All replacement capacitors in iMac G5 PSU burned up

[PCBONEZ]

Re: All replacement capacitors in iMac G5 PSU burned up

Capacitors resist changes in ___________ ?
Inductors resist changes in ___________ ?

'Inrush current' ring a bell?

What happens when you don't limit how fast the caps charge at turn-on?

.

[i4004]

Re: All replacement capacitors in iMac G5 PSU burned up

>think of your coil with shorted turns like a transformer with its secondary shorted.

i would...if it had at least 1/100th of a number of turns of typicall trafo secondary..
<wink>

bonez, i don't think they put coil because of inrush current...for that it's better to use resistor...
coil resists alternating current, so here coil should be resisting remains of the ac that's superponed to dc output..

coils in emi filters can be viewed as inrush limiters(to some extent...there's also resistor there)...how many turns they have?

i think this circuit would work without coil(ie jumper instead of coil) but obviously that wouldn't be too good for the mobo...

flywheel analogy?
it just reminds me of another bloke on another forum(that i likened to you once before); he's comparing satelite receivers to cars...
hehe...

[Wizard]

Re: All replacement capacitors in iMac G5 PSU burned up

Coils/inductors are GOOD idea to place between diode and the capacitors, keep current spikes down and smooths noise so diode and capacitors don't get stressed as much.

Some of them use a 12V or 5V to generate 3.3V that is why I NEED the bloated or failed PSU to analyse.

Cheers, Wizard

[Toasty]

Re: All replacement capacitors in iMac G5 PSU burned up

1) What does an inductor do i4004 ?

2) Here's a hint i4004. - An inductor to a circuit is as a flywheel to a machine. [Come on... I KNOW you know.]

3) Capacitors resist changes in ___________ ?
4) Inductors resist changes in ___________ ? 'Inrush current' ring a bell?

5) What happens when you don't limit how fast the caps charge at turn-on?

Is this a test? Nobody said anything about there being a test!!

Are you really asking these questions or is this just another "let's tease i4004" post?

Toast

[i4004]

Re: All replacement capacitors in iMac G5 PSU burned up

toasty, bonez is not really that much into electronics...if he was he would compare capacitor to flywheel...
<wink>
(that said, he's rather good at reading cap specs pdfs!
to tease him right back)

offcourse he'll now just deny everything and say it's all my fault...hihi...

btw. gotta love your avatar...

[PCBONEZ]

Re: All replacement capacitors in iMac G5 PSU burned up

I'm an electronics tech by trade since 1981.
Specialty is Control Systems for NUC Reactors.
Cross trained in Electrical, Mechanical, and Chemical Engineering.
Certified Electronics Tech and Reactor Operator on 4 platforms.
Certified QA Inspector for repairs/overhauls of Trident Subs.
Been building Hot Rods since I was 17.
Been working on PC's first as a hobby then a side job since 1995.
-
i4004 is a 'shade tree mechanic' that fixes electronics with glue and duct tape.
Not entirely his fault because he lives in part of the world where one can't get good parts.
Problem is he doesn't understand that make-shift techniques aren't needed elsewhere.
.

As he doesn't understand the analogy between electrical and mechanical systems I'd have to say he doesn't understand what inductors do with electrons.

.

[PCBONEZ]

Re: All replacement capacitors in iMac G5 PSU burned up

When you work on PCBs smaller than that on a laptop HDD that cost upwards of $60,000 [in 1980's/1990's dollars] you tend to look at data sheets and specs before you go replacing things.

When you're stuck rebuilding things out of used parts because you can't get the real deal you do what you have to do.

There is nothing wrong with either one within the given situation but the attitudes that come out of that are why i4004 and I never agree.
[And piss each other off a lot.]

.

[i4004]

Re: All replacement capacitors in iMac G5 PSU burned up

hold on matey..you're 'shade tree mechanic' as you work on your car!
<wink>

you know when people put up their credentials?
when they're running out of ammo...hehe...

what you are, or what you do/did doesn't change the fact that i would like to know how coil gets hot, and otoh, you don't seem to have that interest at all, you would just change it...otoh, i would change it too, BUT ALSO i would try to explain why was it getting hot...you know, just like you trying to explain why caps go bad with cheap aluminum etc.
so no, i won't feel bad because you think we shouldn't be explaining this...i say we should...

