Mini Maglite LED 2AA mod

EvilLithiumMan · Sep 4, 2006

With the tailcap and batteries removed, push (hard) on the points indicated:

(I guess you could push directly on the emitter, but I don't like that approach)
The LED module will drop out of the bottom. The retaining cup will come out from the top.

Use a small screwdriver to break down the module. There is an offset line on the module housing (highlighted here). Take note of how the PCB and emitter housing are oriented to avoid the wrong polarity when reassembling:

The sense (feedback) resistor is labeled "R050" - it's the only resistor:

OK - here is where I became tired, stupid, sober or something. Without measuring the R050 resistor, I somehow decided it "had to be .050 ohms". I added a .10 ohm resistor across it:

But after adding the .10 resistor, I became suspicious. How did I know R050 equaled .05 ohms and not 50 ohms? Did I just increase the programmed current by a factor of 500? Well, the mod was already done. I had no practical way of undoing it. And I know damn well the circuit isn't capable of a 500 fold increase. Worse case, I had simply maxed out the converter. The only question now was whether maxing it out would mean the popping the emitter or failure of the converter itself. Hell, I've blown up $150 lights within seconds of ripping them open. For $24, I should worry?

Before the mod, I measured 380ma draw from both Nimh and Alkaline cells. With the mod, it was initially drawing 712ma with alkaline cells and 630ma with Nimh. The light is brighter. I can't tell you how much brighter until my roomate wakes and I can get a hold of his Maglite AA LED. He's usually up at the crack of noon.

I've run the light for 20 minutes in still air. It becomes warm, but nothing terrible. The emitter seems fine. The converter seems fine. Time will tell. I'll try for beam shots later today.

ELM

carnal · Sep 4, 2006

Oh Noooooo..... EvilLithiumMan has MANGLED A MagLed! You'll probably be sued!
The evidence is here with full documentation! You've probably used SOME tool to open it with that mag CLAIMS to have a patent on (you know a screwdriver or something)

I hope Newbie weighs in on the PERFECT sense resistor for this lite (2AA and 3AA).
Guess we'll have to wait for a FULL report on your dastatdly deed.

Great to see an opened AA MagLed.

Great job EvilLithMan!
Your sig line is missing a Magnadoodle you know! I have one for sale if you need one.
Brian

CM · Sep 4, 2006

Ah ya beat me to it. Nice! How good is that thermal path from LED to the body?

The higher current draw with alkalines confirm my measurement also, this is a voltage multiplier circuit which is great for alkalines. The fact the light didn't get very warm suggests that their heat sinking is very poor. I guess underdriving the snot out of the Lux III is the solution to the problem.

Now for the meat. The circuit (if you haven't determined yet) uses the Zetex ZXSC 310. 50 milliohms is correct and is supposed to provide 380mA to the LED.

ElektroLumens · Sep 4, 2006

EvilLithiumMan said:
With the tailcap and batteries removed, push (hard) on the points indicated:

(I guess you could push directly on the emitter, but I don't like that approach)
The LED module will drop out of the bottom. The retaining cup will come out from the top.

Use a small screwdriver to break down the module. There is an offset line on the module housing (highlighted here). Take note of how the PCB and emitter housing are oriented to avoid the wrong polarity when reassembling:

The sense (feedback) resistor is labeled "R050" - it's the only resistor:

OK - here is where I became tired, stupid, sober or something. Without measuring the R050 resistor, I somehow decided it "had to be .050 ohms". I added a .10 ohm resistor across it:

But after adding the .10 resistor, I became suspicious. How did I know R050 equaled .05 ohms and not 50 ohms? Did I just increase the programmed current by a factor of 500? Well, the mod was already done. I had no practical way of undoing it. And I know damn well the circuit isn't capable of a 500 fold increase. Worse case, I had simply maxed out the converter. The only question now was whether maxing it out would mean the popping the emitter or failure of the converter itself. Hell, I've blown up $150 lights within seconds of ripping them open. For $24, I should worry?

Before the mod, I measured 380ma draw from both Nimh and Alkaline cells. With the mod, it was initially drawing 712ma with alkaline cells and 630ma with Nimh. The light is brighter. I can't tell you how much brighter until my roomate wakes and I can get a hold of his Maglite AA LED. He's usually up at the crack of noon.

