[SOLVED] The most ridiculous cooling problem

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

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

User avatar
jvas
Captain Atari
Captain Atari
Posts: 434
Joined: Fri Jan 28, 2005 4:30 pm
Location: Budapest, Hungary
Contact:

Re: The most ridiculous cooling problem

Postby jvas » Sat Apr 15, 2017 7:19 pm

asik wrote:
jvas wrote:The clock patch must be removed in case of the CT60. Not sure about the CT60e. Don't know how it is related to the fan though.

Are you sure about clock patch (please double check).
I used many years ct63 (107mhz stock ct63 cooling system and stock closed falcon case) with clock path, and use ct60e with rev. six cpu, at this time only 66mhz, and 75mhz mhz without cooling system.
But i install big cooling and will try ct60e go to 110mhz :)
I bought new cpu rev. six for ct60e ,without cooling system is stabil 75mhz one hour on quake game.
When i come back to home a try more mhz, at this time is easter holiday....


It is according to Rodolphe Czuba. See Mikro's post above.

mikro
Atari God
Atari God
Posts: 1010
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sun Apr 16, 2017 4:53 am

OK, finally I can report some real progress (!).

Good news is that the CT60e is now completely out of suspicion so I'm really glad I listened to my instinct and been patient enough to explore this issue into great detail without baking / shipping the CT60e away :).

This new information about clock patch got me thinking -- I can clearly reproduce the workaround with fan. The machine works fine, there are no hiccups or crackles, no problems, 030 mode is also fine, basically all it boils down to a mysterious temperature problem. The only difference between my machine and the other (working) one is clock patch. Shielding is the same, CT60e+peripherals is the same, temperature conditions are the same.

So what if the problem is my clock patch but not the usual "my machine is crashing because of it" but that it's the clock patch which has a temperature problem? Mind you, U63 is hidden beneath the shielding. So I orchestrated an experiment like this:

IMG_20170416_135611.jpg

And what happens in this crazy setup? It works (30+ minutes and counting, even on the low rotation setting) ! It's really the 74F0x (can't read it due to glue) or some of its wires which suffers from this (admittedly, still mysterious) problem.

Of course, after all this I'm super paranoid, so I made sure the machine is still freezing within a few minutes as soon as switch off the fan and then again starts working as soon as I switch it on without touching anything else.

Even though it's not totally obvious, it also explains why it worked only in certain positions, with only highly aggressive air flow -- the cold air was bounced against the SDRAM and/or DOM on the motherboard and found its way beneath the shielding. No wonder that nobody could imagine this, it's really not so obvious.

Now, that's all nice but it still doesn't explain why U63 or its "mate" on top of it needs the cold air. But looking at it (now I know where to look):

IMG_20170416_144403.jpg

gives me some hope. The three wires look extremely strange (and fragile) plus we can only guess what's happening in that glued part. First thing I'll try is to change the wires for something more, erm, common and perhaps resolder all the connections to the 74F0x. If that doesn't help, perhaps I should order clock patch from exxos (or try the Czuba's suggestion although this one looks far harder to execute).
You do not have the required permissions to view the files attached to this post.

User avatar
dhedberg
Captain Atari
Captain Atari
Posts: 391
Joined: Mon Aug 30, 2010 8:36 am
Location: Sweden
Contact:

Re: The most ridiculous cooling problem

Postby dhedberg » Sun Apr 16, 2017 9:22 am

Way to go mikro. Mystery solved!
Daniel, New Beat - http://newbeat.atari.org - ACE Tracker v2.0 is out - get it from our web site!

mikro
Atari God
Atari God
Posts: 1010
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sun Apr 16, 2017 9:34 am

mikro wrote:The three wires look extremely strange (and fragile) plus we can only guess what's happening in that glued part. First thing I'll try is to change the wires for something more, erm, common and perhaps resolder all the connections to the 74F0x.

Well, it didn't help. First I removed all the glue / silicon and checked -- nope, then replaced the wires and resoldered all joints on the clock patch IC. Good news is that it's behaving still the same (i.e. I didn't break anything).

