XHP35 output and death test + PWM efficiency test by Texas_Ace

Ok, another installment of the LED test series, think this is the last of the test I have done so far.

Thanks to KawiBoy1428 for donating the LED’s for testing!

First up we have the best output test from an undamaged high bin LED.

Click any image to see it larger

This next test is of a low bin 90cri HD version and the goal with this one was to see if it could handle an FET driver at full power and what the efficiency was like at lower levels. It only looses 11% efficiency at 50% duty. So while the super low modes would not be very efficient the mid and high modes will work just fine with an FET + PWM to limit current.

This last test is purely to see what these LED’s could handle and when it would die. There had been debate in the past that these led’s randomly combust at ~3A, it seems they can actually handle quite a bit more then that.

It is interesting that both the XHP35 and the XP-L2 both max out around ~40W, I am thinking this might be the “limit” for the XP footprint to dissipate heat with the tiny thermal pad. I would love to see the XM-L3 released with the latest setup and see if it can handle more power due to the larger footprint.

EDIT 11-10-17

Here are some more xhp35 tests I ran during the BLF Giggles project. These have higher Vf then prior LED’s for some reason but the 4000k also makes more lumens for some reason, even though it is a D4 bin. These were both tested back to back.

Thanks for doing this test TA. Looks like 2A is the sweet spot and 2.5a to push it hard. I would have thought 3A would be the max usable current but it’s not any better than 2.5a.

Yeah, 2.5A is a good current for the XHP35. up to 3A will give you higher numbers in the first 30 seconds though.

Cool, thanks.

I would say 1.75 Amperes is the ’recommended maximum’.
Above that efficiency suffers considerably.
It’s a 20 Watt LED basically, but does 100LM per Watt at around 15 Watts, which is the same as XP-L2 (!!)

O, wait, i was looking at the damaged XHP35 results… Sorry…

Maybe a dumb question, but looking at the results, does an xhp 35 have any advantage over an overdriven xp-l2?

Not nearly as much as it had over the XP-L.

The biggest advantage is the lack of extreme tint shift and coming in a HI emitter. It also does provide more output.

I experimented with some super tall centering rings to block the tint shift, it does reduce the tint shift but it would have to be a really high ring to block it. The tint shift actually seems to come from the dome and not from the edge of the die like first thought.

Remember he’s using an actively cooled set up unlikely to be duplicated in a handheld light, even a larger one, so in real world conditions they probably do self imolate @3A and diminish output well before that.

In order to try to replicate a flashlight I actually reduce the cooling efficiency of my setup on purpose. For example I do not use thermal grease on the star. The heat sink while large is far from the largest I could have used.

It appears to replicate flashlight use fairly well, at least for the first few minutes.

Be increasing clamping pressure and adding thermal grease I can improve the numbers some but that is for sure well outside what will ever be seen in a flashlight.

From Dales testing he saw higher lumens at turn on at up to 3.1A IIRC, which lines up very closely with the test here.

Silly me, I thought you had a nuclear plant cooling tower hooked up to a small block Chevy with twin turbo’s. :person_facepalming: I’ll be interested to see what people get and how they achieve it.

Not TA’s style. It would have to be a blown 800 inch big block for him :smiley: and even that would be marginal :stuck_out_tongue:

You gotta like his fearless pushing of things to (and over) the limit :sunglasses:


lol, actually I am 100% a turbo guy. I love me a good ole Blown big block naturally but I really love a nice sleeper turbo setup that you never see coming.

Not to mention that turbos are better, make more power and win more races. Thats why all the outlaw guys have gone turbo in recent years, they are just better.

Cars are a topic I could go on about forever. Built a few in my time, also blown a few. You learn from things that go wrong, not things that go right.

As long as you learn something from a mistake or when something goes wrong, then it was a good thing.

I went on a bit of an XHP35 binge today, I built a 1504 with an XHP35 HI H2 3C 80CRI using 26350’s.

It is using a Texas Avenger driver with FET only and PWM to keep current from going to high. Runs about 2.5A on high.

It is not bad, a bit ringy but I am using a white centering ring at the moment, going to print a black one off and see what that does.

Makes about ~1200 lumens OTF in flood mode, forgot what it was in zoom, think it was around 800? I didn’t really pay attention to the lumens since this was a throw build.

Around 250kcd for 1000M of throw. Outside you don’t really notice the rings. Overall it looks nice but gonna see what it looks like tonight. A bit better throw then the DO1 I built but the DO1 had more spill and a better overall beam IMO.

The other build is a Convoy L2 with the same LED and driver both with 26350 and 26650 although I think I will be using the 26350.

It is making a bit over 2100 lumens OTF @ 2.5A also with PWM to keep the current in check.

Tint is good and beam is as well. The interesting thing is that I have an XP-L HI L2 that is the same tint and build (although I don’t think it is a V2 bin XP-L).

Comparing them side by side inside and the L2 has about twice the hotspot size but otherwise the beam is basically the same. Interestingly the throw is a bit better on the XHP35 light (it was around a 8-10% improvement).

I really like the XHP35, I just wish that it didn’t need 4S cells.

I am right now building ThorFire TA13 zoomies with XHP35 HI C2 30G emitters and H2-C drivers from Kaidomain. These should provide between 1700 - 1800 lumens at 1.5A in such host before full warmup, but the real out the front lumens may be ≈70% that figure or so in flood, doesn't it?

zoomies vary a lot in my experience with what you see OTF. At full flood ~70% is reasonable. At full zoom it can be at little at 25%.

Zoomies are fun though, I have an xhp35 1405 that is a blast, like a laser pointer. Not the most useful but super fun to play with.

For real world use I prefer the L2 xhp35 HI or my XP-L HI D01. Well Giggles is the best thrower I have but thats not really a fair comparison lol.

Hey TA, any update on where to purchase the BLF GT emitters? I believe their XHP35A H0 0000 0D0BD440E, I looked at a bunch of different sites, but the minimum order is 1000 and I don’t see myself needing more then 8-9. Do you think Lumintop would be willing to sell a few emitters on the side? I have no problem with paying a slight premium over the normal cost of some XHP35 Hi emitters if Lumintop would be willing to sell some.


I have wanted to do this same thing but have not found a source for them yet. I have been trying to talk Lumintop into selling the LED’s separately but that led us down the rabit hole of trying to find an even better LED.

We did find a 4500k that could be real nice but the samples they sent me were lost in transit and I have not heard anything since. Once things quite down with the GT group buy I will see about the LED’s again.

Thanks TA, I look forward to getting my hands on the GT’s 4000k emitters.

It seems like some of your highest outputs were between 2600 and 2700 lumen. This is a raw led, correct?

In a flashlight where you have losses in the reflector and lens, what do you think the “Out the Front” output would be?

DB Custom was saying the MF04 output might actually be 2700 lumen out the front. I’m pretty skeptical of that. You would need the raw led to put out at least 3000 lumen and I havent seen any measured that high.

Plus, I doubt the MF04 will be using a top bin led anyway.

Do you have any thoughts on this?

The OTF lumens on the XHP35 always seem to be lower then expected and I have no idea why. It just does not take to being overdriven as well as modern LED’s.

LED’s like the XP-L2 / XHP70.2 always surprise me with more lumens then I expected OTF.

Far as actual numbers, I would not even try to guess, too many variables. Better to just see what happens.