Driver to convert 120V AC light to DC power?

I’ve picked up several lights similar to this for cheap recently. They vary from 10W to 60W outputs. I’d like to convert one or two of them to run on an 18650 battery setup.

I’m hoping this won’t be too involved. Since LED’s run on DC power, the 120VAC light that I currently have probably uses some sort of AC/DC transformer, and running off of a DC power source would theoretically be simpler, right?

I’m assuming the LEDs are 12V, but I have not opened up the lights to confirm.

I would like the light to have more than one mode (“low” and “high” would be fine), so I think I need a proper driver, not just a DC-DC conversion circuit. But I’m not sure what kind of driver to use.

Needless to say, I’m trying to keep this mod cheap.

Thoughts?

If this is the wrong place for this post, moderators please feel free to move it.

The 10w lights with a COB LED will run off a 12V source, but actually only need about 9V to turn on, since the “chips” in the COB are connected 3S3P. Higher wattage COB lights are usually ~36V rated, since the “chips” are arranged 10 to a row. You can find cheap drivers on eBay. I don’t know about modes, though. You could use a regular FET flashlight driver with either zener or LDO “high voltage” mod. You can either build one yourself, or buy one from Richard or somebody. I’m sure it will work fine with the 10W rated ~12V COB emitters. You’d have to check the components to see if they will work with voltages as high as ~36V for the other lights. Maybe talk to Texas_Ace about his Texas_Avenger drivers.

Edit: Here’s another idea I just thought about. To keep it super cheap, you could use a three position switch and two high wattage resistors in place of a driver. Switch one way to get one resistor to feed the low mode. Switch the other way to get the other resistor to feed a higher mode. Here is a simple calculator to help you choose the right resistor(s) to use.

Thanks DavidEF, I was just looking up some spec sheets on random Aliexpress 10W, 20W, 30W, 50W, 100W COB type LEDs and noticed exactly what you said - the 10W ones were 12V, the higher power ones were 30-36V.

I figured I could used something like this for $2.21, but it doesn’t give me the ability to do a lower level output:

I could use something like this, and adjust the teeny-tiny little pot to raise/lower the current, but that’s a lot more fiddly than I wanted.

Edit: (in response to your edit): Your idea with the resistors would require that I have a power source that has higher voltage than the LED requires, right? So that would be feasible to put 4x18650 in series to power a 12V 10W light, but not so feasible for a 20W+ LED that needs 30V.

Yeah just remove the crappy power supply from inside the light and wire the LED directly to your batteries.
If your batteries are over 12v then you probably want to use a buck voltage converter like a cheap one from amazon, or a boost converter if your batteries are far below 12v total.

If your batteries are less than led rated voltage: Amazon.com : boost voltage converter
If you batteries are more than led rated voltage: Amazon.com : buck voltage converter
I would recommend at least a 100W one since they are pretty cheap and made in china.
It will probably be a 12v LED but you should check with a multimeter first.

Enderman, with those type of convertors how do you know what type of amperage the LED will draw? Will it pretty much always draw whatever its (maximum?) rated amperage is?

sac02,

Yes, if you want to go the simple route, you will need a source voltage high enough to turn on the LEDs. Otherwise, you will need a boost driver, as Enderman suggested.

I don’t know how many amps your worklight LED draws, but if it’s 10W and 12V you can assume it will be about 1A.
I would always get extra though, like something that can do 4A or something so that it remains cool under load.
As long as you provide the LED the correct voltage it will draw the maximum amps that it can.

Ok, thanks for confirming what I thought was the case.

Regarding the “simple” solution from DavidEF in post #2, (a switch and a couple current limiting resistors used to achieve different outputs): in that scenario is the resistor basically just a heatsink consuming the excess current/power? How can I estimate/calculate how much it would affect runtimes?

Well, the resistor is calculated for the amount of current draw you need/want for the LED. Then, it basically takes up the extra voltage between the Vf of the emitter and the voltage available at the source - at the rated current load. So, to find out how much power it is consuming (wasting) you subtract Vf from Vin and multiply by the current. For instance, if you have a 4S battery source, fully charged, that gives you ~16.8V at the start. If you decide that you want to feed your light 1.0A and your emitter needs 12V to do that, you’d subtract 16.8 - 12 = 4.8 then multiply that by 1.0A for a total of 4.8W of wasted power. As the source voltage goes down, the current will too, meaning that you’ll be losing light output from the time you turn it on. But, since the power loss will be shrinking, you could say that it is getting more efficient as it drains. If you made it 3S, it would be more efficient from the start, but runtime would be a bit lower.

It’s just a cheap way to ensure that you don’t blow up your emitter with too much power, which is what will happen if you give a ~12V emitter the whole ~16.8V without a driver or anything at all. Direct drive works when the Vin and Vf are close, but if the Vin is too much higher, it will kill the LED. So, the resistor keeps that from happening.

Hmmm, makes me wonder if I could drive that 10W 12V LED direct drive off of 3x18650 (12.6V fully charged, 11.1V at nominal voltage). The lower voltage limit when an LED shuts off is a characteristic of the LED itself, right? I wonder how low of a Vin the 12V LED would accept.

I’d rather have a DC driver that does 12V out with modes though. I guess I figured since there are a billion 3.7V DC drivers with modes, that a 12V driver would be no big deal.

Aren’t the XHP70 LEDs 12V? What do those lights use? edit: I see they also come in 6V, I guess they don’t use the 12V version in a flashlight if there’s a 6V version available.

