Group Buy Now Live - BLF GT90 SBT-90.2 powered 5,000 Lumen, 1.8Mcd / 2600m Super Thrower

I will split my order of 4. I have a spare set of carriers anyways.

I was thinking on it and I think the only fair and unbiased option is for me to use a random number generator and the first 6 numbers it picks get first dibs.

Those with multiple kits will have the choice of splitting the order or passing for now and waiting for the remaining carriers.

I will send PM’s out in a little while with payment instructions.

I'm very interested!!!

Interested in the GT90 or the GT90 conversion kit?

Sorry Texas_Ace, I misread the post. I thought it was the GT94 thread.

I have a GT90 already, but I don't want to mod it. Thanks for replying.

K - completed my first GT to GT90 conversion with the kit and it's pretty impressive for sure. I measured 1810 kcd on only 4 30Q's at 4.15V. I used the shorty tube.

Hard to measure lumens but I'd guess about 5250 (mine) at start, maybe ~4800 calibrated lumens.

Charging up VTC6's now, solder blob tops.

Update: With the 8 VTC6's, I did measure a little bump. These are old VTC6's, like from 2017, but not sure how much use they really got used.

  • Measured 1890 kcd taken at 5 meters
  • Approx. 5600 lumens (mine), ~4950 calibrated lumens

Now I'm wondering if some mods may help, like:

  • swapping out the SIR800DP FET for a SIRA20DP
  • spring bypasses
  • 20 to 18 AWG wire upgrades
  • using higher performing cells like the Samsung 25S or SONY VTC5D's

Has anyone tried any of these type of mods?

I would think taking the throw at 10+ meters would bump the reading, but to me, actually at 5 m the hot spot looks pretty good.

I opened 2 more GT front bezels for doing the same conversion kit upgrade. Had to clamp the head down in pine boards and then use a strap wrench on the bezel. One took a while to get open, while the other 2 went easier. Didn't expect it to be this hard - thought I could get by with grippy gloves but it's awkward to get good pressure on. The strap wrench did not slip 1 mm - got a great grip with it. It's the most amount of pressure I've applied with a strap wrench, and this was for a light without glue!

Driver, MCPCB, and centering piece. Shown are the shorter stock screws and the longer ones I chose to use instead - Phillips flat head 2mm x 0.4mm x 8mm. The heads fit snugly in the centering piece hole, holding it down with modest pressure.

The FET is the SIR800DP - not sure if it's a clone or not. The wires are 20 AWG which I decided to stay with. Because they are doubled up, it's probably equal to a single about 18 AWG.

Compared to stock, same width but the 90 is slightly thicker which is a good thing:

With oversized holes, it makes it impossible to center the LED from the centering piece alone:

Flat surface sanding with 2000 GRIT paper shows the MCPCB is not perfectly flat on the bottom. All around the holes the surface was lower, perhaps slight deformities from the hole drilling in soft metal like copper. Usually I would drop to rougher GRIT's and get it smooth, but I figure there's enough surface area smoothed out and differences are minimal.

Applied healthy amount of MX4. The Stock light had maybe 60% coverage:

Fully assembled. The pre-tinned pads are done with something high temp, like lead free solder. Next one I'll be sure to remove as much solder as I can before assembly. Turned the iron up to 750F just to be able to melt it.

I would say the end result has the LED pretty well centered, though I'm sure not perfect. Hot spot looks very good.

Note: the carriers in the kit had most screws untightened, and the brass contact pieces were loose as well. Easy to check and tighten. A couple of screw heads were slightly stripped but could still tighten and loosen.

Strange that the driver seems to have contact points for flash re-programming, but they are on the inside, not the backside.

Nice job!

I didn’t notice the loose screws on the carriers or I would of tightened them before shipping, didn’t even think to check that. That is something people should check though.

The driver was not designed by me, they adapted the GT70 driver to 3V without my help for some reason. It works though and still uses the same parts/firmware (except for some resistors).

The new GT4/GT94 driver does have programming pads on the bottom for lexels programming key, I keep trying to get them to use it in the GT70 as well.

Thanx! I finished my 2nd conversion kit mod. This time, I swapped the FET for a SIRA20DP and swapped the 4 20 AWG LED wires to 18 AWG. Also I re-programmed the driver with Anduril2.

Re-measured #'s and did get a bump in throw and lumens, but not all that much. Actually the BT 30Q's got a bigger bump than the solder top VTC6's -- could be from more compression from the springs because the BT 30Q's are a little longer.

I also re-measured the stock BLF GT, and at 5 meters, it's ~950 kcd while the conversion kit lights I measured at 5 meters are 1,850 - 1,950 kcd. Wondering if 5 m is too close for focusing of the XHP35 HI, while the SBT90.2 does much better at 5 m, at least in the GT reflector.

I also noticed the cells fit much tighter in the 90 carriers than the older GT carriers. Looks like they are intentionally designed this way. Actually I got some BT 30Q's that don't fit in the 90 carrier, but most of my BT 30Q's do fit, fortunately. With the cells that fit in the 90 carriers, the springs are just about fully compressed so probably no bump from adding bypasses.

