Great work. I ordered the GFS16 boards with the hope I can solder such small pieces. If I fail, at least I got BlueSwordM giving it a go. I want it mainly for the reliable power capabilities. Not so much for lights. Has anyone tested the upper limit of these? Total novice here but wouldn’t that mosfet get fairly hot. Especially in the context of its vicinity to the switch. Maybe not such a concern, since you could only run a 50 watt led for so long.
I think the limit for reliability is 30A.
At a Vgs of 3-4,5V, the max resistance if 3,00mOhms.
That means, at 0,003Ohms of internal resistance and a driving current of 30A, that would mean power dissipation would be around 2,7W.
For the MOSFET, that should not be an issue if it is soldered down well. No light has reached that power consumption yet for continuous operation.
Very nice job clientequator, a big 
from me.
It looks like hot air may do the trick. I’m just glad loneoceans didn’t use another DSBGA4 Digital Temp IC if you know what I mean. 
I’m still waiting on Oshpark. Hopefully I’ll have the boards in a day or two.
Thanks for the resistance measurements. This should help out the GXB172 alot.
I’m really impressed with the low resistance.
Thanks for sharing and very well documented clientequator. 
I’d like to give a big thanks to Loneoceans also for sharing his creations with the members of BLF. 
Thank you! Yes this project much easier to solder than Gxb172. if you can do Gxb172, this is no problem.
After using this lighted switch at night and discuss with moderator007, I agree that the design I build is too BRIGHT at night! For the white led on tailcap loneoceans recommend 30k resistor, but this is because he is useing stainless steel switch which has thin gap for light to go through. moderator007 recommend using 100k resistor instead when using clear rubber cap like what I use. Maybe I will adjust it next time!
So, I’ve received my GFS16 PCBs. I now just have to wait for the Omten 1288 switches to come by.
Now, a question.
Where do I actually put the super cap/battery on the PCB that one of the members actually modified?
I would assume it goes where the battery goes (B1, top right pic) but I’m no expert here.
My boards still haven’t made it yet. If you get the super cap working, please post up your results. It might save me some headache.
Oh yeah, forgot to say the problematic PCB is on the left.
Schizobovine used this battery retainer. https://www.digikey.com/product-detail/en/mpd-memory-protection-devices/BK-414-TR/BK-414-CT-ND/2079911
So my guess is that’s it dead in the middle, battery to the round circle and the two rectangles hold the retainer and the switch goes on top of that I guess.
Center circle should be negative and the retainer positive.
Oh, then the build is easier than I thought. I did actually know that the cell was in the center, but wondered how it would get retained.
So, here are the bill of components:
0,70$US each for the boards.
2,33$US for the supercap. Still cheaper than the lithium ion battery, and much safer to solder.
0,38$US for the LDO.
0,24$US for the SMD cap.
0,06$US for the resistors
0,45$US for the schotky diode
0,87$US for the FET
0,40$US for the Omten switch
0,40$US for a single BeCu spring, and 0,80$US for a dual spring
So, if everything goes well, the FET switch should cost, with labor, about 7,00$US.
What battery fits in the holder design by schiobovine? It seem to me that the battery is under switch? This maybe cause the spring to sit out a little high from Convoy S2+ housing because the battery holder is over 2mm thick. In this case, the aluminum case of flashlight may not be in good contact with aluminum of the back of tailswitch. Do you know what battery you are using for that design?
I’m using a supercap for the design rather than a battery, and I’m not going to use a holder for it for the schiobovine design since it’s actually pretty easy to connect it directly to GND:
https://www.skilhunt.com/product/h03-rc-magnetic-rechargeable-led-headlamp/
It’s also safer
And I can use a lower value resistor to charge it since it can be charged faster.
I built one GFS16 fet switch with the rechargeable battery and 2 with the Super Cap that BlueSwordM linked from digikey.
The Super Cap is much easier to work with. It comes drained at .001 volt. It doesn’t care if you short it with tweezers. It with stands soldering very well. You can bend the tin tabs without having to worry about shorting and destroying it. I recommend trying the Super Cap for anyone who might be having trouble with the battery. The Super Cap is almost the exact same size as the battery.
.
