Hi Friends,
Ive been making lights for friends and family and one question that I find myself asking over and over is, “How many amps can I pull before running into problems.” Most of my own lights are pulling high amps, but I am not concerned because I’m paying close attention and I know that if I blow an emitter, I can fix it. I have gifted lights to my closest family, but my budget requires me to charge the rest, so building flashlights for everyone else is not that simple.
So to simplify my question, How hot can I let the emitter get(by measuring at the junction on the mcpcb) and how long can the flashlight be turned on in the highest mode before it reaches this temp?
I know there are a lot of factors, but just let me know what you think. I want to make bright lights, but I don’t what them all coming back to me three months later with fried emitters. You know what I mean? |(
Can you run them on high until the battery drains without any damage to the led or driver. I know light output will be significantly decreased, but that is not damaging.
If building lights for non-flashaholics, the max drive levels should have consideration for worst case scenarios like leaving light on highest mode unattended.
Even if you caution them about managing heat, someone else can borrow their light without knowing or a child might use it.
I would assume something like 3A and below with proper heatsink and copper boards maybe even 2A max.
My P60 lights can run on high for a very long time, around 3.05A - 3.42A regulated. They do get quite warm but never too hot to touch; I put additional copper shims inside.
My C8 runs about 4.42A right now (slightly higher with a newer battery), the wires are thin, no braided springs. It’ll get warm but not hot. Can run on high if held in hand, gets hotter if tailstanding.
As said before, if you give them away I would be more careful. High power flashlights can (theoretically) start a house fire if someone is careless.
Yes… This has been my battle. It’s hard to know where to draw the line and each light is different.
I just did some measurements on an aleto 26650 light I just built. It pulls around 4.6amps at the tail with Blf a6 driver on turbo. I have a 26mm by 5mm copper spacer behind a copper sinkpad. The led is a xml2 u3 1a.
Thermal test results
At 3mins temp = 65C
5mins temp = 72C
10mins temp = 90C
Over 10 not much increase
At 10 mins you can tough the light, but if you held it it would burn. So it’s not holdable after 10mins on turbo but it’s not hot enough to damage the light. I would think that on a smaller light like the a6 or convoy s3/2+ the temps would be much higher, but idk?
It’s surprising to me that your high amp lights do not get hot to the touch when on high. Do you think the heat is getting out fast enough to keep your emitters cool? I wonder if they are getting realy hot in there? Do you have a way of testing this?
As long as the star doesn’t get so hot that the solder attaching the leadwires melts, I don’t mind heat.
I use DTP copper boards, but beyond that heat is fine with me. I figure I’ll upgrade the LEDs long before they burn out from excessive heat.
I like tiny hot-rodded lights so heat is a big issue. Usually getting the heat out from the LEDs isn’t hard. The real trick is getting the heat out of the flashlight body without having a light that is too hot to hold.
I use my lights outside and it is freezing atm. Therefore they’ll never get that hot in the first place. Even if I build and measure them I’d do it in the garage (outside) and it is winter…
Inside the house I have almost no need for flashlights.
Ya… cold weather is nice to our modded lights I live in MI all the way up near Lake Superior. So I understand the cold:) I, unfortunately, am bound inside, but I definately don’t mind my limitations around this time of year!
The answer to your first question is very well documented in Cree’s LM-80 tests, which run all their various emitters at different temperatures and brightness levels, and then calculated the LED lifetime (down to 70% brightness).
What I find very interesting from reading the LM-80 tests, is that the lifetime (or damage) of the LED comes almost completely from heat, not from how bright (or how much current) the LED is run at. So, all you have to do is control heat, and you can make your LED last a long time.
Basically, most of their LEDs can handle temperatures up to 100C, but you might be shortening their lifetime down to a few thousand hours at that temperature. If possible, keep the temperature below 85C, and the LED should last the 50,000 hours it is spec’d at. If you can keep the temperature down closer to 55C, then the LED will last essentially forever without any fading over time.
You’re probably going to run it hotter than 55C if you run it anywhere above a few hundred lumens, but you should be able to keep it under 85C quite easily with proper heatsinking, even if you run it at full brightness. Overdriving it (usually above 3 amps), will probably result in too much heat, and you might be pushing that 100C limit or even higher.
The answer to your last question about how hot the LED will get, of course depends entirely on the heatsink and how well that sink conducts heat away to the environment. And how well the environment sucks heat away (a light outside in winter probably won’t heat up much, nor will it heat up in water).
out of interest and in line with the topic: How and where do you measure this? If we are really referring to the temperatures on the LED board/junction/solder joints (or directly on the LED die) then I am not sure how I would be able to measure this. I have one of those cheap ebay IR thermometers (mine goes up tp 380°C) and if I point it right at the LED die when it is switched on the thermometer shows way over 100°C.
