I have lots of parts & boards on the way, will be building the 15, 17, 20, & more of the SRK drivers. I'll post up a revised parts list for each one once I'm confident everything is right.
Also on the way are a few different big FETs, to see if that changes anything with the gate/pulldown resistors.
Just got in the 20DD w/o the resistors (not V2.0) do you think they will still work…or should I wait to order?
Gonna put together a SRK one…and build the 15DD w/ ground ring tonight, and probably a 3.04A 20mm TexasPyro as well
You'll have to try it and see, I built 3 of the SRK-DD boards and they all worked without the resistors, but the exact same parts on the 17DD (v1.0, missing GND ring) don't.
I would bend the gate pin on the FET up off the board before soldering, then try it first with just a short jumper wire between the leg & pin. If it works like that without a resistor, remove the jumper, bend the pin back down to the board and solder it. If it doesn't work, remove the jumper and put a 100-130 ohm resistor between the leg & pad (and add a high value resistor between the gate pad and ground, the pulldown resistor). If it works without the gate resistor it won't need the pulldown resistor either.
Hey Matt, what’s up with the 72 x 7135 board?
Matts retired. Everyone wore him out.
And you were the last straw.
I've been sort of shocked seeing 18 AWG in the big Fenix lights (TK61 and TK75) with the heavy insulation, making it look more like 16 AWG. The wires going from the contact board to the driver are that thick and less than 1" long. I suspect they have a good reason for that, since everything about the design of these lights is about weight reduction. So they seem to value heavy gauge wires over using an aluminum reflector instead of plastic, for example.
Did I break him? Bugger. I did send him a present this morning but now I have to send a get well card as well? This budget light hobby is certainly expensive. Have you seen the price of cards lately?
I'm alive lol. Just trying to focus on finishing renovating my bathroom so I can sh*t again. Nah just joking. I've been doing in my back yard*.
The 72x7135 board an OL request. Lips are sealed on what he wants to do with it! Note it will NOT fit an SRK. Way too big.
- Matt
* It's an en-suite so we've been using the main bathroom. It's just at the opposite end of the house to our bedroom. First world problems.
Now all of the enlightened forum knows where you take a dump, whew, we were so worried! Can you sign this piece of TP for me, pleaaaase!? 
Now, get back to work!
Speaking of dump, either y’all are rough on a house or y’all bought a dump cause you’ve been fixing on that place for 2 years!!!
(yk I’m just kidding right? Gotta personalize it…hey, look who you’re talking to, the guy that customizes EVERYTHING!)
I don’t know what you could use to dissolve fiberglass leaving behind the dielectric, copper and mask layers. Maybe instead of doing that a person could delaminate the board? As I understand it you could just add a ton of heat until the dielectric layer no longer wanted to be stuck to the fiberglass layer. The other layers might delaminate before the one you want though.
Either way…going to be a “destructive” method [and probably toxic/chemically as heck], getting that top layer masking and traces off the pcb then mounted to a copper slab is the end goal without destroying the integrity of the overlay…getting there is the tricky part…unless you pay a hogillion dollars to have someone MAKE them from the get go
oh and anyone want to see the process for making the boards…as in the actual construction and whatnot…pretty interesting…it’s european…but I don’t think the process is much different here
http://www.eurocircuits.com/index.php/making-a-pcb-eductional-movies
In fact in the last video, you can see some tell tale purple boards…not sure if OSHPark…but makes ya wonder
I’ve purposely de-laminated boards. Some boards are are not bonded as strongly as others.
(This post will de-laminate / self-destruct momentarily)
Oh boy... Not all the boards have their own thread, like the 15mm. Is there any one place I can find the latest, greatest info on the BLF15DD? I just got mine today but they don't have the R4 and R5 resistors, so, am I out of luck? Should I just go ahead and order the latest? Anyone test it yet? Do I have to navigate thru this massive thread?
IIRC the IRLML2502 used on BLF15DD hasn’t yet been reported to be buggy without R4 & R5. No need to jump to order replacement boards just yet.
Also, if they are needed, you could cut a trace & scrape a bit of soldermask off to add R4. R5 would fit across 2 vias. 
Rufusbduck, was trying originally to keep this thread less cluttered. Seems he was forced to abandon that goal.
Hi Tom, I was hoping to have original research done in other threads and just use this one for links but I’m not willing to go tyrannical over it so it is what it is. If you have any questions just ask and I’m sure those of us that monitor this thread will speak up.
After Comfychair finished the initial research on the SRK version( most of which was discussed in pm’s) Mattaus worked up the 15, 17, and 20mm versions and then I posted them here without any real testing on those versions. A few bugs have shown up and since been sorted out but this is new enough territory that unknowns may persist until others can duplicate results.
Dbcstm reported complete success using a customized version of firmware but I myself had iffy results using stock Qlite firmware. He was using one of the Tiny10 boards without extra resistors and I used a Qlite depopulated of 7135’s and needed a 200 ohm gate resistor to obtain stable results. There’s no question in my mind that either board can be made to work and as Helios points out it’s pretty simple to break the gate trace for the 100 ohm resistor and jumper to ground with a 10k if one or both are needed. The irlm2502 seems fine supplying as much current as you are likely to get from any 14500 or 14650 cell but standard programs need to be adjusted downwards for the lower mods as each pwm pulse is an unregulated DD rather than a pulse limited by the sum of the 7135’s. Unless you are limited by space or after a very high drive current(both of which Dbcstm overcame) you are better off with a stack of 7135’s.
I had issues with the Qlite FET as well, including blowing out the positive trace under the MCU. I bypassed this by drilling a hole through the pcb at one of the non used 7135 chip ground pads next to the diode and running a 22ga wire from the top of the contact spring to the emitter in a direct route. Then we learned of the resistor in the pwm gate and that has been working. I built my first one containing both extra resistors yesterday on the BLF 17mm DD board (#1 with no grnd springside) and used a contact board of larger size for this particular light. The driver is functioning just fine, albeit with a bit of whine in the middle modes. I used a 200 ohm resistor between the FET gate pin and pwm pad (vertical orientation), one of the 1912 resistors for the pulldown. Not using moon in this light.
The BLF Tiny10 with irlm2502 FET is running perfectly with no extra resistors, 3.11A from an Efest IMR10440. The 4 base modes work perfectly with no audible. Everything done here for these boards is greatly appreciated, and this particular BLF Tiny10 has made my year! 
While that BLF 17mm DD board is capable of well over 5A with the right cells, I’m having difficulty achieving that in the light. It would seem that the battery tube to head connection isn’t permitting solid enough ground to reach the high amperage. Testing with the same cell applied directly to the back of the pill results in nearly an amp higher readings. I’ve filed anodization, wire brushed threads, don’t know what else to do. (Defiant Super Thrower with a different make of light’s cheap 26650 battery tube and tailcap)
I'll do a separate build thread for each version once I have the boards. Oshpark estimates they'll be back from the fab on the 13th.
I tested the same version of STAR FW with the only change between them being the PWM speed, one at 9.4kHz and one at 19kHz (both using the same CPU clock, 4.8MHz). Both worked the same, except the 9.4k one was noisy in the upper-middle modes, and the 19k one was silent. Some of the early prototypes weren't happy running the fast-PWM code, but the latest versions seem to handle it just fine.
I've since tested the 19kHz NLITE and also luxdrv converted to fast-PWM, those work fine too.
Sorry if it's obvious but there should be NO noise at 100% PWM...