>Problem is he doesn't understand that make-shift techniques aren't needed elsewhere.

who said that? i said that if parts are available at acceptable price, get best parts you can.
and oh yeah, why are you offtopic yet again?
<wink>

your analogy is not impressive to me because..well...capacitor is device that's storing energy more directly than inductor..ie it stores electric energy directly...
flywheel
"A flywheel is a mechanical device with significant moment of inertia used as a storage device for rotational energy."
cap
"This field can be used to store energy, to resonate with a signal, or to link electrical and mechanical forces."

inductor also stores energy, sure, but it's a thing of magnetic/electric energy conversion etc.
it is also not so obvious; cap stays charged for long time; coil will typically release it's energy immediately...

why is your analogy completely flawed? well, coil is not used here to store energy at all. quite the opposite, it's used to BLOCK particular energies(those of ripple/noise).
also, it is equally not used to limit inrush current, because there really is no need for that...output of trafo are relatively short pulses of energy, not some giant inrush that would damage the cap. and the voltages are overall minuscule.
remove coil and see if cap gets damaged.

offcourse, if you have better explanation, i'm all ears, but otherwise, your electronics knowledge got a bit rusty....

>There is nothing wrong with either one within the given situation but the attitudes that come out of that are why i4004 and I never agree.
[And piss each other off a lot.]

you're kinda blowing it out of proportion: we essentially agree, but you're not in my situation, so you have something of an understanding problem.
example; i'm gonna back-down when you prove me wrong(every time) but you won't when i say we can't get those parts at reasonable price.
what i dislike about you the most is that you can't seem to forget the past: move on, this is not about what we can find locally(that discussion is over), this is about coil getting hot, concentrate on that.
we don't have to bicker about the past...unless you insist...
oh yeah, sorry...too long for you to read...hehe....however, i'm not writing for you only..
<wink>

after all, it's a crash course about coils by now...

so remember, the torrid coil, not you or me, nor caps from the other thread...
(i'm currently taking a rest froum our lil "Caribbean Sea expedition"...hehe)

[Toasty]

Re: All replacement capacitors in iMac G5 PSU burned up

...'shade tree mechanic' that fixes electronics with glue and duct tape...

Actually, both PCBONEZ and i4004 analogies are correct. Inductors and capacitors both act as "flywheels". Since they are both connected here in a LC circuit, I'm wondering if the drive frequency is out of spec causing the coil to saturate? Or, is the coil shorted, then it's a simple matter. What has me consternated is that it appeared fine until the caps were replaced.

I just wish someone (hint HINT) would check and see if that coil is shorted to the core. If it is, we can move on to replacing it. If not, as I stated previously, we have a problem elsewhere.

I've moved away from thinking this is an mobo overload condition, with reservations, as the choke located below and to the right of the burnt coil in the second photo in Post # 12 of this thread, is in series to one of the feed points at the lower right. (voltage?) It does not appear to be damaged at all.

I don't recall reading that the voltages were measured when this is (a) disconnected and (b) connected to the system?

There is a similar (if not identical) torOId under the heatsink right next to the cap that nwb did not replace. Both coils are connected to the device mounted on the backside of the heatsink. I'd like to know what that device is. FET, transistor, dual Schottky? Inquiring minds wanna know...

nwd - You gotta confirm that other cap is okay. From what I've seen, it's a 330uF 35v. I'll bet it's out of spec, if not just plain open. It's connected to the power output supplied by the middle size of the 3 transformers, which appears to feed the nearby IC and its associated circuitry.

The cap is sitting in an oven where it is. Yes, it's a pain to get to. You'll have to remove the cap in front of it and perhaps the nearby choke (coil with straight core). There appears to be just enough clearance to get it unsoldered and lift it straight up then slide it out. The replacement cap is going to have to be the same height.

God I hate AC theory....

