Resistors vs. efficiency?

Candle Power Forums

Help Support Candle Power:

[ QUOTE ]
Doug S said:
[ 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.

[/ QUOTE ]

THANK YOU Doug! This is exactly the type of insight I was hoping for! /ubbthreads/images/graemlins/thumbsup.gif

(Not to take anything away from the others who've contributed... I've enjoyed reading the thread as it's progressed, and have learned quite a bit. Thanks to all! /ubbthreads/images/graemlins/smile.gif )

I have a little bag o' Nichia 5mm LEDs (the super bright, 9200mcd ones I think) that I bought via eBay. Over the weekend I breadboarded a few of these in parallel, all fed by a 6V source consisting of four alkaline AA batt's limited by a 45 Ohm resistor. (In other words, the single resistor is between the positive voltage feed and the three paralleled LEDs.) I read a *total* current of approx. 65mA in the circuit with fresh batteries, and the leads of the LEDs didn't get warm at all, at least based on feeling them with my thumb and forefinger. Seems to me this circuit, encapsulated in a flashlight, would yield some nice runtimes? Or would I be better served to regulate via a BB? My goal is to drive the LEDs near full brightness, but using not a drop more power than necessary to do so! and keep running approximately this way for many many many hours...

I want this to be my designated "power failure light" that will be just bright enough to light a room, yet keep running for at least 20 hours, preferably at least several days...
 
As you have engineered it, it should run a couple of days continuous. The BB is a high power, short runtime solution, not what you are looking for in your application. To me "not using a drop more power than necessary" does not make sense for this application. You could use a bunch of electronics to be more efficient than your design but it doesn't make economic sense for your application.
 
Jonathan, is the MM also efficient, like the MM. I realize the MM sandwich is voltage regulated. Bill
 
[ QUOTE ]
Bullzeyebill said:
Jonathan, correction. Is MM as efficient as BB. Bill

[/ QUOTE ]
While the relative improvement of the MM over the BB varies with Vin, Vout, and Iout, the MM is in general more efficient than the BB. The MM uses the MAX1674 IC which has onboard synchronous rectification which the BB IC [LT1618] does not. This improves efficiency.
 
RE: Doug's post
[ QUOTE ]
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!

[/ QUOTE ]
Yes, as Doug said, InGaN LEDs have an interesting characteristic that makes them more efficient at low current levels than at high ones, but you get less light. (sort of like gas mileage - you can drive real slow and use your gas efficiently, or kick your speed up and use your gasoline less efficiently).

I have to disagree with using a resistor for that. If you look at the numbers - you'll waste far more energy in the resistor than you'll gain in LED efficiency. OK - for 65mA into three LEDs - you're at ~22mA per LED. Nichia rates the LED at this current level so you are operating at an efficiency of ~100%. At 6V, and 65mA you're wasting 2/10th of a watt in the resistor for 2/10th of a watt of light. You're 50% efficient - not great.

If you cut this down to 10mA/LED you're at 30mA total. A resistor of 100ohms. You're now wasting 1/10th of a watt in the resistor, and putting 1/10th of a watt into the LEDs to get 12/100th of a watt of light (this is the efficiency boost). So you're still just a hair under 55% efficient with half the light output of 20mA/LED.

Let's be insane. Say you get 15 expensive LEDs at 2mA each for 30mA total. According to tests done on Nichia LEDs, 2mA is the peak efficiency point. Still 1/10th of a watt wasted in the resistor. Now you're getting 1/8th of a watt of light for 1/10th of a watt of power. Total efficiency is still about 56%.

You'd be way ahead of the game to run at 4.5V with a ~25 ohm resistor for efficiency. You'd be nearly 70% efficient.

