Need a 500 ma/ 600ma single cell driver

I cannot find a suitable single cell driver, it has been years searching.

Background: I want to build a second 21700 light that I will actually use. 500 ma, on a 21700 (5000 ma capacity) will fit 95% of my workdays. 600 mA will fit about 50% of my workdays and *1 amp fits about 5%*of my workdays.

  1. Naturally, icing on the cake is more flexibility in lower modes to allow non work use, and a burst mode of up to 2 amps for 1 minute, mostly for entertainment. Nothing so harmful as over 10 watts- I only use reasonable high efficient leds like the xml u4, xpl v5 and v6.

I built my own 2xamc 7135, controlled by a 555 circuit and a dimmer. But everything was too large to be practical. * I love dimmers.* :THUMBS-UP: Then, I built a dd, using a .33 ohm 5w resistors plus a 25 ohm 3 watt rheostat; it too was a bit too big for my tastes.
A suitable driver board with these settings, or board with infinite dimming would be nice.

mnt-electronics has zero use for me, since I want a suitable run time that matches a workday. I I built a driver I would choose the following modes: 2.5 hour, 5 hours, 10 hours, 24 hours, 40 hours, 100 hours, extreme firefly….10 hour mode is most important by a mile, aka .1C

I don’t understand the modes of
Ramping sound good. But the tech told me this driver wouldn’t work. Either the page makes no sense, or he doesn’t know what he is talking about. 540 ma, %18, should work. But not according to the seller…. The other thing I don’t understand is how a FET prevents a light from going over the rated 10W.

Apparently, no one else uses their lights for work in a serious fashion:2000-10000 hours annually per crew?

Do you want a light that only runs at 600mA-800mA or a driver?

If so, you are in luck actually.

There is this light which support 21700s with an additional tube:

And you can get this 22mm triple channel driver, only put 1*7135 +1*7135 for maximum regulation, and FET if you really need it.

You can get it built from Lexel:

He merely mentioned an MCU. So I Google MCU.

This looks promising, but difficult to figure out and get set up with all the tools:
Zilog Document Download So, I recon, based off past experience of 85% efficiency, that in real life, 425 ma is 10 hr, 530 ma is 8 hr. 700ma is 6 hours… 350 mais 12 hours… 1000ma is 4h 15 minutes. 100 ma is 42 hours, 30 ma and 10 ma desirable
Desired ma levels in decreasing order of importance 425, 530, 700, 240,350,100,30,10 ma.
This is the closest board to my dream, but no mention of getting 8 modes of various percentages in the pdf. Steep learning curve and equipment to buy, I bet. Complete with 2 days worth of snags?

I just got tired and sick of it all, so built,this last weekend, a 2x amc7135 + 1 amc7135 (all 350s), 2 switches. Unfortunately, it runs only 6 hours at 700 ma before dimming. So, 85% of expectect runtime, 4250 mah are available for use, using the amc7135. …I want another build, the hassle of clicking between modes is not going to cut it for my guys. I have only had true love for the dimmer switch, but could live with an 8 mode light with these steps, on a 21700.

Got an email this morning. Looks like the LiS battery is a few years off still. Something about their physical size. I know that they swell and contract, which is no big deal for a headlamp.

I use the H17f just like you mention. If it will fit in your host. It has battery side 7135s so it doesn’t clear some retaining rings.

What host do you want to use?

Back to Mountain Electronics, Richard can program this driver: with up to 6 modes that you specify. You specify them as a percentage of maximum current and can specify the number of 7135s you want to set maximum current.

For example, the driver comes with 8x7135s 1x350ma + 7x380ms) as default giving approx 3 amps so you could ask for 2 of these to be removed to give 2.25A max current for your 1 minute burst, and then specify one mode as 22% to give around 500mA, and another mode at 27% to give around 600mA, and other modes to sort.
Of course you could have 3x7135s removed for 1.87A max and specifiy 27% and 32% modes for 500mA and 600mA respectively.
And it is called a moonlight special because the moonlight mode is what i believe you would call extreme firefly, which is another of your requirements met.

There are tolerances to consider so modes would only be roughly right not exactly right, but should be close enough. If you factor slightly lower than required you should get your runtimes with no visual difference in output.

Alternatively you may be up for programming your own modes, or even flashing a ROM with ramping software. I haven’t done it yet (have the electronics so soon!) so can’t help you with that but these threads would be the starting point:

I think Richard thought I was from Mars… I could see desoldering all but 2 of the amc7135 and requesting proper pwm percents. But he was not too helpful or interested.

Chad, I don’t want the fet. It cost me money when guys kill the batteries accidentally in high modes. In fact they kill early to get out of work. Can it be removed or disabled and 6 or 8 pwm modes like I described?

