XHP70 build questions.

Is there a thread of this light? Seems really interesting to me.

An integrated shelf is best, but there is no reason why a pill can’t be made to transfer heat 99% as well. Just that after doing what you need to do to make a pill work like a shelf it’s no longer a pill.

Just take any given pill and use a very high quality thermal epoxy on the threads and you will have the thermal transfer you need. Now of course you won’t have a removable pill, but you don’t get that with the shelf either.

After that it all comes down to the size of the area that you are transferring heat to. The Fins on a light are really only doing one thing, they are increasing the surface area of the light so more air can reach the area that needs to be cooled.

The second option is to build a heat sink into a light. This only works up until the heat sink gets hot (sometimes seconds) and then the heat sink actually holds heat in the light so even after you have turned the light down it’s keeping the emitter heated. A combination of both can be used effectively but rarely does this work if it’s not combined with some way to also transfer the heat to the outside of the light or use some non-passive way of cooling the heatsink, like a fan or other cooling agent.

Well anyway, don’t like your pill? Buy some good thermal epoxy and fix it. :wink:

The thermal conductivity of Arctic Silver Thermal Epoxy is 8W/mK.

The thermal conductivity of aluminium is 205W/mK.

I realise there is contact in the threads, but it is mostly the tips of the threads, and more on one edge, when tightened.

Edit: yeah, fins are essential for dissipation.

As I said the integrated pill is best, but thermal epoxy is going to fill in the air gaps that are acting as in insulator. So it’s not a comparison of aluminum to epoxy, it’s a comparison of AIR to Thermal Epoxy or whatever you fill that gap with. Removing the air that is acting as an insulator with just about anything is better than leaving it there and it’s often going to transfer heat faster than the surface area of the light can remove it. Once you have reached that point it doesn’t matter if it’s 8 or 8,000w/mK.

I can take a light that is barely transfering heat and turn it into a light that gets too hot to hold in a flash. Now if the outside of the light is getting hot then thermal transfer is taking place. So it doesn’t matter what you do, or how you do it. If the outside of the light is getting really hot, really fast, you have done all you can do.

Now you can limit yourself to lights with only the shelf if you want. But why should you when there is a proven and effective way of making lights with pills transfer heat faster to the outside of the light, than the outside of the light can transfer that heat to the air? The surface area of the light is the limiting factor. So let’s say that with a shelf it transfers heat to the point where the outside of the light is scorching hot in 10 seconds and with the epoxy it takes 10.1 seconds. How much have you gained with the shelf vs the pill with epoxy? According to you it’s 205/8 but trust me, that is not how it works.

But if you don’t believe me then don’t bother with it and stick with the limited selection of lights that have shelves. What do I know, I have only built a thousand lights or so. :wink:

Ps, I’m not trying to be rude, but this is one of my pet peeves.

I agree with some of that, but I still think the difference between pill+thermal Epoxy and unibody is higher than you suggest.
I know it’s not 205/8, and I tried to express that at the end of my previous post, but, I still think it’s noticeable.
Also, once I’ve got an MTG2 at 7A on a noctigon, with Arctic Silver paste to the unibody shelf, I am very much chasing marginal gains. To me, what you referred to as the 10s vs 10.1s is still a noticeable difference, though I still think it’s more like 15-20s. Maybe it’s overkill, but I stand by unibody shelves. I’ve had a pilled UF C8 and a Convoy L4 both running XM-L2’s at 2.8A, and the L4 got hot twice as fast as the C8.
I do have and use some pilled lights, they just don’t see much power (a couple of cheap supermarket bought lights, a Ti3, that sort of thing)
I know you have so much more experience than me, but discussion and learning to improve our knowledge and our light building is the whole point of BLF! :stuck_out_tongue:

That’s fine and since you see that difference as worth it, then it is worth it to you, one second or a dozen doesn’t matter really. Only your preference does. For me I just don’t see the point is chasing marginal gains if it means that I’m limiting myself when it comes to lights to buy. If it’s the same basic light at the same price or close to it then I will go for the shelf every time. Especially in larger lights. The smaller the surface area the less difference it makes. The thing is that I have found some really great reflector designs on lights without shelves and I wouldn’t turn my nose up at them because they have a pill. Some of the really cheap lousy designed lights happen to have some of the better reflectors. (Don’t even get me started on plastic vs aluminum reflectors)

Think of it like a drag racer with three inch tires. It doesn’t matter if it’s 200hp or 1000hp the limiting factor is the surface of the tires, not the horse power. But a BIG light is like having really Fat tires on. You can now use that 1000 horse power. With a small light you are going to heat that outside fast with even basic thermal transfer.

