Regulated SL Prop 4AA LED?

[milkyspit]

I read lots of glowing testimonials on these pages for the SL Prop 4AA LED, and I love mine as well, but nearly all have the same complaint: the light output drops-off steadily, beginning even with the first few minutes using fresh batteries. This is also not a good candidate for lithium AA's due to the fact that the LEDs are already being driven pretty hard. (I burned-out 3 of the 7 LEDs in one of mine by using Lithiums, in fact.)

What about regulation? I've not done any regulator mods (yet) myself, but this would seem like a great light to accept a small regulator board in the space behind the bulb module, and regulation would seemingly eliminate the primary complaint about the light. Superlight, anyone? /ubbthreads/images/graemlins/cool.gif

I've got to believe someone's already tried this? Which module(s) would work best with this particular flashlight?

 

[Doug S]

From what I have seen of the module, this would be an easy light to add a linear constant current regulator to. Set to 30mA/LED or 210mA total, it should be possible to get around 10 hours of constant current before the slow roll-off begins using alkaline AAs.

 

[wualta]

Would you find the tradeoff of lower brightness for constant output of nimh rechargeables useful? This change is at least simple and quick.

Walt Brand

 

[milkyspit]

[ QUOTE ]
wualta said:
Would you find the tradeoff of lower brightness for constant output of nimh rechargeables useful? This change is at least simple and quick.

Walt Brand

[/ QUOTE ]

Walt (what's wualta?), I have a personal bias against NiMH and avoid them if possible. Psychologically speaking, I have a hard time accepting the concept that the NiMH will lose its charge day by day even if I don't use the light. I'm more geared toward things like lithiums, where even years later the light works as if the batteries were inserted yesterday. I know many flashoholics have great success with NiMH and I'm probably being a little paranoid, but in the end I go with what makes me comfortable. /ubbthreads/images/graemlins/icon23.gif

Alkalines, of course, have the advantage of extremely low cost while retaining decent shelf life, although they sadly aren't rechargeable. /ubbthreads/images/graemlins/icon23.gif /ubbthreads/images/graemlins/icon23.gif That's where the regulation circuit comes in! And I can keep the brightness near that of fresh batteries, at least in theory. Muahahaha... /ubbthreads/images/graemlins/buttrock.gif

 

[shankus]

I was thinking about this same thing, two nights ago.

I measured the current going to my LED module, with fresh alkalines, at 580mA.

I have a fried MadMax board that I fit-checked in the light, and it fits fine.
The MadMax board even still has the connectors for the luxeon mount board still installed, and that makes it larger than just the converter board.

I put the board along side of the cells, on the Streamlight logo side, below the sticker that shows proper cell installation.

The output leads from the board could be soldered directly to the bottom of the LED module (both Negative leads soldered together, since they're common), a small insulator epoxied onto the Positive lead, and then the Positive input lead epoxied onto the insulator.

As for the technical details, I'm unsure. If a BadBoy 500 board, for instance, were used to drive the module at a constant current of 500mA, would that work? I ask because the module is designed to work at around that current, but driven by 6V. I wonder what the loaded voltage would be from the BadBoy. I just haven't played with LEDs, and I don't know how they react with a constant current, at varying voltage. I suspect that since an LED is a current oriented device, that it would work fine.

Another choice would be a dialed-down MadMax board, but it is a voltage regulated board, as opposed to a current regulated one.

 

[wualta]

[ QUOTE ]
milkyspit said:

(what's wualta?)

[thinks:] He's playing right into my hands, heh heh. [aloud:] Imagine you're a female Miami native. Now say my name.


> Psychologically speaking, I have a hard time accepting the concept that the NiMH will lose its charge day by day even if I don't use the light.

Ah, then this will be an emergency light, and in that case I agree with you. But many of us use our lights every day or at least once a week, and have fast chargers (for our other AA-operated gear, like digicams) and so always have fresh nimh on the boil.

A flashlight would have to go unplayed-with for weeks before anyone would begin to notice a falling-off of runtime when using nimh.

This assumes good, high capacity cells and a good charger that treats each cell individually. A lot to assume.

> I know many flashoholics have great success with NiMH and I'm probably being a little paranoid, but in the end I go with what makes me comfortable. /ubbthreads/images/graemlins/icon23.gif

[/ QUOTE ]

Not a problem. If you wanted to run a flashlight off a series string of lemons with copper and zinc strips jammed in 'em, I'd say go for it.

