Maybe the driver thinks you are double tapping instead of a single? Try taking it apart and doing a good cleaning. An intermittent connection might be tricking the uC. If the head is loose enough the wiggling can make it jump to momentary mode and back. When it does jump back it will resume the last brightness. If you pocket carry I would make sure it is not near the point where it can toggle back and forth. But I would start with a good cleaning too since it won’t hurt.
thanks for confirming
I have seen the switch refuse to turn off, 3 times in a row, producing only lowest mode, over and over. I thought it was due to some part inside being warm from using maximum
but Im getting similar off click acts like double click, with CR123. And I have read reports of the light not turning off and going to low in old threads…
all that to say, 11.2 is no my first choice
I would rather use V10 drivers…
I still hope its just a stiff spring syndrome with the v11.2 lights
but unless I can find a softer spring… the 11.2 lights are imo, entertaining, but not predictable (thats PC talk for unreliable), hate to say it… but, a spade is a spade
correct, the stiff spring lights, which are all v11.2, confuse a single tap as a command to go to minimum… the only way to avoid it is to press deliberately, firmly, and not too fast, afaict
otoh, the softer spring models, all v10, are a joy to operate, and predictable…
my lights are clean and new, not in my pocket, just toyed with on my desk.
I bet they sacrificed a debounce capacitor for the sake of board space. You can get away with that but you need to account for it with extra uC code. Still not as good as a capacitor though IMO.
Yep clean it is… I didn’t like how hard it was to actuate and how gritty it felt when doing so…. I set out to improve this and maybe solve the problem … but it was very clean out of the box I was impressed.
The pill was taken apart and completely polished before re-flowing new LED. Even the spring was polished inside and out on/with the buffing wheel and the sharp ends of the spring de-burred. ( relieve mechanical friction).
. The inside of the spring was polished using a Leather wheel and then a buffing wheel on a Air Pencil die grinder.
. Of course the Piston and Bezel ring got polished also, bling mostly 
. The contact ring was de-burred (the edges were sharp/burred and the split gap too) with a rubber diamond embedded hand hone/stone if you will. Maybe the contact ring was hanging up? It is a split ring.
. It actuates buttery smooth now…even tried out different piston springs…
. I found I could turn the driver/pill/ reflector assembly by twisting the driver, also I could push/pull the driver/reflector assembly when assembled in the head. I added a thicker o-ring to the Lens/ reflector assembly to tighten it all up, thought maybe the axial slop might of had something to do with it when actuating the piston? Nope…
. maybe a glitch in the driver/mcu…like a timing issue? I noticed sometimes the ramping takes some time or hesitates to start to ramp up or down and sometimes it takes several hit’s to shut off…
TBC…
beautiful work
my guess is the bottom of the contact ring hanging up is a likely cause for the light reading a double tap from an intended single press to off
making the contact ring travel smoothly without hanging up seems like an priority focus.
so, I pulled the battery out of my ex11.2 and just pumped the button a few hundred times while watching a movie… its much smoother now, and does not hang up…
Im hopeful not to see any more unintended failures to switch off, going to low instead by mistake… thumbs crossed
the v11.2 driver has a lower low and a lower parasitic drain than V10, but uses more parts and has three added features that work from OFF
I have never had a v10 fail to turn off, nor mistake a sigle press for a request for lowest mode, the way my v11.2 lights have done.
I suspect there is a component on the V11.2 driver that gets confused, in a way the V10 does not… unconfirmed…
I believe the springs on the contact ring are also softer on V10 than v11.2… possibly allowing a less ambiguous single contact signal, when pressing once for off
speculation
I note the EX piston has a longer travel before the contact ring is fully compressed than my D10. this suggests the D10 pistons are slightly longer, and there is a smaller gap between the end of the piston and the contact ring.
Funny never had a failure to feed… I mean turn on… if it was mechanical you should have failures to feed… the LED…
from off
I get failure to feed reliably if I dont press the EX piston down deep enough
I also get momentary on and off if I dont hold the EX piston down long enough
from on
I get failure to turn off, and I get low instead
I suspect caused by my switch press and release timing being too fast
that sounds like a good way to also move the contact ring closer to the piston edges
So there is a shallower travel to make contact… I think that is a good thing… imo the EX piston travel is longer than necessary for a crisp switch action
the EX piston travel is longer, more like a DAO trigger
the D piston feels shorter, more like a SA trigger
Never had a failure to turn ON from OFF …OR
. Turn ON after setting the ramp to any Level….
. OR after a battery change where it turns on in LOW then I ramp up…
. I have had it making contact to ramp, it would dwell for a bit, then start to ramp…
. Never a failure to turn ON… just OFF
. Got any pics of the other side of these drivers… I had mine out but forgot to take pics…
no, I have not figured out how to pull the driver out of the pill
any tips welcome
I used a skinny pin punch and went thru one of the anti rotate holes where the reflector centering device locates into on the pill….low level light and magnifying visor recommended
. More or Less what I’m seeing is not a mechanical contact issue… like the split ring is missing the signal ring on the driver… in my case it is NOT, or else it would or should have failures to turn on… and it should not take 2-3-4 hits before the light turns off in any mode… imho…
I would like to learn more about the LiFeP04 option for these Piston Drive lights.
