How to get max amps out of one li-ion cell

Does direct drive mean connecting battery directly with led?

If I direct drive with an 18650

or stack 7135’s

I tried to test this myself but my DMM is broken. I get .6A when I direct drive. And my eyes cannot tell difference of brightness past 3A.

direct drive (DD) means that there is no regulation going on. what the battery gives is what the LED is going to get.

for max amps you wanna go DD, copper braid the springs, thick wires. the lower the resistance the higher your amps will be. this also means also your battery must be able to give the high amout of amps. for example: i got some old laptop batteries, they only give about 1.5A. while i have ncr18650b’s, they can give ± 4A i think. but then you have to make sure you have a good thermal path so you dont burn your led. also when you go DD you will not have any modes or any regulation. a XML on a ALU board will not give alot more lumens when going over 3A-3.5A, this is because the ALU board cant take away the heat from the led fast enough. if you want more lumens for the amps you put in it a copper board is the way to go.

i hope i shed some light :wink:

The term “direct drive” is used in two ways on flashlight forums:

  1. In the past it was more common to see lights built for “direct drive” with nothing but batteries, wires, a switch, and an LED. People may still be doing it, but it’s not common. There are no modes, there is no protection for the LED, there is no protection for the batteries, and brightness is not constant as the batteries drain.
  2. Now we use a special form of driver to add brightness modes and (sometimes) undervoltage protection for the batteries. There is still no protection for the LED from overvoltage or overcurrent, and the brightness is still not constant or regulated as the batteries run down. This type of driver uses an FET, which is a high-current transistor. Examples are the East-92 drivers and more recently, the BLF17DD.

It’s normal not to see that much difference as drive current increases. Search around this forum for more information on tweaking your DMM to get accurate current readings, but a good start is very short and heavy wires.

Let me add a little more info that is pertinent to the thread title:

  • To get max amps to the LED you don’t necessarily need to get the most amps out of the batteries. Buck drivers are capable of driving an LED very close to it’s limit - they take a high voltage source (such as two batteries in series) and step it down to feed a low voltage source with more amps.
  • To get the most amps out of a single cell you can use a boost driver such as the one discussed in this thread. There are limits to how these should be used, but what they do is take a lower voltage on the input at a high current draw and make that into a high voltage on the output with lower current. You can cause them to draw a lot of current from a single cell.

It’s not the aluminum that’s the problem, it’s the pcb substrate on cheap factory star’s that is the problem…

It takes some modding to remove the substrate on those type stars, or buying a noctigon or sinkpad that is purpose built to remove that substrate

djozz did a test and under 4-5A the aluminum with substrate removed did just as good as the more expensive noctigon or sinkpads

But for the most part yes, a stock star is much less efficient thermally than a purpose built direct thermal path

Good call

oh did not knew that. we learn every day :stuck_out_tongue:

Actually (unless I missed it) he did not test an aluminum board with the substrate removed. He tested Aluminum sinkpads (they have no substrate there and are HASL / pre-tinned), he tested direct-copper boards, and he tested a pair of copper MCPCB’s that originally had a substrate under the thermal pad - one intact and one with the substrate removed and replaced with solder.

Your point is still valid - he showed that the performance of aluminum in without a substrate under the LED’s thermal pad is good. Removing the dielectric under the thermal pad and flowing solder into the gap is more of a problem than doing the same with a copper board.

get Sony VTC4/VTC5, or NCR18650PF for a start :smiley:

In Direct Drive (literally wiring the emitter directly to the cell) you can get up around 7A to the emitter with the top cells. Like the Samsung 20R or Sony C4. The C5 and Efest 35A come in a bit under that and then everything else falls lower.

The BLF17DD FET driver comes close, running up around 6.6A with those top cells IF you have a copper star and the springs bypassed top and bottom.

BUT, Cree emitters are the limiting factor there. A Luminus emitter has a much lower forward voltage and allows considerably more current to be pulled from the cells. How much more? How about 16A from the top 18650’s to an SST-90. Even an AW IMR14500 can give over 7.5A. And when using the BLF17DD with an AW IMR18350 to run an Luminus SBT-70, how about 11.12A.

So it greatly depends on what emitter you’re using. An XP-G2 can do up in the 4.5A range but if you de-dome it you’ll be lucky to get 3A from the same top cells. Lot of variables.

And yes, the Luminus emitters pull that kind of current from a Single cell. Crazy stuff!

Good info Dale, thanks for posting. I’ve seen some of your posts on running the Luminus emitters (as well as posts from other people) but I hadn’t really connected all the dots you spelled out in this post!