Walt Brand

 

[Kill-O-Zap]

NiMH self discharge: Hmm... not entirely convinced... have had a set of old Panasonic 1600's in a kid's toy (Pixter) for many months, maybe even over a year, and still going. This toy does not get a lot of use (it stinks!), but that's not the point; the point is that the self-discharge issue seems to me to be overblown. If the possibly dimmer output on NiMH's bothers you, you could replace the resistors with a lower value.

MadMax/BB: these run on 3V (except BB driving 5W Lux), and the SL runs on 6V, so you would have to reconfigure the battery wiring from series to parallel, probably not hard to do. BB500 is probably too much for these 7 LED's, that would be over 70ma to each LED rated at 20ma. (I am assuming of course that the resistors have got to go). You might be able to get away with BB400, but I would go with MM ratcheted down a few notches. Or MicroPuck (350ma total, 50ma per LED, perfect). And I WOULD use NiMH's even then.

Lithiums: if you don't mind paying for Lithiums, how about decreasing the resistor values and running on 3 lithiums and a dummy cell?

Aside: running white LED's on 4 AA's is an inefficient design. You could run whites on 3 alkaline's with a smaller resistor value and get exactly the same output for exactly the same amount of time as 4AA with higher resistance, so you're basically wasting one AA per battery change, and lugging around 1oz more than you need, with a 4 cell design. With hand-picked low Vf whites, you could even do just TWO lithiums with no resistors, for the ultimate in efficiency, though probably not as bright as the massively overdriven propoly is.

Just my random ramblings. YMMV.

 

[BuddTX]

I am pretty sure that the Streamlight folks have made several modfications to this light since it was introduced.

The newest versions have different resistors and other improvements in the light.

 

[The_LED_Museum]

Mine has 22 ohm 5% resistors on each LED. There is no "extra" resistor on the outside of the LED module that some SL Propoly 4-AA owners have reported. So I assume I have one of the newer models... but when you "assume", that makes an "***" out of "u" and "me". /ubbthreads/images/graemlins/blush.gif So I guess I'd better not do that. /ubbthreads/images/graemlins/eek.gif

 

[shankus]

[ QUOTE ]
shankus said:
I measured the current going to my LED module, with fresh alkalines, at 580mA.

[/ QUOTE ] Do you use NiMH in your Streamlight 4AA LED?

The LEDs in the Streamlight Propolymer 4 AA are overdriven quite a bit. The 580mA my module was drawing equates to 82.9 mA per LED.


As far as the BadBoy & MadMax boards, the BadBoy can have Vin as high as 9V. The MadMax says 3.1V max (Limited by Vf). Wayne, at the Sandwich Shoppe could say for sure, but I'm wondering if the MadMax could take the 6V as well. The Streamlight module is fine with 6V applied, so I'm calling forward voltage, in this case, 6V.

Either way, the BadBoy 500 could be used, although it would underdrive the module.


I also considered a parallel cell arrangement. I don't know how to disassemble the switch assembly on the 4 AA. The cells can be loaded into the body in parallel, no problem. But there is a switching problem to solve, once they're in.

 

[milkyspit]

[ QUOTE ]
wualta said:
Ah, then this will be an emergency light...
Walt Brand

[/ QUOTE ]
Actually, if (when?) I successfully regulate the SL it won't be an emergency light, though I'm sure it would serve me well in a power outage. My real motivation is to keep the brightness at or near that of fresh batteries, for as long into the runtime as possible, something the unregulated light doesn't do very well. With regulated output, suddenly this light would become a workhorse.

No, I guess my aversion to NiMH has to do with the fact that I will occassionally NOT pick up any given flashlight for weeks or even months at a time, based on my specific needs at any point in my life. But when I need the light, I want it to work as I expect (which typically means close to its full potential) immediately! Or perhaps it's just a mark of laziness, in not wanting to have to open the housing and swap cells very often. /ubbthreads/images/graemlins/sleepy.gif Or it might be a mark of advanced mental illness. /ubbthreads/images/graemlins/help.gif

Ease of use counts for something, too, and I've already got too many maintenance tasks in my daily life. Those evil magnopolists got the ease-of-use thing right with the rechargeable 3D incandescent I got for Christmas, where you just click it into its wallmounted charger/holster and it waits for you to come calling again.

