Direct drive 18650
- Thread starter James3
- Start date
Help Support Candle Power Flashlight Forum
Fireclaw18
Flashlight Enthusiast
- Joined
- Mar 16, 2011
- Messages
- 2,412
Direct drive should work fine with either emitter. An XM-L2 should pull around 6 amps on a fresh cell and output around 1700 lumens at the emitter.
Wow, does anyone really do that? I mean that is relying on the limitations of the battery to limit current without any control whatsoever. We're talking something like 18 watts initially.
timsmile
Newly Enlightened
I think the XML2 Mini Turbohead made by Oveready is using this set up exactly. XML2 direct driven by protected 18650.
Direct drive is one way to maximize output and make the brightness directly proportional to the amount of juice the cell can supply, typically, the light dims as the cell drains below its capacity to supply full power.
The regulated lights can maintain a constant brightness, but, that constant brightness will be dimmer than the maximum brightness it would have had if a direct drive.
So, direct drive is best for pedal to the metal full output for relatively short bursts (Depending on the cells, for a few seconds, to maybe 20 minutes or so...).
If you have 2 of the same light, one regulated, and one DD, and you turn them on at the same time, the DD will be brighter for a while, maybe 10 - 20 minutes if you have good cells, and, then, start to dim. As it dims, it will dim down to what your regulated light's output is....and, eventually, keep getting dimmer and dimmer as it sucks the juice box dry.
If you look at HJK's cell test charts, you'll see about what that might look like proportionally. Each cell has "sweet spots" etc, and will be able to maintain different amp levels as its voltage drops. Some can maintain higher amps than others can, at a given voltage...so, if going DD, you do want to look hard at the cells for it, to pick a good match performance-wise.
The regulated lights can maintain a constant brightness, but, that constant brightness will be dimmer than the maximum brightness it would have had if a direct drive.
So, direct drive is best for pedal to the metal full output for relatively short bursts (Depending on the cells, for a few seconds, to maybe 20 minutes or so...).
If you have 2 of the same light, one regulated, and one DD, and you turn them on at the same time, the DD will be brighter for a while, maybe 10 - 20 minutes if you have good cells, and, then, start to dim. As it dims, it will dim down to what your regulated light's output is....and, eventually, keep getting dimmer and dimmer as it sucks the juice box dry.
If you look at HJK's cell test charts, you'll see about what that might look like proportionally. Each cell has "sweet spots" etc, and will be able to maintain different amp levels as its voltage drops. Some can maintain higher amps than others can, at a given voltage...so, if going DD, you do want to look hard at the cells for it, to pick a good match performance-wise.
James3
Enlightened
- Joined
- Mar 6, 2013
- Messages
- 260
Thanks guys,
theres going to be no driver board in the light at all, just a spring attatched to a PCB board with '+' and '-' wires straight to the emitter, if that makes sense? Then some increased heat sinking and 22AWG wiring.
None regulation isn't essential for me as I won't be using the light for long periods, mostly short 30 second or so bursts
What would the runtime be with the protected NCR 18650?
theres going to be no driver board in the light at all, just a spring attatched to a PCB board with '+' and '-' wires straight to the emitter, if that makes sense? Then some increased heat sinking and 22AWG wiring.
None regulation isn't essential for me as I won't be using the light for long periods, mostly short 30 second or so bursts
What would the runtime be with the protected NCR 18650?
Run time to what output?
Most flashlight makers said run time stopped when the light was down to 10% of full output, so a 100 L light was down to 1 L, etc.
What percent out put is acceptable to consider as part of your "run time"?
Another point - you said its a protected cell. The protection may trip on a DD set-up.
You might want an unprotected cell, or, to run it with an IMR 18650 instead.
The IMR will also allow the light to be brighter than the panny you mentioned.
(The XML2 is a better choice than the XML U3 btw)
Last edited:
James3
Enlightened
- Joined
- Mar 6, 2013
- Messages
- 260
Run time to what output?
