Arc LS + AA twisty + 14500 li-ion = bright

Arc LS + AA twisty + 14500 li-ion = bright

What a timeless design, the ARC LS first run is still in the running. Add a 14500 AA sized li-ion to 1x AA twisty to ARC LS first run. you are overdriving to a estimated 700+ milliamps based upon how fast light heats up.

Thanks to pbarrette who did an analysis of ARC LS first run's circuit board and determinded that it would support up to 2amps. He ran 6volts with no damage so far to first run's board + lux.

Before I started, was fairly sure the board would not make the magic smoke. If the luxeon toasted, it could be replaced.

Measured 14500 at 3.85v, inserted into AA pak, ran for 1 sec, still OK. tried 2 sec, then 10 seconds, then 1 minute, then 5 minutes.... getting warm.

First run lives being overdriven with 14500.... so far anyways. the jury is still out on longivity.

My MM+ R2H one watt lives happily driven at aprox. 1amp, so I figure it worth a try to see if ARC LS first run makes it.

So far so good..... brightness is at least BB500 R2H in firefly. If you do this and toast your LS don't blame me or ARC. Mine is doing fine so far with no problems.




ARC LS w/14500----------------------Firefly BB500 R2H

Have you changed the sense resistor size or is it a stock board?
Chris

cy:

Where do you get the 14500 AA sized li-ion battery and how much?

Thanks.

Everything on the ARC LS first run is stock, including the 1 watt LD lux. Please note the circuit analysis/6v test was done on the first run board, not the LS1-LSH series.

Cheers to foursevens for doing a group buy and offering them for $6ea. I'm still trying to decide which li-ion charger to go with.

Using a bare li-ion cell, you need to be aware of discharge and charging issues.

If you installed a two stage switch in the twisty. Then you would have a very useful combo, almost same size as ARC4.

Hi all,

Also, please note that my testing involved an odd 6v (6.6v fresh) 500mAh military use battery which was not designed for high current draw applications. This means that the amount of current the LS drew was somewhat limited by the battery itself.

I measured the current through the Lux in my LS first-run on this battery at 835mA. The 2A rating I mentioned is the max current rating of the boost IC used in the LS board I have, but I haven't actually tried to push 2A through this circuit yet.

Basically, with the ArcLS first-run circuit, if your battery exceeds the Vf of the Lux, the boost circuitry is virtually nullified and the circuit acts as a direct-drive circuit with a schottky diode in series.

If you plan on going this route, be aware that the Lux current draw is the issue at hand and that because of this, Lux die heat is the major concern. I, personally, think that the Lux will die before the circuit does.

If the Lux does start to burn out on you, then you should turn off the light as soon as possible. Without the Lux there as a load, the cap will take the full brunt of the load and the board could quickly fry at that point.

A good idea may be to replace the 1W Lux in the LS with a Lux3 to give a better safety margin.

Just remember that the main issue here is heatsinking. Get the heat to flow away from the Lux as quickly and efficiently as possible, and you should be good for occasional, intermittant use. If you want to run this setup all the time, you may be better off replacing the board with a step-down convertor like the DownBoy or similar.

pb

Pb, I agree with you that heatsinking is the primary issue here. The provision for heatsinking on the Rev 1 ARC is pretty poor. I disagree that the emitter failing open would harm the board in any way. Those stepup switcher ICs on the board [two running in parallel] are set for a fixed output of 5V. With the emitter failed open, the output caps charge to 5V and stay there.
Due to the poor heatsinking of the Rev 1, this near direct drive with a li-ion is a recipe for very short emitter life.

PB,DougS, thanks much for your valuable feedback /ubbthreads/images/graemlins/bowdown.gif

I've only done up to 15min runtimes with no problems. The first run LS is heating up, but not unmanagable hot like firefly gets running MM+ R2H.

Everything is potted fairly solid on first runs giving a decent thermal path to the body. There is a fair amount of mass/heatsink in first run head + AA twisty.

