Modifying the FandyFire Rook, what to do with this driver? *Pic Heavy*

I got the little FandyFire Rook from FMcamaroZ28 and really like how it’s a minature king. It’s got a single XM-L emitter and a nice reflector so I decided to change out the emitter for something a little more exotic, and since I had a second Luminus SST-90 I figured “Why not?” I don’t want to run this one max out like the one in the C8, so I figured I’d have a look at the driver and see if I could maybe just bump it’s output a little, it’s got 2 levels, Hi and Lo and a hidden strobe.

The aluminum star under the XM-L is Very thick! Probably 2.5mm, haven’t measured it yet but it’s much thicker than a Noctigon. So I’ll have to shim up a Noctigon with some copper sheet in order for the emitter’s height to be correct. I’ll also have to open up the reflector to fit over the large dome.

But popping the driver out, I got a surprise! It’s got an FET on board, AND 6 7135 chips! Now what? Figured I’d take pics and let y’all tell me how to go about this one…Comfy, you watching? :wink:

Is this little yellow part melted? It looks blackened on one end, and it appears to have a puddle of black melted plastic at the board junction, what’s up with that?

An inductor?

One of 6 7135 chips, 3 widely spaced on this side of the vertical board, 3 closely spaced on the other side…


An unknown MCU sitting close to the apprently burned yellow component…

So, with both an FET and 7135 chips, how do I bump this one up to between 3 and 4 amps? Simply stack some 7135 chips? Or take a different approach altogether? I’ve never seen a driver like this one before. Any help will be greatly appreciated! I’m looking for that 3-4A range so the SST-90 will still be comparatively efficient and making some 1100-1200 lumens.

Thanks a bunch for your help!

Sorry for the sloppy pics, tried the digital zoom on my G1X…handheld at 16x it did pretty good I thought. :wink:


Nice macros. It must have two different circuits. One boost circuit for 1.2 volts (4P Ni-mh,etc) and one for 4P 14500 Li-ions. I would guess the 7135's are for the Li-ion circuit.

Measure current and if you get 2.1A for one cell type or the other, that should tell you, and go from there.

Wow, just when I thought I was getting good enough to recognize what components do what this comes along. That’s a lot of crap in there. I’d agree with ima4wheeler’s assessment.

I’ll be following this build, I’m sitting on an SST-50, I need to find a use for it and get an SST-90 while I’m at it.

Ok, got it back together with a feeble first attempt. The reflector is fitted to the SST-90 and that emitter is in and running. FWIW, a smallish smooth reflector doesn't give a very pretty beam with the SST-90. Had to go with an OP in my C8, which drastically improved the beam profile. Don't have that option here, so I might try the O-L version and mist it lightly with some clear overcoat. No going back once that's tried! lol

I stacked 4 7135's. This ends up with no apparent gain running 3 14500's (it's 3 not 4) The big SST-90 is making 631 lumens OTF. I'll try to get an amps reading, should be up around 3.5 if the stacked chips are doing anything.

I went ahead and pulled it back apart, stacked an R056 resistor on top of the R100. Now, with my 3 14500's on the charger, it'll pull 3.31A from a 20R. Don't have a lightbox reading yet, but will have when the cells on the charger are done. Should be up over 1000 lumens now. :)

Well, it was just underperforming, SO I fixed it. Not underperforming anymore. :)

I stripped the driver, removed everything. I built a brand new BLF 17 DD FET driver and piggybacked it onto the base board, just using the board as a contact. I have the MCU flashed with the same STAR firmware as my M6, but with 5 modes instead of 7. The UI allows a long press to go straight into Turbo mode from off. Or a normal press goes straight into the low mode, from where a slow press will go back off. Very nice, I have Richard to thank for this firmware and I very much appreciate it!

So, the FandyFire Rook now utilizes the SST-90 starting in Low at 33 lumens (batteries arent fresh, or all 3 the same kind for that matter) then 193 lumens, then 557, then 976 and Turbo is a whopping 1842 lumens at start (mind you the cells, not fresh)

It gets heat moving into the head post haste! I need to get the cells charged up, then remove the spring board from the tailcap and run an amperage test to see just what it's pulling, but I can guarantee it's way the heck up there!

The toasted yellow component is a tantalum capacitor. They are fond of blowing up, violently. Like small fragmentation bombs. :bigsmile:

Thanks Helios, it is now deceased and in the circular file. As are the rest of the components that made up this driver originally.

I got the cells charged, did a spring bypass at the tail, and saw a tentative 11A+ from the 3 14500 cells. This translates to a start value in the lightbox of 2239 Lumens!! :) At 30 seconds it's doing 1763 and at 60 seconds it's still putting out 1666 lumens, albeit the head is quite hot!

So, I'd call the upgrade to the FandyFire Rook a complete success! Circumvention of tantalum tantrums complete. ;)

Nice! :bigsmile:

Sounds nice now. Let’s see some pic’s of that driver & emitter!

Based on the location, that's probably a simple limiting resistor. Or, it was probably a simple limiting resistor.

I did the same thing to my $12 UF-T50 clone:

Planning to do the same thing to a Roche F6, which is nice, but could be better.

Comfy, I tried to keep that driver, really I did, but stacking 4 chips made no difference I could find with either Alkalines or Li-ions. So I stacked an R056 on top of that R100 as well, again no discernible difference. Then I noticed it singing. And with the fact that it is known to drain batteries with a parasitic drain, well, it had to go.

Oh, your doing a Roche F6?! :bigsmile:
I’ve been quite interested in the Roche F6. And its mod potential.

I pulled the F6 apart and looked it over, but put it back together for now. All I know so far is that the switch will be tricky unless I strip the driver and use it as a contact board/switch bracket (which I hate doing, since the driver would still be functional if I could work around that), and that it loses all the modes if you jumper the sense resistor. Oh, and the brass retainer plate thing is reverse thread, so righty-loosey lefty-tighty.

With measurements I could create an oshpark switch bracket or board if your interested.

There was space for 2 more 7135’s on that board, see the empty slots for them

You could have ordered a string of 10-7135 380mA chips from FT and replaced the 6 there and add two more for a total of 3.04A

But the FET board is going to pump ALOT more power and with proper firmware give more levels better PWM and dump the annoying flashies

For some reason I saw those ground pads as affiliated with the LED + and - symbols, didn't even notice it was extra space for chips! I have a few hundred chips, that wasn't an issue. I stacked 4 additional chips but there was no visible result, not with either type cell. The lightbox couldn't tell it either. I don't know what was up with it, but it's pieces parts now. :)

Two little locating nubbins are 2mm diameter, exactly 90* from the switch location. The stock switch is thru-hole mounted (2 for the bracket, 2 for the contacts), dunno if there is a SMD style that would put the button in the same location but if there is, that'd be much easier to work with. PCB thickness is the same as the standard Oshpark parts.

edit: There's not much vertical space in the driver compartment, only 5.5mm available between the roof and the inner surface of the outer board. It'd probably work a lot better to add the switch mounts & locating nubbins to the 20DD driver...

Wait... are the 14500s in series or parallel?