Hot caps

[Rainbow]

Re: Hot caps

It seems to work fine even when it's really close. The only problem is that you need to ignore the laser beam. There's even backlight (nice white) for the LCD that can be turned on using a button.

[davmax]

Re: Hot caps

I agree IR thermometer is the best Mbd component temperature measurement method. The heat generated in the Mbd due FET's etc can be measured and compared with the electro can temperatures. Yes heat can flow from the electro to Mbd also heat flows from Mbd to capacitor. The can temp is typically great than the Mbd copper layer. As far as the electro is concerned the ambient temperature is the motherboard copper layer temperature in that it is considerably more than the air ambient temperature.

It is always good to note the operating temperature rise of freshly installed electro caps.

[lucius_mckaine]

Re: Hot caps

guys! just use a mercury thermometer after letting the caps running for 5 mins or more

hottest caps the the duds.

[Elitist]

Re: Hot caps

Goodness, gracious me! Full circle. This is what I wrote on p.1 :
"....the ones that are warm are the duds! "

[999999999]

Re: Hot caps

OK! I(squaredR) in free air. But as anyone who has done recap-ing on a motherboard knows, the can is connected to the ground plane of the interlayer which behaves as a massive heatsink, so ripple, etc. dissipation doesn't tend to raise the can temp. by much, normally. Unless there's something wrong. If you've got G-Luxon or OST written on the sleeve, they might just as well have printed QED on it!
The tongue-test has served me well over the years in all sorts of situations...

I have done plenty of recappings and can tell you the ground plane is seldom if ever as hot as a poor cap can get, and replacing that higher ESR cap with a lower ESR cap will result in the replacement running cooler, because the cap itself generates less heat. The actives are typically 'sunk to the power plane anyway, not the ground plane. That doesn't mean this heatsinking has no effect, certainly it does make it worse for a nearby capacitor if that cap were running a little warmer than ideal already.

Normally what happens is the subcircuit's design puts too much ripple current on the capacitors, and there is either insufficient numbers (for an additive ESR reduction) or sizes of them and cap family ESR is too high, for long life as used in the design.

So in a way you are right when you wrote "unless there's something wrong", because there is something wrong in that a design was not using the right parts (or numbers of (caps)) instead of cheaper parts, OR if a power plane or ground is that close, it shouldn't be, or a (motherboard) design was reused for a newer higher current CPU (like a Tualatin) but kept the slower two stage regulation that subjected the caps to that much more ripple current.

So it is possible some caps fail from local heat, but caps do most definitely heat themselves up from ripple current, when brand new and years before they fail. Too often we try to blame only one variable though, while it's often several interacting to cause a high temp (if not a defective cap outside of these other issues).

They often fail because they ran hot, not running hot because they failed (in most cases, though it is certainly not impossible for them to get even hotter suddenly before a violent death). This is easy enough to see if you overclock CPUs, you can feel some caps getting really hot, and know you have one of four choices- ease off the speed and/or voltage some, get more cooling on the caps, replace the caps, or accept it'll have a shorter life. Replace the caps and do nothing else and if you chose better grade of caps (like MBZ to replace a G-Luxon) you can feel they don't get as hot. This is with a new board, long before the caps are worn out.

[Big Pope]

Re: Hot caps

IR thermometer is the way to go. There are some cheap ones - I have this one:
http://hh17.vip.sina.com/product/1322.htm
(about 35 euro here).

Great, i can buy this one with cheap price in my side.

I will buy one soon.

[yanz]

Re: Hot caps

So it is possible some caps fail from local heat, but caps do most definitely heat themselves up from ripple current, when brand new and years before they fail. Too often we try to blame only one variable though, while it's often several interacting to cause a high temp (if not a defective cap outside of these other issues).

very true. in the end, finding what's wrong is an art. that's why i like it.

So, leave the board 'running' for five minutes and stick tip of tongue onto top of biggest/suspect caps - the ones that are warm are the duds! Caps should not be much above ambient - they are energy stores, not energy consumers.

gee, using your tongue to sense the temperature, are you a reptile?

i never try nor have a bravery to try. i understand though that tongue is much more sensitive than a finger. but my fingers are super sensitive anyway