how much harm is this capacitor causing my led?

[darron]

hi all. i have a question. i'll try to keep it simple...




simply, how much damage occurs to an led when it gets exposed to a 22uf cap charged to 9v?




why?:

i'm putting a 22uF cap over a square wave LFO signal which lights an LED/LDR to smooth out the clicking (not coming from the actual clock or bleeding in, just he sudden volume change).


NOW... i'm having this switchable, and with the arrangement i've made the cap consequently gets exposed to and quickly charges to 9v through a 2k5 resistor while disconnected. when switching back in it drops the charge over the led and makes a quick flash

it doesn't seem mega bright, warming, or damaging. but will it break down the junction over 500+ repeats for example? i've flicked the switch a heap of time to test.

i can avoid this whole problem by using a switch with an extra pole but i'd like to keep this if it's really not a problem.

am i probably completely safe, or is it a bit too sus' and bad practice?

would the cap make a difference? ie avoid low esp cap?


thanks for your thoughts. i made a simple question complicated.

[GibsonGM]

So, what you need to know, really, is:  can a charged cap source enough current to damage the LED, right?  Good question.

I would THINK that if there was going to be damage done, you'd have found it already.   Is there a series resistance associated with your LED?  If so, then you should be safe.  Or perhaps the ESR of the cap is enough (?).   A lot of "probably" involved here!   

You're in kind of unknown territory (to me at least)...I would add some series resistance to the LED if there is none, and that should do it.   Do you have a schematic of this, just for the heck of it? 

[R.G.]

Mike is right, perhaps especially about the amount of 'probably' involved.

LEDs are commonly run with high currents for short times in displays where they're multiplexed. However, they are specified for a maximum 'do not exceed' current in that kind of use. The common 20ma limit for LEDs in the T 1 3/4/5mm round package is based on thermal considerations - how much heat can get out. The junctions can actually take a lot more current, but you burn them up thermally if that's continuous. So the duty cycle on the LED has to be below 20ma time averaged.

In your case, there is no good way to specify what the current goes to. It's limited by the ESR, the wiring resistance, the lead and wiring resistance inside the LED, and the material and doping levels of the semiconductors in the LED. The best way to nail things down is, as Mike said, put an external series resistance on the LED. That eliminates some of the unknowns by making them irrelevant.

First principle of engineering: that which cannot be controlled must be made irrelevant.

[PRR]

LM3909 was a chip that flashed a 2V LED from a 1.5V dry-cell by pumping with a transistor, a 300uFd capacitor, and a 12 ohm resistor.

LM3909's product-life was even shorter than LM381; no sign on TI's website. Jameco has a mangled datasheet via FTP:

ftp://ftp.jameco.com/archive/Obsolete-TechDocuments/24192.PDF --500K PDF via FTP... slow!!!

I can testify that the "1.5V flasher", using old and surplus LEDs, flashed over 5 years 1/second, over 100 million blinks. I used it for "security alarms". It didn't actually do anything, but nobody messed with a blinking light and a "High Voltage!" sign.

The typical Peak Current in the suggested applicaton is just 45mA, which at low duty sure won't stress a 20mA LED. I wouldn't much worry about 100mA at 10% duty-cycle.

[darron]

Thanks for all of your inputs guys.

I think I'll be safe then. The good point that was raised is that LEDs do have peak currents as well which I was thinking, which I guess people consider when running them with AC. The duration is incredibly short and not repeating. It's interesting that the flash isn't as instantly fast as I would think.

I was looking for some feelings from the more experienced guys and I'm really glad I got them Thanks