41

Nice video, Matt! This gets me even more excited about my XHP70.2 mod! :partying_face:

I had everything ready to drop into the GT but toasted the emitter while reflowing. :person_facepalming:

Currently waiting on a new 80CRI 4000K 70.2. I’m hoping to share beamshot comparisons with my friend’s stock GT when I get mine modded…

42

I think you’ve just created the ultimate search light for search and rescue operations VOB. Sounds like a side gig for ya to me. I’ve already shown my light to my emergency manager counterparts and they’re amazed by it’s presence. I have three GTs so might have to have one modded.

43

What driver are you using?

TA was thinking the FET driver is not a good match for that particular emitter.

44

The driver is an ok match for that emitter as long as the carrier springs are not bypassed. That was the setup I used in the GT’s I built.

It is the sinkpad aluminum mcpcb that I was worried about with that emitter. A copper mcpcb should be ok.

45

I got TA’s recommendation before gathering the parts. :+1:

I’m running one of his GT drivers and an SD75 copper MCPCB modded for 12V. I won’t bypass the springs.

46

would definitely be interested in a kit

something like:

MCPCB with xhp 70.2(dedomed or domed option) already on it with the option of either 1) resistor or 2) fet driver

47

XHP70.2 has been available for over a year.
XHP70 has been available for much longer too.

The XHP35 was chosen because it was a good balance of lumens and throw.
The goal of the BLF GT (as per the name) was mainly throw but to still have a usable spot size and amount of lumens, which the XHP35 does perfectly.
A poll was done and the XP-L HI and XP-G2/3 options were discarded.

The manufacturer was not going to do any dedoming or dome shaving so the XHP50(.2) and XHP70(.2) would have gotten much worse throw than the XHP35 HI since there is no HI version of those two LEDs.

48

I think the xhp70.2 was not available in large quantities until July or so (iirc). The emitter for the GT was chosen well before that. That was the point I was trying to make about why the 70.2 was not chosen. I probably should have listed all the reasons instead of giving a short answer.

As far as why the xhp35-HI was chosen over an emitter capable of longer throw or shorter throw, I agree with you for the reasons you stated. It’s a good compromise, etc…

Enderman, if someone we’re wanting to get much longer distance from a GT at the sacrifice of lumens and a small hotspot, do you think the Olsen black flat would perform well?

Those are only 700 to 800 lumen, but should punch out a really thin, pencil beam of light.

49

The black flat is the best option for throw, but with the conductive center pad it is a bit of a pain to work with because you need to insulate it while also heatsinking it.

There’s a new LED that should be available within a year or two that is basically the same but with a neutral center pad: OSRAM OSTAR® Projection Compact, KW CSLNM1.TG

Going to XHP70.2 is the opposite, more lumens less throw, unless you swap the driver and get way more heat and only a few minutes of battery life on max.

To get better throw with even more lumens the CFT90 BLF GT from vinh is the best option, just very expensive.

50

This may be a silly question, please excuse my ignorance in such things but, are they going to make an XHP 70.2 hi? Having watched the video several times now it does seem to lose the hotspot quite badly, which is what after all the GT was made for (throwing). The light may go as far? but it is spread right our in comparison. VOB did you show the sliced 70.2 with the original driver (with resistor mod) sorry if you did but I must have missed it.

51

When he slices the dome he is turning the LED into a HI manually.

For raw throw distance, the XHP35 is still the one you want but for a more usable beam at more reasonable distances. I like the XHP70.2 more.

52

I don’t think that Cree will ever make a domeless version of the 70.2. There’s no real market for it. The 70.2 was made to be used in street lights and maybe home lighting. It wasn’t built to be used in flashlights at all. Someone correct me if I’m wrong on that.

Most people find the larger-sized hot spot an advantage because it gives you a larger field of view.

Big die emitters like the xhp70/70.2 and MT-G2 typically don’t make good throwing flashlights, but if the reflector is big enough, it can throw pretty decently.

