PDA

View Full Version : Another electronics question



gregl
05-26-2010, 12:43 PM
I posted a few weeks ago a question about LEDs, and I thank you again for the great answers. But now I have another question. I bought a bag of blue LEDs from a Chinese Ebay listing. (Mistake #1. But they were only 10 cents each compated to a couple of bucks each from Radio Jerk....)

I find that the LEDs are all over the place on brightness/voltage. For example I have one that is nice and bright at 2.8 volts while drawing just over 20 mA but another requires 5 volts to get to the 20 mA draw and an acceptable brightness.

So the question is: is this variation normal for LEDs or do I have a bad/cheap batch?

MotorradMike
05-26-2010, 12:46 PM
You don't really have a batch. If you want them matched in any way you can buy them like that but as you said, they cost more.

gregl
05-26-2010, 12:49 PM
Thanks, Mike. I never seem to learn the old lesson that you get what you pay for.

Evan
05-26-2010, 02:40 PM
If they are that far apart on voltage and current then what you have is floor sweepings, known as fallouts. Blue LEDs have much more variability than lower frequency types and they are static sensitive as well. Stray charges that you might be carrying can fully or partially destroy the junction resulting in greater variation. You should take normal anti static precautions.

gregl
05-26-2010, 03:10 PM
Thanks, Evan. As I said, these were way cheaper than any others so that makes perfect sense. I should have just gone to the local electronics supplier and paid the price for good ones.

soap
05-27-2010, 12:41 PM
I didn't read your other thread so forgive me if this is old info. :-) An LED is a current-driven device. Are you using some type of constant-current supply, or a constant-voltage supply with a current limiter? (Such as a resistor.)

If you're using a constant voltage with no current limiter, a tiny variation in voltage can be a huge swing in current through the LED, which will make them perform inconsistently. The small leaded LEDs are very sensitive to their drive current.

dr pepper
05-27-2010, 12:55 PM
I agree with the above, you have the ones that lit up on the test bench but didnt pass.
If you want them to be a similar brightness you can either drive them off a simple transistor constant current source, or even simpler if the supply voltage is high enough you can connect them all in series with a current limit resistor, then the current through them all will be the same, but the brightness still might vary a little if they are seconds.

You can buy constant current leds, which usually are matched and will remain a constant brightness over a wide supply voltage, and usually do not require a resistor in series.

macona
05-27-2010, 01:58 PM
Some leds are intended for 5v operation without an external resistor.

Also die composition will effect the voltage the LEDs. Most high brightness Blue, green, and white leds run at a much higher voltage than the old red, yellow, and green leds of the past.

Check out http://ledmuseum.candlepower.us/

Bad link, heres the good one:
http://ledmuseum.candlepower.us/led/

gregl
05-28-2010, 12:17 AM
I didn't read your other thread so forgive me if this is old info. :-) An LED is a current-driven device. Are you using some type of constant-current supply, or a constant-voltage supply with a current limiter? (Such as a resistor.)

If you're using a constant voltage with no current limiter, a tiny variation in voltage can be a huge swing in current through the LED, which will make them perform inconsistently. The small leaded LEDs are very sensitive to their drive current.


Soap: I'm testing them with a 3 volt CR2032 button cell. A few work fine, most do not. I also have a home-made model RR power supply that has transistors in it (as I said in my previous post, my knowledge of electronics is very limited) so I don't know what it's doing except that if I put a meter on the output, the LEDs that don't light up on the 3v. cell will light up at 5v. and draw 20 mA. The ones that do light up on the battery light up at 3v. and draw 20 mA on the power supply.

For my application, I want to run them from a pair of AA or AAA batteries. Several folks in my earlier post suggested ways to do this, but I figured I could just test them with the button cell. At least that's what some of the vendors do in their eBay photos.

Thanks for your help!

RKW
05-28-2010, 12:30 AM
Rule #1 ... Never buy anything from the Crap Shack!


But they were only 10 cents each compated to a couple of bucks each from Radio Jerk....)

Evan
05-28-2010, 12:33 AM
You can get away with using a 3v button cell because it can't supply enough current to fry the LED. Hook up 3volts worth of AA or AAA and it will be very bright for a short time and then turns into a DED (Dark Emitting Device).

darryl
05-28-2010, 03:50 AM
Perhaps you might test each one with a 12v power supply and a resistor in series, such as a 1000 ohm or so, wattage not important. Look at the relative brightness, and the voltage across each led. Ones that are a fairly close match can be put together in series, ones that light at the lowest voltage might be best for battery operated situations, and ones that take a lot higher voltage and don't shine as bright can be used in non-critical applications singly. Those are the ones that probably will fail early, unless they are a different led that are self-regulating and designed for direct operation from 12volts. One other factor to be aware of is the light dispersion pattern. Some cheap leds don't align the beam very well, and the light might actually come out the side of the lens, for instance. I've found a lot of Christmas tree leds are bad for this.