The IC is (S)74F04N (FPX0407, 9405VF for those who know what it means :-)), from what I see it's the exactly same IC as Exxos uses: http://www.exxoshost.co.uk/atari/last/f ... /index.htm however wiring is bit different (photo taken before my changes):

smaller.jpg

(perhaps due to the PowerUp 2 box which takes care of some things).

EDIT: OK, it seems that this clock patch is 100% identical to the Atari one (version with 74F04), except:
  • PIN 8 doesn't go directly to R222B but: PIN8 -> blue wire -> the box -> yellow wire -> R222B
  • There's a white wire coming to U68 (and three others going to the right of the NVRAM IC)
I'm really tempted to try disabling this 'detour' and connect to R222B directly. Although I can't imagine why this detour would make the 74F04 angry and hot. ;-)

So now there are two possibilities:

1. The 74F04 has gone bad for some reason and it just needs replaced and everything will be fixed
2. Falcon can no longer work with this kind of clock patch and needs something as the one by Exxos (maybe just the wires to the box are too long, who knows).

Fortunately, I can try 1.) while waiting for 2.). Only thing which really saddens me is that I'm going to lose the PowerUp 2 in case of 2.) -- it doesn't make sense to have it with the CT60e of course but in case I'd like to keep the CT60e and sell the Falcon, it would require additional work to put it back again. Perhaps that means I shouldn't ever sell it. ;-)
You do not have the required permissions to view the files attached to this post.
Last edited by mikro on Sun Apr 16, 2017 10:34 am, edited 4 times in total.

EvilFranky
Atari Super Hero
Atari Super Hero
Posts: 828
Joined: Thu Sep 11, 2003 10:49 pm
Location: UK
Contact:

Re: The most ridiculous cooling problem

Postby EvilFranky » Sun Apr 16, 2017 9:37 am

Glad you found the root cause, this would have made me go insane.

mikro
Atari God
Atari God
Posts: 1010
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sun Apr 16, 2017 9:56 am

Now the real dilemma is order of things to try:

- Should I just remove the PU2 and try it out (can do immediately, downside is all the soldering if I ever decide to put it back one day and also it's very unclear how this should affect temperature of the 74F04)
- Should I keep the PU2 and try to replace the 74F04? (easiest route but have to wait to purchase a new one)
- Should I just order Exxos' clock patch and remove everything (biggest chance to success as this setup is proven to work but also requires a lot of wating /exxos is still working on his house/)

What would you do?

EDIT (so many edits..,): after more careful look it seems that Exxos' clock patch is basically the same as the Atari '04 patch with buffered resistors. So whatever effect those resistors do, it's certainly not lowering the temperature. ;-) This would point to the '04 inverter as the culprit (or, less likely, to the PU2 box). I think I'm going to try disabling the PU2 box, if that doesn't help -> purchasing new 74F04.

User avatar
Ektus
Captain Atari
Captain Atari
Posts: 168
Joined: Mon May 24, 2010 2:58 am
Location: Germany
Contact:

Re: The most ridiculous cooling problem

Postby Ektus » Sun Apr 16, 2017 10:49 am

Isn't there a gate (or two) free on the 74F04? You could try to just switch over to another gate (needs soldering of two wires on the 74F04).

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: 1010
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sun Apr 16, 2017 11:26 am

Warning, this post contains traces of optimism and possibly false hope!

mikro wrote:(or, less likely, to the PU2 box)

So I've gone this path first, as mentioned in the previous post. But I didn't remove whole PU2 as the U68 pin is output anyway + the others make sense only for overclocking (i.e. they are dormant when PU2 is off).

So what I've done: cut the yellow wire in the middle (so it fits nearly to the mm to U63's pin 8), desoldered the blue wire from U63 and placed the yellow one there instead.

Result: Quake is shooting monsters at 85 MHz for about 20 minutes now and the most surprising part: the reset problem is gone !!! (OK, I must confess I wont believe this claim until I see it boot into 060 mode on every startup for following 10 days...)

Ran also SDMATEST.PRG, no sound cracks as well.