Also a trick that I use is to increase the size of the Pill itself. Let’s say the pill is only 1/8th in thick, well even if it was perfect like a shelf it’s still only able to transfer heat through that 1/8 inch spot. Now add a copper washer or something similar to the bottom of that pill and make it 3/8 of an inch thick and all of that up tight against the side. Again this is something to do on any light, shelf or pill. I don’t normally bother with smaller lights. I do this with Sky Ray Kings all the time. I have turned the pill into a shelf and then added a half inch under that. They get smoking hot FAST. I also have a SRK that has a 1 inch copper pill. No other light in my collection gets hot faster than that one, but it stays hot too and is a lot heavier. But even though it’s noticeably hotter faster it’s still only a few seconds faster and I have built an identical light with the only difference being the pills and tested them.

But I do get it, you notice that the light is transferring heat faster and that is a concern for sure. Morehotterfasterbetter. Plus it’s a lot less work if you simply buy one in the first place with a good shelf.

Yeah, I think for me, I’ll do whatever I can to keep the emitter happy, even if that means some occasional scorched fingers.
I use my L4 as an example, because the dedomed MTG2 was originally coated in ‘LED Seal’. One day, I was doing current tests on turbo for my DIY driver. Suddenly there’s a black spot in the centre of the beam, and the light starts smoking. The LED seal on the emitter had burned and charred. It then took me about a week of supervising it in lighter fluid to remove the charred silicone. Turns out I had used my generic thermal paste by accident instead of my small supply of Arctic Silver. Once the MTG2 was resurrected, I put it back in the L4 with the right Arctic Silver, and presto, no emitter fire.

In turbo, it gets too hot to hold in 30s, but for 3000lm I’m certainly not complaining! :evil: It needs more fins to dissipate the heat, so I’m going to machine some more fins in the head of the L4, when I can borrow a lathe.

Moral of my story: get heat away from the emitter as fast as possible, and worry about dissipating it later.

Oh God I hate plastic reflectors.

It sounds like there was something on the led seal like solder flux that chared the led. This is not an uncommon thing to happen to any type of led. I doubt very much the generic thermal paste had anything to do with it.

I have never had that happen with an MT-G2 I have fried them completely with too much current though. A single MT-G2 can’t handle three high amp batteries at once.

I have had that happen with SST-90’s though. Too much heat and part of the emitter will turn brown. So I wouldn’t say it can’t happen with an MT-G2 just that I haven’t managed to do it yet, give me time though. I do think you are right though, probably something on the emitter did it.

I have a Yezl Y3 that is set up like that too. 3,000 lumens and gets hot FAST, but I love the light.

I ask not to get me started on plastic Reflectors. :bigsmile: The truth is that Most of the time the plastic reflectors are cheap and not really well done at all. However it’s not the plastic that is the problem, it’s the cheap coating they use. They are trying to save money on the plastic so it’s not a big jump to think they also did the coating as cheap as they can. A really good high end plastic reflector is a whole world of difference. So are some of the cheaper ones. The ZY-T08 was my favorite light to mod when they had the plastic reflectors. They changed them to aluminum and messed them up and now they don’t focus correctly. The Fenix thrower, I forget the name but the biggest reflector one they have, is plastic and it’s an incredible reflector. 600kcd is easy to hit with it and it’s light weight too! Love that reflector.

But most of the time, the plastic ones are just cheap junk and the coating flakes off. But if you have a plastic reflector that is actually focusing correctly don’t automatically assume that a similar sized aluminum reflector will be an improvement. If it’s dimensions are off by just a little bit it might not focus nearly as well.

My modded L2 (with heavy thick brass pill) overheats with XM-L2 on DTP at 4A in about 8 minutes. Once heat soaked, it takes a good while to cool down before I can use it again. While its superior to something C8 sized, it simply lacks the surface are to wick away that much heat over a sustained period of time (as do the many other lights I own in that size and power range). The L2 would probably make a great 2 minute “wow light” if driving an XHP50 or 70 @6 or 7A. Sure, we can still operate in the lower power settings for sustained periods and have turbo available when needed. Its just not as much fun. :stuck_out_tongue:

Pic stolen from JM. :bigsmile:

Lights with integrated shelf are only effective if they are designed properly, with a sufficiently thick shelf and surrounding material to conduct the heat through the entire head surface areas. Any thin material in those areas will cause a thermal bottleneck (especially where employing thin material connecting the shelf and head) which prevents most of the heat from traveling through the entire head. When that happens, heat remains trapped in the lower head and neck, heating only a fraction of the head while often cooking the battery tube and cells. Great examples of this are the large Acebeam/Supbeam lights, with their well noted botched lack of thermal considerations.