Host? I don’t need no sticking host. My bezel :
Though, I might make my own from a copper sheet, next build, then I can solder it a copper fitting, thence on the battery tub. I directly epoxy water proof. The open design cools way better, especially with an aluminum reflector, which acts like part of the heat sink, when using thermal epoxy to physically attract it to the star. Life is so much better being host dependancy free.

You can program it to never USE the fet.

Edit, You can from 1-7 modes, and a double tap mode, 8 total.
You can less if you want, I have group set for just 2 modes and turbo on the double tap.


So, 7 or 8 modes? But efficiency not so good over 380, due to led droop.

Check if something from Zebralight works for you. Their buck-boost drivers are pretty much the most efficient drivers used in flashlights but they are not available for DIY.

Edit: If you want to DIY I think the LD-A4 from led4power may be a good option since it uses constant current rather than PWM, so the efficiency would be better. However a good buck driver may still be a bit more efficient .

Maximum current may be set from 2A to 3A and the mode groups are:

  • 100%
  • 9% - 100%
  • 16% - 100%
  • 2,5% - 20% - 100%
  • 2% - 10% - 30% - 100%
  • 1,5% - 7% - 18% - 35% - 100%

For example with the 16% mode and 3A max you would get 480mA. With 2.5A and the 20% mode you would get 500mA.

Since 10h>8h>6h etc are the target. 31mm smo xpl, or 26 mm xplhi, to lay down appropriate lux which 20 mm cannot achieve, even with xplhi , unless pumping an amp.

Desired ma levels in decreasing order of importance 425, 530, 700, 240,350,100,30,10 ma. (aka., I calculate, pwm of: 60,75,100,50,34,14,4%1.5% - - - off 2x amc7135, 350 ma each, but different if 380 ma each)

A good led in 2019,should reach close to 202 lpw (373 lumens at 700 ma, but higher efficiency below 560ma). I don’t want to slip down (loosing 25-30% of the light) especially to 130 or 140 lpw range, because the the driver is slamming higher current than necessary then pwm it back down to a level actually useful. Just to allow a wow mode that might be used 1 minute a day, but could be lived without.

I don’t think mntelectrons is an option. He has no interest, no matter the price.

I can see adding a second amc7135 (with own slider switch) in parallel with the Dr. Jones, since it has a good selection of drive levels under 380ma. The Dr. Jones driver is poor in the area of 381 to 600 ma.

I would rather figure out how to just buy a driver, desolder a few amc7135 chips to maintain the brightest light during regular usage, and program my pwm settings , mainly the 10 hour runtime, then the 8 hour. Then, working the way downward in terms of current, up in terms of runtime.

Lexel says: _“What you need is a biscotti driver with custom mode set flashed and 1 AMC removed×371…

Ask the programmers in forum if they can give you a hex with your modes.“_

But, do I have the tools, the cables to do my own flash? And who are the hex programmers here who I need to talk to?

That’s strange, he was very helpful when i requested the exact same thing, a couple of moonlight specials with 6 modes specified by me. Has he stated he’s not interested or could it just be miscommunication going on?

I think you have misunderstood the specs of the driver, there are already 8x7135s on the board. From the page you link to:

“Low modes (up to 13%) are driven by a single AMC7135 IC (380mA), which allows very low moonlight/firefly modes and makes those modes also quite efficient (the LED is more efficient at lower currents). Medium and high modes up to 3.0A are driven by additional 7*AMC7135.”

You also mention a slider switch, and have mentioned dimmer switches before. These are useful if your controlling factor is the amount of light being emitted as you can control it precisely with the slider but you won’t know how much power is being used and therefore have no accurate way of knowing how much runtime you have which seems to be your controlling factor here. Unless i’m missing something the only way to specifically control runtime is through preset modes where you control how much power is being used at any given time. And don’t forget the modes would control how much current is being delivered to the LED which is different to the amount of current being drawn from the battery due to losses in the system, you would need to factor this in for accurate run times.

As mentioned above these threads are a starting point for programming your own modes:

The first thread explains the process and what programmer/cables etc you need, the second is an example of someone doing it themselves.

The H17f is only available through Mtnelectronics? The reason I have no interest in its 530ma is that I didn’t buy a 202 lpw led just to slam it with 3 amps, for a 100 lpw result, or such. Though 530 ma might be the ideal setting for the 21700. Yes the pwm will mitigate the heat, but I bet just adding a second driver to the Dr Jones, would offer a noticeable brighter 530ma.

I spent last hour reading the treads on flashing. But a bunch of the software is missing, like the firmware .C repo. Amtel 5. Doesn’t anyone have a Google drive folder with all the needed sw for windows 7 that works?