Get a Madmax or build the equalvalent, set the voltage output to give you ~60mA through the LEDs, and run it off of two "D" cells. Now you're 70-88% efficient, and you'll run for probably 220 hours based on the "D" cell discharge curve and the fact that the MADMAX will operate down to 0.8V per cell or lower. Heck 2AA alkalines will run this configuration for 20-25 hours. You'll enjoy fairly consistant brightness (as opposed to the dramatic light level drop in a resistor'ed light), and you'll have two fresh spare cells that you would have spent in your four cell light heating up a resistor - that you could use for another 20-25 hours. Can you tell I'm a regulated light fan? If you use rechargeable cells, feel free to make resistor'ed lights since the waste is less painful when you can recharge the cells.
 
[ QUOTE ]
php_44 said:

I have to disagree with using a resistor for that. If you look at the numbers - you'll waste far more energy in the resistor than you'll gain in LED efficiency. OK - for 65mA into three LEDs - you're at ~22mA per LED. Nichia rates the LED at this current level so you are operating at an efficiency of ~100%. At 6V, and 65mA you're wasting 2/10th of a watt in the resistor for 2/10th of a watt of light. You're 50% efficient - not great.



[/ QUOTE ]
I would have to argue that your response has wandered away from the original question posed in this thread which was directed towards unregulated resistive ballasts such as the Streamlight 4AA. Within the confines of the original question it is not relevant what percentage of the heat is generated in the LED vs the resistor. What is relevant is the total integrated light that is generated as the batteries' chemical energy is expended. Regardless of the resistor value [including 0] the same amount of charge will pass through the LED if enough time elapses. We both agree that if that charge passes through at a lower rate [current] more light will be produced over time. The intended use of this light is that of an emergency light for long term power outages. I note that the original poster is from New Jersey. All the New Jersey jokes not withstanding, this is not a third World country. Power is pretty reliable. 1 resistor = $.02, 1 MM = $20. It is higher unlikely that a lifetime of power outages could justify the difference.
Yes, I am a big fan of true constant current regulation too. As an engineer, however, there is no way I could justify it for the specific application cited.
 
[ QUOTE ]
Doug S said:
I would have to argue that your response has wandered away from the original question posed in this thread which was directed towards unregulated resistive ballasts... Within the confines of the original question it is not relevant what percentage of the heat is generated in the LED vs the resistor. ... The intended use of this light is that of an emergency light for long term power outages.
It is higher unlikely that a lifetime of power outages could justify the difference.


[/ QUOTE ]

A good debate can be very illuminating /ubbthreads/images/graemlins/wink.gif

Actually, the original question posted is:
[ QUOTE ]

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?

[/ QUOTE ]

I intended my post to be an illustration using real numbers that yes abosulutely the power efficiency is lower. I also posted two methods of obtaining a better power efficiency.
<ul type="square">[*]Run at 4.5V with a cheap resistor, and be prepared for the steeply declining brightness as the battery drains. This can be ~70% power efficient.
[*]Use an inductor based switch mode regulator, and have reasonably constant brightness. This can be ~70-90% power efficient.
[/list]

You're right, the same current flows through the LED for a 4 cell or a three cell resistor'ed light. The difference is the 4th cell is solely there to heat the resistor. A three cell solution saves you burning up one battery. Save another battery by using an inductor based switchmode regulator - since that third battery is mostly just heating up the resistor given that InGaN LED's only need a hair over three volts to operate. I'll call the cells dead at ~1V.

For emergency use, the choice of tools is a very personal decision. I happen to prefer a light with even and reliable brightness, even if the cost is higher and I never recoup the cost of saved batteries. Personal choice.

I've been through two fairly long blackouts due to storms. During those storms I brought out my collection of three regulated lights, and one resistor'ed light. The regulated lights were 1 LED 2 cell homebrews with clikie switches and one ARCLS grey. The unregulated light was a homebrew 4 LED turtle style light (3 cell) with a clikie switch. All I can say is, guess which ones the other four members of my family grabbed to find the bathroom? It was one of the regulated lights. Guess which lights were grabbed to use as low level room lights to read by? At first the unregulated light, but after it burned for about two hours, it was left on the kitchen counter in favor of the consistently bright regulated lights. To me a LED flashlight that rapidly dims during emergency use is not as valuable a tool as one with consistent brightness. All this said, I'd rather have a resistor'ed LED light than any incandescent light for long term lighting.

Look at these run time plots:
Sample Resistor'ed light - light level falls off You only have peak brightness for 10-20% of the time the thing still makes light. The rest of the time you're operating at half or less of the original light level. At what point would some folks chuck the half used batteries out?

Sample 1 cell semi regulated light - even light output Even light output. Yes once it falls off you're out of light, but you've used all the battery power you purchased and can throw them away without guilt.
 
[ QUOTE ]
php_44 said:
A good debate can be very illuminating /ubbthreads/images/graemlins/wink.gif



[/ QUOTE ]

Indeed it can! I hope that we produce more light than smoke. I'll rise to the bait and flog this horse a bit more. How's that for a mixed metaphor.

[ QUOTE ]
php_44 said:
Actually, the original question posted is:
[ QUOTE ]

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?

[/ QUOTE ]

I intended my post to be an illustration using real numbers that yes abosulutely the power efficiency is lower.

[/ QUOTE ]
And I have argued that absolutely the power efficiency is higher .
I suspect that we differ on the definition of power efficiency. To me, for the context of the question, power efficiency would be a measure of the amount of desired output per unit of power input to the system. To me, the desired output is light and the system is the circuit external to the battery. It appears to me that you are defining efficiency as the amount of power that gets to the LED as a percentage of total power supplied by the battery without regard to what happens to the power once it gets to the LED. This seems to be the basis for our difference in interpretation of whether the resistor results in greater or lesser efficiency.

[ QUOTE ]
php_44 said:
You're right, the same current flows through the LED for a 4 cell or a three cell resistor'ed light. The difference is the 4th cell is solely there to heat the resistor

[/ QUOTE ]

Not quite. If you are using just a resistive ballast [again, the context of the original question] with a battery chemistry that has a sloping discharge curve [like alkalines] the added cell permits a significantly flatter discharge curve before dropping off to what might be considered unacceptably dim levels.

php_44, both you and I would probably seek a similar solution to this need but that is because we are both Techno-geeks when it comes to lighting. I cannot, however, recommend the techno-geek solution to the need as presented on either economic or relability grounds.
 
Totally off topic or kind of anyway, I almost feel stupid for asking this but this very subject came up today and I had to explain the best I could.
One of the flashlights that have come into the hands of the Mad Dog soon to be Light Company has a form of dimming I have not seen in a flashlight. Depress a button and the flashlight dims more the longer you hold the button. Besides this function the flashlight also has blinking mode in two different rates of time. I think the dimming is through resistance and a EverLed lamp would total be a lose as the efficiency drops with the EverLed as the voltage does and to have the resistance on top of that just seems stupid.

Right?
I knew I would feel stupid after asking this never mind.
 
Well, as the author of the original question, I'd like to say I'm totally blown away by the direction this thread is going! Lots of great stuff in your debate, php_44 and Doug S. Yes, truly illuminating! /ubbthreads/images/graemlins/bowdown.gif

Between both your most recent posts, I've got several thoughts, so please read on. If you'd be so kind, I'd appreciate your comments on the issues I raise below... and apologies in advance for what might become fairly involved!

First of all, I'll mention that I too am an engineer, though in my case I started as EE and ended up in CS, and now am a programmer. I think like an engineer I guess, but don't have all the skills you do in translating my ideas electrically-speaking into reality! Also, I do indeed live in New Jersey, though it's quite rural where I live -- think cows and cornfields (yes, there is such a part of NJ) -- and unfortunately we can count on 1-5 power outages every year with duration ranging from a few hours to a couple days. Last summer I began deploying resistored Nichia-based LED flashlights in various rooms for each power outage, and simply leaving them on all night. My favorite light for this purpose has been the TurtleLite II, given its extemely long runtimes. It's not cheap, though, so I began thinking about designing my own mod in an effort either to create something better at roughly equal cost, or something equivalent at significantly lower cost.

My design goals are completely impossible: I want a light costing $0 that runs forever with no loss in brightness the entire time. However, I'm willing to compromise a bit on these constraints. /ubbthreads/images/graemlins/wink.gif In practice, I want a light that will extract every last mA of energy from its batteries and turn it into USEFUL light for as LONG as possible. By this I mean that a light with 200 hour runtime is of no value if it can't help me see the furniture in my room, read a book, etc. On the other hand, there's little value for this particular application if the light runs out in only a few hours, plunging me into darkness. Yes, I could replace the batteries, but it would be so much more convenient if I didn't have to! Just some extra peace of mind, and more flexibility in usage.

Given the success of the TurtleLite II (see Craig's review at LEDMuseum.com if you don't know this light), I began my journey with the belief that 4AA is a workable configuration for my goals, plus it offers the advantages of cheap batteries (22 cents of less bought in bulk), easy storage (small size), and nice form factor (the side-by-side arrangement makes for a light that can be set on its butt candle-style, to splash light off the ceiling and illuminate an entire room). I also began my search thinking in terms of resistors since that's how the TurtleLite II does it, but I couldn't shake the concept that the resistor might be generating a waste product (heat) which I'd rather channel into additional light if at all possible. Hence my original question.

[ QUOTE ]
php_44 said:
I have to disagree with using a resistor for that. If you look at the numbers - you'll waste far more energy in the resistor than you'll gain in LED efficiency. OK - for 65mA into three LEDs - you're at ~22mA per LED. Nichia rates the LED at this current level so you are operating at an efficiency of ~100%. At 6V, and 65mA you're wasting 2/10th of a watt in the resistor for 2/10th of a watt of light. You're 50% efficient - not great.


[/ QUOTE ]

php_44 seems to have anticipated my next step: namely, if the resistor really DOES cause the light to run far less than 100% efficient, which seems clear to me now, are there alternate designs that would provide greater efficiency while staying true to the overall design goals of extreme runtime (1+ days) and useful light (bounce off ceiling, minimally illuminate a typical room)? That's what I'd like to nail down next, if you guys would be willing.

[ QUOTE ]
Doug S said:
1 resistor = $.02, 1 MM = $20. It is higher unlikely that a lifetime of power outages could justify the difference.


[/ QUOTE ]

Costwise, you're absolutely right, Doug. I guess I'm becoming a CPF enthusiast like both of you, though, because I'd be willing to incur the greater cost for a regulated light so long as it truly will accomplish my functional goals better than would a resistored light. I will note that the light(s) I ultimately create will get some use besides the power failures themselves, so to some extent I can justify a little more expense. On the other hand, no $50+ light is likely to be acceptable for this application because I'd need to spend that money several times over to have enough lights for all the rooms in my house (rough count would be 6-7, or 4-5 since I have two TurtleLite II's.)

[ QUOTE ]
php_44 said:
You only have peak brightness for 10-20% of the time the thing still makes light. The rest of the time you're operating at half or less of the original light level. At what point would some folks chuck the half used batteries out?


[/ QUOTE ]

Personally I'm one of those people that chucks the half-used batteries because the light is too danged dim. Maybe I wouldn't if I had started with twice as much light as I really needed, but with Nichia LEDs I generally want all the light I can get out of a given light. One thing I do with my TurtleLite II's is use AA lithiums in them, which I intend sort of as a "poor man's regulator." I exploit the flatter discharge curve on the lithiums to simulate the regulated effect, plus the lithium gives me better shelf life and superior cold weather performance. I don't really NEED those additional benefits, though; the key point is the poor man's regulation.

A truly regulated light sounds appealing because I could get essentially the "lithium" effect with much cheaper batteries. That to me is where the economy comes into the equation! For me it's the cost of the regulator vs. the additional cost of repeatedly buying lithium AA's rather than the alkalines.

Trouble is, the only regulators I've read about didn't give the number of hours I need! Even the NewBeam in my Mag seems to run only 10-15 hours, and that's just not enough for what I want to build. Can I use a MM and get 24+ hours out of 4AA at the brightness level I want? Also, since MM is voltage, not current regulated, will it truly keep my light bright enough at the 24+ hour mark, or will I encounter the same dimming problem I'd have with a resistor?

