouble is the thermometer in the CPX shows CPU temperatures around 30 degrees. When I touch the CPU/heatsink, doesn't seem to be very far from the truth, definitely not hot, I'd call it "warm" at most. Fan is already spinning as crazy (so called "Boeing jet engine" mode) so... what's left here to replace?

I was nearly sure that the CPU must be wrong, that for some reason it stops functioning properly even on much lower temperatures (it's the 100 MHz mask). Luckily for me, I met one Amiga guy and he had a spare 060 to test. Did test it and what? Same freaking issue! So this isn't about CPU or my application of the thermal glue, it's something else.

But what? I briefly tested all ICs on the motherboard, nothing suspicious, only one DAC seemed to be a bit hotter but blowing air to that one down didn't help at all.

So what's left? CT60e. Neither of its ICs seemed hot and yet testing brought different results:

IMG_20170408_165059.jpg

Didn't help.

IMG_20170408_163533.jpg

Didn't help.

IMG_20170408_163245.jpg

Did help!

(but placing the fan in the left most position, next to the CPU, didn't help)

So obviously, something next to the first Xilinx IC needs cooling. I thought "OK, I'll lower voltage for the fan so it's not so noisy, write Willy about it and learn to live with that, hopefully it will get sorted out before I get SuperVidel". Lowered the voltage, installed everything and bam! Doesn't work!

The painful truth is that I need really, really strong air flow coming out of that fan (blowing *to* the PCB, not from it) else this trick doesn't work. Now I'm nearly deaf from that noise but -- lines.app is alive and kicking for half an hour now (usually the freeze happens after a couple of minutes).

So... now look at your CT60e and tell me, what, WHAT ON EARTH may need such strong cooling there, esp. if it isn't hot at all?!
You do not have the required permissions to view the files attached to this post.
Last edited by mikro on Mon Apr 17, 2017 12:15 am, edited 1 time in total.

User avatar
Ektus
Captain Atari
Captain Atari
Posts: 192
Joined: Mon May 24, 2010 2:58 am
Location: Germany
Contact:

Re: The most ridiculous cooling problem

Postby Ektus » Sat Apr 08, 2017 11:05 am

Maybe it's not the heat itself, but some heat-induced tension that moves a cold solder joint from "working" to "not working" and back? Take a magnifying glass or better some USB microscope and carefully check the solder joints in that area.

Regards
Ektus.
Schneider CPC464 (long retired), Atari Mega ST4 (retired), Falcon+CT2A, Falcon+CT63+CTPCI+Radeon 9250, Milan040+SCSI+Rage

User avatar
dhedberg
Atari Super Hero
Atari Super Hero
Posts: 536
Joined: Mon Aug 30, 2010 8:36 am
Contact:

Re: The most ridiculous cooling problem

Postby dhedberg » Sat Apr 08, 2017 11:21 am

I think Ektus could be right, check the solder joints.
Otherwise, could it be that the fan is indirectly cooling something else in that position, or could it be the pressure from the air-flow?
Daniel, New Beat - http://newbeat.atari.org

mikro
Atari God
Atari God
Posts: 1284
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sat Apr 08, 2017 11:28 am

Ektus wrote:Maybe it's not the heat itself, but some heat-induced tension that moves a cold solder joint from "working" to "not working" and back? Take a magnifying glass or better some USB microscope and carefully check the solder joints in that area.

Not bad thinking at all! At some point I had a similar idea but disregarded it as impossible. ;) Too bad I don't have a microscope. We do have such equipment at work but I'm too new there to ask for favours like that.

I have one last attempt in my sleeve: convince gazak28 to bring his Falcon and try the CT60e there, that would rule out any doubts for good. As a last resort, I will return the CT60e to Willy then. :-(

mikro
Atari God
Atari God
Posts: 1284
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sat Apr 08, 2017 11:31 am

dhedberg wrote:Otherwise, could it be that the fan is indirectly cooling something else in that position, or could it be the pressure from the air-flow?

Yep, I suspect something like that, too. Because it's way too random -- it is somehow related to CPU's temperature/frequency (85 MHz is way less prone to freeze than 95 MHz), sometimes it crashes even with the fan on that ICs and sometimes the fan seems to help on the CPU after all.

Only thing I can say for sure that if it's freezing cold out there (say 15 degrees) and the fan is on the heatsink, it's rock stable.