Toast

[Toasty]

Re: All replacement capacitors in iMac G5 PSU burned up

coil is not used here to store energy at all. quite the opposite, it's used to BLOCK particular energies(those of ripple/noise)

I disagree. Please remember that this is a square wave oscillator and the frequency is upwards of 30kHz. The LC circuit is "kicking" the current here. BOTH the capacitor and the inductor are theoretically working together (resonance) when the chopped voltage goes to zero. As the current ramps up, the coil passes the current and also feeds the cap. Then when the current shuts down the inductor resists the "backflow" of current that the cap is providing. You then, in effect, have an RMS value of DC current.

Are you 2 done shaking your "boys" at each other now? LMAO

Toast

[nwd]

Re: All replacement capacitors in iMac G5 PSU burned up

I stopped by the hardware shop today and then Radio Shack and borrowed multi-meters at each place to test coil. When set to ohms, I got a reading touching both ends to the core (not the wire) so core is metal. When I touched one end to the inner part and one to one or the other of the leads on the backside of the board, sometimes I would get a reading and sometimes not. Not sure if I just don't know how to use a meter correctly or if that represents an intermittent short. I plan to hit a decent electronics shop 30 miles south tomorrow to try to find a replacement and get the unreplaced capacitor tested and changed if possible. Is it possible to test it in situ? Does it need to be tested with power supply plugged in. Does the power supply need to be connected to the MOBO? I already have the replacement.

[Toasty]

Re: All replacement capacitors in iMac G5 PSU burned up

I stopped by the hardware shop today and then Radio Shack and borrowed multi-meters at each place to test coil. When set to ohms, I got a reading touching both ends to the core (not the wire) so core is metal. When I touched one end to the inner part and one to one or the other of the leads on the backside of the board, sometimes I would get a reading and sometimes not. Not sure if I just don't know how to use a meter correctly or if that represents an intermittent short. I plan to hit a decent electronics shop 30 miles south tomorrow to try to find a replacement and get the unreplaced capacitor tested and changed if possible. Is it possible to test it in situ? Does it need to be tested with power supply plugged in. Does the power supply need to be connected to the MOBO? I already have the replacement.

Good Job! Now we are getting somewhere!

Any reading between the core metal and the leads on the backside indicates the presence of a short. The coil must be rewound (it can be, since it is iron powder) or replaced. I'd go for the replacement.

You can test the cap in the circuit with an ESR meter as described elsewhere in the forums. I wouldn't worry about testing it though, since you already have a replacement. Just get it changed.

No, the power supply should be OFF and the cap discharged.

If you can get the voltages checked out at the electronics store, that would be a peach!

Cheers!
Toast

PS: The info needed on the coil would be:
-Number of turns
-inside diameter of core
-outside diameter of core
-width of core
-the gauge of the wire (a micrometer may be necessary as they use an oddball gauge of wire sometimes, like #11)
Any HAM/Amateur radio operators or shops in the area where you are? That would be the best place to go for the coil. Find yourself an Elmer. ...Try the Yellow Pages...

[kikkoman]

Re: All replacement capacitors in iMac G5 PSU burned up

i'd just replace that cap. as toasty said, it's gonna be quite some fiddling about, but there's no way this cap could be okay.
as for the replacement inductor: if is really is an iron core you could re-insulate and rewind it.
if you buy a replacement, i'd go for a value between 10 and 15 µH.

[PCBONEZ]