This is the SC (super cap) soldered in the same position as the battery. The only difference is the SC is backwards with the casing soldered to the smaller outside pad.
In this pic you can see the SC sticks up above the switch and will hit the original S2+ washer or lighted GFS18. That’s the reason Loneoceans added the oblong hole in the GFS18 led PCB. If using the washer it will need to be modified with a hole for the battery or SC to stick up through, then some way to hold the washer in place so that it doesn’t spin hitting the top of the battery and ripping it off the PCB when installing the switch holder ring. Maybe glue or epoxy.
.
.
I came up with a different solution that could work without hitting the original washer or having to have a hole in anything. I only tried this with the SC because it seems a little smaller and is a whole lot easier to work with.
.
.
The tin pads can be bent if very gentle, so that the battery can lay back towards the switch. Its a little tricky to get the right bend, and a little tough to solder on the outside pad. It can be done though. I bent the the center tab of the SC on about a 45 degree angle and the arrow looking tab on the case I just cut off as it’s not needed. Laid the omten switch in position and checked clearance. Done that a few times bending the tab back further until I had good fitment. Once the SC was lower than the switch shelf I placed the SC in the right position for contact to the pads with no shorting once soldered. I then applied a small drop of UV glue in one corner of the SC tin tab on the big PCB battery pad and turned on my UV light and presto its held in place. Soldered the big pad first with a small soldering tip and then very carefully soldered the case to the small outside pad, with just enough solder to make a good connection. Doesn’t need much as this is very low current.
.
.
Now it will clear the washer or GFS18 with out shorting or having to have a hole for clearance.
.
I used the Q2 LFPAK33 and added the larger Blue spring. If you use the Q1 PowerPAK SO8 then this will make the switch assemble taller but adds more clearance of the SC or battery because the switch shelf is higher. This option would be the easier to build without having to lean the SC or maybe just slightly to clear the washer or GFS18.
I changed R1 to 500 ohms and the SC takes about 1 to 2 minutes to charge up before I could operate the GXB172 in turbo. This is after the battery is installed in the light and screwed together in the off position. The lower modes would work but turbo flashed 3 times indicating low battery voltage, after a minute or so of being off, turbo began to work when the SC became charged. If the light was drawing a few amps I suspect it would work within a minute if the switch is in the off position. On the second SC GFS16 I built I used a 220 ohm for R1 and after screwed together and being off, turbo will work between 30 to 60 seconds. Might could go even lower but this 200 ohm range seemed very usable to me.
Yep, that looks like really nice.
BTW guys, I’m just waiting for the Omten switches to come to start building these FET switches.
Using the bigger Q1 might save you some work Blue. Makes the switch shelf higher to clear the SC.
I built this second one with the leaning SC this morning. I been playing with it off and on all day, it’s worked very reliable.
Very cool fet switch now with a super cap.
I built a GFS16 with a battery and the led GFS18 last night. It works but doesn’t act right, probably something I done in the process of the build. The GFS18 seems to be working correctly, something in the GFS16 isn’t hardly right I guess. I tested the led GFS18 on my power supply and the led color change to voltages were dead on, cut off voltage was dead on too.
Moderator007, this is very cool! Did you see any difference between using SC and battery (after they both charge?) I am wondering what the run time is for SC compare to lithium battery before each charging.
No difference in performance. Just takes a minute to charge. If you unscrew the tail switch and leave it off several minutes it does drain back down and needs to be screwed back on and set in the off position for about one minute, then it works threw all the modes. Once it’s charged a few minutes and turned on, I have clicked and changed modes over a 100 times quickly in a row, it just keeps switching modes like normal.
We have to compare the quantity of energy between them first.
The lithium ion cell has 10,8J of energy.
The super capacity has 0,4356J of energy.
It can store 25x more energy, meaning it can last 25x longer.
One advantage of using the super cap though is it can be charged extremely quickly.
A full charge can be achieved in 9 min with a 100Ohm resistor.
A partial charge will take around 30 sec-1min.
Thanks for the info Blue, gives me something to go by now.
All I know is, It works great and thanks for sharing your idea.
Dont these super caps self discharge must faster than a lion battery? Just comparing the capacity is not enough. Caps are probably way more reliable in the long run though.
These switches may not be a great idea for people who physically lock out their lights frequently.