Extra jumper for LED+ wires added and reflowed SIRA20DP shown:

18 AWG wires:

I started with 600 GRIT and still could not get the full surface sanded smooth. so I took a large punch and hammer and gave it a couple taps on the top. The bottom then got real easy to sand smooth. Now there's only non-sanded surface areas near the edges and around the holes:

Stock MCPCB from the GT (XHP35 HI):

18 AWG wires shown:

I believe this is bouncing off the low cloud cover. I took a few shots at different settings and this one looked the best reproducing what I was seeing - beam looks NW, nice bright hot spot with a halo around it. It's been misting and drizzling all evening. Measured 1,975 kcd taken at 5 m on this one:

The old BLF GT seems to be sooo outdated now.

Are there any conversion kits avaliable? Or at least only the centering ring, led and driver? I suppose I can convert the carrier if its the only thing I am missing for the conversion. Thanks guys!

Someone dropped out so I think I will end up with an extra kit at this point but will have to wait for the carriers to show up to get a final count. Still waiting on those, they were supposed to be shipped “next week” 3 weeks ago lol.

Finished the 3rd conversion kit upgrade. This time made the following changes;

  • removed the window on the SBT90.2 LED
  • removed R1 and R2 resistors on the driver - they are not used as voltage dividers and add to parasitic drain

With the window:

Window removed:

Driver changes;

  • swapped 20 AWG wires with 18 AEG
  • swapped FET from SIR800DP to SIRA20DP
  • swapped 10K R15 with 5.6K ohm (to make switch LED brighter)
  • removed R1 and R2

Polished MCPCB and shelf top to 2000 GRIT. Also flattened out the MCPCB with a couple taps on a large pin punch (flat end):

With this kit I measured ~2,100 kcd (~2900 meters)

I compared the 2 kit modded lights on the same set of carriers and cells, with one light having the SBT90.2 window removed, and measured about a 4.2% gain on the ilght with the window being removed -- everything else the same (driver FET, LED wires, etc.). This wasn't very scientific, limited to just one trial, and being it was done on 2 different LED's, 2 different lights, and I measured the one with the window first, and didn't recharge the cells. But still, it indicates there is probably an improvement. If I were to guess, it may be closer to 5%, maybe higher.

The switch LED did not get much brighter - hardly noticeable - not sure why it did not end up much brighter since the resistor value was nearly halved. I suspect the switch boot cover is not so transparent, not as much as a Q8 for example.

Carrier mod from 4S to 4P. I did two carriers the same way.

Decided to use old Q8 springs (large ones) since they are longer and narrower. Hot air station worked great on removing the springs and brass buttons. Used thin copper sheeting to make the low profile bridges, and 20 AWG teflon wire.

For the springs mounted over the prior button contact pads, I scraped off the traces to expose the copper in order to solder the springs down. Wherever you see soldering direct to a via, the coating was scraped off to expose the copper trace. Used 22 AWG bypasses.

The original GT carriers can hold longer batteries than the 4P stock GT90 carriers, so a stiffer spring arrangement can be used. Since there's less spring compression, the bypasses are a better option. Thinking these should perform a little better than the stock GT90 carriers.

Did you re-flash the firmware on the driver when you removed the R1/R2?

The firmware is the same firmware as the GT70 which does use the R1/R2 arrangement as far as I know, they just adjusted the resistor values for the lower voltage.

5% gain from removing the window is what Vinh said he got as well, so sounds perfectly reasonable to me.

Yep - forgot to mention I flashed Anduril2. Voltage readings are accurate, least in the ballpark. Anduril2 has voltage cal built in so not a worry.

Ok, that makes more sense. :+1:

What do you think - would a bypass on the spring under the brass cap help much? I'm think'n it might on these modded setups. It looks like a high quality, (beryllium?) spring but it's carrying a ton of amps through it. We are probably over 25 amps and thinking the output curve on the SBT90.2 is probably flattened out anyways (djozz LED test), so probably not much of a factor.

Hhmmm... Maybe it would help to hold higher amps for longer

On a setup like your trying to get every last lumen, it might help a little, but we are talking a very small gain if I had to guess.

For anyone else, I would say it is not worth it.

I don’t see it causing a noticeable gain and we are already at the point of diminishing returns on the SBT90.2 anyways.

I personally am not worried about it.

What's pretty cool about this FET driver setup is that we are probably pulling about 30 amps or so, think VOB measured 27 amps, but 30 amps equates to only 3.75 amps per cell.

Really 3500 mAh cells would be the best option to go with. From tests I've done with the Q8 in the past, the 35E's did pretty well. Was thinking of trying 25S's or VTC5D's but probably not much point.

I am finding measured variances on different cells. For example, I got some 30Q's from 2016 and they are showing their age. Coming off the charger, aging cells will drop fairly quickly to 4.18V/4.19V even though the charger is set to 4.22V. These cells are getting lower #'s then newer 30Q's on the GT90.

SONY cells seem to age/degrade quicker than Samsungs, at least it's been my experience with them. I've seen my VTC6's and VTC5A's and VTC5D's all degrade quicker than Samsungs.

I own over 30 lights, from mini to gt and ms18, the only light that kinda makes sense the light with a active cooling system.
I would have more lights but they are almost all the same, too damn hot and unable to sustain high outputs.
Build cooling systems for your lights instead of bypassing springs and modding to get them hotter.