Re: All replacement capacitors in iMac G5 PSU burned up

Before I moved down here my "shade tree" had 4 stalls and an office with a fireplace.
Garage was built out of the gutted out 'old house' but was detached from the 'new house' where we lived.
I rebuilt/modified cars from the frame up.
Can't do that much anymore due to back problems and arthritis so when I moved I sold off the paint booth, sand blasting gear, compressor, arc-gas-mig welders, hoists, presses, cherry-picker, frame straighter, mill, drill press, and all that crap figuring I'd replace it later if I ever got interested again. So far I haven't though I've replaced some of the more universal tools anyway.
.
Credentials. - At least I have some

~~~

An inductor resists changes in current flow [electron movement] just as a flywheel resists changes in mechanical movement.
-
When movement is first applied a the system the flywheel [inductor] restricts the sudden change in movement by forcing it to increase slowly to it's final value.

A capacitor on the other hand does the opposite.
When movement [voltage causing current] is first applied the capacitor is 'empty' and acts like a vacum to electrons increaing the flow [current] suddenly at first then tapering off as it charges up.

This is why there is a torrid toreaid toride TOROID placed before the output filters.
It's not part if the PI filter, it's a current surge limiter before the PI filter.

If the torrid hemorrhoid dinknoid TOROID is shorted then caps trying to charge instantaneously could easily cause it (or something else) to overheat, melt, break.

So, if the TOROID was marginal and then you put in nice STRONG caps [that are capable of charging much 'harder' than than the old ones] that may have been enough to push it over the cliff.

It's entirely likely that MOSFETs, IC's, and/or main xfmer were damaged too which is why I suggested recapping a unit that at least works.

Core saturation is irrelvant if the coil is shorted.
There are effectively no windings for the magnetic fields to act on.

-

I would like to point out now that a nitpicker is a picker of nits.
Nits are head lice.

.

[i4004]

Re: All replacement capacitors in iMac G5 PSU burned up

>This is why there is a torrid toreaid toride TOROID placed before the output filters.
It's not part if the PI filter, it's a current surge limiter before the PI filter.
If the torrid hemorrhoid dinknoid TOROID is shorted then caps trying to charge instantaneously could easily cause it (or something else) to overheat, melt, break.
So, if the TOROID was marginal and then you put in nice STRONG caps [that are capable of charging much 'harder' than than the old ones] that may have been enough to push it over the cliff.

hehe...you're aware of one thing: this instantenous charging of caps happens only ONCE: when you turn it on.
so what you're suggesting is that one voltage increase (from 0 to 5 or 12v..and via short pulses, ergo not dramatically fast anyway) will cause such thick wire to overheat?
or are you suggesting the core metal(incase it's shorted to the core) can't take that amount current?
heh...

toasty, i don't really see it as lc circuit(like oscillator), but as a combination of cap that "grounds" ripple, and coil that "blocks" the ripple.
ie like this
http://en.wikipedia.org/wiki/Capacitor-input_filter
on this pic

we can see "full" pi filter, and one without another cap.
why?
well, probably because just one cap and one ind. suffice on those rails...

>I just wish someone (hint HINT) would check and see if that coil is shorted to the core. If it is, we can move on to replacing it.

that's just fine and dandy, indeed, "shorted" coil(still hard to call it shorted because it only has what, 8 turns..heh) doesn't serve any purpose and should be replaced.
BUT what interests me the most is why is it getting hot.
ie, in pi filter circuit that's otherwise functioning correctly, why would the shorted coil get hot?
i offered 2 explanations:
-eddy currents formed in the core
-cap replacement cause more noise/ripple in the circuit and this overstressed the coil
( http://www.jonnyguru.com/modules.php...ory4&ndar_id=8 )
so after that they INDEED become somewhat of an resonating oscillator...hehe...
(perhaps a thing of fuhhyu having less capacity than stated..perhaps low esr there is just not good->perhaps it leads all the noise back to the trafo and makes a loop, actually amplifying it...while caps with somewhat higher esr[offcourse, if esr goes too high they just are not "grounding" noise at all] dampen this noise down...perhaps )

---
i think pcbonez and i would probably be pretty damn good friends if we were near each other(i think he loves cars more than electronics(?)....i like riding bikes so he would be fixing my bikes...hehe..)...not to say we would always agree, but in tose cases we could drink some beer at the local bar...of wait...he rarely drinks...oh well...in that case we would eat...

and i think our "thread-spiking" is actually making them more interesting!
<wink>

[PCBONEZ]