I forgot to mention the other end, the switch. While it’s all well and good to want to drive a light harder we also have to consider that many stock switches might not handle over 4A. Certainly the springs need to be bypassed. I don’t recall ever seeing a switch fail internally, except for a McClicky. I fed a McClicky over 10A once and the internal spring that keeps the components separated for the off position failed, couldn’t turn the light off. But again it’s springs. By soldering in a new current path the springs don’t take the heat of current load and survive much better. :slight_smile:

I am playing with the Cypreus now from Sinner with it’s switch that looks like a Solarforce Reverse clicky, pushing 6.47A through it and so far so good. If it holds up much longer I may put the AW cell back in and see if it can stand the 11A delivery. Yeah, I know, glutton for punishment and all that.

Luminus emitters can take VERY high current loads and will work a cell to it’s limits. Might want to consider some built in limitations if going the Luminus route. Even I am thinking of pulling the driver back out and re-flashing the MCU to limit Turbo. Thinking about it. Doesn’t mean I will! :stuck_out_tongue:

Wow thanks for all the info. I will try remember this.
I will definitely use copper for anything I DD.

On the topic of switches… I only found 1.5A max rated switches 1.5A switch
I just searched up “McClicky” and that looks nice, but too big for my needs. Are there clicky switches that offer around 6A? ): or at lest 3?

I got a few $1 clicky switches, took them apart and replaced all the thin metal parts with copper sheets to lighten the load, but I don’t think that will hold 6A which I am using at the moment.

I once asked a guy who was good with electrical engineering about Amps and wire size. The guy says Amps=Wire size and voltage doesnt do anything to wiresize.
But, there is this switch that has 2 different ratings: switch
it says “6A/250V 10A/125V AC” Is this because of the lightning arc that it makes everytime it is switched on and off? I am not sure.

Apparently that arcing becomes a much more major issue with DC currents than with AC. So much so that it’s very difficult to find DC rated pushbutton switches that are rated to carry decent amperage. The full blown DC current can lead to welding of the contacts and a circuit that cannot be shut down.

From what I can gather, a very strong spring with quick shifting of the parts is essential, more so in separating the components for off than bringing them together in the on position. This is, again my take, because the arc wants to follow as the parts are pulled away from each other. So a large strong spring to quickly spread the contacts would be ideal, but we usually have small spaces for our switches and as a result, small bodied housings with very delicate springs. Don’t know if there’s a win in there for us anywhere, except going to the electronic switches with independent power supply for the electronics.

That too is being worked on, with a max rating in the 30A range. Won’t be cheap though.

I see…. I can kinda guess how spot wielders work after an incident happened to me. I have a 24V lithium battery pack and I accidentally shorted the two wires and they fused together. So what switches does the sst-90 use? must be one of those household wall switches lol

I use the Tofty 10A+ for the C8 that houses the SST-90 and my M8 with 12A going to an MT-G2. And the C8 with MT-G2 and the Crelant 7G5 pulling over 6A.

Very robust switch, made with copper plate and 3D printed housing.

That’s the first time I’ve heard of de-doming affecting current draw.The increased throw and reduced lumen output are

optical effects AFAIK.Any idea why,change in forward voltage?heat dissipation?Those two are inter related anyway.

Does the XM-L1/2 suffer the same phenomenon?Always ready to learn.

Very difficult to get high amps from a de-domed XP-G2. I had put one in the Sinner Cypreus 18350 EDC and just wasn’t getting the lumens or current draw I expected to see. It was running off the BLF17DD FET driver, so I thought it should be much higher than 2.8 or so Amps. The beam was nice, output appeared decent, but the lightbox showed only in the 500 lumen range. So I pulled it and put a domed XPG2 S2 2B and the numbers went where they should be, 4.37A for 955 lumens at start. Then I sliced the dome off. Not the gasoline de-dome but simply cut the dome off with the bond wires still embedded in silicone. I’ll be danged if the numbers didn’t drop right back down to 2.91A and 527 lumens! Was trying to tighten it back up without losing the tint. So I pulled the XP-G2 altogether and put a de-domed XM-L2 U2 1A in it. None of the changes were easy, as I had 2 Noctigons and the emitter all reflowed to the copper pill. Made it a PITA to do changes. The XM-L2, de-domed, was pulling 5.48A for 1228 lumens in the small light. Couldn’t leave well enough alone though, pulled that and put the SBT-70 in it. lol

I just finished up a build the other day with 3 dedomed XP-G2’s in series from 3s cells threw a 17dd Zener mod running 5A per emitter.

The magic in that statement CK is “Zener mod”. You’re giving more voltage than the high Vf of the de-domed emitters.

You checked the amperage AT the emitter? Never seen one run that high to my recollection, in any light. So you’re pulling 15A at the tail? What switch are you using?

So 2 out of 2 XPG-2’s dropped current when de-domed,but with extreme currents :bigsmile: The XM-L2 didn’t budge.

Something strange is going on,I’ll have to check out the Sinner Cypreus EDC thread-cool light BTW-time for some bedtime

reading(gone 10PM here)and food and :beer: