Power Supply Capacitor Question

[ronkar]

For the EE's out there...

I have a hefty old 12 volt power supply which no longer puts out clean 12 volt DC. My presumption is that one or more filter capacitors no longer function. No scope here, but my DMM shows a significant AC component on output. The unit looks as if it was well built, so I don't want to just chuck it. Not sure if I want to bother repairing it right now, but I do have questions.

The main capacitors are rated at 33 volts, 10,000 microfarads (Sprague 208238). These are huge units, about 4 1/2 inches long, and 2 inches in diameter with screw terminals. I see on the internet that modern capacitors with a similar voltage and capacitance rating are available relatively cheap, but they are far smaller, and they have only solder leads. The new ones are so much smaller that something does not seem right...unless modern materials and construction have rendered the old dinosaurs obsolete.

My question is, can a new, small electrolytic capacitor be substituted for the original physically large units which are wired to the bridge rectifier? Or am I looking at an instant Also, the old caps are marked with a red spot on one screw connector, which I can only guess is the + terminal. Is this correct?

If an extremely inexpensive replacement isn't possible, I will put this thing back in the basement. The case, transformer, etc. could be used for a different project on another day when finances are a bit more conducive to experimenting.

 

[Mr Happy]

How old is old? I remember back in the 70's I had some similar capacitors that were rescued from computer power supplies (the big kind of 1970's computer in room-sized cabinets). It would not surprise me at all if technology has shrunk them in the intervening period. I think modern equivalents would be fine...except why did they put the screw terminals on them in the first place? I'm guessing they must have been designed for large ripple currents.

I know that when I used to short out those old capacitors by putting a screwdriver across the terminals it welded the screwdriver to the screws. It was a good way to destroy a screwdriver (and probably didn't do the capacitors much good either).

In the good old days they didn't have switching power supplies. For big power supplies they got a humongous transformer, rectified the output, and put huge smoothing capacitors on it. It's sad those days are gone because those transformers were rather run for melting iron nails and the like.

 

[greg_in_canada]

Capacitors have gotten better. But you will want to make sure that the new caps have a sufficient ripple current rating. I don't recall exactly how to calculate it but Google found this http://sci.tech-archive.net/Archive/sci.electronics.design/2005-10/msg00144.html

If you don't they will heat up and fail prematurely (maybe even with smoke).

Greg

 

[ronkar]

Ripple current, the missing parameter. Overheats the capacitor if excessive. Larger caps dissipate more heat due to larger surface area, hence the reason for the physically large caps in the original design. Don't ask me to do the math. I was always better at putting one word in front of the other, instead of one number in front of the other. Too many equations, and my brain freezes up.

Thanks guys, for pointing me in the right direction. I think this project will be on hold for a while. I see on the internet that a new 10000 microfarad 50V electrolytic of approximately the same dimensions as the existing filter caps can be had for around $14 USD each. No telling what shipping or minimums the supplier would require, however. Not a bad price, but for now I have an 11.5V switching supply (from an old computer) that is powering my hobby charger well enough. I think I'll save the bucks for something brighter.

The physically small caps I first mentioned might not fare well in the sealed metal box. Maybe with active cooling, but not in dead air. Not worth taking the chance on.

 

[SemiMan]

What is the current rating of your power supply?

Capacitors have gotten much much better and depending on the current, newer small ones will work better. They have lower equivalent series resistance which keeps the internal heating lower and also leads to higher capabilities for ripple current in general. Worst case, you can use more of them... which spreads out the ripple current assuming you wire them such that the resistance in wiring per capacitor is approximately equal.

Semiman

 

[zipplet]

Also if you can try to get 105C temperature rated capacitors for longer life, and low ESR types (equivalent series resistance). I like SemiMan's idea of using multiple capacitors and it's an idea I have also used myself

 

[poguy]

How much ripple is there? Could one diode on a full wave bridge rectifier have gone bad?

Cheap source of capacitors are those 1 farad "stiffening" caps used by loud car audio enthusiasts. Don't know ripple current rating. I got a couple used ones on craigslist.

Also try hamfests. A lot of those old "computer" supplies go cheap. Electrolytics are commonly found there, too.

 

[London Lad]

I would second that. Check that one diode in the bridge hasn't failed. Use your ohm meter, post here if you need help.

 

[J_C]

The size issue is only one of mechanically mounting them where they'll be durably restrained from movement, and electrically isolated of course.

With any decent quality capacitor you won't have a ripple current problem at 120Hz off a rectifier. That is a very undemanding use of modern capacitors, the high performance ones found in switching circuits are total overkill.

The amount of capacitance you need depends on what you'll use it for. It could be entirely legitimate to replace these two with a single 1,000uF capacitor for example if your load isn't bothered by a little ripple and is at 1A or below, but I agree with prior comments that you should check the diodes for possible failure and it does make sense to equip with enough capacitance to handle the full load the transformer is capable of in case the use/reuse changes in the future.

One of the factors with different sized caps or multiples is how you plan on mounting them, since it seems you may have clamps and ring terminals instead of a PCB to solder them to? If you have a nice long piece of copper clad it become easier, etch away the end so you can affix it to the chassis with L-brackets, etch or dremel a break down the middle to end up with two huge +/- solder pads though it would be easier to solder if you had proper thermal reliefs instead.

Something like these would be a pretty low cost swap,
http://www.goldmine-elec-products.com/prodinfo.asp?number=G8716
http://www.bgmicro.com/index.asp?PageAction=VIEWPROD&ProdID=9958

Adding multiple in parallel to reach a target of about 1,000uF /amp you need. Actually with an oversized transformer for the load the capacitance goes down a bit as transformer regulates better but it's of little use to think about when caps can be had so cheaply, the time to devise an alternate mounting is the larger bother though I suppose you could always bundle a bunch together, put some caulking between them to make a big log and strap that down, upside down.