I've run the light for 20 minutes in still air. It becomes warm, but nothing terrible. The emitter seems fine. The converter seems fine. Time will tell. I'll try for beam shots later today.

ELM

Your image is not showing up.

Took mine apart. Destroyed the keeper in the process (used floor press to push the LED module out.)

I measure the resistor to be .5 ohms. You reduced the resistance to .1 ohms. That set the output higher. Nice.

The chip looks like it is a Zetex ZXSC310

I want to connect a potentiometer to the resistor points, so I can vary the resistance, measure the power and brightness.

If the circuit can put out 712mA that is a good circuit. Perhaps it can be bumped up to 1000mA!

Might be worth it to take out the circuit, pump it up, put the LED and circuit into something else that can dissipate the heat better. The LED and the circuit have a value equal to the cost of the flashlight ($23).

Wayne

CM · Sep 4, 2006

ElektroLumens said:
Your image is not showing up.

I'm trying to do the same thing, remove the LED module, but it's not budging.

If I can get mine apart (without ruining it) I can measure the resistance.

I want to connect a potentiometer to the resistor points, so I can vary the resistance, measure the power and brightness.

If the circuit can put out 712mA that is a good circuit. Perhaps it can be bumped up to 1000mA! Might be worth it to take out the circuit, pump it up, put the LED and circuit into something else that can dissipate the heat better. The LED and the circuit have a value equal to the cost of the flashlight ($23).

Wayne

-The Zetex 310 based circuit is not capable of 1A from alkalines (or even NiMH).

-You will have a tough time measuring 50 milliohms also with standard ohm-guessers.

-For the cost of the light, you're better off buying a Badboy circuit for $13 rather than transplanting the circuit into a better host.

-The BOM cost for this converter is probably less than $3. Zetex plus transistor is less than a buck.

ElektroLumens · Sep 4, 2006

CM said:
-The Zetex 310 based circuit is not capable of 1A from alkalines (or even NiMH).

-You will have a tough time measuring 50 milliohms also with standard ohm-guessers.

-For the cost of the light, you're better off buying a Badboy circuit for $13 rather than transplanting the circuit into a better host.

-The BOM cost for this converter is probably less than $3. Zetex plus transistor is less than a buck.

A single Zetex 310 chip cannot do 1 amp, but stack 2 of them, yes, they can. I have one on my desk, made by Mike Jordan, which pumps out 1 amp from 2AA.

Badboy circuit: $13
Shipping: $?? probably at least $3 if he ships First Class
LED: $ Lets say $6 - $10

Total: $22 (on low side)

Or,

Go to Wally's World and buy a MiniMag LED for $23. (In oregon no taxes.) No wait for items to come.

LED is not soldered in or attached, thermal grease is all there is.

Cost is about the same.

A zetex 310 circuit is extremely efficient. The Micropuck also uses a Zetex chip, and is most efficient circuit I've tried.

I do not see a capacitor on the board, or a transistor, both called for in the 310 chip circuit diagram. It puzzles me that it can work with no input or output capacitor.

wj

vortechs · Sep 4, 2006

ElektroLumens said:
A zetex 310 circuit is extremely efficient. The Micropuck also uses a Zetex chip, and is most efficient circuit I've tried.

Hi Wayne,

I don't want to sidetrack this thread but I noticed your comment about the MicroPuck (2009A, 2009-HI). Where is a good spot to discuss the MicroPuck? Should I create a new thread?

ElektroLumens · Sep 4, 2006

vortechs said:
Hi Wayne,

I don't want to sidetrack this thread but I noticed your comment about the MicroPuck (2009A, 2009-HI). Where is a good spot to discuss the MicroPuck? Should I create a new thread?

I'm not sure where to discuss it? If it is in relation to a specific mod, I suppose it would go in this forum.

I'm not the moderator of this forum, so I have no say so as to whether it's okay or not.

Wayne

CM · Sep 4, 2006

ElektroLumens said:
A single Zetex 310 chip cannot do 1 amp, but stack 2 of them, yes, they can. I have one on my desk, made by Mike Jordan, which pumps out 1 amp from 2AA...

Yes, but we're not talking about stacking them. The Mag LED 2AA only has one on board. Apples and oranges.