There are plenty of 12V drivers made. The XHP-35 is 12V only, no 6V, so drivers have been made to accommodate that especially. There are even some boost drivers that will feed 12V from a single Li-Ion cell. You can have all the modes you want (within reason). And, of course, there are buck drivers and then linear drivers as well. Last, but not least, there are FET drivers. Those are basically direct drive without even so much as a resistor to “regulate” the output, but they use PWM to make lower modes. So, the drivers are available, and some very functional firmwares are also available. The only reason I even mentioned the resistor and switch get-up was because you said you were aiming for cheap.

Oh, do you mind linking me to some? I tried searching for as many different keyword variations of 12V, DC, driver, flashlight, boost, etc. that I could think of but got nothing. I guess my Google-fu is not very good.

Edit: I just googled “xhp35 driver” and have found like 3 options. I just wasn’t searching the right terms I guess.

First open the light and double check with a multimeter what voltage is going to the LED when it’s on.
If it’s 12v then you can definitely run 3x 18650 with no driver, as long as you have the LED on a heatsink it will be fine.
However if you measure like 10v or something a lot less then it would be risky to wire the LED directly to 3x18650.

Just for completeness: I run a 25W Floodlight off a 4S LiPo with this: http://www.ebay.de/itm/272266593764

For 20W LEDs with the typical 30-36V, there are ready boost drivers: https://www.aliexpress.com/item/12V-24V-20W-LED-Driver-for-20W-30-36V-700mA-high-Power-20w-led-chip-transformer/32629863487.html
Of course you can use this driver for 30, 50 or 100W LEDs, to increase lifetime and efficiency plus reducing the heating effect a bit.

Or there are ~12V 20W LEDs: https://www.aliexpress.com/item/20W-Cool-White-15000K-2000LM-12V-15V-1500mA-SMD-LED-Light-Parts-For-12V-battery/2044346526.html

Why not? Pretty much all those cheap boost or buck converters have a voltage control pot. That one sure appears to. Many have voltage and current limiters (that one doesn't). Either one alone can control output though. Current controls it a bit better but either work. Of course with those little screwdriver pots you won't be adjusting it while you walk. It's a a set it up and then use it kind of thing. But you can replace the pot with one with a knob. You better have the kind with both controls in that case so you don't blow the led. Then you set the voltage to the upper limit you'd want, and use the current limit for operational control.

I'm not sure you shouldn't be using a buck instead of a boost or what voltage ranges you need. Listen to others here about what voltage that led requires. In general I think it's better to use a buck. If you start with twice the voltage you get to pull half the current off the batteries and that means you can use low drain, high capacity batteries, or still use high drain batteries but with even less heat loss. Compared to using a similar voltage to the LED with a buck or linear or direct driver, (that's not really true actually, current per cell is the same for the same number of cells) but.. it also does mean you don't get voltage sag as the batteries deplete or even when they're full under high current (and that can be a big deal). You can get really great performance this way with long lasting controlled ouput. In that setup you can even get full-battery direct-drive level output, even when the battery is half empty.

In the OP, I made some assumptions, like the assumption that the LED would be a discrete component and driven by some separate driver/transformer. There’s a saying about what happens to U and ME when one ASSume something…

Here’s the guts of the light:

The AC/DC transformer is integrated into the chip, so I probably won’t be driving it so easily with my battery-pack-and-flashlight-driver idea.

I’ve got another floodlight arriving in the mail tomorrow that has this type of chip instead:

I’ll open it up and see if it looks like a better candidate for a simple DC conversion.

@L4M4 - yup, those were some of the type of cheap drivers I was starting to look at. Since I really wanted to try and have modes, I had found this:

That’s pretty much what I had in mind when I started. $8 isn’t too bad. But now that I’ve seen the inside of the stock light, I don’t think it will work. Maybe better to just buy one of these as linked above and use the LD-51 driver.

@Flintrock - I guess I didn’t realize that voltage control could effective drive an LED to various levels. The adjustable voltage-only regulators are cheap (like $2) but I didn’t think they’d be effective. The adjustable current and voltage regulators are more like $6 from what found so far (same type that L4M4 linked above).

And at $6 price tag, I think I’d rather pony up the $8 for the LD-51 driver that can switch modes with the click of a switch instead of turning a trim pot.

The thought that’s brewing in my head currently:

  1. Get a floodlight housing in this dual LED format to use as a host
  2. Install two LEDs
  3. Install two LD-51 drivers
  4. Run it off of 2S2P pack of Panny 18650B.

Sure they will. It's like direct driving an led off a 4.2V battery except you can use more reasonable voltages even. No it's not current controlled, but at 4.2V there's a limit to how much current an LED will draw anyway, and at 3.9V that limit is lower.

In a resistor current and voltage are just proportional. In a diode it's a curve, not a line, but it's still a relation. The reason current is better is because that curve moves up and down a little as the LED gets hot, so just controlling voltage can still create a little output fluctuation with temperature. There's a concept of thermal runaway, because they tend to draw more not less current as they get hotter... which makes them more hotter etc, but in practice this doesn't seem to go nuclear as far as I can tell, not with the xm-l2's I've used anyway. Maybe I read there are some lesser used leds where it could, but I suspect still it's an issue with people who try to push the limits in the first place.

The thing with only one control though is you'll need a voltmeter every time you adjust it because you don't have the second control as a backstop. You don't want to set it too high. You'll blow the LED.