I have 14500's forsale on my groupbuy that is still open.

GROUPBUY

$6 shipped each.

Hi Doug,

The board that I have in my LS first-run only has 1 IC on it, so I don't think it's a parallel circuit setup.

I'm more worried about the current increase due to the lost load of the lux than overvoltage. I assume that the IC has a thermal limiter, but there's always the possibility that it could kick in too late. The cap will want to draw current as quickly as possible to get to a charged state and it may do that too fast for the IC and the schottky to keep up.

Either way it seems like a bad situation for the circuit to be in, so it's probably best not to take your chances.

I agree with Cy in the first-run's ability to dissipate heat. It looks like the Lux's heatsink lug is epoxied to a large hunk of aluminum in mine. That is then forced snug against a shelf in the head and epoxied into place with what looks like thermal AS.

I ran my test with a cpu heatsink+fan strapped to the head, but when playing with the light previously, the head and body got warm, but not horribly so. And the PCB heatsink seemed to be at approximately the same temp as the head, so that interface seemed to have a good thermal path.

The path to worry about the most is the one from the Lux to the PCB heatsink. That's the one that is likely to have the highest thermal resistance. It's also the one that is the most critical for dissipating the heat from the lux before it fries.

pb

[ QUOTE ]
pbarrette said:
Hi Doug,

The board that I have in my LS first-run only has 1 IC on it, so I don't think it's a parallel circuit setup.

I'm more worried about the current increase due to the lost load of the lux than overvoltage. I assume that the IC has a thermal limiter, but there's always the possibility that it could kick in too late. The cap will want to draw current as quickly as possible to get to a charged state and it may do that too fast for the IC and the schottky to keep up.

Either way it seems like a bad situation for the circuit to be in, so it's probably best not to take your chances.

I agree with Cy in the first-run's ability to dissipate heat. It looks like the Lux's heatsink lug is epoxied to a large hunk of aluminum in mine. That is then forced snug against a shelf in the head and epoxied into place with what looks like thermal AS.

I ran my test with a cpu heatsink+fan strapped to the head, but when playing with the light previously, the head and body got warm, but not horribly so. And the PCB heatsink seemed to be at approximately the same temp as the head, so that interface seemed to have a good thermal path.

The path to worry about the most is the one from the Lux to the PCB heatsink. That's the one that is likely to have the highest thermal resistance. It's also the one that is the most critical for dissipating the heat from the lux before it fries.

pb

[/ QUOTE ]
pb: the board that you have described is not a first run board. If you found it in a light marked "first run" it is a Rev 1 Hybrid. The Hybrid used the first run body and the later rev circuit. You are correct that this circuit does not like to be open circuited. The issue *is* overvoltage, not overcurrent. BTW, it is a myth that this circuit has overtemperature protection, what Peter says not withstanding. The IC used, does quit operating when it reaches extreme temperatures but this is not an engineered, documented, IC thermal protection feature.
The Lux is attached to the board with double sided thermal tape as I recall, not epoxy on the hybrid boards. You are right that this is a fairly high thermal resistance joint.

The 1618 used in the rev2 LS (Hybrid, LS2, LSL, LSH) has a built in thermal shutdown that cuts out at about 150F. This is not well documented, but the feature is there and I have tested it repeatidly. The IC gets hot because it is near the LED and inductor. If the LED gets too hot, it will shut off the output. This saves the LED but you still need to shut the power off because the circuit will continue to draw some power. I still have not heard of any other flashlight that has a thermal cutout for the LED. Of course, some mods based on the 1618 can have this effect, depends on how close the LED is to the chip.

With the Arc4, we further developed the idea of protecting the LED from thermal damage.

It is rare for the LS to go into thermal shutdown. It takes an unusual combination of voltage and run time.

Peter

When I got 14500, initially I planned to cut the 2AA TSP half to house it for LSH-P. But I find it a bad idea.