ElektroLumens said:
Badboy circuit: $13
Shipping: $?? probably at least $3 if he ships First Class
LED: $ Lets say $6 - $10

Total: $22 (on low side)

Or,

Go to Wally's World and buy a MiniMag LED for $23. (In oregon no taxes.) No wait for items to come.

OK, you got me there. But I'd rather have a sandwich, better thermal path from LED die to flashlight body, particularly if you're going to boost the current to the LED.

ElektroLumens said:
A zetex 310 circuit is extremely efficient. The Micropuck also uses a Zetex chip, and is most efficient circuit I've tried.
wj

If the zetex is the most efficient circuit you've tried, take a look at the TI 6100x series contollers. Or the synchronous converters from Linear Tech. They will make you real happy

ElektroLumens said:
I do not see a capacitor on the board, or a transistor, both called for in the 310 chip circuit diagram. It puzzles me that it can work with no input or output capacitor.
wj

The circuit will run without a capacitor. Zetex calls it the "maximum battery life solution". See page 8 here on the 310 data sheet:

http://www.zetex.com/3.0/pdf/ZXSC310.pdf

You can put in a capacitor at the output and zetex calls that the "maximum brightness solution" Mag chose the maximum battery life solution. It's cheap, requires only the Zetex ZXSC310, an NPN transistor, an inductor, and sense resistor and LED (5 components, probably around 1-2 bucks in quantity without LED)

Mag deserves credit (and discredit) for the design. It's about the lowest cost LED driver solution and it works. I used it in my first mod about 6 years ago :nana: and that light is still running though my kids now own it.

vortechs · Sep 4, 2006

ElektroLumens said:
I'm not sure where to discuss it? If it is in relation to a specific mod, I suppose it would go in this forum.

I'm not the moderator of this forum, so I have no say so as to whether it's okay or not.

Wayne

I found a good thread to discuss the MicroPuck in general. Let's go here: https://www.candlepowerforums.com/posts/1579089#post1579089

NewBie · Sep 4, 2006

Okay, I'll have to admit I am amused now. MagLite finally found the Zetex circuit we were using back in 2002-2003?

MrAl has a great many posts discussing the circuit.

Oh, and before I forget, remember, with a boost circuit, in the boost mode, the current going in, is more than the current to the LED (thus the LED gets less current than what you measure on the cell end of things).

EvilLithiumMan · Sep 4, 2006

I got a hold of my roomate's 2AA - my light is brighter, but just barely. All the work was hardly work the effort, it seems. Worse still, my buddy's 2AA has got a great tint; mine is tinted blue by comparision.

As noted elsewhere, no "moon mode" - light flickers for about 15 seconds and then turns off. Looks like about 1.3V is the miniumum operating voltage.

If the emitter has only thermal compound on it, I will swap in a SX0H. The one's I have all seem to be great.

Thanks for the Zetex info, CM.

Wayne and others - don't worry about highjacking - anything is welcome here - where ever the thread goes, it goes.

Update - the Mini Maglite LED 2AA does indeed have a "moon mode" - just leave the light on after it has appeared to shut off. The cells continue to drain and shortly thereafter, "moon mode" is activated.

CM · Sep 4, 2006

NewBie said:
Okay, I'll have to admit I am amused now. MagLite finally found the Zetex circuit we were using back in 2002-2003?...

From Mag's Website:

"...If these products have been a long time coming, it's because the MAG-LED™ Technology has taken a long time to develop to the point where it is feasible to make an LED flashlight that meets Mag Instrument's high standards of quality, durability, style and function.

Now, after years of research & development, testing and refinement, Mag Instrument is committed to introducing, in early 2006, a line of MAG-LED™ flashlights..."

The R&D, testing and refinement of... a circuit right out of Zetex's datasheet. I'm more than amused

ViReN · Sep 5, 2006

CM said:
From Mag's Website:

"...If these products have been a long time coming, it's because the MAG-LED™ Technology has taken a long time to develop to the point where it is feasible to make an LED flashlight that meets Mag Instrument's high standards of quality, durability, style and function.

Now, after years of research & development, testing and refinement, Mag Instrument is committed to introducing, in early 2006, a line of MAG-LED™ flashlights..."

The R&D, testing and refinement of... a circuit right out of Zetex's datasheet. I'm more than amused

if I am not wrong, Arc AAA also uses Zetex 310... yet it is successful and efficient... why bash Mag?