To illustrate the difference that reflector size can have on throw, I will compare the GT’s 120mm reflector to a Convoy L6 63mm reflector. Both using the same emitter at 8,800 lumen.

GT: 1,328,000cd or 2,300 meters
L6: 436,000cd or 1,320 meters

Big difference, huh?

As the reflector size to die size ratio gets bigger, the hot spot becomes smaller and more focused.

53

Thanks for the explanation guys. When/if this kit materialises, and I hope it does, I think I’d be inclined to slice it too.

54

Tough call. About all I can say is that we lost 11.48% output removing the dome on the 4.5A current test.

To answer the question about sustainability I did a test with the final version of the light from turn on to 90 seconds. The output is quite stable because of the large copper mcpcb, however the light does heat up very fast. I took the temp readings on the lowest part of the “neck” piece. I would say realistically at 60 seconds you should power it down, by 90 seconds touching it gets a little sketchy.

Speaking to the “dark cross” issue. I know others have said their reflector created a dark cross using this LED. I tried the reflector from my first batch of the group buy, as well as the two that arrived recently and did NOT get this from any of them. However, the second two reflectors were not tested until after the dome of the LED had been removed. I also tried using 2 other thicknesses of isolator with the reflector I used in my video. 2.5mm is thickness from the factory, so I tried a 2mm and a 3mm. None of these produced a dark cross, but they did seem to effect the corona and hot spot a little. At a quick glance I would say that the 2mm one reduced the corona, and the 3mm one increased it.
.
.
The ring I used to de-dome was just a metal washer. Not certain what size it would be in a package but its 0.050” thick with a 0.404 center hole. This actually leaves a little more material on the LED than I like, but it does guarantee you will not cut the bond wires when slicing.

I had not seen the data graph you posted prior to uploading my video. I was a bit wondering if my own numbers could be high. I’m using a new light meter and have only had about an hours work to put into it calibrating. Looking at your chart I see that at 12V 9 amps you got 9900 lumens with the dome on. So “IF” the declination rate for a dome off LED is the same at 9 amps as it is at 4.5 amps –11.45% in my case that would make us both track about evenly right? 9900 x .8848= 8759 lumens… Yes, no???

Lexel, I love your solution for an MCPCB. This would be much better than using a 20mm IMO.

This light vs CFT90, from what I understand the CFT is generating between 1.5 and 1.8 million CD, so more throw, but the ouput is lower, I think like 5000 peak lumens. Then theres the price I suppose. If I’m not mistaken the CFT90 is like $120 bucks…

I’m not offering a mod service for the GT right now, if that changes I will post. Vinh might be willing to do this mod also… I keep pointing people in his direction and will do so till he says to stop :student:

55

The other GT’s I built did not have a dark cross, the middle of the hotspot was just dimmer then the rest.

Yes, our numbers match up very well. I took some before readings with the dome on and the TA driver and it was just over 10k lumens on my sphere at peak. Although we used a bit different LED’s, so who knows how much that plays into things.

Overall though the lumen numbers are very comparable.

The throw numbers are a fair amount higher then what I got, I posted them in the GT modding thread, they were around 750-800kcd IIRC though with the dome off.

But then I am using a 4000k emitter vs your 6500k and my throw numbers here in humid Texas City are generally lower then yours.

56

My problem is I melt things when I try to solder. Being able to buy a drop in kit would be fantastic.

57

This is simply not true
2 good 26650s like in a Convoy L6 can push a XHP70.2 to about 16A
this equals 8A at 12V

lets have a basic look into your own measurements

Lets target a current close to maximum sustained output
max. turn on 10500 Lumens dropping to 9000 in 30 seconds is just bullshit looking on heat and LED lifetime in a light not very practical use,
even at 7A the LED gets tortured enough almost at 300% of the rated current, usual XPL builds run only at 200%

so lets see at 14A means 6.94V on the LED needed to reach this current

to adapt this to our 12V LED board 13.88V @7A
now we got AWG18 wires 2 times 7.5cm this is 3.15mOhms
lets add for the MCPCB 1.5mOhms
so we got 33mV voltage drop
lets add some more conduction losses on the driver so we end up at 14V at the drivers output to run the LED

lets have a basic look at Samsung 30Q and 35E battery discharge curves

we got 2P Carriers so on a buck driver at the point dropping out of regulation we got 3.5A per battery
at this current discharge after 30 seconds voltage per cell
Samsung 35E 4V x4=16V
Samsung 30Q 4.04V x4=16.16V

on a single carrier we get 7A per cell
Samsung 35E 3.85V x4=15.4V
Samsung 30Q 3.93V x4=15.72V