My LSH runs about 0.3-0.4A from the 123 battery. But when connected to 14500 Li-Ion, it takes about 1.2A and it is way above acceptable. (4V * 1.2A = 4.8W)

There is nothing mentioned in 1618 datasheet about the behavior when the Vin is higher than Vout. But regardless 1618 stops oscillating or not, the high voltage (3.7-4V) runs to the LED directly through the inductor and schottky diode. Even 1618 shutdown due to the thermal protection, the current still runs to the LED and it could fry it.

One thing to note is that those "faulty?" boards sold on eBay recently appear to draw a lot less current than the regular ones. For example, two of those boards in my hand for testing hardly light up at 3V, and draws only about 400mA from fresh charged Li-Ion. No wonder they look bright yet they don't die off with 14500.

I wish to find out (perhaps from Peter) if these boards on eBay are actually designed to be this way or this is the reason they are rejected.

-- dj

Hmm, DJ, I measured the emitter current recently(TWAK@561mA) on one of the ebay pcb's from Peter by first calculating the current draw of that modded pcb .050/ .089 = .561A where r1=0.15, r2=0.22 total resistance in parellel is .089

I then measured the voltage of the old CR123 battery I have which is 2.3 volts to determine what resistor value to use this case, 4 ohms 2.3v/.561mA = 4 ohms using a 1 watt 4 ohm resistor , then place it in series with my fluke76 and measured the current draw at the LED terminals which is .560 -.559mA which is what it should be I think /ubbthreads/images/graemlins/thinking.gif I tested this particular board because I was curious /ubbthreads/images/graemlins/smile.gif it was a random pick of the litter and wanted to verify my mods Thanks to Doug_S for helping me with this test procedure /ubbthreads/images/graemlins/bowdown.gif I have more of these arcs and concluded they are working but now would verify all of them just to be on the safe side /ubbthreads/images/graemlins/smile.gif

also follow up question, what does it mean when Arc goes dim momentarily then flashes 3 times then lights up as if nothing happened at all?

How about an update? /ubbthreads/images/graemlins/smile.gif

I have a Rev1 LS, and 1AA pack, I love this combo, and was wondering how this all turned out.

Is your LS still running? Have you tested this much more?

I hope to get some of these cells soon, and was thinking of trying this out.

I know its not really recommended, this sounds like you guys are pushing the LS HARD.
But, it seems that it might still be ok, just shortening the LED life.

-John

I've been running the Arc LS first run (flat gold board) w/14500 in 1x AA twisty.

This combo smokes! output is equal to ARC4X. Not as much throw, but certainly as much lumens out the front end.

I've run this combo for the last two weeks with no problems.
Been running another ARC LS first run with R123, no problems.

I need to start logging batt draws and for which light.

[ QUOTE ]
cy said:
This combo smokes!

[/ QUOTE ]

I think that is what John is worried about /ubbthreads/images/graemlins/wink.gif
The biggest potential problem I see with this setup is that the cell voltage may be pulled below the recommended minimum for a Li-ion before appreciable dimming is likely to be noticed. I suppose that if one is diligent about keeping the cell topped up it might work out OK for some usage patterns.

thanks for the update cy.

heh, yea, smoking is what I was worried about. /ubbthreads/images/graemlins/smile.gif

Im not real worried about over discharge, for the most part, I just want to do this cuz I like the look and feel of my LS with 1AA pack, and to have a rechargable version, then saw this, and figured, ooooo wow, I can even have it brighter. /ubbthreads/images/graemlins/smile.gif

I will be getting some of these batts later for some Arc AA mods, so, I will try it out then.

-John

BTW - I've got some new AA li-ions 850mah! (also some 700mahs)

web page

my bad... smoking is a bad term to use.

I've had no problems with overdischarging so far. At the first sign of dimming I imediately yank the spent cell and measure voltage. I've been getting 2.54v at the lowest. no problems charging up to full capacity from there.