User avatar
1st1
Atari Super Hero
Atari Super Hero
Posts: 672
Joined: Mon May 07, 2012 11:48 am

Re: The most ridiculous cooling problem

Postby 1st1 » Sat Apr 08, 2017 12:06 pm

Have you swapped your DIMM module?
Power without the Price. It's not a bug. It's a feature. _/|\_ATARI

1040STFM in PC-Tower (PAK68/2, OvrScn, 4 MB, 1GB SCSI, CD-ROM...) * 2x Falcon 030 32GB/14MB+ScrnBlstrIII * 2x TT030 73GB/20MB+Nova * 520/1040STFM * 520/1040STE * 260/520ST/+ * some Mega ST * 2x Mega STE 500MB/4MB+M.CoCo * Stacy * STBook * SLM605 * SLM804 * SLM605 * SMM804 * SH 204/205 * Megafile 30/44/60 * SF314 * SF354 * 5x Pofo * PC3

mikro
Atari God
Atari God
Posts: 1284
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sat Apr 08, 2017 12:15 pm

1st1 wrote:Have you swapped your DIMM module?

Yep, have several of them. But in my experience, freezes doesn't happen due to bad RAM, bad RAM usually leads to crashes. I wish it were so simple. :-(

joska
Hardware Guru
Hardware Guru
Posts: 3624
Joined: Tue Oct 30, 2007 2:55 pm
Location: Florø, Norway
Contact:

Re: The most ridiculous cooling problem

Postby joska » Sat Apr 08, 2017 12:28 pm

mikro wrote:Only thing I can say for sure that if it's freezing cold out there (say 15 degrees) and the fan is on the heatsink, it's rock stable.


One solution would be to move to Norway. It's below 15 degrees 90% of the time.
Jo Even

Firebee - Falcon060 - Milan060 - Falcon040 - MIST - Mega ST - STM - STE - Amiga 600 - Sharp MZ700 - MSX - Amstrad CPC - C64

mzry
Captain Atari
Captain Atari
Posts: 271
Joined: Tue Jan 26, 2016 12:39 pm

Re: The most ridiculous cooling problem

Postby mzry » Sat Apr 08, 2017 2:14 pm

No1: Is your PCB bending to push down the bottom connection pins? I have mine half way up to prevent any bending
No2: Did you clean your connection pins with contact cleaner? You would be suprised how much oxidisation appears
No3: Did you try putting some RF insulation between the PicoPSU and the RAM? I find RF intereference an issue

mikro
Atari God
Atari God
Posts: 1284
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sun Apr 09, 2017 2:16 am

So far, Joska's suggestion of moving to Norway would be the best solution. ;-)

Today I made another experiment -- lowered the CPU@66 MHz, put the fan back on the heatsink and let Quake run for the whole night. When I woke up, it was still running.

Changed fan's wire to lower voltage (I have it prepared in such way that I don't need to touch anything), let it run for a few minutes and the freeze again. Changed again to 12V (and awful noise) -- running for another half an hour now.

The fact it's less prone to crash at lower frequencies (=> lower CPU temperatures) and that it actually seems to work at 66MHz if fan is spinning like crazy (or during the night when generally temperatures are low enough even for 95 MHz) still makes me suspect some weird problem with heat and not the 'nice' bugs like dry solder joint or so. On the other hand, I know next to nothing about electronics so if it's possible to change a joint's properties in 35 degrees while working OK in 28 degrees, then yes, I can't rule this out either.

If it weren't for the other CPU I'd be 100% sure it's a CPU's problem but... it is not. So the title stands, "the most ridiculous cooling problem", oh my.
Last edited by mikro on Sun Apr 09, 2017 3:52 am, edited 1 time in total.

mikro
Atari God
Atari God
Posts: 1284
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sun Apr 09, 2017 2:33 am

mzry wrote:No3: Did you try putting some RF insulation between the PicoPSU and the RAM? I find RF intereference an issue

Btw what do you use for this?

User avatar
TheNameOfTheGame
Atari God
Atari God
Posts: 1044
Joined: Mon Jul 23, 2012 8:57 pm
Location: Almost Heaven, West Virginia

Re: The most ridiculous cooling problem

Postby TheNameOfTheGame » Sun Apr 09, 2017 2:53 am

Contact Willy and swap the board for a new one. Sounds like a defect.

mikro
Atari God
Atari God
Posts: 1284
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sun Apr 09, 2017 3:50 am

TheNameOfTheGame wrote:Contact Willy and swap the board for a new one. Sounds like a defect.

I did but no reply so far.

User avatar
Ektus
Captain Atari
Captain Atari
Posts: 192
Joined: Mon May 24, 2010 2:58 am
Location: Germany
Contact:

Re: The most ridiculous cooling problem

Postby Ektus » Sun Apr 09, 2017 5:16 am

mikro wrote:
Ektus wrote:Maybe it's not the heat itself, but some heat-induced tension that moves a cold solder joint from "working" to "not working" and back? Take a magnifying glass or better some USB microscope and carefully check the solder joints in that area.

