Resistors vs. efficiency?

[milkyspit]

Does the use of a resistor in an unregulated LED flashlight such as the Streamlight 4AA LED reduce its power efficiency due to some of the energy being lost in the resistor as heat? I'm afraid I switched out of my Electrical Engineering major halfway through college, and now questions like this haunt me. /ubbthreads/images/graemlins/icon23.gif

 

[Doug Owen]

The short answer is 'of course, it uses energy (it gets warm and there's only one source of energy....), and therefore lowers efficiency'.

However, added resistance does provide current regulation of sorts (that is it changes the control of the current), allowing longer drive times at a given level (or close to it) as the battery ages.

If it wasn't an improvement in overall performance (as most customers would define it), Streamlight wouldn't do it. It seems the added cost for true current regulation is not worth the return, therefore a *compromise value* resistor seems to be the call.

Doug Owen

 

[Zeppert]

Doug Owen, I ordered the new L4 from surefire. From what I gather so far it is suppose to have a "digital regulator" where the it stays at a power so long, then goes to a lower power for addition time, and then yet another lower setting for more time still. This would have true current regulation right as oppose to compromise value. Correct?

 

[shiftd]

I would say that is true regulation, even though it is only half regulated. A fully regulated light will run on constant output over time and then suddenly dim out and die over a short time, leaving the user in the dark.

just mt .02

 

[Doug Owen]

I don't know the light, but from what you say, no.

I suspect it's the same sort of duty cycle control the EternaLights and my nifty new PT Eclipse use. That is the lower light level is done by rapidly switching it on and off. Try waving it around in a dark room (or swinging it in a circle on a piece of string), you should see a series of dots not a dimmer streak on the reduced light levels.

The digital bit is the on and off, ones and zeros.....

True regulators measure the current (or voltage) delivered and adjust accordingly. Think of it as a 'smart series resistor'.

Doug Owen

 

[Doug Owen]

[ QUOTE ]
shiftd said:
I would say that is true regulation, even though it is only half regulated. A fully regulated light will run on constant output over time and then suddenly dim out and die over a short time, leaving the user in the dark.


[/ QUOTE ]

I guess we disagree. To me regulation includes the concept of adjustment, not just simple change. I change the position on the gas pedal to keep a constant speed as the road conditons change, *regulating* my car's speed.

Duty cycle changes to change light levels are more like changing gears.

Doug Owen

 

[shiftd]

Well, as far as I know, there were 3 types of regulation: fully regulated, half/partially regulated, and "cheaply" regulated (aka, by using resistor). I have never consider duty cycle before as I don't really know what it was.
to be honest, I don't understand fully the statement above, so could you digress more about it?
about the car, changing the pedal is a simple change even though the process may be complicated. Like you said, you change the position; you did not adjust it.
I also never said that the regulation is just a simple change. I am only trying to divide kind of regulation into sub division that is easily understandable.

I may or may not answer your statement. If I was off the topic, please forgive me. I don't know what we are disagree upon.

 

[milkyspit]

[ QUOTE ]
Doug Owen said:
The short answer is 'of course, it uses energy (it gets warm and there's only one source of energy....), and therefore lowers efficiency'.

However, added resistance does provide current regulation of sorts (that is it changes the control of the current), allowing longer drive times at a given level (or close to it) as the battery ages.

If it wasn't an improvement in overall performance (as most customers would define it), Streamlight wouldn't do it. It seems the added cost for true current regulation is not worth the return, therefore a *compromise value* resistor seems to be the call.

Doug Owen

[/ QUOTE ]

Doug, I should mention that I wasn't in any way ragging on the Streamlight 4AA LED; in fact, I own one, and it's one of my favorites! /ubbthreads/images/graemlins/thumbsup.gif I just mentioned it because I knew it contained resistors, and I wanted to offer an example of something with a resistor in it. My question really is more in terms of understanding some of these design concepts. Someday soon I might try my hand at designing my own light, and I really want to understand the basics first. /ubbthreads/images/graemlins/confused.gif

 

[milkyspit]

One thing I'm thinking about the duty cycle concept is that an on/off cycle of sufficient speed, like in terms of thousands or even millions of times per second, would look indistinguishable from more classical methods of current regulation, in the same way that compact discs (with 44,000 or so samples per second) sound essentially identical to analog music. When you also factor in the discharge time of the phosphor in white LEDs, it probably makes things even more indistinguishable. I believe most of the current regulation chips presently available work in this way.

 

[Doug Owen]

[ QUOTE ]
shiftd said:
Well, as far as I know, there were 3 types of regulation: fully regulated, half/partially regulated, and "cheaply" regulated (aka, by using resistor). I have never consider duty cycle before as I don't really know what it was.
to be honest, I don't understand fully the statement above, so could you digress more about it?



[/ QUOTE ]

Love the 'cheaply regulated'!

I guess I missed. I'll try again.

My contention is regulation is the changing of something based on an output and how close the output is to the ideal. A closed loop system. A voltage regulator in your car looks at the output voltage and drives the alternator field depending on the output being above or below the target (remove the connection by which the output is measured and no control, no regulation).

