Martin circuit - blown capacitor?

[cgeiser]

Hi, I'm new to CPF - this is my first post. I've been working on a dynamo powered LED light and have a bulging capacitor that looks like its about to blow. Does anyone have any advice on what might be causing this? Here's the details on the build:
XP-E triple (Q5 bin?)
Martin circuit 2
4700uF 16V capacitor

What I'm thinking is that maybe as people at work (maiden voyage was a commute to work) may have spun the wheel, but not fast enough to meet the Vf, all the power was just building up in the cap?

Any suggestions to get this running before I move on to Circuit 8,9 or 10?

Thanks in advance,
Che

[MikeAusC]

When electrolytic capacitors let go they do so with dangerous explosive force. You could easily lose an eye.

Replace the capacitor and put a 15 volt Zener Diode across it.

Capacitors will only blow from reverse voltage or overvoltage.

[Steve K]

Replace the capacitor and put a 15 volt Zener Diode across it.

perhaps a bidirectional zener, so that you don't effectively short the cap for half of the AC cycle.

or two 15v zeners wired in series, anode to anode, or cathode to cathode.

I haven't used Martin's circuits, but this does seem to indicate that the cap you used was not rated for the job. Might have been a problem with the voltage rating, or it could have been an issue with the current ripple rating. Some caps aren't designed to handle a lot of current. If you go to the manufacturer's web site or sales literature, you should be able to look at their various models of electrolytic caps and see how they differ in intended application. For instance, Kemet has a variety of resources to use when searching for the right capacitor for your application:
http://www.kemet.com/kemet/web/homep...weben/products

good luck,
Steve K.

[cgeiser]

Thanks for the tips. I'm curious, why the Zener diode(s)? I thought the bridge rectifier would not only transform the AC to DC but also ensure the voltage direction is correct. The circuit is as described by many others with the bridge rectifier output feeding the cap in parallel with the LED.

I looked back at previous posts and this simple ckt was often described with only one LED. Do I maybe need a higher voltage cap since I have a higher Vf with the 3 series LED's?

Thanks again,
Che

[MikeAusC]

You are right. After the rectifier there can be no negative voltages.

The maximum voltage across three LEDs will be 11 volts

[Steve K]

okay... I just looked up which circuit this was and what the cap is doing. It's just a smoothing cap, so it only sees the voltage across the LEDs. It should only see the 11 or so volts across the LEDs, so there's no obvious issue with the cap's voltage rating. The rating for ripple current could still be an issue.

The cap itself could be a problem if it's not a good quality cap purchased from a reputable source. Even large electronics manufacturers have had problems with electrolytic caps failing early, and I think (??) it's been a problem with cheap caps, or possibly counterfeit capacitors.

Any chance that there's an intermittent connection at the LEDs?? If there was an open circuit, you could see much higher voltages across the cap.

Steve K.

[cgeiser]

I believe the ripple current may be my error. The ripple current of the cap used must be higher than the max current that will ever flow through it, correct?

Thanks again, I've no real electronics experience so this is all learning for me and I REALLY appreciate all the great help available on this forum!

[cgeiser]

Hold that thought. It should not be a ripple current issue as the ripple current rating for the cap I used was 1480 mArms. This was a 16V 4700uF cap, Digikey part number 565-1299-ND.
I don't think the LED's had an intermittent connection though there was a nick in one of the wires which could have caused a short to the light body (3/4" cu pipe light body). Could that short have caused the problem?

[Steve K]

I don't think the ripple current could be more than 0.5A, so the cap would appear to be adequate for the task. The datasheet for the cap is available through a link on the Digi-key page:
http://www.chemi-con.co.jp/e/catalog...g-e-110701.pdf

There are some guidelines for high reliability that recommend using an electrolytic cap rated for twice the voltage of the application, but even at the rated voltage, the cap should have lasted 1000 hours.

I don't suppose you ever had the bridge rectifier and cap connected to the dynamo without the LEDs connected? That's about the only way that I can imagine the voltage rating would have been exceeded.

Personally, I haven't used smoothing caps in this application, so I can't make any recommendations based on direct experience.

regards,
Steve K.

[ray6520659]

Any chance there is a bad electrical connection between the smoothing capacitor and the LED? Any plug in connectors or similar between the smoothing cap and the LED? The only thing that regulates the voltage output of the dyno in this circuit is the load of the LED. If there is a bad connection to the LED, not only is the power to the LED limited but the hub is not loaded properly and can produce higher voltage than the 16v smoothing cap is rated for.

I learned this the hard way. I was using the same circuit with a single Cree XPG. I had some cheap Radio Shack plug in connectors between the smoothing cap and the headlight. My smoothing cap was rated for 50 volts, all that was available at my local Radio Shack. I actually destroyed 3 LEDs in three rides. Led would work for a while, eventually would burn it out. The connectors weren't making a good connection which reduced the electrical load on the hub. This let the hub charge the smoothing cap to a higher voltage. When cap got enough voltage to overcome the resistance in the bad connection, hit my LED with too much voltage and cooked it. I got rid of the connectors and soldered the wires, haven't had a problem since.

Hi, I'm new to CPF - this is my first post. I've been working on a dynamo powered LED light and have a bulging capacitor that looks like its about to blow. Does anyone have any advice on what might be causing this? Here's the details on the build:
XP-E triple (Q5 bin?)
Martin circuit 2
4700uF 16V capacitor

What I'm thinking is that maybe as people at work (maiden voyage was a commute to work) may have spun the wheel, but not fast enough to meet the Vf, all the power was just building up in the cap?

Any suggestions to get this running before I move on to Circuit 8,9 or 10?

Thanks in advance,
Che

[cgeiser]

Thanks again everyone for the advice. The connections were all soldered and intact, however there was a cut through the insulation on the wire coming from the dynamo to the bridge rectifier - I don't think that could have caused the cap issue. I've replaced the cap with a different brand but similar specs and it seems fine after a 34 mile commute (the first was selling after 16 miles).

So it lolls like it may have just been a bad cap.
Thanks again,
Che