[ QUOTE ]
php_44 said:
Sample 1 cell semi regulated light - even light output Even light output. Yes once it falls off you're out of light, but you've used all the battery power you purchased and can throw them away without guilt.


[/ QUOTE ]

The sudden darkness effect is troublesome to me, too. I want a light that's predictable, and one aspect of this would be to know when the light's running out of juice well before it goes dark. I really like the red indicator in my NewBeam, and the "moon mode" in some other regulated lights. Would a MM or some other regulation circuit allow me to get either of these so I could anticipate the need to replace the batteries somewhat? Keeping a mental note of the number of hours logged with each light is not an acceptable solution to me.

Would another approach be a direct drive 1W Luxeon driven by 2 or 3 AA cells? I'm thinking in terms of the AA light InReTech sells. A day or two after turning such a light on, would it still produce enough light off alkalines to minimally illuminate a room? (I do like the fact that Nichias don't really color shift as they dim, whereas the Luxeon probably would... right?)

I should also insert the random comment here that I have a number of dummy AA cells, so I don't mind using 3AA or even 2AA in a 4AA housing, though I would only do so if the resulting light met my runtime and "persistent sufficient brightness" criteria.

[ QUOTE ]
Doug S said:
php_44, both you and I would probably seek a similar solution to this need but that is because we are both Techno-geeks when it comes to lighting. I cannot, however, recommend the techno-geek solution to the need as presented on either economic or relability grounds.


[/ QUOTE ]

If all the above criteria were met (1+ days life with acceptable bounce-off-ceiling room illumination even toward the end of that time, plus some indication that the battery is going to need new batteries before it goes dark [a few hours' warning perhaps?]), I would consider the resulting light reliable for my application. Now the question for you folks is: what's the most satisfactory solution for my need?

By the way, you guys rule! /ubbthreads/images/graemlins/thumbsup.gif /ubbthreads/images/graemlins/thumbsup.gif /ubbthreads/images/graemlins/thumbsup.gif
 
[ QUOTE ]
milkyspit said:
...Even the NewBeam in my Mag seems to run only 10-15 hours, and that's just not enough for what I want to build. Can I use a MM and get 24+ hours out of 4AA at the brightness level I want?...I really like the red indicator in my NewBeam, and the "moon mode" in some other regulated lights.

[/ QUOTE ]
This is where you have lost me completely. It sounds like the NewBeam already meets your criteria. If you can get 10-15 hours out of 2 AA's in your NewBeam, all you have to do is change the batteries once to meet your 24+ hours of light from 4AA's requirement and it already has the low battery indicator. All you have to do is keep spare batteries handy. Or is this one of those flashaholic quests for exactly the light you want, ie 24 hours from 4AA's without having to change the battery?

Either way the info. on resistors and regulation is quite enlightening,
Pat /ubbthreads/images/graemlins/grin.gif
 
[ QUOTE ]
treek13 said:
[ QUOTE ]
milkyspit said:
...Even the NewBeam in my Mag seems to run only 10-15 hours, and that's just not enough for what I want to build. Can I use a MM and get 24+ hours out of 4AA at the brightness level I want?...I really like the red indicator in my NewBeam, and the "moon mode" in some other regulated lights.

[/ QUOTE ]
This is where you have lost me completely. It sounds like the NewBeam already meets your criteria. If you can get 10-15 hours out of 2 AA's in your NewBeam, all you have to do is change the batteries once to meet your 24+ hours of light from 4AA's requirement and it already has the low battery indicator. All you have to do is keep spare batteries handy. Or is this one of those flashaholic quests for exactly the light you want, ie 24 hours from 4AA's without having to change the battery?