So lets assume most people got 8 35Es to run their GT
so we got 16V on the batteries

as a BLF member with a buck driver we bridge our carriers springs and especially the button spring with 0.3mm² wire
I would estimate now from both carriers to the driver we get about 0.15V voltage drop ending on driver input with 15.85V

now we got the numbers to start the buck driver calculations

drivers resistance at the point it drops out of regulation to DD
Inductor 4.2mOhms
Sense resistor 22mOhms
MOSFET 6mOhms
33mOhms
Voltage drop at 7A at the point entering DD 225mV

lets calculate the depleted battery capacity where we drop out of regulation
remember LED plus voltage drops in wires and driver PCB 14V
Driver Voltage drop 225mV
Carrier coltage drop 0.15V
=14.4V
divided by 4 to get each cells voltage = 3.6V

So we have to look where the discharge curve again

upper blue line is Samsung 35E at 3A discharge current so we have to go a bit lower than this line to read the capacity at 3.6V
and we end up dropping out of regulation on Turbo when 1.25Ah per cell are used

now to get the time we can run the light on Turbo if its properly cooled
Our driver calculation says on full battery we draw 3.09A per cell this includes driver losses
at the point dropping out of regulation we draw 3.5A so on average we used in that period about 3.3A

1.25Ah divided by 3.3A are 20.5 minutes before we drop out of regulation
and remember LED plus driver loss is roughtly 100W, I the GT will have a hard time stay cool enough 20 minutes with 100W heat

58

Yes, it is true. Both myself and VOB measured actual lights in the real world with this FET driver and with a 70cri LED got a bit over 9A. Add in the extra voltage drop from the buck driver and it would be under 9A for sure.

What an L6 got doesn’t mean anything. An L6 does not have 10 springs causing voltage drop.

Facts are facts, we both tested it and both got the same results.

All the paper numbers in the world does not change it.

Real world > Paper numbers.

Even the stock buck driver at 6A would fall out of regulation in a major hurry as I found when testing throw numbers, the numbers kept dropping and I thought it was because it was getting warm, turned out it was because it could not maintain regulation.

After that I gave up on the buck driver for high output in the GT. Simply no point to limit the max output and gain nothing in return.

I don’t want to argue, I have too much stuff on my plate as it is.

59

Wow, my similar power L6 gets pretty darn hot at 60 seconds so I usually don’t run it more than that (typically not more than 30 seconds). I was really expecting the GT, with its much larger mass, to go a lot longer than 90 seconds. Thanks for letting us know.

60

Maybe you should read how to bypass springs!

I have real world TN40S with 4 XHP35 at 8A build with this new driver parts and it runs on 8A long enough to get pretty hot

The stock driver is build for 2.5A of course it falls out of regulation very quick

Reasons:
Input capacitors too low, if you ever made a Boost driver you know what I mean you can lower or short the current resistor output will hit a wall

Too high switching frequency for 6A with a 10uH inductor
Ripple current 0.3A the datasheet gives a formula and says minimum 1A
Almost 5 times more inductor resistance
Inductur reaches saturation and DC limit 5/6A
4 times more switching losses in the stock MOSFET
2.5 times more conduction loss in the stock MOSFET

As I wrote bypass 2 carrier button springs and take 8 full high drain batteries, you need no clamp meter you will see the LED turn angry blue when you turn it on

You are always talking about falling out of regulation, a DD is falling from max current the second you turn the light on