Dale, you've got to stop making us all jealous with those lights you're building!
I'm "only" getting 5.85A absolute max at the tail with a A17DD-S08 in an X6 powering an old, good XM-L2 U2 1A on a Noctigon, using 20 AWG LED wires, bypassed springs. It's really hard to tell, but out of the 3 types of BLF 17's I used: v1.0 original BLF - Mattaus, the C_K v3.1 version, and the A17DD-08, the only one I could get over 6A out of was the BLF v1.0 originals. I did not compare these head to head -- all were in different lights/builds, so I can't be certain of anything, but really thought I'd get over 6A in the X6 with the 20 AWG wires, and the lock down retaining ring - seems to have a really good ground (batt -) path.
Multiple LED's is the easy case with low Vf's. The single LED setup is the worse case, but is the best real world way to test the resistance of the driver that I know of. I may have to build up a BLF 17 v1.0 just to see if it can do better. Why I need higher amps is not because of the domed XM-L2 U2 1A, but I'm getting a max of only 3.7A in the X6 (w/A17DD-S08, 20 AWG wires) with a de-domed XP-G2 S2. I have a Convoy C8 with a de-domed XP-G2 R5 that gets 3.95A at the tail on a 7135 based driver.
I saw your post about the XP-G2 over in RMM’s thread. I’ve been assuming that you’ve simply got LEDs with a high Vf.
You don’t really have to put them head to head IMO (although that would be ideal). Instead just measure the Vf with your DMM. If you compare the two XP-G2 setups and find that the Vf is similar between the DD setup @ 3.7A and the linear driver at 3.95A… mystery solved.
After seeing all the smiley faces from Dale-mode mods I sort of assumed that the FET had become a non issue but if it turns out to be the case that the FET could be improved I would gladly pop for 10 of the more expensive ones to spread around for testing to see if there is justification for a GB for enough of them to bring down the cost plus shipping.
im in for 5-10
I’ve got 24 boards on the way and a few left on hand. 
I, uh, seem to go through drivers a lot lately. hahaha
Again, let’s not jump the gun. The datasheets clearly show the current FET in a good light, and as far as we know “better” alternatives are all more expensive. IMO the thing to do here is test with what we’ve got vs what we’re trying to meet/beat: the Vishay 70N02. An extra dollar or so per driver really is quite a lot. I’m not saying not to test the more expensive FETs at all, but step one is to see what we’ve got now!
I had bought 3 Convoy M1’s for comparing components before. I can use one of these to put an A17DD-S08 in and compare side by side to the 70N03 version driver. Shall I do so?
Edit: More or less they’re already set up that way. One has the prototype of this driver, one has Richards Moon Qlite, and one has C_K’s Ver 3.1 driver. All I have to do to compare the A17DD-S08 to the 70N03 driver is swap emitters in the proto model. (It’s got an XP-G2 where the 70N03 has an XM-L2 U3 1A)
Sounds worthwhile to me!
Ok then, tomorrow I will build a fresh new A17DD-S08 driver and a fresh new BLF17DD Ver 3.1 driver and put the same UI in each. I will then put the same XM-L2 U3 1A emitters in each M1 (from same batch and purchased together) and use the same cell for testing. In this way the test bed will be as level as I can make it. I won’t even put any of my well squeezed amperage rocks in either light. 
Sounds nice if i ever get to the point where i have all of this figured out so i can start flashing firmwares myself, i will look forward to try it out if its ready then, Good luck
Checking the Vf of the emitters will verify the levelness of the test. Maybe also compare output at lower cell voltages.
I agree wight about not jumping the gun, just making an offer to our mega mod members who might want to eliminate this as possibility.
I built up some drivers last night, not done - not much time lately. I'll do a direct swap in the same X6 of drivers, then see (hoping this eve). I'm testing with an old XM-L2 U2 1A now to eliminate the de-domed XP-G2, so hopefully that's a better test case than the de-domed XP-G2.
Really could all be the Vf of the specific LED/MCPCB I'm seeing - dunno, but would be nice to get an understanding. In my other builds there are lots of variables - some with e-switch, some with the LED+ wire bypass of the driver (not always the best), different LED bin's/tint, etc. The only obvious thing in comparing the 3 driver designs is the original BLF v1.0 seems to have the thickest ground ring, but that's probably not exactly true - the others could have a thicker one, but not all exposed.
Are you referring to the original, or the latest? This is exactly why I didn’t want Mattaus to reset the version numbering in the way he did. You say “original” but in the same breath say “v1.0” which is the designation for the latest version which Mattaus has released.
I can build the original and compare it to the A17DD-S08 if that would be more prudent than the Ver 3.1?
By original , I mean the one with no ground contacts on the spring side, the original , aka, first board.
I don’t really have an opinion on that.
Ok, I’m about to build 2 lights…one with an A17DD-S08 and the other with a BLF17DD Ver 3.1. I will use AVRGCC44 with the A17DD-S08 for the 22,000 R1 to work at 3.0V and 2.8V and the AVRGCC45 with the BLF17DD Ver 3.1 for the 19,100 R1 to work at 3.0V and 2.8V. All other settings are the same.
This has 7 mode levels at 0, 3, 9, 27, 73, 140, and 255.
I’ll be back!
What is a BLF17DD v3.1? Am I missing something here?
The way to go is to use different drivers in the same light, with the same emitter and the same LED wires. That way there is only 1 variable changed. Using different emitters is a bad idea when you're trying to test for 0.2A difference.
Done! I have 2 identical M1’s but different driver styles. Before assembling the emitter/star into the pill (I reflowed them both) I checked Vf on a Sanyo FJ and each had 3.61Vf at 5 seconds on direct drive. These are XM-L2 U3 1A on the new 20mm Noctigon that has temperature sensor pads and the contact pads are closer to the edge (wider stance)
Both have the large bronze phosphor spring from Intl-Outdoor, upside down, and a copper disc soldered to the wide end for ease of use with various cells (flat tops or button tops) No spring bypass due to the spring type and no spring bypass in the tail cap of either light.
So, what kind of tests do we want? Vf on the same cell fresh off the charger each time?
Amperage on the same cell, freshly charged each time?
Amperage in the lower modes simultaneous to the Turbo reading? Or separate from that reading?
Lumens in each level, again on the same cell and freshly charged for each light?
What am I missing?
The BLF17DD Ver 3.1 is the last variation on the 70N03 FET driver, rearranged by Cereal_Killer for an easy Zener modification with Zener pads on the board. The difference between 3.1 and 3.0 is that the Voltage Monitor set-up is taking it’s reading straight off the batt + pad for the spring, before the diode and or Zener. Dan didn’t change the silk screening so it still reads Ver 3.0, but the trace can be seen coming off the resistor bank for the Voltage Monitor.
I am not organized enough, engineer enough, patient enough or willing enough to use the same emitter on different drivers and do all that testing. I have 2 emitters out of the same tape, side by side, that both test at the same Vf on the same cell. That will have to do or I’m done here and will just use the 2 lights and call it even.