Not bad thinking at all! At some point I had a similar idea but disregarded it as impossible. ;) Too bad I don't have a microscope. We do have such equipment at work but I'm too new there to ask for favours like that.

I've got a cheap USB device but you could also try a normal digital camera in macro mode. Or take the CT60e with you to work and check it there during a break. Just ask nicely, it shouldn't matter how long you've been there. Worst thing is they can say "no".


Regards
Ektus.
Schneider CPC464 (long retired), Atari Mega ST4 (retired), Falcon+CT2A, Falcon+CT63+CTPCI+Radeon 9250, Milan040+SCSI+Rage

mikro
Atari God
Atari God
Posts: 1284
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sun Apr 09, 2017 6:14 am

Ektus wrote:I've got a cheap USB device but you could also try a normal digital camera in macro mode.

Hmm. Another issue is that I have no clue what to look for -- how a bad solder joint looks under microscope?

User avatar
Ektus
Captain Atari
Captain Atari
Posts: 192
Joined: Mon May 24, 2010 2:58 am
Location: Germany
Contact:

Re: The most ridiculous cooling problem

Postby Ektus » Sun Apr 09, 2017 9:37 am

mikro wrote:Hmm. Another issue is that I have no clue what to look for -- how a bad solder joint looks under microscope?

Browse Google Images with the search term bad solder joint or cold solder joint.
Mainly, cracks in the solder, insufficient wetting of pin or pad, short circuits between pins, slightly lifted pins of SMD components. Dull surface of solder where it should be bright and shiny.

Regards
Ektus.
Last edited by Ektus on Sun Apr 09, 2017 10:46 am, edited 1 time in total.
Schneider CPC464 (long retired), Atari Mega ST4 (retired), Falcon+CT2A, Falcon+CT63+CTPCI+Radeon 9250, Milan040+SCSI+Rage

mzry
Captain Atari
Captain Atari
Posts: 271
Joined: Tue Jan 26, 2016 12:39 pm

Re: The most ridiculous cooling problem

Postby mzry » Sun Apr 09, 2017 9:55 am

mikro wrote:
mzry wrote:No3: Did you try putting some RF insulation between the PicoPSU and the RAM? I find RF intereference an issue

Btw what do you use for this?


You can buy some Aluminium foil tape from Jaycar. I use plastic insulation tape first, then put foil over. Just be careful as it is conductive. I just insulated the bottom of my psu and the sides, left the top because of the heat.

joska
Hardware Guru
Hardware Guru
Posts: 3624
Joined: Tue Oct 30, 2007 2:55 pm
Location: Florø, Norway
Contact:

Re: The most ridiculous cooling problem

Postby joska » Sun Apr 09, 2017 10:04 am

mikro wrote:Changed fan's wire to lower voltage (I have it prepared in such way that I don't need to touch anything), let it run for a few minutes and the freeze again. Changed again to 12V (and awful noise) -- running for another half an hour now.


Maybe the clock generator is bad? What happens if you put a heatsink on it (small DS1085 between the Falcon power connector and the rightmost Xilinx)?
Jo Even

Firebee - Falcon060 - Milan060 - Falcon040 - MIST - Mega ST - STM - STE - Amiga 600 - Sharp MZ700 - MSX - Amstrad CPC - C64

mikro
Atari God
Atari God
Posts: 1284
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sun Apr 09, 2017 11:48 am

joska wrote:
mikro wrote:Changed fan's wire to lower voltage (I have it prepared in such way that I don't need to touch anything), let it run for a few minutes and the freeze again. Changed again to 12V (and awful noise) -- running for another half an hour now.


Maybe the clock generator is bad? What happens if you put a heatsink on it (small DS1085 between the Falcon power connector and the rightmost Xilinx)?

Yep, I plan to buy some heatsinks and try ICs one by one -- it will provide much finer resolution what's helping and what's not.

User avatar
TheNameOfTheGame
Atari God
Atari God
Posts: 1044
Joined: Mon Jul 23, 2012 8:57 pm
Location: Almost Heaven, West Virginia

Re: The most ridiculous cooling problem

Postby TheNameOfTheGame » Sun Apr 09, 2017 5:12 pm

Ektus wrote:
mikro wrote:Hmm. Another issue is that I have no clue what to look for -- how a bad solder joint looks under microscope?

Browse Google Images with the search term bad solder joint or cold solder joint.
Mainly, cracks in the solder, insufficient wetting of pin or pad, short circuits between pins, slightly lifted pins of SMD components. Dull surface of solder where it should be bright and shiny.

Regards
Ektus.