Hm. Didn't see this coming. ;-)

User avatar
dhedberg
Captain Atari
Captain Atari
Posts: 391
Joined: Mon Aug 30, 2010 8:36 am
Location: Sweden
Contact:

Re: The most ridiculous cooling problem

Postby dhedberg » Sun Apr 16, 2017 2:21 pm

Great! If you continue like this you'll soon receive a doctorate in electrical engineering. ;-)
Daniel, New Beat - http://newbeat.atari.org - ACE Tracker v2.0 is out - get it from our web site!

joska
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 3383
Joined: Tue Oct 30, 2007 2:55 pm
Location: Florø, Norway
Contact:

Re: The most ridiculous cooling problem

Postby joska » Sun Apr 16, 2017 2:30 pm

jvas wrote:The clock patch must be removed in case of the CT60. Not sure about the CT60e.


If I understand it correctly this is because the CT60 also has a built-in bus speeder, and Rodolphe devised his own clockpatch for that.
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: 1010
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Sun Apr 16, 2017 3:11 pm

joska wrote:If I understand it correctly this is because the CT60 also has a built-in bus speeder, and Rodolphe devised his own clockpatch for that.

This is not 100% true. As mentioned here: viewtopic.php?f=97&t=31501 the clock patch and bus acceleration are two different things, even in the context of CT60 fitting.

What R. Czuba proposes is just a super-simple SDMA fix on its clock line, keeping (reverting) the SMD resistors. After doing that you have an option to add the bus speeder stuff but it's not related ("devised for that").

It's really confusing, in recent days I've counter four (!) different solutions how to do the clock patch (ignoring projects like Phantom or my PowerUp 2) -- three of had been issued by Atari (!!!) with no clear indication when to use what.

User avatar
wongck
Ultimate Atarian
Ultimate Atarian
Posts: 11583
Joined: Sat May 03, 2008 2:09 pm
Location: Far East
Contact:

Re: The most ridiculous cooling problem

Postby wongck » Sun Apr 16, 2017 10:48 pm

You need for thermal paste (or any media in between) to transfer the heat from the CPU/IC to the heat sink.
Without it or just placing a fan/heatsink on the CPU/IC will not transfer heat away from the COU/IC fast enough.
Like heat needs to travel across the air gap (which in fact air being an insulator) is definitely not going to happen.
My Stuff: FB/Falcon CT63+CTPCI ATI R7500 14+512MB 30GB HDD CF HxC_SD EtherNEC/ TT030 68882 4+32MB 520MB Nova/ 520STFM 4MB Tos206 SCSI
Shared SCSI Bus:ScsiLink ethernet, 9GB HDD,SD-reader @ http://phsw.atari.org
My Atari stuff for sale - click here for list

mikro
Atari God
Atari God
Posts: 1010
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

[SOLVED] The most ridiculous cooling problem

Postby mikro » Mon Apr 17, 2017 12:13 am

It's alive, it's alive, it's aliveeeeeee!

... said mikro when entered his Atari room and noticed that after 8 hours running at 95 MHz (with fan on the CPU this time, lowered voltage) the Falcon is still all right.

So in the end:

- for some reason the PowerUp 2 box was giving the 74F04 really hard time with the CT60e up to the level it couldn't work without fan after a while

- for some even more mystical reason, the PowerUp 2's box is too late/slow/whatever for the CPU clock pin on R222B leading to inability to boot in 060 mode

Interestingly, I have one more Falcon with PowerUp 2 at home in Slovakia and I clearly remember it was giving me headache with CT60 too. It was so frustrating that I was considering selling the Falcon because of that. But it's true it was different, I could only observe the reset problem (which, after a time, got even worse, couldn't start in 060 at all). Who knows, maybe the 74F04 had fried in the end.

Thank you everyone for all the hints and remember -- be patient and don't give up so easily! (plus, if you really love your Atari, learn to solder -- without this, newly acquired, skill I would be lost)

ThorstenOtto
Atari maniac
Atari maniac
Posts: 89
Joined: Sun Aug 03, 2014 5:54 pm

Re: [SOLVED] The most ridiculous cooling problem

Postby ThorstenOtto » Mon Apr 17, 2017 2:59 am

Congratulations ;)

acsi
Atari User
Atari User
Posts: 39
Joined: Thu Jun 12, 2014 11:20 am

Re: [SOLVED] The most ridiculous cooling problem

Postby acsi » Mon Apr 17, 2017 3:44 am

Congrats.

I've been watching this with interest and need to look at my own board when I finally dig out the beast as I get sound glitches and noted somewhere in the pages you noted that this is one symptom of the clock patch being incorrect or non existent.

User avatar
rian_ata
Captain Atari
Captain Atari
Posts: 269
Joined: Mon Apr 24, 2006 10:00 pm
Location: Netherlands
Contact:

Re: [SOLVED] The most ridiculous cooling problem

Postby rian_ata » Mon Apr 17, 2017 10:16 am

acsi wrote:Congrats.

I've been watching this with interest and need to look at my own board when I finally dig out the beast as I get sound glitches and noted somewhere in the pages you noted that this is one symptom of the clock patch being incorrect or non existent.


Yeah, me too. I have sound glitches in certain demos. I think my issues are patch / DSP related too. However my technical skills are not top notch to do the fixes myself, so I need to look out for help somewhere over here in the NL.

User avatar
checksum
Atarian
Atarian
Posts: 7
Joined: Mon Apr 17, 2017 9:36 am

Re: [SOLVED] The most ridiculous cooling problem

Postby checksum » Mon Apr 17, 2017 12:13 pm

Nice to see that the problem is solved. Hope that peoples that have problems with CT60e will take a look at the patchs present on the motherboard.

A good source of explanation is always Rodolphe website: http://www.powerphenix.com/CT60/english/Fitt_nosold.htm. I patched my MB 2 years ago after running SDMATEST... and using the patch explained on Rodolphe's page. The irony of this, is that i leave Atari world in 1993/94 mostly because of this hardware bug and i never released ProTracker DSP. And now it's working like a charm even on a CT60e...

mikro
Atari God
Atari God
Posts: 1010
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: [SOLVED] The most ridiculous cooling problem

Postby mikro » Mon Apr 17, 2017 12:47 pm

Checksum, welcome aboard, having another DSP coder around is always pleasure!

joska
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 3383
Joined: Tue Oct 30, 2007 2:55 pm
Location: Florø, Norway
Contact:

Re: The most ridiculous cooling problem

Postby joska » Mon Apr 17, 2017 8:52 pm

mikro wrote:This is not 100% true. As mentioned here: viewtopic.php?f=97&t=31501 the clock patch and bus acceleration are two different things, even in the context of CT60 fitting.


No it isn't. Bus acceleration and clock patch is tightly connected. The clockpatch is all about timing. As soon as the bus is accelerated your current clock patch may or may not work.
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: 1010
Joined: Sat Sep 10, 2005 11:11 am
Location: Brisbane, Queensland, Australia
Contact:

Re: The most ridiculous cooling problem

Postby mikro » Mon Apr 17, 2017 10:28 pm

joska wrote:No it isn't. Bus acceleration and clock patch is tightly connected. The clockpatch is all about timing. As soon as the bus is accelerated your current clock patch may or may not work.

Agreed. I meant that you can do Czuba's clockpatch without thinking about CT60 bus acceleration, Checksum has just mentioned it -- he did the clockpatch part to fix the crackles but he didn't have bus acceleration in mind.

User avatar
checksum
Atarian
Atarian
Posts: 7
Joined: Mon Apr 17, 2017 9:36 am

Re: The most ridiculous cooling problem

Postby checksum » Tue Apr 18, 2017 10:14 pm

mikro wrote:
joska wrote:No it isn't. Bus acceleration and clock patch is tightly connected. The clockpatch is all about timing. As soon as the bus is accelerated your current clock patch may or may not work.

Agreed. I meant that you can do Czuba's clockpatch without thinking about CT60 bus acceleration, Checksum has just mentioned it -- he did the clockpatch part to fix the crackles but he didn't have bus acceleration in mind.


Right. Just mention this, because it's in the troubleshooting part of the CT60 concerning stability.


Social Media

     

Return to “CT60 / CT63 Area”

Who is online

Users browsing this forum: No registered users and 2 guests