Re: All replacement capacitors in iMac G5 PSU burned up

hehe...you're aware of one thing: this instantenous charging of caps happens only ONCE: when you turn it on.
so what you're suggesting is that one voltage increase (from 0 to 5 or 12v..and via short pulses, ergo not dramatically fast anyway) will cause such thick wire to overheat?

Not exactly.
I'm suggesting high start currents over a few starts burned something else and then something else being burned caused charred toroid.

~OH NO~
This thought didn't occur till just now but 'something else' may not be on this PCB at all but on PCB[s?] with the loads.

i think pcbonez and i would probably be pretty damn good friends if we were near each other(i think he loves cars more than electronics(?)....i like riding bikes so he would be fixing my bikes...hehe..)...not to say we would always agree, but in tose cases we could drink some beer at the local bar...of wait...he rarely drinks...oh well...in that case we would eat...

and i think our "thread-spiking" is actually making them more interesting!
<wink>

>> good friends <<
Maybe.
Sometimes I think so and sometimes I'd like an ICBM to fall on your head.

>> loves cars more than electronics <<
Electronics was maintenance part my job in the power plants.
Operations part was operating the plant.
Cars was my hobby/side job/home based business.
I started transition from cars to PC's for 'side jobs' in the mid 90's.
Sold all my 'car tools' off 9 years ago when I retired from service and don't do cars any more except light stuff or for my own cars.
Until I meet wife#2 I was doing 'rent a tech' mostly for big projects that require a security clearance which meant I was changing cities every 3 to 6 months. Moving all the time does not make for a good situation to be building cars so now it's all PC stuff in between contracts.
I only take contract jobs occasionally now.

Yes, I rarely drink any more and when I do it's never beer.
Mixed drinks when out to a nice restaurant is about it.
.

Toasty,
Could please you talk to those towers and get them to work together to slap that pesky mosquito?

.

[Toasty]

Re: All replacement capacitors in iMac G5 PSU burned up

Toasty, Could please you talk to those towers and get them to work together to slap that pesky mosquito?

I've been working on it, but it's very elusive...

Somebody on the 107th floor managed to snap a photo of it the other day. We haven't been able to identify the species yet. Can anyone help?

Here's the photo:

[Toasty]

Re: All replacement capacitors in iMac G5 PSU burned up

~OH NO~
This thought didn't occur till just now but 'something else' may not be on this PCB at all but on PCB[s?] with the loads.

That's where I was going when I asked nwd to test it "no load" and then "loaded" with the mobo connected. Since it only got hot when it was loaded, I was more convinced there where other problems on the mobo.

But, since the choke in the Pi filter isn't cooked also, I changed my mind, as stated "with reservations".

I need to correct the location of that choke. I was flipping back and forth trying to follow the circuit path, and got my right confused with my other right.

...as the choke located below and to the right of the burnt coil in the second photo in Post # 12 of this thread, is in series to one of the feed points at the lower right...

The correct locations are directly below and at the lower left

Toast

[i4004]

Re: All replacement capacitors in iMac G5 PSU burned up

>Somebody on the 107th floor managed to snap a photo of it the other day. We haven't been able to identify the species yet. Can anyone help?

pretty good!


>Not exactly.
I'm suggesting high start currents over a few starts burned something else and then something else being burned caused charred toroid.

not many critical components there i would say. you have rect. diodes and a filter, essentially...
feedback loop probably acts with a bit of a delay anyway, so....and the viltage never really goes too high(if it would, caps would scream..)

>Sometimes I think so and sometimes I'd like an ICBM to fall on your head.

i never really think hurting you would fix anything..just too stubborn...
<wink>

>But, since the choke in the Pi filter isn't cooked also, I changed my mind, as stated "with reservations".

you're aware that every output has it's own choke, right?
ie that if one rail(say 12v) is problematic(on the mobo) that doesn't mean others are too, ie if one coil burns it doesn't mean all should burn...hehe..

STILL, overload by mobo still doesn't impress me as a way toproduce heat on the coil: traces on the pcb are waaay thinner and they were not getting hot, so it's not like we have current capable of heating this up so much...
but we may have the frequency and/or noise to do it...

let's see what happens when he changes it....
will we place any bets on the outcome?
hehe...

[PCBONEZ]

Re: All replacement capacitors in iMac G5 PSU burned up

O.I.C.

According to i4004 there is no way for that coil to burn.

Yet it's burned.