Cap go BOOM!

[primalphunk]

I built an ultra clean power supply from GGG and since it was my first time to mess around with a build using ac power I refused to plug it in until I could get a friend with much more knowledge than I to help me out.  Just trying to be as respectful as possible to all those highly motivated little greenies flying around in all those wires.

We took the lid off of the enclosure and plugged the sucker in.  At first everything looked to be groovy. So my friend, let's call him the Professor, held my enclosure up a little closer to his ear so he could hear the transformer humming along...then pop goes the weasel and it starts smoking... :shock:

Now the Professor is very familiar with this stuff, not quite RG but very capable nonetheless.  Even he was a bit startled by this turn of events at first as evidenced by his eyes getting really wide for a sec.  Then he realized,  that it was just the big 1000uF cap blowing it's lid.  After seeing the smoke we quickly unplugged it.  If I had tried to plug this thing in and then troubleshoot it by myself the little thing with the cap blowing probably would have made me piss myself. :oops:

Turns out I should have used at least a 35volt 1000uF cap rather than the rather puny 16volt one I originally used.  I found one I had laying around and exchanged it for the blown cap and everything seems to be working fine thus far.  Heheheh...definitely a learning experience.  

I'm gonna be testing it again soon but my advice for anyone else who builds this is to make sure that you use at least a 35volt rated cap for the 1000uF.

peace,
James

[ninoman123]

I had a Ruby breadboarded once and blew up a 10 v rated cap. I didnt see that it was rated for 10v. It left a brown spot on my breadboard. Anyways yes always check the voltage. Unless u wanna blow up a cap.

[Paul Perry (Frostwave)]

Yeah, I've done this... running a 16v AC supply thru a diode to a 25V electrolytic. What could go wrong? well, suppose you have a BIG 16v plugpack & there is bugger-all load, then it miight be more like 20v. And then the peak voltage is 1.4 times the 20, so you have 28V on the cap.
Oh dear :oops:
Sure, most times you would get away with it.. but, you can't really complain if you put 28 across something with 25 printed on it...

[cd]

I haven't blown a cap, but I have toasted resistors - ever see one of those catch fire on a breadboard?  Jeez!  I was breadboarding a little preamp and got the pinout of a transistor wrong - thought it was EBC but it was ECB.  Plugged it in and two resistors on the board started burning up.  Turned off the power and triple checked all connections, everything was fine so I replaced the resistors and tried again - poof!  It wasn't until I tore the whole thing down and built it up again, burned another set of resistors, then I realized duh! wrong pinout!!  

Cap-wise, I don't think I've ever used a close rated cap in a pedal effect, like a 10V cap on the 9V line.  25V or 100V rated pieces are so cheap and plentiful, and space is rarely an issue so I always go overrated.

[ninoman123]

Yah i burned a resistor once, it burned a little spot on my finger. I dont know how I did it....

[brett]

80W amp. + and - 40V rails and 4 x 4700uF 63WV caps.  There's 3 leads coming out of the transformer (+,- and gnd), so Murphy's Law told me that the chances of making wrong connections to the PS board were greater than the chances of correct ones.  I made them wrong and blew 2 x 4700uFR caps with a small(ish) explosion and hot, white stuff ejected (by steam??) all through the guts of the amp.

Unbelieveably, my parents allowed me to continue my electronics hobby (I was 15 or 16 at the time).

[primalphunk]

I still feel like I don't know enough about what I'm doing to trust myself to pull off a project like this all by my lonesome...Thank goodness for the Professor.  I think I can trust his superior knowledge and it's kinda cool to just be able to call him up and get his help. I checked out the bill of materials that I originally downloaded from GGG and it is a fairly old one(last revised in 2002) that only called for 16volt caps.  JD is pretty much on top of things so when I went back to his site today to see what the current bill of materials is I saw that he had updated it some time ago and it now calls for 35 volt caps.

The two other electrolytic caps on my build are still just 16 volt ones but I think they are okay because of where they are located.  The Professor thought they would probably be ok but does anyone know for certain if that is right?  I've also got a fuse on the primary side of the transformer and one before the dc jack daisy chain to be as safe as possible.

peace,
James

[Transmogrifox]

if they haven't blown yet, you're probably ok.

In my electronics labs it was a popular prank to plug electrolytic caps in backward so that they would blow and make everybody in the lab crap their pants. :twisted:

[ninoman123]

Transmogrifox - that sounds like fun.  I scared my gf doing that one time.  :twisted:

[Paul Perry (Frostwave)]

The two other electrolytic caps on my build are still just 16 volt ones but I think they are okay because of where they are located.  The Professor thought they would probably be ok but does anyone know for certain if that is right?  

If they are after a 15v reg, they will be OK, wherever they are.

[ESPguitar]

When we're at it:)

I have also buildt a power supply.. It's one tranny at 18Volt and one at 12 volt, both regulated, one to 18V and one to 9V..