If it were me, I'd say the 'good' ones are the ones that match most closely in voltage drop and brightness, and have a consistent beam pattern. The others are most likely to be too far out of spec to be good parts. I haven't been aware of the voltage sensitivity problem that Evan mentioned, but I'm not doubting that. In general, you can't go wrong by observing some methods of static control when playing with electronic parts. Sliding around in a plastic chair and handling electronic parts is the kind of thing to be avoided.

MotorradMike
05-28-2010, 07:36 AM
You can get away with using a 3v button cell because it can't supply enough current to fry the LED. Hook up 3volts worth of AA or AAA and it will be very bright for a short time and then turns into a DED (Dark Emitting Device).

:D
Send those ones to me.
I've always wanted to build a flashdark that projects a cone of darkness!

gregl
05-28-2010, 10:38 AM
Ah MotorradMike, would that be to keep in the glove box of a British sports car?

Darryl: Thanks for the suggestion. I'll try that.

But guys, I'm confused about this current thing. If I apply the formula I find on the net (http://www.kpsec.freeuk.com/components/led.htm, scroll down for the formula) for determining the resistor needed, I take the supply voltage and subtract the LED input voltage and divide that by the LED current. But if you have 3v. for the LED and a 3v. input, the formula gives me a value of zero for the resistor regardless of whether it's a button cell or two AA batteries. I'm still trying to understand this. THANKS again for all your patience.

darryl
05-28-2010, 09:08 PM
Greg, that's essentially correct. When the battery voltage is only just enough to overcome the led's 'forward voltage', it doesn't have the means to jam a lot of current through the led. No resistor would be required.

If we talk in ideal terms, if the load (the led in this case) has a forward voltage spec (which in this case is the minimum voltage that needs to be applied before any current will flow) then any battery or power supply voltage fed to that load would not push any current unless the supply voltage was higher than that spec. The moment it becomes higher, a current can flow. This current would be infinite in the ideal sense at this point, except that in the real world the power supply couldn't output that much current, nor would the load suck that much because the forward voltage is not strictly a go/no go situation, but a variable depending on the current that flows. The more current is forced through an led, the higher the voltage across it becomes- to a point, and that point borders on burnout.

So if your battery supplies 3v for instance, and the led begins to glow at 2.7 volts, and draws rated current at 3.3 volts, you can see that feeding it 3v directly is going to result is light output, but less than rated current flowing. That would be a safe situation for the led, even though there's no resistor in the circuit.

But now lets say that the led draws rated current at 2.8v for instance. Feeding it 3v from the battery is going to mean it's drawing more than rated current, and a resistor would then be needed to bring that current down to a safe level.

The value for the resistor is complicated by the fact that the battery voltage is going to drop as it discharges. Say it drops to 2.5 volts- that could either mean that the led doesn't light up at all, or that it's dim. Removing the resistor would brighten it up a little if it's dim, but it won't help it to light up if it isn't already lit at that lower voltage. Keeping a fairly constant brightness on the led isn't easy when the supply voltage is so close to the forward voltage drop of the led.

What gives the light output is the current flowing, and what determines the brightness is the amount of current flowing. For a consistent brightness you'd want to be able to control the current and not worry about the voltage. The only problem here is that controlling the current to a fixed value can only be done electronically, and any such circuit is going to need some minimum voltage to operate with. The simplest circuit is a series regulator, so you would need about 6v minimum from the power supply. Because the led current is so low, relatively, you can still get very good life out of batteries, so wearing out four cells instead of two is not big deal.

Another electronic method of supplying a consistent current to an led is some kind of switching power supply circuit. This is more than a person at home would want to spend time cobbling up, but this kind of thing is available in some flashlights. Some use just a single cell, usually AA or even AAA, and have to 'upvert' the voltage and respond to the led current to control it. Even for me, an electronics guy, this is too much to fab up just to have a simple led light, so I'd go with the easy two part circuit and minimum of four cells. By the way, this circuit would at most require an area of about one centimeter square and about a half centimeter in height. One three lead part, and one two lead part.

Lew Hartswick
05-28-2010, 10:52 PM
When the battery voltage is only just enough to overcome the led's 'forward voltage', it doesn't have the means to jam a lot of current through the led. .
Now there is a REAL electronics technical term if I ever heard one. :-)
...lew...

gregl
05-30-2010, 06:05 PM
I've been out of town for a couple of days and away from the net, hence the delay.

THANK you Darryl for your complete and patient answer to my question. So it sounds as though the power supply for an LED should start out with a voltage higher than the rated LED voltage and then be brought back down in order to provide a constant, or relatively constant, voltage to the LED. So the notion of running a 3v. LED off a 3v. battery is OK in theory, in practice it's not the best way to go. That explains why the little keychain lights I bought use four button cells in series.

I think I've got it now. I'll get some resistors on Tuesday and check these LEDs on a 12v. source as you suggested earlier. In the end, I'll have to change the design for my gizmo since I already bought 2-cell AA battery boxes. Perhaps I'll go with a 9v. battery, since four AAs is starting to get large for my project. I want something that will run for at least 60 to 100 hours before needing a battery swap.

Thanks again, everyone.