NextLight · Sep 5, 2006

EvilLithiumMan said:
Update - the Mini Maglite LED 2AA does indeed have a "moon mode" - just leave the light on after it has appeared to shut off. The cells continue to drain and shortly thereafter, "moon mode" is activated.

The sun sets well before the moon rises.

It seems once the 310 stops oscillating, it latches "on" with the transistor draining the battery thru the coil. When the batteries drain further, the 310 unlatches, turning the transistor off. The coil then acts like a ~.5 ohm resistor; The LED is direct driven at ~ Vbat. Not much light, but it is better than nothing I suppose.

NextLight · Sep 5, 2006

ViReN said:
if I am not wrong, Arc AAA also uses Zetex 310... yet it is successful and efficient... why bash Mag?

With decades of high quality US manufacturing and as the creator of America's most widely regarded 2AA incandescent flashlight enters LED game late... Could it be we expected more from Mag than simply copying the cheapest possible 3 year old circuit driver design from a manufacturer's data sheet?

The LED-driver assembly is a marvel of mass production mechanical engineering. Too bad they didn't pay as much attention to the circuit's engineering, and that they mucked up the form factor of the platform(s).

The beam is marvelous, and the brightness is decent. We can make it better. I don't care much for the package the AA LEDs lights come in. I want a 1 cell Ledean cutdown, threaded for the old tailcap, and will stick in a 14500. (hint, hint)

Bag Mashing; Who,me?

I'm not waiting for Mag's
NextLight

CM · Sep 5, 2006

ViReN said:
if I am not wrong, Arc AAA also uses Zetex 310... yet it is successful and efficient... why bash Mag?

As I said, Mag should be given credit for coming up with such a low cost solution which fits their target market perfectly. As far as the quote, that is so pretentious as to be hilarous. We've been using this circuit for years and as Newbie stated, there's enormous material written about it on CPF years and years ago. My first mod (around 2001--I remember well since that's when I met Peter Gransee and I proudly showed him my home brew Luxeon LED light) used the Zetex. But times have changed and most everyone has moved on to more advanced circuits.

On to the Arc--The Arc AAA does not use Zetex. They use something that no one would think of using for that application. Arc is a lot more innovative and creative than that.

kevinm · Sep 7, 2006

EvilLithiumMan said:
Before the mod, I measured 380ma draw from both Nimh and Alkaline cells. With the mod, it was initially drawing 712ma with alkaline cells and 630ma with Nimh. The light is brighter. I can't tell you how much brighter until my roomate wakes and I can get a hold of his Maglite AA LED. He's usually up at the crack of noon.

I've run the light for 20 minutes in still air. It becomes warm, but nothing terrible. The emitter seems fine. The converter seems fine. Time will tell. I'll try for beam shots later today.

ELM

Hey ELM,

OKay, it's probably something you have thought of, but were you measuring the LED current draw or the board current draw? The jump from 380 to 700mA should be a 40% jump in light output with a Lux III. Also, any that I've run at 700mA have gotten really warm really fast. Is there something weird about the circuit?

Thanks,
Kevin

greenLED · Sep 7, 2006

ViReN said:
if I am not wrong, Arc AAA also uses Zetex 310...

The ArcAAA doesn't use the Zetex chip.

Final comment: on SMT resistors the "R" represents the decimal point.

chimo · Sep 7, 2006

Finally found this thread.

I can confirm as well that the ArcAAA/AA does not use a Zetex. It uses a much simpler chip. :sssh:

Looks like Mag spent most of their "design time" writing up a patent app instead of actually designing. Using the Zetex is a bit of a let-down. At least in their drop-ins, they made a "little" more effort.

Paul

Mini Maglite LED 2AA mod

Enlightened

Newly Enlightened

Flashlight Enthusiast

Flashlight Enthusiast

Flashlight Enthusiast

Flashlight Enthusiast

Enlightened

Flashlight Enthusiast

Flashlight Enthusiast

Enlightened

*Retired*

Enlightened

Flashlight Enthusiast

Flashlight Enthusiast

Enlightened

Enlightened

Flashlight Enthusiast

Enlightened

Flashaholic

Flashlight Enthusiast

Similar threads

Retired