XHP70.2 P2 4000k Output test by Texas_Ace - it's over 9000!! lumens and still going strong!

It is like the smaller one you have there. I have found it to be a pretty good analog for common flashlight sizes for most LED’s. This XHP70 is the first one I have even considered the heatsink being an issue with.

I tried to drill into a larger one like you have about 6 times and I have 6 broken drill bits to show for my effort.

If someone can drill and tap a heat sink like that for both 16mm and 20mm stars (with fan of course) I would be happy to take it off your hands.

The larger one has a vapour chamber in the center, if you drill into that one it will stop functioning properly.
There are coolers far better than any of the ones intel makes though, but something good like a NH-D14/15 from a used site like craigslist.

Yeah that is what I thought. I was going to say that I would not touch it, but Enderman said it first.

Also, for testing even something like an SST90, Noctua’s cooler would be a bit overkill. For these big LEDS overdriven, it would leave a comfortable margin for pushing the emitters.

Finally, to drill them would be a bad thing as Texas Ace, as it would break the heatpipes.

However, as Barkuti said, mixing thermal paste with thermal silicone works very well as to fix a very high power LED to a suitable heatsink, if you have a ratio of 5-7% thermal glue, and 95% thermal paste.

I’m currently doing that with my XHP 50.2 in my Imgur link, and it works wonderfully, until the smaller heatsink can’t handle it passively at about 30W.

It’s not overkill when your LED is using 100-300W though, like the CBT140, CFT90, XHP70.2, etc

Yeah, edited the comment for reflecting this.

The large heatpipe copper bases are very useful though: very powerful multi-emitter constant brightness floodlights, or just a monster light, like the Storm of Ra.

But for testing, they are a bit problematic if you want to mount them using screws.

Oh, the cooler I was using just had a copper slug, did know they moved to vapor chambers now.

I am not that worried about it, this one works great up to around 100w and the only LED I have tested yet past that is the xhp70.2.

Only the large intel heatsinks (like BXTS13A) have the vapour chamber in the middle, the small stock ones don’t.
The thing is that the more power the LED uses the hotter it will get with that tiny heatsink, and as we all know, higher temp = lower output.
I can guarantee you that the small heatsink you used had the XHP70.2 at over 100C or possibly even 150C.

It was not 150c that is for sure after touching it during the test. I am guessing around 100c. Which seeing as the LED’s are binned at 105c, is still perfectly fine. Which is why I have not worried about upgrading.

“Binned at 85 °C”

And you can get 10% or even more output by running it cooler than 85C.

Ok, I was thinking of another model, either way my goal is not to get max lumens on a bench, it is to get realistic real world lumens so people can know what to expect in a light.

My numbers have proven to do this well over time.

I don’t think anyone building a light with a 200W LED is making a “regular” tube flashlight with horrible cooling, like a ‘realistic’ flashlight.

Nice work Ace. I really love the way your graph looks. What program did you use to make that? Also have you done a control test like this with XHP35HI?

Don’t worry, none taken at all my friend… Were all in this together!

Thanks! The graph was done with Excel.

Yes, I have a few dozen LED tests now, you can see them in my sig. A member here made a graph that can compare them if you click the link to see all the tests.

Just curious about the difference in life hours for the 70.2 vs. Hi 35 with peak watt ratings.

talking about the time until the LED is dead or battery life?

The LED life itself is generally far longer then any of us will use a flashlight so it is not something many of us worry about.

Battery life will be less with the XHP70.2 of course since it uses more power.

Someone with a CNC mill could make a nice custom watercooling block from copper. :smiley:

You don’t need all that. I brazed one together from copper using an oxy-acetylene torch 17 or 18 years ago, back before you could buy them. Heck, for the last 15 years now you could just buy a copper water block. Get a bucket of water, put a little antifreeze in it to block algae growth, drop in a submersible pump and away you go. Keep the flow rate high and it will take a very long time before the reservoir starts to warm up. Almost completely silent as well.

Guys what’s the best XHP70.2 emitter for ending up with a nice neutral white without much tint shift after dedoming?

I have yet to find an XHP70.2 that does not have significant tint shift with dedoming.

To end up with a NW tint after the de-dome always ends up with a green tint shift in my experience.

I don’t even mess with de-doming them anymore on my own lights.

OK, this sounds really serious.
Without dedoming the CRI90 emitters seem to be the best choice right? Or do they all have the same shift pretty much?