To experiment with any host, just turn your flashlight to turbo until it gets hot and feel which parts of the light get hot and which parts don’t. From reading the threads, I suspect many people never pay any attention becasue they dont use their lights for much more than whitewall hunting or short bursts before tossing it on the shelf with dozens of others. Keep in mind also, that cells act as heat sinks until they become heated, at which time the entire light takes far longer to cool down once it becomes completely heat soaked. After the light heats up, a second sustained blast in turbo becomes heat soaked far more quickly than the first. Both are unacceptable scenarios for people who use their well driven lights for extended periods of time in the higher power settings… like me. :bigsmile: Active thermal management in the way of forced air cooling will hopefully become more readily available and affordable some day soon. :money_mouth_face:

IMO, from the best TIM available to using something like bacon fat, independent thermal conductivity tests in CPU’s are close enough to indicate that the proper application of TIM to be far more important that choosing the proper TIM itself. This is another subject that has already been beaten to death across the internet. :exmark:

That’s a couple of great points there Flashpilot.

Using too thick of a layer of any thermal grease/epoxy whatever will create a bottleneck right there under your emitter. Your second point that was also really important was to look at the area on your light that is heating up. I have had a few light where I took a good 20 minute walk with a light and didn’t know I had gotten it hot. Got back home and burnt my hand on the head of it. That was good though, it meant the heat was going up and away. Other lights are not so great, I love the yezl Y3 but it heats up right there in the driver area. If you don’t put some kind of thermal grease in the threads, the head will be the last thing that heats up.

So basically what have been said is that no host is has the thermal capacity to house the XHP70 at Maximum power?

I have a xph70 in a solarforce M9 running around 4000+ with a 12amp LD2 . Single mode as I messed up the soldering. That got too hot to handle but no damage to the emitter or driver.
I have toned it down a bit now with a 9 amp LD2 and 3 modes.
Btw I used a thermal sticky pad designed for cpus.
While it wont light up something a mile away, it will light up everything that you can see.

__I have been following this thread and I just wanted to add a few thoughts.

Yes, contrary to what some people think, thermal paste is a very POOR conductor of heat compared to the metals they are mating. That’s why the least amount possible is to be used. I just finished rebuilding a SRK that failed when all 3 XM-L’s literally slid off their aluminum stars. Each LED STAR was floating in a thick layer of thermal adhesive.
It was mentioned that the thermal conductivity of Aluminum is 205 and artic Silver is 8. Aluminum is 25 times better. As big a difference as that is, it pales in comparison to the difference between Artic Silver and Air. Air is 0.024, artic Silver is 333 times better than air. (microscopic air gaps between the mating surfaces) That’s why thermal paste is so critical.

The most important factor concerning heat flow between 2 metal surfaces is contact surface area. If the thermal paste is infinitesimal thin then a pill can be almost as good as a solid head.
Where a pill has an advantage is that it can be replaced with one made of a more exotic material, say copper. Then there is a distinct advantage
Someone mentioned, I think the OP, that the most important thing though is to get the heat away from the emitter as fast as possible. That’s why just a copper star is so effective. Surely the most bang for the buck.

That was me! :stuck_out_tongue:

Yeah, imagine a world without direct copper PCBs, where XM-L2s catch fire at 4A, and MT-G2s start to diminish in output above 3A.
It’s a distopian nightmare…

I put an XHP-70 in this light, 6V and 7+ Amps and it hardly breaks a sweat.
42 Watts isn’t exactly pushing it, but diminishing returns sets in anyway.

Read this and see what a small host is capable of

Does anybody have experience with this cant seem to find any info about that website and light?
Would it be reasonable to install 3 MT-G2 in this host? And what driver?

I think I have had four of them total and they are an amazing light but not really set up to do 3x MT-G2. The problem is the battery carrier is 3x 18650 so that’s a bit of a problem. The only drivers around that I know of that will function for any time at all while driving 3xMT-G2 Are Direct Drive Drivers and It just doesn’t work out with 3x18650. If you wire the emitters in series a DD driver doesn’t have enough power if you wire the emitters in Parallel it’s too much and the emitters are killed.

The Trustfire J-19 is the best light that I have worked on to make a 3xMT-G2 build actually work.