I placed some heat sinks at the regulators and right beofre the regulator it is a 1000uF cap rated at 35 volt.. But the regulator get warm as hell, will that fry the cap??

Thanks,

RB

[brett]

But the regulator get warm as hell, will that fry the cap??

Heat in the regulator is due to it *reducing* the load.  When the regulator is working hard, it helps the cap, unless it blows.  

If the regulator is getting altogether too hot, choose a more rugged design (eg TO220 case instead of a TO92), or put a heatsink on it (there's lots in old computer power supplies), or feed it a few less volts so that it isn't cutting so much voltage.

cheers

[Paul Perry (Frostwave)]

I can't see any reason to ever use any regulator smaller than a TO220, because there isn't any noticeable price difference. Plus the TO220 at least gives the option of a heat sink (WARNING!! the metal tab is connected to one of the pins!!!! don't accidentally short it out!)
It is an unfortunate thing, that the main filter cap is sometimes put so close to the regulator, that heat from the regulator shortens its life.
Regulators are one thing that it may be worth buying 'name' brands like National and Motorola, not all regs have automatic overload shutoff, and some have poor heatsinking.

[ESPguitar]

But the regulator get warm as hell, will that fry the cap??

Heat in the regulator is due to it *reducing* the load.  When the regulator is working hard, it helps the cap, unless it blows.  

If the regulator is getting altogether too hot, choose a more rugged design (eg TO220 case instead of a TO92), or put a heatsink on it (there's lots in old computer power supplies), or feed it a few less volts so that it isn't cutting so much voltage.

cheers

The regulator is a TO220 and i also buyed a heat sink for the TO220 regulator..

Help the cap what?

And for Paul i have inserted these plastic things to put between the regulator and the heat sink..

So i won't toast the cap..?

RB

[ESPguitar]

I can't see any reason to ever use any regulator smaller than a TO220, because there isn't any noticeable price difference. Plus the TO220 at least gives the option of a heat sink (WARNING!! the metal tab is connected to one of the pins!!!! don't accidentally short it out!)
It is an unfortunate thing, that the main filter cap is sometimes put so close to the regulator, that heat from the regulator shortens its life.
Regulators are one thing that it may be worth buying 'name' brands like National and Motorola, not all regs have automatic overload shutoff, and some have poor heatsinking.

It's two 1000uF caps right before the reg.
It's national semi. i think..
And i also have inserted Raychems-Polyswitch ight after the trannys..

So i don't think i will have any problems :wink:

[ESPguitar]

Another question.. Why did R.G draw in three caps in this values 1000uf, 1000uf, and 10uf..

Isn't 1000uf to high to cut the "bass frequenses that comes with the power?

Shouldn't i change one of the 1000uf caps to lower value?

Anyone who can give me a formular so i can calculate how big (ca) the caps should be?

Thanks,

RB

[Paul Perry (Frostwave)]

Calculating the size of a filter cap isn't simple.
Things that affect it include the losses & impedance of the transformer, the regulator drop out voltage (that is, the amount of headroom the regulator needs, about 3v for cheap common ones) and peak current draw by the load. The problem is, the cap is only charging for a fraction of the AC cycle.
It's one of those cases where it's easiest to just see what works (or take the advice of RG!).
It doesn't hurt to read the data sheets for particular regulators, either. Sometimes you find somethng suprising...

[ESPguitar]

Calculating the size of a filter cap isn't simple.
Things that affect it include the losses & impedance of the transformer, the regulator drop out voltage (that is, the amount of headroom the regulator needs, about 3v for cheap common ones) and peak current draw by the load. The problem is, the cap is only charging for a fraction of the AC cycle.
It's one of those cases where it's easiest to just see what works (or take the advice of RG!).
It doesn't hurt to read the data sheets for particular regulators, either. Sometimes you find somethng suprising...

Thanks :wink:

RB

[birt]

my grandpa made that white stuff fly all around the kitchen by blowing up a really big cap can :lol:

i never blew up a cap, i only smoked a couple of old transformers

[brett]

Isn't 1000uf to high to cut the "bass frequenses that comes with the power?

In general, the bigger the better.  
You'll often want a small non-electrolytic cap (e.g. 0.1uF) in parallel, too (one before and one after the regulator doesn't hurt).  

As Mr Frostwave said, there are good technical reasons for these things, but because they work in practice, it's easier to just do it and save the brainpower for other things.

cheers

PS Cutoff freq (Hz) = 1/(2.pi.R.C), where R is the resistance in ohms and C is the capacitance in farads
For a simple, single-pole filter, the theoretical "roll-off" above or below the cutoff frequency is 6dB/octave.
e.g. A fuzzface has a 0.01uF output cap and a 500k pot. So the "cutoff frequency" (which is the -3dB point) = 1/(2 x pi x 500 000 x 0.000 000 01) = 1/0.0314 = 32 Hz.