Either way the info. on resistors and regulation is quite enlightening,
Pat /ubbthreads/images/graemlins/grin.gif

[/ QUOTE ]

No need to be lost! I really like the NewBeam (incidentally, it's a direct replacement for the bulb in Mini Maglites), but there are some problems for my purposes: (1) it's relatively expensive; (2) I want the 4AA batteries to all be in a flashlight at once so I can run it 24+ hours without changing anything (the lights will be scattered around my house during a power failure, and I'm trying to achieve leave-it-and-forget-about-it convenience); (3) not sure this part will run on 6V (in other words, 4AA) without self-destructing; (4) even if it would run on 6V, I'd have to do a fair amount of work to make it fit into a 4AA flashlight since it's not really designed to fit there; (5) not sure exactly how efficient NewBeam is compared to MadMax or other regulation circuits.

Bottom line on the NewBeam is that I think it makes for a GREAT everyday light for walking through the house after dark, checking on the kids, etc., but it's not quite there for what I want out of my power failure lights. Frustratingly close, though! /ubbthreads/images/graemlins/icon23.gif
 
Sounds like you need an Eternalight. 3 AAs driving 4 LEDs with various electronic dimming modes that extend run times to the hundreds of hours. Most models will stand on the tail. There is also a "find me" mode. Check it out at the LED Museum http://ledmuseum.home.att.net/eterna.htm
 
Actually, given the various cost constraints, the run time goal, and the cost of batteries, I think that your requirements would best be met by linear regulation.

You don't really want _efficiency_; you want consistency. Use a 5 cell supply, and a linear regulator configured to give 60mA to 3 leds. You will get about 45 hours of light of constant brightness, with a $1 regulator and $1 in batteries.

The 'efficiency' will be poor, because you are starting with 7.5V of battery and 'throwing away' half of it, but you need the headroom to allow for battery voltage drop.

There are simple linear regulator designs floating around CPF and the wider web. Look for something low dropout, or just add another cell to your pack.

-Jon
 
Scott,

You have some demanding criteria - a true light nut /ubbthreads/images/graemlins/wink.gif. You might be able to get close with electronics, but that would be costly unless you would like to build your own. A 4AA light with the cells connected in two parallel strings of two cells for 3V at double capacity would supply a NewBeam type of light for about the amount of time you'd like. Same with a MM module and a few Nichias if you set the voltage right with the trimpot. But this is is costly for lots of lights. I've built some inductor based switchmode mod's and although the parts are inexpensive it is time consuming - the MM's are a bargain if you value your time.

Three Lithium AA's and a resistor is a cheaper and easier mod, but as you said the battery cost is higher.

You could try 4AA light with the cells connected in two parallel strings of two cells for 3V at double capacity along with Red or Amber Luxeon (or Spider) LEDs and a resistor. These require only 2.5V to run and will produce light with a lot less. Long run time - odd color, light falloff.

Eternalights are smart lights with energy conserving features (excellent lights for their market), but are roughly equivalent to resistored lights in terms of light falloff - they are not regulated, and more costly than NewBeams or MM's.

Wish I had a your ideal solution - but as always:

1. Long battery life
2. Constant brightness
3. Low cost

All of the above are obtainable, but only two at a time!
 
[ QUOTE ]
Jonathan said:
Actually, given the various cost constraints, the run time goal, and the cost of batteries, I think that your requirements would best be met by linear regulation.

-Jon

[/ QUOTE ]
If you want one level of techno-sophistication above resistive ballast, Jon's suggestion is excellent and is the one that I would use myself for your application. I personally would use 4 cells rather than 5 though. At these low current demands you have adequate voltage "headroom" with 4 cells deep into their total discharge. There is also the practical matter that 4 cell holders are everywhere and 5 cell ones are hard to find.
 
Doug and Jonathan have an excellent suggestion if you're willing to work a little.

There are numerous devices you could use with minimal extra parts. One example device that you might look at would be National's LP2951ACN. This part will operate down to 2.1V and with an input to output differential of only ~0.3-0.4V. So you could power 4-5 Nichia whites from 6V till the batteries got down to 3.1V + 0.4V = 3.6V or 0.9V per cell which happens to be the point when most folks consider an alkaline cell to be exhausted.

