GFS16 - Battery Indicator Tailcap Light & 1mR FET Tailswitch! (Rev B)

Tried to build one, and finally managed after destroying the first two batteries (puffed up, I assume because shorted). This was despite using kapton on both the battery and tweezers (in case it slipped off the battery a bit like it did for the first one).

The issue is that the side of the battery is still + so you have to scoot it as far as possible away from the negative contact on the board, else you can trivially solder bridge the battery (the fate of battery #2).

Third try was sorta the charm (as in it worked), but when I broke out the milliohm meter, it never did better than 33-35mOhm (it was near 50mOhm if when I clamped to the top of the 45% IACS unbridged spring from MTN). I compared a bridged plain Omten 1288 (and not one of my better ones, either) and got 19mOhm.

And then while trying to install in a light, the negative contact ripped right clean off the PCB from the most minor pressure against the battery (it was encased in two layers of kapton, btw). Iā€™d liked to have run a test with spring bypassed for the most fair comparison, but after 3 batteries and nothing to show for it, I was done.

If youā€™re using a S2+ host, youā€™ll need to trade your metal washer bottom plate for something thinner, or the retaining ring for the spring assembly wonā€™t sit flush with the shell (risking a short, depending on how much of the bottom of your battery is a contact).

If you could use the through-hole legs, this would be MUCH easier to assemble and try to use ā€” and much less likely to rip apart just trying to install it.

This is a neat idea, but the battery choice / mounting is too fragile, making it practically unusable. Pretty disappointed. :frowning:

My (attempted) design change. Not yet assembled or tested, but idea was to switch out the LFPAK56 FET (so only uses the LFPAK33 FET) for a coin cell retainer that the same sized cell should fit (4.6mm). The pathway to load it has to be free(ish) of components, and I tried to keep the components out from under the switch otherwise. Comments/feedback welcome (trying not to be just a negative review, but rather a suggested direction for improvement).

https://oshpark.com/shared_projects/paPrb43Z

EDIT: Fixed error in first revision (via that was too close to the spring contact area)

very good job schizobovine! I hope to see this project work for you.

In your new pcb, does this mean the switch is sitting on top of the small battery? Will this be bad for battery?

when my gxb172 driver is finished, may try to build this tailswitch also!

So, is it possible to only get the FET switch without the battery needed at all?

That would be really nice, as I only need very low resistance.

this is not possible

Oh sorry I meant without a rechargeable cell.

Would it be possible to just use a non rechargeable cell?

It is possible, but you may need to replace cell when it is empty. Rechargable battery avoid the need for this. I also have pcb order for this to finish my s2+ flashlight, hope it work out since schizobovine had a few problem with the batteryā€¦

Do you think a super capacitor could run the fet for very long?

It could be recharged quickly and most of the time my small convoys are on less than 10 minutes continuously anyway.

I wouldnā€™t write it off, Iā€™ve got a project Iā€™m in the midst of right now running the entire driver at the tail of the light off a capacitor bank charged up through the bleed resistor(s).
If I can run an entire driver off caps I would think someone could run just a FET, and just full-on/off in the same manor.

50ohms of resistance at the head (needs to be rated 1w or using a bank of 1/4 or even 1/10 Watters) and ~350uF at the tail has been working on the breadboard without issue. Time will tell when I switch to a real PCB with smaller package caps but folomov did it so Iā€™m betting I can too.

I bring this up here cause like I said, I donā€™t think it should be impossible or even difficult at all to run the switch FET without a battery, just need to do a lil backyard EEā€™ing.

Oh also lumintop does it in the Ti Tool AAA (but their implementation isnā€™t all that great TBH)

I slept way too late yesterday after doing tons of work, so sorry if I sounded tired.

Come to think of it, couldnā€™t we just use a small super cap instead of a lithium-ion battery?

In this small size, and a lower value bleed resistor, it should work very well actually.

Ah! Found the battery from a Canadian supplier:
~~https://www.e-sonic.com/product-detail/seiko-inst-battery/ms412fefl26e.html~~

Nice :slight_smile:

Scratch that. You literally need to order 100$ worth of stuff to get it actually shipped to you.

Again found a replacement :slight_smile:
https://www.mouser.ca/ProductDetail/Seiko-Instruments-Micro-Energy/XH414HG-IV01E?qs=vrQyzjTsotmXvC7BtxG71Q%3D%3D

So, Iā€™ve finally decided to make about 10 of these FET switches.

Iā€™ve finally understood of how this works, but one thing is still unclear to me:
Is it possible to use a super capacitor with a very high value resistor to prevent overcharging to the cap?

Itā€™s the only component that Iā€™m unsure if I should actually order.

Super caps are usually too big to fit tail switch assembly!

This one is the exact same size as the battery lonoceans linked though, which is very nice :D:
https://www.mouser.ca/ProductDetail/Seiko-Instruments-Micro-Energy/XH414HG-IV01E?qs=vrQyzjTsotmXvC7BtxG71Q%3D%3D

And it can actually be shipped to any country, unlike a lithium-ion cell :slight_smile:

It looks like the capacitor is rated for 0.3 volts more than the battery and 50X the charge rate as well. It should be fine the way it is

The stored energy will be about 6% of the battery, might want to bump up the charging rate.

Is there a schematic somewhere? It would be nice to charge the capacitor to the full 3.3 volts if it doesnā€™t cause other problems.

Yes there is:
https://www.mouser.ca/datasheet/2/360/seiko_XH414HG-IV01E-1202874.pdf

According to post 24 it is being charged to 3.3 volts and the charging is limited with a 2.2k resistor in series with a diode, looks like changing
the charging rate is pretty easy.

The spec sheet for the battery says it has ā€œcontinued stable capacity characteristics even after the battery is overdischarged down to 0.0V. ā€
So maybe its possible to drain the battery to keep it from shorting out and destroying itself while its being installed.

Ok then.

Iā€™ll be ordering the PCBs first, then the components.

Which PCB is the best, or rather, the most appropriate?

Iā€™m quite bad around PCB design, so I donā€™t know if the Rev B is the most appropriate, or what shiczobovine did.

If your going for the lighted indicator fet switch then rev.B has a lot lower quiescent current. I think loneoceans also added a second fet pad in the middle of the spring in rev.B. Are you building fet switch with the indicator or just the fet switch?

Just the FET switch. I like lighted tailswitches, but I donā€™t want to do additional work.

I just want to build 2 FET switches for me, and sell the remaining 10 to people who canā€™t, or donā€™t want to build these amazing switches.

So, would the Rev B be the best bet?

Iā€™m ready to order the PCBs, but not the components yet.

Iā€™m probably going to sell the rest with my dual BeCu springs so other people can enjoy it.