Most flashlight makers said run time stopped when the light was down to 10% of full output, so a 100 L light was down to 1 L, etc.
What percent out put is acceptable to consider as part of your "run time"?
oh Im talking about the high end runtime 1000lm+
With what IMR?
The Sony, if its the one I'm thinking of (Unprotected/IMR), would be the only one that might work well in this application, and heat would probably be a limiting factor before dimming, say ~ 5 - 10 minutes as a range to get down to 1k L, but its MAXIMUM will just scream out the photons.
Fireclaw18
Flashlight Enthusiast
- Joined
- Mar 16, 2011
- Messages
- 2,412
Use 20 gauge wire to the emitter and solder braid the battery compartment spring(s). Run it on Samsung 20r INR 18650 ... which according to RMM's tests is the lowest resistance cell currently available on a full charge. It will do 6 amps easily... but that will decline rapidly as voltage drops. If you want more capacity, switch to a Samsung 25r... 5% more resistance at full charge, but after a few minutes it will pull ahead of the 20r.
Instead of going direct drive, a MUCH better option is to use Comfychair's FET mod. This is a simple mod that can be done on any 7135 based driver. Remove all of the 7135 chips and replace them with a single FET. Result is near direct performance on max... but unlike direct drive it also retains the ability to go into lower mods, which makes it much more practical. Using the FET mod, Comfychair and RMM tested this type of driver with Samsung 20r cells to pull over 6 amps with a single XML2, 7.4 amps with triple XPG2, and 11 amps with an MTG2 .... impressive for a relatively small 17mm driver. The FET doesn't even get hot.
The FET mod works with any 9.4khz PWM firmware that will fit on the driver's microprocessor, so you can get quite a fancy UI. I prefer DrJones lumodrv firmware. Gives me shortcuts to min, max, and 25%, 2-way ramping with 17 brightness steps, voltage indicator, hidden strobe modes, hidden momentary mode, and low voltage stepdown.
Note that if you run a small light with the FET mod or direct drive on a low resistance cell it may get hot very quickly. If this is a problem, you may wish to switch to a higher resistace Li-Co protected cell. This will lower the output and the heat production, while increasing battery life.
Last edited:
yoyoman
Flashlight Enthusiast
I tried to find the post where Oveready explained the proper cells for their mini turbo head dd/xml2. I couldn't find the one I was looking for - it used plumbing to explain the concept - but some of the other posts in that thread do talk about what cells to use and what cells not to use. They advised not to use IMR cells because it would fry the emitter. I'm not an expert and suggest looking that thread for some ideas.
Fireclaw18
Flashlight Enthusiast
- Joined
- Mar 16, 2011
- Messages
- 2,412
A small 18650 light with proper heatsinking shouldn't fry the emitter even when direct driven if run for brief periods of time. Don't tailstand the light or leave it unattended. Carry it in your hand, so your hand can help wick off extra heat. When the head becomes too hot to comfortably touch, turn the light off (or if FET mod, turn to a lower setting).
I don't have the Oveready so can't comment on its heatsinking beyond what I see on the website. I see that the LED is mounted on a "solid copper base". But it is unclear if that means that there's direct copper under the central heatsink pad on the LED. For best results and highest output with direct drive or FET mod, a light should use a direct to copper star. Unlike a conventional star, there should be no dielectric layer between the star's center pad and the center pad of the LED. Examples include Noctigon, Sinkpad or the house brand from Illumination Supply.
Truly insane enthusiasts may even punch a hole in an existing star in the shape of the central heatsink pad and then have a giant copper pill with a pillar that goes into this hole. But results have shown that doing so doesn't give much improvement over just having the LED mounted on a high-quality direct copper star.
Last edited:
Could you explain how to do this exactly I wanted to add 4 more 380ma chips to my intl-outdoor XM-L2 T4 5B1 neutral led p60 but this actually sounds easier and I would like to try something similar to direct driving it. Would I have to replace every chip with a FET and which one's would I buy could I get them from a radioshack or special order?