This part number is for a reasonable 8 Pin DIP (the older IC package with leads far enough apart that most folks can solder then with no problems.) You'd only need an inexpensive 5uF 10V or more electrolytic cap connected between the output and ground, and a 100K trim pot to set your output voltage (brightness). This would be conntected between the output and the feedback pin. You might want to set the voltage a little higher than needed and have a small resistor in series with the LEDs to accomodate temp changes in the LED forward voltage. (Any collective thoughts on that?).

The chip consumes almost no current itself (same with the trim pot). The chip costs ~50cents, the pot about the same, and the cap 25 cents. You'd need to solder it yourself, but it would be cheap and give constant brightness. Better efficiency than a resistor, not as good as a switch mode regulator. If you want to be fancy, hang a red LED off the ERROR pin with a resistor to the battery and you'll have a "LOW battery" light. /ubbthreads/images/graemlins/wink.gif
 
[ QUOTE ]
Doug S said:

If you want one level of techno-sophistication above resistive ballast, Jon's suggestion is excellent and is the one that I would use myself for your application. I personally would use 4 cells rather than 5 though. At these low current demands you have adequate voltage "headroom" with 4 cells deep into their total discharge. There is also the practical matter that 4 cell holders are everywhere and 5 cell ones are hard to find.

[/ QUOTE ]

Funny world, this. My solution to this is *three* NiMH cells and 'Mr. Al's' LDO regulator' (parts about a buck). I typically get 50 or 51 hours at a rock solid 30 mA from 1750 mAH NiMHs. At an honest 90% efficiency (typically .3 of 3.6 V across the regulator, 3.3 for the LED). Seems to work quite well driving a few in parallel. Right now I'm fiddling with 4 white Nichias in parallel at 7, 30 or 120 mA selectable with a small slide switch. The idea is 'two f stop steps'. Yes, there's a bit of chage left, but I get most of it and can't discharge the cells far enough to get into trouble.

Anyway, some time back there was a thread about night lights. I ended up cobbling up a two D cell holder, a microswitch on it's end pressing down on the table and a couple of resistors. When you pick it up the microswitch returns to 'normal' driving the Spider LED to 75 mA (full blast) which is very bright (but red). When at rest, there's another resistor supplying a very small current (under 2 mA with new cells), still bright enough to navigate. I forget the exact numbers, but using the first half of the alkaline D cells (down to 1.2 Volts per) was something over a year with 10 minutes of high and 24 hours of low per day. This was back before Thanksgiving as I recall, it's still ticking happily on the headboard. I've often thought the high was actually too high for indoors, and I ended up dropping the high current down to more like 25 mA. I think the mark II should have some adjustment for this. I did some runs down under this voltage (2.4 total), as Vf is under 2 Volts at low current, light is still usable, but it's less than half the original brightness (the unofficial CPF runtime point), so I figure I'll happily swap them out next winter when I notice it's getting dim.

Aside from that it's super simple to use, even half asleep. Bright enough and wide beam enough to make a useful night light, but automatically jumps to high power when you pick it up (and lay it on it's side or top) Total cost a couple of bucks each. If you don't have access to the spider (very neat part, BTW), a couple of ultra bright red LEDs in parallel should work fine, the key is the low Vf .

I gave some random thoughts a while back about a single cell holder with a small piece of PC board material (or some other thin sandwich of two plates and an insulator between them) soldered to it. You could slip it into the 2 D cell holder (under the plus end) when the battery went flat and install a similar 'half used' cell in it and drain all 3 down under .8 each. Somewhere between here and there is the land of diminishing returns.......

A year's battery life, *being on full time*, should be long enough to satisify most folks, one would think.

BTW, I too enjoyed lurking on the spirited discussion on what's most efficient. I hope we have more of that in the future. I'm willing to do my part to stir the pot.....perhaps we need to consider the *system*. LED, driver, and *battery*. Shouldn't efficiency really be measured in 'light per battery', that is isn't that the way the user rates it? 'How often do I have to put batteries in?'. Since total energy delivered by the battery is highly dependent on rate of discharge, I suspect this becomes a major factor (or suggest it should be).

Fun stuff, fer sure.

Doug Owen
 
You must log in or register to reply here.

Latest posts

Back
Top