You make good suggestions, but of course any of those on a new product is a QA failure and defect. Also messing with it may void the warranty. Up to him but my opinion is probably let Willy have it for exchange or repair.

User avatar
Ragstaff
Atari Super Hero
Atari Super Hero
Posts: 609
Joined: Mon Oct 20, 2003 3:39 am
Location: Melbourne Australia
Contact:

Re: The most ridiculous cooling problem

Postby Ragstaff » Sun Apr 09, 2017 7:16 pm

If cold solder joints are a possibility, perhaps try reflowing the board in an oven? 200 degrees Celsius, 8 minutes...

joska
Hardware Guru
Hardware Guru
Posts: 3624
Joined: Tue Oct 30, 2007 2:55 pm
Location: Florø, Norway
Contact:

Re: The most ridiculous cooling problem

Postby joska » Sun Apr 09, 2017 7:44 pm

Ragstaff wrote:200 degrees Celsius, 8 minutes...


Sounds like a recipe for disaster...
Jo Even

Firebee - Falcon060 - Milan060 - Falcon040 - MIST - Mega ST - STM - STE - Amiga 600 - Sharp MZ700 - MSX - Amstrad CPC - C64

User avatar
willy
Atari nerd
Atari nerd
Posts: 45
Joined: Fri Apr 05, 2013 2:38 pm

Re: The most ridiculous cooling problem

Postby willy » Sun Apr 09, 2017 9:58 pm

Hi.
I got asked by micro to reply in this topic, but i couldn't do that earlier.

My first think is - it dos *not* make a sense.
You got one of the first ct60e's. It was hardly tested before i send it out to the world. I never use fan in tests, and run a random TBL demo for few minuts - at least.
68060 is an old style CPU, and Tjmax (max junction temperature) for proper operation is 110℃ ... so including a total inacurracy, any temp under 80℃ should be OK. All other IC's should work flawless up to minimum 70℃

I would rally want to trace this issue, so if you can do the following.

Remove the xtra fan, and set the CPU fan to minimum (or remove it). Start some demo or somethinh to test the stablility.
Take the screwdriver shown on your 1'st picture, hold in on the shaft, hold one of your hand any any metal element of falcon ports, and using your other hand, try to knock on every ct60e chip. Hard (but not so hard, consider your falcon..).
Don't worry, you are trying to trace down the worst possible issue. Even if it will completely stop working, it will be much better than unknown problem.

Let me know results.

@Ragstaff: ... no comments ...

BR
willy.

mikro
Atari God
Atari God
Posts: 1284
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sun Apr 09, 2017 10:23 pm

willy wrote:I got asked by micro to reply in this topic, but i couldn't do that earlier.

Welcome, glad to have you here. ;-)

My first think is - it dos *not* make a sense.
You got one of the first ct60e's. It was hardly tested before i send it out to the world. I never use fan in tests, and run a random TBL demo for few minuts - at least.
68060 is an old style CPU, and Tjmax (max junction temperature) for proper operation is 110℃ ... so including a total inacurracy, any temp under 80℃ should be OK. All other IC's should work flawless up to minimum 70℃

Yep, this matches what I thought -- it's not possible to see this kind of bug in such scenario -- when cold started, the machine runs at 95 MHz without any fan for at least 20 minutes, under heavy load. After that, I can make it freeze every 1-5 minutes (depending on the CPU frequency set). So yeah, seeing the CPU at 30 degrees definitely doesn't sound like something is overheating.

Remove the xtra fan, and set the CPU fan to minimum (or remove it). Start some demo or somethinh to test the stablility.
Take the screwdriver shown on your 1'st picture, hold in on the shaft, hold one of your hand any any metal element of falcon ports, and using your other hand, try to knock on every ct60e chip. Hard (but not so hard, consider your falcon..).

This is a good one, will do as first thing when I come home.

Don't worry, you are trying to trace down the worst possible issue. Even if it will completely stop working, it will be much better than unknown problem.

Completely agree. I'm happy to do any troubleshooting because sending the board back to you would knock me off of any CT60 activities for at least two months so anything is better than that (don't forget I still have the possibility to use it during nights :D).

ThorstenOtto
Captain Atari
Captain Atari
Posts: 154
Joined: Sun Aug 03, 2014 5:54 pm

Re: The most ridiculous cooling problem

Postby ThorstenOtto » Sun Apr 09, 2017 11:22 pm

mikro wrote:would knock me off of any CT60 activities for at least two months


two months is plenty of time to move to norway ;)

SCNR. I wish you all luck with this issue.


Social Media

     

Return to “CT60 / CT63 Area”

Who is online

Users browsing this forum: No registered users and 3 guests