Here's a crazy thought: what about some sort of induction-driven charging system like those Sonicare toothbrushes, where putting the light in a charging sleeve induces current flow inside the light and recharges the batteries, all without needing to open the housing? Things that make you say, hmm... /ubbthreads/images/graemlins/thinking.gif

[ QUOTE ]
wualta said:
Not a problem. If you wanted to run a flashlight off a series string of lemons with copper and zinc strips jammed in 'em, I'd say go for it.
Walt Brand

[/ QUOTE ]

What kind of runtime would the string o' lemons provide? /ubbthreads/images/graemlins/smile.gif

Incidentally, though off-topic, one cool thing about the SL is that I can cover it with dish soap, toss it into the kitchen sink and handwash it with my dishes when it gets filthy. You've got to admit, that's darned cool. /ubbthreads/images/graemlins/cool.gif

 

[milkyspit]

[ QUOTE ]
Kill-O-Zap said:
NiMH self discharge: Hmm... not entirely convinced...


[/ QUOTE ]

My experience is based entirely on NiMH notebook computer batteries, which would frequently be dead as a doornail if the notebook sat around for a month or two without being used. Someone in that realm once estimated that a NiMH battery in storage loses something like 2% of its charge daily, and in my experience that certainly did seem like a reasonable estimate. Lithium-Ion notebook batteries can sit for months on end and still deliver the majority of their original charge. Haven't tried the NiMH AA's; maybe they're very different.

[ QUOTE ]
Kill-O-Zap said:
Lithiums: if you don't mind paying for Lithiums, how about decreasing the resistor values and running on 3 lithiums and a dummy cell?


[/ QUOTE ]

I might first try using 3 lithiums and a dummy cell *without* changing the resistors, on the theory that 3 lithiums might approximate 4 alkalines midway through their discharge cycle... but the lithiums should CONTINUE to deliver at that level over most of their service life. Am I whacked, or does this make some sense? I'll try it regardless in the interest of science. /ubbthreads/images/graemlins/smile.gif

[ QUOTE ]
Kill-O-Zap said:
Aside: running white LED's on 4 AA's is an inefficient design...


[/ QUOTE ]

This makes a lot of sense. My only thought would be to wonder whether 4 AA will deliver longer useful runtime overall, despite perhaps delivering LESS useful runtime per cell (i.e., lower per-cell efficiency)?

 

[Kill-O-Zap]

Milkyspit:
[ QUOTE ]
My experience is based entirely on NiMH notebook computer batteries, which would frequently be dead as a doornail if the notebook sat around for a month or two without being used. Someone in that realm once estimated that a NiMH battery in storage loses something like 2% of its charge daily, and in my experience that certainly did seem like a reasonable estimate.

[/ QUOTE ]
It's a widely accepted fact that this self-discharge exists, and I've seen that 2% figure before. I can't explain the fact that one set of batteries of mine has not exhibited nearly as bad a self-discharge, but nevertheless, there it is. /ubbthreads/images/graemlins/confused.gif

[ QUOTE ]
I might first try using 3 lithiums and a dummy cell *without* changing the resistors, on the theory that 3 lithiums might approximate 4 alkalines midway through their discharge cycle...

[/ QUOTE ]That's a good idea. Now that I know how overdriven these LED's are on four fresh alkys, I'm inclined to think three Lithiums w/o changing the resistors may be just the ticket. It won't be as bright in the beginning, but really it was insanely overdriven to begin with, and this will be more reasonable.

[ QUOTE ]
My only thought would be to wonder whether 4 AA will deliver longer useful runtime overall, despite perhaps delivering LESS useful runtime per cell (i.e., lower per-cell efficiency)?

[/ QUOTE ]
I realize it's counter-intuitive, but my claim (as yet not substantiated by experiment) is that the runtime curve would be identical, i.e. no increase in useable output or anything else with more cells+more resistors. I base this conclusion on the fact that if you run the same current through 4 cells or three cells, then in either configuration each cell will discharge at exactly the same rate, and the voltage will decrease at exactly the same rate. This conclusion startles even me, it just seems that you should get *something* out of an extra cell... Maybe someone can show how that would be the case.

shankus:
You are right, the BB can run off 6V, and will then step that up to something >6V as is the case when driving a 5W LS. I had assumed that if putting in a regulator you would want to do away with the resistors. Leaving them in results in a lot of inefficiency, but if you don't mind that, then I don't see why it wouldn't work. One thing you might want to check with Wayne on is whether Vin=Vout=6V is supported by the board.

