Sure, here is the link to an IR appnote DT94-8 http://www.irf.com/technical-info/designtp/dt94-8.pdf that explains it. Basically if the battery is reversed, the mosfet gate voltage is positive with respect to the drain source voltage, so the mosfet remains off ( I think I have that right). You want to make sure the Drain pin of the pfet is on the B+ side.
Relic had brought up reverse polarity protection on my DIY 1A buck/boost driver. Here is my post on it here DIY 1A BUCK - BOOST DRIVER.
I’m glad to see so many great minds working on this! Now in anticipation of this being figured out, in theory can a quad XPG2 be built with this driver?
XP-G2 is not available in colors. Do you want to individually control different tints or something? Regular drivers can already do quad XP-G2 in parallel or 2S2P configs or 4s1p configs (maybe need to tweak component values for such a high voltage as 4s1p).
This is reverse polarity protection on the S1100....
I was thinking of 4x XPG2 in RBG and a white emitter with control of individual channels. So RBG strobe with each color strobing independently, then white only in say 2 or 3 levels. Essentially a puke light that can function as a real light as well.
Great Pic, that’s about as conspicuous as it gets 
.
There are no such things as R, G, or B XPG2s.
http://www.cree.com/LED-Components-and-Modules/Products/XLamp/Discrete-Directional/XLamp-XPG2 (click the ‘Variations’ tab)
My Cheerios taste like pee! Seriously though, what! Comfy, I had no idea. What are the alternatives or runners up then?
XB-D (red, red-orange, amber, green, blue, royal blue), XP-C (royal blue, blue, green, amber, red-orange, red), XP-E (royal blue, photo red (whatever that is), blue, green, amber, red-orange, red), XP-E2 (royal blue, blue, green, PC amber, amber, red-orange, red), XR-C (royal blue, blue, green, amber, red-orange, red), XQ-E (red, red-orange, PC amber, green, blue), XR-E (royal blue, green, blue), XT-E (royal blue).
Maybe others, you can look through this stuff just like I did: http://www.cree.com/led-components-and-modules/products/xlamp
I only have experience with color XP-E / E2’s and royal blue XT-E’s, the XT-E is a different animal, it’s got a different material for its dome and is much brighter at the same current that royal blue XP-E2’s. Now that you mention it comfy I want to get some other colors of them and do a light in them, l have to check on where I can find them.
Haven’t mentioned this yet but I have plans do a triple emitter setup using 3 XML-COLOR’s for a total of 12 die’s.
Back to the quad setup I plan on RGB XP-E2’s and a white XP-L (if the dome can be modified to fit the optic, otherwise an XP-G2) on a quad board from the Russian guy.
Of and thanks for the pic CC, so that’s a P channel FET? I found a contact board from a solarforce P60 I had to get the schematic from, is an AO3400 enough or do I need a DPAK FET? What’s the difference from N to P?
Well no one sells any color or XT-e other than royal blue and only a few XP-E2 options are available 
AO3400 will explode if you put more than around 8 amps through it, and probably be at risk of failure at 6.5A. The package is just too small to deal with that kind of power, even if the on-resistance is low enough to allow more current through. All the current to the driver has to flow through the FET between drain & source.
Same issue with why I don't think they're what you need to drive the SST-90s. They'll only work if you use something else to limit the current through each one to ~6A or less (and using PWM of less than 255 won't cut it, the part of the cycle where it's on still pulls the max unregulated current, only the average current is reduced).
DTU40N03 is a N-channel, it's switched on by applying positive voltage to the gate pin, and the load goes on the negative path of the circuit (negative = N). A P-channel is the opposite, applying ground to the gate allows current between drain & source on the positive side of the circuit (positive = P). To copy the S1100's design you can use either Vishay 70N02 or AOD510. Both those have voltage/current limits well up into the safe range to handle the full driver current.
Wow that was a better answer than I anticipated, Comfy. If I remember correctly the XPE2 is a good thrower too. I don’t know how that will be affected by a TIR cluster…
Ok XPE2 it is and thanks for the link I’ll look to see if there is more.
Someone here built a 3XP with XP-E2s (I forget who), but yeah, it's not too shabby as a XP-G2 substitute. Even the little XP-G2, with the domes still on, and in the 20mm TIRs, is pretty floody. About like a domed XML2 in an OP reflector. XP-E2 helps a lot with that, even if the total output is down a little.
For direct Fet and linear drives, a small signal PFet (only a handful of ma’s required) in place of the diode, would reverse battery protect the micro per post #77 et al. The S1100 is a buck driver and they chose to protect the whole circuitry with a large PFet (large amperage) that handles the whole of the flashlight output. I don’t think that needs to be the case for this driver.
I’m with the others, in that having a direct drive Fet is going to smoke the non SST color led’s, without some form of current limiting.
Yeah, what a downer. I was looking through the link that Comfy posted and realizing that. The XRC has nice color options but the lumen output for red and green are craptastic.
Still looking…
XPE2 is all right, 350 mA = roughly 40 lumens in blue, 100 lumens in green and 60 lumens in red.
Can I you ask why this is the general opinion? I dont understand. The emitters are rated at 27A each, no battery can come close to 90A and even if it was powered from a PS (or by magic) the FET’s are rated well below the emitters (and if they die thats one thing, they’re <$1/ea, also note there are only 6 remaining green and blue LED’s available to buy @ $45/ea so I really cant afford kill one of them for more reasons that just cost). Should I go with different fet’s (different size or rating)? What if I was going to drive one with a BLF driver, would you be saying the same thing? I know white SST-90’s run great with a 17DD at about 8-9A.
I know how to use the program, I know what pin’s hook to what components but I dont understand the theory of why everyone is saying this is going to fail. Can someone PLEASE explain in LAYMENS TERMS why this wont work? I definitely dont want to throw away $130 worth of emitters but I honestly dont understand, I’m not trying to argue with you guys, you all seem to think the same way (well some people say the FET’s will die, now your saying the emitters) so you must be right. Please guys, help me out with the theory of it… Also I’ve asked several times but havent got an answer, where would the current limiting resistors go? On the FET’s source leg (like how it was done on the old style 15DD) or between BATT+ and LED+ (like on the JB SRK drivers I’m so familiar with) or where? Some other method? I really want it to be a DD FET driver for this one, otherwise if I only wanted a couple amp’s each I would use XP-E2’s running off 7135’s like in my currently running RGB light (and save $120 in the process)
Again I’m sorry for asking the same questions over and over but I would greatly appreciate if someone would give me an answer in simple english, not just post some datasheets and highly techincal talk I cant follow. I love you guys, I’m thankful your trying to help, its not you guys, its me, I just dont get it, I’m not an electronics guy, I use the same laptop I bought new in 2008, and I didnt graduate high school. I’m not any sort of engineer nor do I pretend to be, I learn by doing, not by reading some theory but you’re all so convinced this is going to fail I would really like to know why so I dont have to learn by failing.
Because a SST-90 will pull more current through the little FET than they can withstand. They (the FETs) will burn up or pop right off the board and make lots of stinky smoke. The LEDs will be fine, that's never been part of the equation. The AO3400 can only dissipate ~1.5 watts of heat, and going over around 5.5 amps will kill them.
And, the risk to the LEDs mentioned if you switch to a FET that can handle the current, is because of the absolutely huge difference between input voltage and the LED's forward voltage. The red one for instance could never survive being connected direct to a high drain cell at 4.2 volts, and that's essentially what a FET driver does.
Have you looked at the datasheets for all these parts you're considering?