The difference between 'stepping on the gas' to some fixed point and accelerating normally up to the proper speed and maintaining it.

Duty cycle is something like the name sounds. If you flash the light on and off, say on for a hundredth of a second, off a hundredth, then on again, the average is 50% and the rate high enough you don't see the flicker, it just looks 'half as bright'. If it's on one hundredth, off for three, before on again, it seems 'a quarter bright'.

Doug Owen

 

[Doug Owen]

[ QUOTE ]
milkyspit said:

Doug, I should mention that I wasn't in any way ragging on the Streamlight 4AA LED; in fact, I own one, and it's one of my favorites!


[/ QUOTE ]

Hey, some of my best flashlights are Streamlights, too.

Still not sure I'd want my daughter to marry one, however. If I had a daughter, that is....

Doug Owen.

 

[Doug Owen]

[ QUOTE ]
milkyspit said:
One thing I'm thinking about the duty cycle concept is that an on/off cycle of sufficient speed, like in terms of thousands or even millions of times per second, would look indistinguishable from more classical methods of current regulation, in the same way that compact discs (with 44,000 or so samples per second) sound essentially identical to analog music. When you also factor in the discharge time of the phosphor in white LEDs, it probably makes things even more indistinguishable. I believe most of the current regulation chips presently available work in this way.


[/ QUOTE ]

Human 'persistence of vision' is something like 1/30 of a second, this is why films (at 16 and 32 frames a second) and TV (at 30) are seen as continuous. Flashing faster than this doesn't look any smoother but does start to introduce other problems at very high rates, typically a few hundred thousand per second is tops.

No, I think most of them are probably linear. The switching ones (turn fully on and off rapidly rather than part way on continuously) need an external energy storage element (typically an inductor) to work.

Doug Owen

 

[php_44]

Just my 2 cents....

As far as efficiency - a resistor current limiter:

If you have a LED light and the battery voltage is very close to the voltage required by your LED(s), then a resistor for "regulation" (more accurate to call it current limiting) doesn't waste too much energy. For example a 3 cell (1.5v X 3 ) would be 4.5V total. A white LED might like 3.5V to run slightly overdriven. Say you had a 7 LED light - each LED at 40mA. The energy usage would be:

7x40mA = 280mA. At 3.5V this is 0.98 Watts.
Your 4.5V battery is delivering 280mA @4.5V = 1.26W
...So on fresh batteries this light is wasting ~1/4 W in the resistor or 23%
But after the batteries are half used up - the waste will fall. So also will your light's brightness.

However - if you have a 4 x 1.5V cell light - same #LEDs, the waste gets a lot worse. You'd be wasting almost 3/4W on fresh cells or 58% of your cells energy!!!!

A real regulator:

An inductor based switching regulator typically is 65-90% efficient (ie- they waste 10-35% of your battery's energy) These are typically most efficient with fresh cells and get less efficient as the cells wear out. The difference is that generally you get more consistent light output between fresh and nearly dead cells, and you can utilize nearly every penny's worth of energy from the cells. I prefer regulated lights as I don't use rechargables. I have one home brew three cell "resistored" light and the output drop as the batteries age annoys me. I tend to use my home brewed regulated lights more, and plan to rework my "resistored" light.

By the way - nearly all lights with switch mode regulators have a capacitor on the output of the circuit to filter the current to the LED and eliminate any flicker (though you'd never see it anyway) and minimize RFI. This is done mostly because the regulators prefer nearly DC feedback to regulate properly and operate in a stable fashion. Also, you're lless likely to exceed the peak LED current which an unfiltered switch mode supply might do.

 

[Jonathan]

Doug raises a _very_ good point in differentiating between duty cycle modulated dimmers and regulation.

A duty cycle modulated dimmer acts by using an electronically controlled switching element (often a MOSFET for battery powered lighting applications) to rapidly turn the power to the load on and off. By adjusting the percentage of time that the load is 'on', you change the power being delivered to the load.

This duty cycle modulation can be _very_ efficient. In theory, during the 'on' period there are no conduction losses, and during the 'off' period there is no current flow and again no loss, and the and the transitions are instant and loss-less. Real duty cycle modulators have real losses, but efficiencies >99% are easily obtained. However it is _not_ regulation.

You can also obtain dimming by using some sort of variable resistance. In fact, a MOSFET _can_ be used in this fashion by supplying it with a control signal that is appropriately chosen to be between 'on' and 'off'. When you use a variable resistance as a dimmer, you _always_ lose considerable energy in the resistance.

For _regulation_ you need to measure a controlled parameter, and use this measurement to adjust your output to maintain your desired target value.

A _regulator_ _might_ use PWM modulation in order to adjust the output, or in _might_ use linear (resistive) modulation, or it might use some other modulation technique entirely.

Furthermore, a regulator _might_ produce a flashing output, or a regulator might be combined with non-regulated control loops. A regulator might not even include internal storage elements.

