Simple resistor swap?

Flashlight Modding and DIY Parts
1

Hello,

I’ve obtained a flashlight which uses a XHP70 and it’s not enthusiast oriented at all.

It’s powering the LED from a 7.4V battery pack of 2S2P 18650 but it’s doing it just with resistors. High mode is just a fat ceramic 10WR47J power resistor and appears the LED gets 1.37A at 6.13V after that which is substantially less than the 30W the hardware brand uses in the advertising.

Shouldn’t be too much risk in swapping in a 0.33 10W resistor, seeing how that goes and maybe look at 0.22 and 0.1 if wattage is ok?

Torch is plastic but there’s an 8 inch aluminium reflector and a heatsink adding up to about 350g of metal for heatsinking/cooling.

Was thinking about a driver but there doesn’t appear to be any downside in just reducing the existing resistor?

2

Yes, that should be fine, if you think the heat sinking is adequate. Also your resistor might get burning hot as you increase the current.

Can you post a link or picture of the light?

3

Here are some pics



It is bulky and there is a lot of enclosed air, mostly the hollow reflector but also 4x 18650 are in the back and in the same space you could fit 20.

Room to mod I guess :slight_smile:

I have detached the big aluminium reflector and powered it with a CC CV bench power supply up to 7V 6A which is daylight inside and it doesn’t seem to warm up but back in the torch body it will be insulated by the glass lens and the plastic shell so maybe needs ventilation.

4

Alternative 10W power resistors arrived :slight_smile:

So original high mode with this XHP70 LED is

0.47ohm = 1.37A, LED 6.06V (8.3W) Resistor 1.39V (1.9W)

Set power supply to 8.4V to imitate fully charged 2S lithium cells (checked afterwards and actually the battery pack of the torch charges to 8.23V)

0.33ohm = 2.56A, LED 6.11V (15.6W) Resistor 1.8V (4.6W)
0.22ohm = 4.10A, LED 6.20V (25.42W) Resistor 1.75V (7.18W)
0.1 ohm = 5.80A, LED 6.35V (36.83W) Resistor 1.42V (8.2W) ignore all struck out numbers, I later find out they’re inaccurate

*Measurements were not particularly well done but they appear to be in the right ballpark.

I guess next step is picking one, soldering it in, closing up the torch and seeing how the heat builds up over time.

**Actually next step is doing a lot more testing, reassembling LED, resistor and power supply resulted in dramatically higher current flowing, willing to bet my scruffy test leads are inadequate and I’ve put it back together with lower resistance from connections and wire.

5

The internal resistance of the battery pack can play a big role in the resulting current. So I would do the measurement with the battery instead of the power supply to make sure it’s not very different.

6

Not too sure about using 8.23V off the battery yet. Feel safer twiddling the potentiometer from low so I can see the current rising on the power supply display.

Going back to the 0.33 ohm resistor and scuffing the crocodile test leads a little has almost doubled the current to 4.88A at the same 8.4V and the resistor is burning off 7.8W

0.1 is getting 5A on only 7V
0.22 is getting 5A on 8V

So my initial measurements were consistently off by a significant amount probably due to temporary connections being too poor for the power involved and it looks like 0.22 and 0.1 ohm resistors let through too much current if I don’t want more than 30W running off the battery.

Now wondering how much current the original 0.47 resistor is really passing since its measurement was in line with the others at the time.

Quick check with 0.33 and 0.22 in series for 0.55 (close enough) and power supply on 8.2V (as the battery reads) is 3.10A with a LED voltage drop of 6.05 for 19W… inflate it a bit because of the extra resistance and it’s over 20W.

Amazing, increasing the wattage of what I already have just by measuring it again.

Oh well :zipper_mouth_face:

Guess the goal of trying to reach 30W is much less ambitious if the torch is already doing over 20W out of an advertised 30W. Perhaps with soldered connections it increases again.

Probably still go ahead and replace the 0.47 resistor with a 0.33 but after more checking.

7

The light has thermal mass but does not shed heat. Once the reflector and “heatsink” get too warm, the LED itself will start to cook, so the light might be “safe” for a moderate amount of time, but not for very long. That, or it’s scaled-back so much that just the warm air inside would “leak” heat to the outside.

There’s a reason Al flashlights get warm if not hot to the touch. The body is shedding heat to the outside world.

8

Yes, the large 2mm thick aluminium reflector is inside two more 2mm aluminium cups (they are visible in the picture) plus the heatsink on the bottom is a lot of metal and surface area which moves heat from the LED, spreads it to the trapped air and then the heat gets out through the plastic shell.