When you say "BadBoy 500 could be used, although it would underdrive the module" well, that's one way to look at it; the other way is that the module is apparently INSANELY overdriven on 6 Alkys and BB500 would somewhat less insanely overdrive it...

 

[wualta]

[ QUOTE ]
milkyspit said:
...My real motivation is to keep the brightness at or near that of fresh batteries, for as long into the runtime as possible, something the unregulated light doesn't do very well.

...Or it might be a mark of advanced mental illness...

Here's a crazy thought: what about some sort of induction-driven charging system like those Sonicare toothbrushes, where putting the light in a charging sleeve induces current flow inside the light and recharges the batteries, all without needing to open the housing? Things that make you say, hmm...
What kind of runtime would the string o' lemons provide?


[/ QUOTE ]

Hmm. This SL Polypro light must be a battery sucker of the first water, if you can notice a significant drop in brightness during typical flashlight-type use. I'm glad I didn't get that one. I almost did.

Mental illness? If you're surrounded by milk spitters, that's normal.

LemonLED runtime? Well, it would depend on whether these lemons were the regulated variety or not.

The induction-charging idea is an excellent one, and I don't know why it's not offered on small flashlights.

Walt Brand

 

[milkyspit]

[ QUOTE ]
Kill-O-Zap said:
I realize it's counter-intuitive, but my claim (as yet not substantiated by experiment) is that the runtime curve would be identical, i.e. no increase in useable output or anything else with more cells+more resistors. I base this conclusion on the fact that if you run the same current through 4 cells or three cells, then in either configuration each cell will discharge at exactly the same rate, and the voltage will decrease at exactly the same rate. This conclusion startles even me, it just seems that you should get *something* out of an extra cell... Maybe someone can show how that would be the case.


[/ QUOTE ]

I haven't thought through this completely yet, but it seems to me that it works something like this: As voltage drops the current of course drops, too, but given the nature of alkaline batteries, the voltage will drop less rapidly as current flow diminishes, thereby producing somewhat of a levelling effect on the voltage. If the voltage at this point of diminished current flow is still sufficiently high, the LEDs will continue to shine relatively brightly (4AA), but if the voltage at this point is a bit too low, the LEDs may grow much dimmer or even go out completely (3AA). LEDs have a particular turn-on voltage below which they won't even light, according to my off-the-cuff experiment using a variable voltage on a bench power supply. Decreasing current, on the other hand, dims the LEDs fairly linearly pretty much all the way down to 0mA. This may not be techically correct, but it's what I've observed.

Clear as mud, eh? /ubbthreads/images/graemlins/wink.gif

 

[shankus]

[ QUOTE ]
Kill-O-Zap said:
I had assumed that if putting in a regulator you would want to do away with the resistors. Leaving them in results in a lot of inefficiency...

[/ QUOTE ]
I had assumed that you guys were talking about regulating it with the module intact.

Now, splitting the module open makes other options. Removing the resistors, or replacing them with smaller current limiting resistors is an option I hadn't thought of. It looks as though it may be possible to put the regulating circuit in there, as well.

 

[Kill-O-Zap]

milkyspit,

First, I apologize about taking this thread off topic like this; since I've never mentioned this crazy idea before, nor have I seen it discussed before, I should have posted to the electronics forum to get hashed out (and perhaps shot down in flames /ubbthreads/images/graemlins/smile.gif). Having said that, I will say one more thing about this, then if I have any more to say, I'll move this subtopic to Electronics.
If I understand what you are saying, then you are not disputing the claim that in both configurations the batteries discharge at the same rate, but you are saying that 4 AA's that are quite depleted would have greater voltage vs. 3 that are equally depleted, and therefore would still be able to light the LED's at the point where just three would not? But is that taking into account the fact that the 4 AA's have a higher value resistors (and voltage drop) to contend with? It seems to me that ultimately the LED's see the same voltage either way. My skill with the equations involved here is too limited; I can't prove what I'm claiming; primarily I'm stymied by the fact that the LED's decrease their load as the battery voltage drops, but while I can't convince myself that that does not affect my argument, neither can I show how it defeats it.

I would love to see an experiment done... but I don't have a 4-cell light to do it with, and you may understandably be reluctant to mess around with the resistors in yours.