I’m not sure on this but I believe the temperature measured at the face of the die is a bit higher than the junction temperature?
I am measuringing with a thermocouple applied to the mcpcb pressed up against the die. I believe this way of measuring will give fairly acurate results, but I don’t realy know
Hi Guys, i made some compare XML2 T6 on Noctigon vs typical Alu PCB with dielectric layer under led. I made it for 4A and 5A.
It was very visible that DTP is much more effective, MUCH MORE .
At 3A aluminum PCB keeps led temp about 100*C, at DTP it’s about 50-60*C.
Aluminium PCB led died about 200*C at 6A. Noctigon led still works :).
Here is thermal view:
I ran a couple lights to the breaking point, getting 13+ amps from a copper sinner with copper heatsink it ran for a little over 5 minutes on turbo before the traces in the noctigon DTP pcb burnt out. I moved the leads to a different set of traces on the other side if the board and everything was working fine. I had my multimeter/temp probe hooked up to the head and the head got to 200*F within 3 minutes. I had to swap cells because the 2000mAh cell wasn't heating up any more, and even after that the host never went much more than 200*.
**edit, I should mention that the tailcap was around 150*F, but the cell itself only got up to 130*. I ended up having to change out the tail switch as well, it started acting funny and being really touchy.
I tested a Convoy S2+ With an aluminum spacer running around 11.5A in the same fashion. The host stayed cooler longer, but the solder melted on the pcb after 4 minutes and the + lead popped off. The Convoy only reached 185*F after 3.5 minutes and wouldn't get any hotter.
Who's to say what the internal temperatures got up to. From my own testing, 115* is perfectly fine, 125 is getting uncomfortable, 130 you have to move it around in your grip to hold it. 135 you can't hold it at all, turn it off!!
It takes resetting a 60 second turbo timer to get up to 135 in a solid copper host.
I am not willing to put a copper spacer in any lights I build to sell to someone even if they claim to be responsible and I know they are, accidents happen. For myself, I want them lumens!!! :) It is possible but harder to get the aluminum hosts scalding hot where you would get burnt just barely touching it.
Given that solder melts ~200°C, and cotton doesn’t burn until you get past ~250°C (IIRC), and even Diethyl Ether doesn’t ignite until you reach 160°C, my guess would be that you won’t be starting any fires.
The solder holding the parts together will sacrifice itself for your Safety. Unless you’re in a Diethyl Ether environment, in which case, well… it’s been nice knowing you!
All this is hot enough to burn fingers, so the rest is subjective.
OTOH, even Yoopers get cold, so maybe turning a flashlight (or two) on & dropping it in your pocket(s) might be useful…
As to your ‘X’, if you pass more current, you’ll need more batteries. Most modern LEDs have exceeded the best-available batteries’ ability to feed them (aka “DD”), so the question becomes “How long can you go between battery changes, and still get ”enough” light OTF?”
But then you’re staring into the Great Void: “How much light is ”enough”, assuming such a state exists?”
In the end, I’d have to go with the solder suggestion. As long as the solder holds, rock it! Blisters heal, and it helps to have lots of practice charging and managing batteries.
Someone on BLF burned themselves pretty good with a light on in their pocket, i forget who so i don’t have a link.
Keep in mind a turned on light in a blanket will become a fire hazard, as enthusiasts we understand the risks and don’t do anything stupid, but a Convoy S2 for example at 2.8A will get hand burning hot, tail standing it will get hotter because there is no human to absorb heat (which is surprisingly effective) and with no physical contact you won’t know how hot is it.In a blanket or closed insulated place (drawer, glove compartment, coat pocket) bye bye house.
All LED lights should be coming with temperature regulating circuitry that dims or stops output when it overheats, but that is expensive and not widely available. I believe there was a custom driver a while back on BLF that has temperature dimming that stated you could put a light in a blanket without risk. I forget who was selling it and if its still available.
The next best thing is to have a time limited turbo or low max setting, say under 1A
Dr Jones has just finished a driver that has full thermal regulation. It both decreases and increases current according to the internal temp of the light. So after the light cools off some it adds a bit more current to raise the output back to the desired mode output setting. This is quite amazing when you think of it. The waiting list for this driver is VERY long right now or so I hear.
Good to know, though i can see why its a long waiting list, i want it now
The one i was thinking of was someone else, a newer member who had asked advice and was in Europe maybe?
I have gifted lights to my closest family, but my budget requires me to charge the rest, so building flashlights for everyone else is not that simple.
I live in MI all the way up near Lake Superior. So I understand the cold:) I, unfortunately, am bound inside, but I definately don’t mind my limitations around this time of year!