I was talking to an old timer, who told me he never used 3.7v LiIon in them, only CR123 and 3.0v LiFeP04
here are my notes so far, mostly questions
Does LiFeP04 need a special charger? Yes
Can I use my Olight Universal charger? No
Do I want Protection? Yes
What charger should a LiFeP04 newbie buy, that can also charge eneloop and 3.7v LiIon?
here is a kit that offers unprotected cells, and a maybe dedicated charger… neither seems ideal
Got LiFeP04 and a light meter?
Can you report regulation in EX11.2 with that power source?
LiitoKala Lii-100
Edit: from the lii-100 listing on AE “Automatically identifies 1.2V Ni-MH / 3.7V Li-ion battery automatically. Charging mode for 3.85V b Li-ion and LiFePO4 batteries must be selected manually”
I’m not having issues with the switch at all on my EX11.2. There is a slight delay before ramping, but it’s very consistent for me. I’m using a protected 16340 though, which adds some length in the tube - I think it’s possibly a bit helpful there. My only concern so far is how quickly it gets warm at max. But no funny behaviors.
Of course, I don’t have a light meter to check if the output has variance as demonstrated by jon_slider
Your Lucky… should go play the lottery… Yep the problems are real…
Lights not shutting off, UI getting wonky, Battery drainage, Ramping anomalies ect.ect.
. I go to the back of these threads and read towards the front for the good stuff…
Some of the posts mention repairing nitecore piston drive lights - I didn’t read all the way through, does anyone mention how? Might be of use.
I read the whole thing… no mention of a repair, just replacing some lights solved some of the more severe issues, such as 24 hour battery drain. I had one of those, and yes, replacing solved that problem.
There is no reported fix for the lights that go to low instead of turning off.
here are my selected quotes from that thread
===
https://www.candlepowerforums.com/vb/showthread.php?312316&p=3771800&viewfull=1#post3771800
[QUOTE=okwchin]The driver is only regulated between 1.7 and 3.1 volts. Anything above and below that the torch runs un-regulated. So running a NiMH, your running way below regulation, so your not getting full brightness, while on any rechargeable lithium, your getting a HUGE direct drive current. So accordingly, max on NiMH is disappointing, while Min on Lithium rechargeable is crazy bright.
The old D10 engine was regulated between 1.3 and 3.8 volts or something like that I vaguely recall from my testing. FAR more appropriate for the type of cells were using, and therefore produced far more predictable results. This D11.2 engine is just… annoyingly lacking in key finishing elements of the design.[/QUOTE]
[QUOTE=bedazzLED]OK, I’ve tried and tried to like these two lights but I have finally had enough. They have both been back and I have replacements. Although the battery drain is not the problem now, the lights are still just plain old flaky. The PD sometimes works, the shortcuts to hi/lo more often than not don’t, sometimes turning it off actually turns it on to its lowest setting instead off, ……
I love my EX10 and D10, but I’ll be stuffed if I’m going to spend any more time with these two lights. Nitecore is off the list for me now, even though these lights look good and are good when they work.[/QUOTE]
===
my conclusions are
I like the v10 firmware without blinkies and lockout, and especially the D10 host, for its short crisp button response
the ex10 ti that started this journey for me, has a mushy button by comparison. I can cause failures to switch off, or on, or to be momentary on/off, if I press too fast or too shallow. otoh, the springs are not too stiff, I find it pleasant to use.
I find my EX11.2 buttons mushy and overly stiff by comparison.
There are failures to switch off, where I get low instead of off. From off, I sometimes get strobe or lockout when trying to turn on again…
The frequency of the switching off failures goes down if press more slowly and deliberately.
I no longer use 16340 LiIon, after seeing the runaway ramping, I switched to CR123.
CR123 drop to 5 lumens max output at 2.3v, which tells me any LiIon used should be protected, and or voltage monitored responsibly by an alert operator if using unprotected cells.
I expect 3 volt LiFeP0 4 will run regulated, unlike 3.7v LiIon
my EX11.2 have regulated output on CR123
I really like that they go down to 0.01 lumens low.
My D10 w sw45k is flawless in terms of function and button feel, using AA Eneloop. The low is 1 lumen, I wish it was lower.
My EX11.2 w 219b 3000k is a good nightstand light loaded w CR123, and the lowest level is more than low enough for my firefly needs
My EX11.2 w 219b 4500k is a fun little light, it may find a home in my car, loaded with CR123
I think the ramping UI is very cool.
The Fix……Need a 15mm FW3 version driver that you can hand fit tightly (since it’s a press fit driver) to about 14.7mm IIRC……
update on my EX11.2 lights switching, and output regulation
since I stopped clicking like a madman I have not had any unintended results with my EX11.2
It has turned on and off as requested. No accidental low modes instead of off, now that I have learned to use the switch deliberately.
To me, they are great little CR123 lights, nice stable regulated output…
now that I have learned not to use 16340.
Easiest light to mod to 219b, of any I have done, with a great beam.
example of the Total Package… clip, glow marker, and a firefly of 0.01 lumens… tics all my indoor and nightstand boxes
Happy Mothers Day! 
some conversion output comparisons, the battery is not fresh, Im most focused on the percent change between LED options…
ex11.2 #1
stock xpg
cr123 @ 2.91v
low <0.005
max 145
shaved 2700k lh351d
cr123 @ 2.91v
low 0.01
max 116
20% less bright than xpg
219b 4500k 9080
cr123 @ 2.91v
low <0.005
max 138
5% less bright than xpg
219b 3000k 9080
cr123 @ 2.91v
low <0.005
max 106
27% less bright than xpg