For example, the 'Willie Hunt LVR-3' uses a PIC based PWM modulator, and looks quite a bit like an ordinary PWM dimmer. However it measures the input voltage and uses this to adjust the PWM period, maintaining a _constant_ RMS output voltage. There are no storage elements in the circuit, and the _output_ voltage is not sampled. However when used to supply an incandescent lamp, this circuit results in very stable light output over the life of the battery. I would call this a regulator.

For LED applications, a microcontroller based current regulator would be quite useful. This device would _measure_ the output current, and then adjust a PWM modulator in order to keep a capacitor charged and maintain this output current. This would be a regulator with a storage element (the capacitor). Finally, in order to _dim_ the LED, one might reasonably PWM modulate (at much lower frequency) the otherwise _regulated_ output, in order to cause the LED to flash rapidly. One would do this to maintain color temperature while dimming the LED. I would consider this a _regulator_ combined with a PWM _dimmer_; the regulator part using feedback to maintain the desired drive current, and the dimmer being part of an unregulated control loop.

-Jon

 

[Doug S]

Jon, as always, another clear, carefully crafted post. /ubbthreads/images/graemlins/thumbsup.gif

 

[php_44]

RE: Jonathan's post...

Good explaination. As far as application - I'd suggest that direct PWM dimming and/or regulation of incandescent lamp brightness from a battery voltage higher than the lamp's rated voltage - would not significantly negatively impact that lamp's life. Within reason, tha lamp will operate just fine as long as the RMS voltage it is presented is close to what it is rated for. (I wouldn't suggest operating a 12V lamp from 1500V at ultra low duty cycle, but operating a 3V bulb from 6V at 50% duty cycle would be fine).

LED's are a different animal. They need current limiting. <ul type="square">[*]Direct PWM of voltage source to an LED would destroy it pretty quickly in contrast to an incandescent lamp.
[*]Direct PWM of a regulated current source to an LED would allow effective dimming.
[/list]
A look at most semiconductor maker offerings will show several types of circuits designed to regulate LED current.
<ul type="square">[*]Many are inductor based switching power supplies and are very efficient current regulators.
[*]Some are linear and are no more efficient than a resistor.
[*]Some newer ones are "switched capacitor" designs that use capacitors to store then dump charge into LEDs in a regulated fashion. They are not as efficient as the inductor based designs due to the number of switches they require, and capacitors' low impedance necessitates the use of resistors to limit current to the LED(s) (and waste energy).
[/list]

 

[Bullzeyebill]

Then, what about micro pucks, and regulators used in MM and BB sandwiches. Some are noted as current regulated, and some voltage regulated. Are these efficient designs?

 

[Jonathan]

[ QUOTE ]
php_44 said:
RE: Jonathan's post...

Within reason, tha lamp will operate just fine as long as the RMS voltage it is presented is close to what it is rated for. (I wouldn't suggest operating a 12V lamp from 1500V at ultra low duty cycle, but operating a 3V bulb from 6V at 50% duty cycle would be fine).


[/ QUOTE ]

Actually PWM operation of lamps requires that you adjust duty cycle as the inverse _square_ of the supply voltage, because you need to maintain constant _RMS_ voltage, not constant average voltage.

/ubbthreads/images/graemlins/smile.gif

-Jon

 

[Jonathan]

[ QUOTE ]
Bullzeyebill said:
Then, what about micro pucks, and regulators used in MM and BB sandwiches. Some are noted as current regulated, and some voltage regulated. Are these efficient designs?

[/ QUOTE ]

I don't know about the _efficiency_ of the micropucks, but I believe they are yet another sort of beastie: I believe that they use switching techniques and an inductor to produce a smooth DC output, but I am pretty certain that they are _not_ regulated. In other words, they are boost switching converters, but not regulators.

The BB devices, on the other hand, are true regulated boost converters, and I happen to be _very_ familiar with the circuit, having used the same chips in a different application. For these the efficiency will be somewhere between 70% and 90%, depending upon the input and output voltages. For operating a single LS from 3V input, I'd bet they are in the 85%+ efficiency range.

-Jon

 

[Doug S]

[ QUOTE ]
milkyspit said:
Does the use of a resistor in an unregulated LED flashlight such as the Streamlight 4AA LED reduce its power efficiency due to some of the energy being lost in the resistor as heat? I'm afraid I switched out of my Electrical Engineering major halfway through college, and now questions like this haunt me. /ubbthreads/images/graemlins/icon23.gif


[/ QUOTE ]
OK, actually answering the question you asked:
Yes, the resistor consumes power but in the case you cite, the more power the resistor consumes [relative to the LEDs, not necessarily in absolute terms], the *more* efficient the light becomes! It may not be as bright but it is more efficient in terms of light output per input of power. The reason for this is that the LEDs become more efficient as the current is reduced. If you wanted the highest possible lumen-hours per set of batteries the resistor would need to be high enough in value that the 5mm LEDs in the Streamlight were getting only a couple of mA each.