Was just testing my old LED par cans tonight before someone picks them up tomorrow (in theory!)...

Had been running a pair for good half hour or so, they got quite warm..... anyway... i just heard a loud *bang* and a puff of pewtrid smoke billowed out of the vents of one of them... I opened it up to be greeted with this:-





Bloody cheap capacitors. That shouldn't have exploded like that :/ and I think its only a 5v line too *sigh*

Hope it hasnt knocked anything else out now....



Cheers,

David

Had a big one in an amp do that a couple of weeks ago

Looking at the brand "JAKEC" - I couldnt find a "bad" mention of them on badcaps, in fact i'm sure its just some cheap chinese brand..

Interesting there are 4 of these 3300uF 16v caps in series dotted about the PCB.. the output on the transformer is 11v AC.. so once thats been through a rectifier and then smoothed out I'm sure It'll read over 16v on the multimeter...

Another one on the board is also bulging.. so will replace that - and *hope* that the rest of the unit is ok... but will obviously check the diodes, etc first! :) (Might work just as well without the capacitor however)...

Cheers,

David *bang*

Might be worth uprating the voltage just in case.

11V AC on load will give a DC voltage of around 14V, probably a little more if the transformer is unloaded. That's a bit close for me!

The ripple frequency might be a bit high as well so a low ESR cap won't hurt.

Interesting you say this, because at the time all 306 LEDs were off -> more likely to be a higher chance of it going above 16v.

Would 4 caps in series lower the ripple frequency a bit more do you think? I've since forgotten my capacitor theory from electronics at uni.. (probably one too many of these since: :)

I've got my faulty DEQ2496 open at the minute, which doesnt start instantly and crashes more times than I can think of, but I have a feeling that its more to do with dodgy caps in it so going to replace the PSU ones to see if it cures it! Nothing ventured, nothign gained and all that (and low ESR ones I think here too).

Cheers,

David

If the 4 caps are in series that's a bit weird, the chinese are all for saving pennies and I'd have thought it would be an expensive way to get the values they need! Obviously the ESR is quadrupled, voltage rating quadrupled and capacitance quartered in that case! Are you sure they arent in parallel, that would make more sense!? Or maybe there are sneaky diodes in between the capacitors acting as a voltage doublers.

Series smoothing caps need equalising resistors across them to make sure the voltage is divided equally between them so they volts-share properly.

Ahh dammit I meant they're in parallel, apologies, having a thicko moment there!

So that'll be 3300uF then...

Then this is to get a good low-ESR value, some amp manufacturers eg C-Audio do (did) this for a similar reason.

YOu need to change all of them for the same type in this case but I guess you know this already!

Yeah will order the "min" quantity of of 5 - but ordering 25v versions (cos they're cheaper than the 16v versions! Supply/Demand maybe? )

Also, not so sure of the low ESR - is it needed *that* much in an LED par can, unlike, say a power amplifier?

Cheers for the info!

David

Im not sure, LED par cans drive the LEDs with a high frequncy pulse rather than a constant current, ie PWM to get a dimming effect, so the current out of the caps may be drawn in short sharp bursts hence the need for low esr. And a capacitor that has to provide short sharp bursts must also be charged up just as quickly by the power source, low ESR means lower internal heating so the capacitor doesnt vent hydrogen and explode...oops too late!!

I must admit though that the temperature that the LED "array" got to when all LEDs were on was very warm indeed..

Cheers,

David