Note that the 202 lpw from your leds is at 250mA and that’s not counting driver or optics losses, assuming you are talking about Texas Ace’s results for the XP-L or XP-L2. Efficiency at 500mA or 700mA will be lower.

Even with a PWM-less linear driver like led4power’s you will only get 75-80% average driver efficiency at 500mA as the Vf is 2.8-2.9V, so the 0.8-0.9V to 3.7V is wasted as heat. I believe a buck driver capable of using a single cell like the LD-29 would get good efficiency if you could set the modes you like and lower the LVP voltage accordingly, but I’m not sure if there is a driver that makes that possible.

Looks like the trend nowadays is using a boost driver to run a quad-die LED like the XHP35 or the XHP50.2 so you can get high efficiency as each die runs at low current even at relatively high output. Using that approach Zebralight gets for example 14 hours at 196lm according to a review There is some technical discussion of boost drivers at [Buck and Boost Drivers, Testing, Modding, and Discussion]

Still, I think the cheapest way to solve your runtime problems would be training your subordinates to swap cells when the battery gets low rather than just leaving the worksite (if you are losing money because of short flashlight runtime as you described the problem is human, not technical).

I sometimes use 9 guys. A simple math is that each swap cost $10, which is $90/day. Even one person, this is about $3000 a year. Not to mention, customers often pay by hour and get antzy with swaps. Forget the swap, ain’t going to to happen. Way easier to just build lights that run 10 hours or more.

I measured my light, xpl v6 at 700ma, came out 373 lumens in relational output to the tk30 and Hp11. All my other lights come in at 270 to 310 lumens. 240 for warm.

Richard at Mtnelectronics said explicitly that the Dr. Jones driver would not work, and he was not interested in programming a useful drive level. Obviously, the Dr driver should work. However it is a question of how inefficient it is at 530 ma.

All my lights since 2014 are 2s18650 on a dial buck Puck, using the avid heat sink, thermal on a computer style heat sink. 4000 candela 9h15 minutes. I get 600 lumens, as I often wear a comparable wristlight 300+300 lumens. I supplement with a 110,000 lumen metal halide. Unfortunately, the corded light is limited and can’t see texture, and half of what is needed.

The buck Puck design beats any light sold. I could have easily built a 3 cell light (1 18650 cell in the front), but didn’t want the headache of an extra cell per man to charge… But I finally got around to different design. 21700 in front, with 2 ports: optional rear pack and vdd pwm (motor dimmer). As an edc, 350/700/1000 ma as an edc (31mm smo, which I love). Built in 1amp charger and cellphone out charger. Option to attach a rear 21700 w/motor dimmer. 9 volt snap is what this prototype uses. The switch can double as the port. I haven’t got the dimmer in the mail, but don’t have much faith that these boards will be suitable and not waste energy as heat. Building my own 555 pwm cost me $30. I only have one… Going from buckpuck to amc7135 (2x21700) , I lose about 1.5 hours to 2 hours , compared to a 3s18650 cell setup. Fortunately, the 2s21700 2 amc7135 gets 12 hours. Edc portion is 85 gram, plus strap and 21700. Probably will put the charger in the detachable rear, next time to shed 20 grams or so…. I will wait more time, to test what I have first….

If the external pwm dimmer gets too hot, I might try the Dr Jones driver. I would program my own, if somebody simply had one directory with all the tested software. This route looks like a lot of incompatibility dead ends.

It may be my imagination, but I like the lg 21700s better than the Samsung. Gut feeling so far.

I directly soldered wires to the amc7135s, soldered and thermal epoxy to copper and mounted to aluminum. Not sure if I could access the chip vdd for pwm, if I bought on a board. The good thing is that I don’t cook my amc7135s like my convoys.

Hello degarb.

For highest efficiency, you want an emitter like the XP-L2 in the highest bin possible and a buck driver set to 500mA max, or even a triple setup for maximum efficiency.

Since 700mA is a low current, you will not have a problem with a driver like this:

It’s an a constant current buck driver, and you can easily set it to 500mA operation with a simple resistor change.

Actually, just ask Kaidomain if they could set the max current to be 500mA. The lower modes will also be lower because of this.

Finally, get a P60 drop in like this, and ask Kaidomain to change the Nichia 219Cs 4000k to XP-L2 5000k, and the driver to 700mA operation if possible:

With this, you will get the absolute highest efficiency possible, and with a 3500mAh battery, or even 2, you will have some crazy runtimes even on max.

Kaidomain has some Xp-l2 HD on 20 mm. Plus, they have rare smo reflectors in a sufficient, optimal size. I like the xplhi in a 26 mm, or xpl in a 36mm, for sufficient lux per watt.

Kaidomain offerings is really exciting. My curiosity if I could work with the xpl2 beam, is killing me.