Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Bolkonskij]

After I got my copper heatsink I replaced it with the (thin) aluminum one that came with my MDD when I got it. Well, well ... after removing the heatsink I discovered THIS:

http://images.macintosh.garden/2021/09/15/kebab-board.jpg

Apparently I drove it a bit too hard encoding all those Quicktime videos for Cornica. Today on the menu: roasted Sonnet upgrades :-(

Surprisingly, it still worked flawlessly! Went to carefully wipe the "smoke" off with isopropyl alcohol. Didn't come off 100%, but looking way better than in the initially taken photo above. Then repasted the CPUs with some Arctic Silver MX-4. After that I put the new copper heatsink on - which will hopefully avoid something like this in the future. Connected it to the Cinema Display and booted it - a chime. The welcome  screen appears. I'm in OS9. Whew! I'm sure ya'll can relate to the emotional rollercoaster ... :-)

Lesson learned: Don't save money on a quality heatsink. Especially with G4 / G5s ...

[FBz]

Sheesh!

Man, that’s a whole lotta hotta.


What’s with the aluminum foil looking appearance on the two chips? Is that Arctic Silver MX-4? Looks more like a metal foil covering. And the blackened areas look like extreme heat transference from an overly hot heatsink positioned just above them. Just how torqued-down (tight) was that heatsink that you removed?

Might also want to check that all fans are definitely working properly. How “full” is that case with HDs? I’d pull any remaining conventional spinning drives and only use SSDs on the back bus. (Nothing possibly obstructing air flow on the front bus from optimum airflow to the large fan to cool the CPU / heatsink.) AND quite possibly, say a permanent good-bye (remove /replace) any old Apple ADC powered Cinema Displays. Increased demand and thus an interior heat increase overall from a hotter PSU under that ADC load.

Looks as if you’ve already closely approximated the “China Syndrome” …so may be time to lighten the internal load overall and optimize cooling to possibly preserve or extend the life of that Sonnet. AND remove all unnecessary expansion cards, not directly related to your primary tasking. What’s that old equestrian adage? “Rode hard and put away wet”?

OR:
“Come to kindly terms with your ass, for it bears you.” -John Muir

Very loosely related errata, and just for fun…
https://alumni.berkeley.edu/california-magazine/fall-2017-bugged/zen-and-art-bug-repair
(Interesting Steve Jobs passage there too.) “…so it would never go out of fashion”.

[IIO]

i must have something like that in my atari. one day all of a sudden a strange smell filled the room so that i had to open the window.

[Bolkonskij]

The MDD has no HDDs for less heat / maximum air flow, just a single SSD inside. It gets regular cleaning with compressed air. Fans do work. heatsink was screwed tightly (I onced toasted a cpu with that, believe me, I make sure now). Really, the issue here was probably using the machine on a hotter day in conjunction with that (thin) aluminum heatsink. Bear in mind, this is an overclocked dual 1.67 Ghz Sonnet upgrade (the one from Knezzen) and it runs pretty stable, but obviously it needs a better cooling than the stock heatsink provides.

Speaking of stock heatsinks, can anyone confirm that dual 1.25 Ghz MDDs were actually equipped with the thin aluminum heatsinks? I thought Apple had already switched to the thick aluminum ones due to heat issues?

[peeperpc]

Speaking of stock heatsinks, can anyone confirm that dual 1.25 Ghz MDDs were actually equipped with the thin aluminum heatsinks? I thought Apple had already switched to the thick aluminum ones due to heat issues?

The thin aluminium one is the newer one. It dissipates heat better than the thick one. I have tried both.

[robespierre]

The black soot is from a burned tantalum capacitor (see the formerly yellow boxes at C2 and C4). They do this when electrically overstressed, it has nothing to do with the heat sink at all.

[FBz]

“They do this when electrically overstressed, nothing to do with the heat sink at all.”
                                                                                                               -robespierre
Thanks!

So, is it possible to simply check those tantalum capacitors (C2 and C4) using a
multimeter between ends for: (1). resistance, and/or (2). continuity
/ and thus… function? (Compared with others also present?)

AND, what about that LT 0621 “Buck Regulator”? Is that residual soot from the
tantalums overstressing or has it also suffered “overstressed” damage?

Like to see the backside too…
but might “rest” that Sonnet for a while and substitute a stock Apple 1.42 GHz
“out of an abundance of caution” /until the rare Sonnet can be checked.

[robespierre]

So, is it possible to simply check those tantalum capacitors (C2 and C4) using a
multimeter between ends for: (1). resistance, and/or (2). continuity
/ and thus… function? (Compared with others also present?)

Generally no, standard multimeters don't have the ability to measure capacitors in-circuit. There are specialized instruments called ESR meters that can do that (they inject a high-frequency signal, above 100 kHz, which gives a localized response of the capacitor without being affected by other components).

The dry tantalum capacitors are like little fireworks waiting to ignite: they contain a sintered mixture of tantalum and manganese dioxide, that can act as a type of thermite if it gets into "thermal runaway". The cap's normal operating temperature is up to 125°C, and ambient temperature around the heatsink is sure to be less than that. They are, however, sensitive to voltage fluctuations outside their design limit: even a brief spike beyond the safe level can cause them to ignite.

The switching regulator chip is likely fine; the black soot came from the capacitor. The actual silicon IC inside that package is very small: if it blew there would be a tiny chunk ejected from the top of the package. It would be a good idea to clean off all the soot with alcohol and a swab, including underneath as much as possible.

[Bolkonskij]

aha, thanks for the clarification, robespierre! Yeah, i found the tiny chunk ejected from the top of the cap during cleaning.

So you'd attribute the problem to a failing PSU? I had mine die on me about a year ago and since replaced it. Maybe it took the caps with it?

Anyway, I'll look into getting it professionally replaced.

[FBz]

Yes robespierre, greatly appreciate the very thorough and concise explanation. Thank you.

*I was partially implying that with a multimeter, that it might be possible to detect a “complete & total” failure of individual capacitors - IF no continuity could be detected whatsoever, even while still in-circuit.

My ESR meter originally touted (before purchase) the ability to measure capacitors “in-circuit”… which was later found to be partly true, except that one leg of the capacitor had to be free / removed from the circuit board in order to do so properly. Now, I just remove questionable caps completely and then test with ESR meter.

AND Bolkonskij, Knez can possibly guide / help you overclock a somewhat lesser CPU for use while the Sonnet’s “in the shop”. I’ll be looking into this too but got sidetracked yesterday with discovery of overclocking a Quicksilver’s System Bus. (Off to pick up those resistors now.)