I do not say it loses heat in the best possible way but it seems to be so large with so much air inside that it does not heat up easily. The plastic does not touch the sides of the reflector cups or the heatsink, the metal has maybe 2cm+ of air all around it plus the big hollow space in the battery area.

Maybe the pictures do not look very big but the torch is 26cm long and at the lens it is 22cm wide and tall.

I have left it running for hours on high indoors to see if it would heat up the shell. It got slightly warm at the closest point above the LED heatsink and nothing more.

Because I have been looking to increase the output I have considered making ventilation holes and maybe adding a fan inside the case so the heat can be more easily transferred to air which is not trapped. I will try putting a thermometer on the heatsink to get a real number for how hot it gets inside the case when left on for hours.

Ok, pushed kitchen thermometer into heatsink fins from the back and taped up the gap, will now cook for 1h on high:

About 30 mins in: https://i.ibb.co/FbvGyLD/tmp.png
After 1 hour cooking: https://i.ibb.co/rdBQN97/tmp.png
After 1 hour temp without touching heatsink: https://i.ibb.co/0h0BXjJ/tmp.png
Room Temp at end: https://i.ibb.co/hs55Shr/tmp.png

9

Part of the original Arc LS design (and that was a 1 watt light) was that you were supposed to grip the aluminum body tightly in your hand, so that your normal blood flow would carry heat away from the light. That was described as a natural liquid cooling system. I do that with my D4v2 nowadays sometimes, and it does let me run the high modes longer without the light getting too hot.

10

I reckon this will end up with me spending maybe 200 on making an overpowered light I don’t need and never will.

Maybe tools don’t count but still, so far, a bench power supply and now a DC clamp (kaiweets ht206d) because I had a look and couldn’t believe how much you could get them for these days without having to pay Fluke a stack of cash.

So after replacing my dodgy inline ammeter measuring, I can finally say these are decent numbers measured while on the 7.4V lithium battery pack:

2.57A @ 6.00V for 15W on high

1.31A @ 5.80V for 7.6W on low

Soldering in replacements for the high mode 0.47 resistor, 0.33 gave about 18W which wasn’t enough of a gain to care about.

0.22 was 3.49A @ 6.39V for 21W

So one left and…

0.1 was 4.78A @ 6.38V for 30.5W

30W was the goal so we’re gonna stay with 0.1 ohm, at least run temperature checks and make sure it doesn’t misbehave when left on high for hours.

The resistor has a drop of 0.47V with 4.78A, making it a 2W heater, far less than expected.

Numbers were quite a bit different running on the 7.4V battery compared to the power supply. As EasyB said, there’s the internal resistance of the battery not accounted for. The resistors were burning off a LOT more on the power supply.

In the end it was a trivial mod. Things might change if china ever sends the enormous battery pack I ordered off aliexpress :innocent:

*Well what I was thinking of was changing to a driver instead of a resistor but if my numbers are correct, burning 2W with a resistor to control 30W through the LED is already fine?

11

In my mind, I’m willing to spend more money when I’m making a mod or project because it’s a learning experience. I don’t just get an object, I build knowledge and get more satisfaction.

12

Don’t mind me.

Just charging up the 140Wh battery pack that arrived from china :smiley:

This will fit into the same space that the current 4x 18650 pack is sitting in. There will still be space around the new pack.

Next is to look at replacing the resistors with a CC buck converter with 1-2 power levels.

Can’t be sure of the date format but could be they sorted that battery pack the day after I ordered it and it’s been crawling to me for almost 3 weeks.

Oh yes and this pack will add about 500g, making the torch a chunky 2kg. It’s not really that bad when you know how large the torch is. All the mass is basically heatsinks/giant reflector and batteries.

13

Risk it for the biscuit and go no resistor.

14

Because it’s a very large torch I reckon I can get away with two plug and play modules.

All the regular flashlight drivers are out of the question because they expect a flashlight tube for a heatsink and aren’t expecting 4S lithium cells.

So we are going with a big board with heatsinks and running at less than maximum power for good temperatures because thermal throttling is for small torches not a 2kg torch.

High ~5A

This board gets advertised at 9A, 10A, 12A but I hear it’s actually 8A: DC CC 9A 280W Step Down Buck Converter 7-40V To 1.2-35V XL4016 Module (tested under 12V) - YouTube

No matter, it should be very safe at 5A.

Low ~1A

And a little board for ~6W mode that will have about… 24h runtime on the new battery :smiley:

15

The electronics arrived :slight_smile:

Smaller than I imagined. Obviously I had the measurements but staring at product images inflated how big the measurements were in my mind.

It’s a 4x4 battery pack. The two boards together are a little thicker than adding another row of 4 batteries.

Think I might just get some board and fix both down on it for ease of securing inside the shell.