 

[asdalton]

[ QUOTE ]
Kill-O-Zap said:
If I understand what you are saying, then you are not disputing the claim that in both configurations the batteries discharge at the same rate, but you are saying that 4 AA's that are quite depleted would have greater voltage vs. 3 that are equally depleted, and therefore would still be able to light the LED's at the point where just three would not? But is that taking into account the fact that the 4 AA's have a higher value resistors (and voltage drop) to contend with? It seems to me that ultimately the LED's see the same voltage either way.

[/ QUOTE ]

The difference between the 3-battery and the 4-battery setups comes from the fact that the LED is (approximately) a fixed-voltage device. Resistors are not; the voltage drop across a resistor is proportional to the current running through it. Resistors will reduce the current flow through the LED, but they cannot stop it--so long as the battery voltage is higher than the Vf of the LED.

Let's say that a particular LED has a Vf of 3.6 volts. The batteries must supply at least this value in order for the LED to work at all, regardless of what resistors you use. For three alkaline batteries, that means that you can only run each battery down to 1.2 volts. This wastes a lot of the battery's capacity. Four batteries, on the other hand, would allow you to run each battery down to 0.9 volts, which is essentially dead.

Using four batteries won't give you a flatter discharge curve overall, but more of each battery is used. This can translate into higher initial brightness or longer runtime, depending on the flashlight design. Also, a four-battery LED light can often use rechargeable batteries, which probably couldn't supply enough voltage in a three-battery flashlight.

 

[Doug S]

Let me expand a bit on what Andrew has said about the merits of 4AA vs 3AA. In keeping with the original topic of this thread, let's look at the case of a linear constant current regulator driving the array of 7 LEDs. Assume that 3.6V to the LEDs is required to maintain the desired current. Assume the CC regulator is perfect in that it has 0 dropout voltage. This is close to what is actually attainable which is <100mV. If less than 3.6V is available at the desired current then the light falls out of regulation, i.e., current begins to decrease. For 4 cells we need at least 0.9V/cell but for three cells we need 1.2V/cell. It is instructive to look at a specific scenario. Assume the CC regulator is set to 0.25A. Look at the Duracell AA datasheet in the link below. The third graph shows the voltage curve for various constant currents including 0.25A. At a current of 0.25A it takes 4 hours to reach 1.2V and 8.5 hours to reach 0.9V. As you can see, for this setpoint current the light will be in regulation over twice as long with 4AA vs 3AA. In general, the advantage of 4 vs 3 cells is greater for high currents and high LED Vf.

http://www.duracell.com/oem/Primary/Alkaline/mn1500.asp

 

[Kill-O-Zap]

Well, I went ahead and started that thread in Electronics, but I cant respond there to these latest appends here without replicating all the history, so here goes one more append on this subtopic here.

Ok, if there's a hole in my argument (and probably there is) surely it has something to do with the ever-increasing resistance of the LED as the cell voltage drops, just as you guys are saying, but I'm not quite there yet. The ever-increasing resistance of the LED increases the efficiency of the circuit (in terms of ratio of power burned in LED's vs. power burned in the resistors) but both configurations benefit from that phenomenon. What's still missing for me is some clear explanation as to why 4 cells would benefit more from it.

Andrew,
I think the point you are making is the same as milkyspit, and it still appears to me that that argument does not take into consideration the fact that if you're going to use 4 cells, and keep the same initial output as with three cells, you need a lot more resistance, and that when each cell drops to .9 volts, the resistors will drop some of that 3.6 volts, and you may be no better off than you would have been with fewer cells and a lower resistance. [BTW, 3.6 V is not the point where the LED will no longer light, but the point where it is at full power]. It is true that you cannot overdrive the LED's with 3 NiMH's, or even properly drive them if you are also going to also support 3 Alkys (i.e. add resistance).

Doug,
Although this thread is about "Regulated" my claims were made specifically wrt. to unregulated, resistored designs, where increasing the number of cells means increasing a resistor value, and therefore decreasing efficiency. Of course, with a 100% efficient CC regulator, adding anything on the input means you get that back out on the output, in the form of longer runtime. The charts you reference, to me, only show that an individual cell is more efficient at lower loads, but I claim that in a 4 cell series setup, the load on each cell is the same as the load on each cell in a 3 cell setup, if the LED drive is kept the same.