Just a warning S21B does not go along with 20->22mm adapter from KD. The adapter is too high and you cannot screw the retaining ring. It accepts only genuine 22mm drivers. Until you find lower adapter or you file it down.
rms,
First of all can't say how the driver cut-off works (never experienced it yet), but the word says cut-off, which is self-explaining.
Cut-off in drivers is one of those things which overall needs room for improvement. In my opinion it is overall set too high (read my P.S. in #3881 post here).
There's still too much actual ignorance about li-ion cells, and besides positive beliefs in this matter I have some straightforward experience. Like leaving a couple scavenged ICR18650-26C cells straight connected to a white led with thin wires for many days (with no driver - closest config to a direct drive setup, less resistance and faster result). Consider the minimum Vf of a white led is around 2.5V or close. This means a li-ion battery connected to a simple direct drive, linear or buck driver driving a white led just cannot get overdischarged. Back to my experience, after disconnection the cells climbed well above 2.5V, which is to be expected.
Another thing I can narrate is my latest experience with my li-ion converted hair clipper. Powered by two TurdFire IMR14500 700mAh (LoL!) cells in parallel, weeks ago I noticed them going weak and losing charge quickly (very high self-discharge). Since I needed it, while I ordered and waited for new cells I recharged the appliance every few days. At some point I forgot to recharge it for many days, and when I noticed it wouldn't even turn on or barely move without load. And still, I confidently dared to recharge it. Nothing wrong happened (it still works right and today or tomorrow will give the old cells their last run). Why? Because li-ion batteries need time to chemically decompose at low voltages. Can't say how much, but must be quite lenghty. I know what I'm saying because I also dared to recharge a couple of the above mentioned ICR18650-26C cells from 1.7V or around. It was a pack of 6 from an old laptop battery and I recharged and tested them all without problems. At last I raised all cells to ≈3.9V and observed their self-discharge over a few weeks. Two cells lost some voltage, and so I discarded them. I kept the rest.
But well, you are free to believe as you like - and deal with the results or consequences.
P.S.: On a side note, overall low lifespan for these (now old) cells. Sold better binned and improved also under other wraps: Shockli and Vapcell, for example. Heter is their OEM manufacturer.
Mon, 11/23/2020 - 13:58
Thanks! 
Had 4pcs 30Q discharge to 0v and left like that for 4 days, no hope whatsoever in reviving that low.
That’s quite possibly correct what Barkuti says but it’s a different story if a light has parasitic drain in an e-switch. That’s different altogether though, only a protected cell will stop that issue
0V for 4 days Sammie 30Qs? Well, you can make a test. Recharge them in some safe place to some not too high voltage, observe their self-discharge over a lenghty amount of time, even properly test their internal resistance. If cells still behave normally they could be used. But this is just because of “science”, in practical terms all of the involved effort is worth way more money than the price of a new pack.
I charged them at 300mA but they instantly discharged back to 0v so I binned them. Had a 4X18A driver fault drain them
Ftfy.
I got the accessory sample for lighted switch ,
But there are two shortcomings, the tail is not waterproof, the flashlight loses its inverted function.
That’s not a problem. I’m sure.
Thx Simon
Just received my S2+ with KY CSLNM1.FY Will upload a comparison between a 2700k SST-20, KY CSLNM1.FY, and XP-E2 amber when I get the chance.
i get some samples from Getian company.
12V 4A driver for L6 is in design ,too
Very interesting.
Is the input voltage going to be 2S or is one cell supported too?
Wow, is there a chance to see the Convoys S11 and M3 with this emitter?
By the way, can you make a long back cover for the 21700 batteries for the Convoys S11? 26650 already withdraw from the market, to buy them all the more difficult and expensive.
This is awesome, chapeau bas Simon!
Comparison of the S2+ KY CSLNM1.FY
All taken on my phone, with fixed white balance to daylight.
KY CSLNM1.FY (S2+) - note the broad spectrum has some colour rendering
SST-20 2700k (S2+) - high-CRI
XP-E2 Amber (KDLITKER E6) - monochromatic
Overall impressions of the S2+ with KY CSLNM1.FY - “beautiful” golden/orange beam, with some colour rendering. Warmer than the 2700k SST-20, but not as warm as the XP-E2 Amber. 5A driver is completely over the top for this emitter, and it gets too hot within a few minutes even on 35% output, thus I’ll need to use it on 10% or 20% for night photography. This emitter would better suit a 3 or 4*7135 S2+.
While I understand there are no 2700K SST-20 tests afaik, the 3000K and 4000K high-CRI tested emitters peaked at 6 and 7A respectively, both with quite high Vf at these currents. So, 5A can be considered fine. With a full battery or close you will be driving the 2700K SST-20 close to maximum performance, but as I say still fine. Concerning the temperature thing, it basically is the same with any other emitter. Tip: you can reduce the amount of dissipated heat by charging your batteries to a slightly lower voltage (3.9 - 4V). This also enhances their lifespan.
30% is 1.5A, should be fine. The drivers also feature temperature protection anyway. If I were to have it, I wouldn't complain at all. Just learn to use it wisely and properly, and enjoy. 
The post was discussing the KY CSLNM1.FY. The mention of the SST-20 was just a comparison.
Also, for my photography, I need constant brightness as otherwise I would have to change the exposure for each photo, thus I need to brightness levels that won’t trigger the temperature control.
I rushed a bit with the answer. :-)
The CSLNM1.FY is a phosphor converted emitter, this means it must behave like the CSLNM1.TG standard white flat.
Now I understand your issue, the @#$% aggressive temperature protection. The solution is here: Convoy 17/22 mm SST40 Driver temperature protection removal. It involves removing the onboard surface mount NTC resistor. Once located, push it by one side with a hot soldering iron tip until it launches. Salute once done, LoL! 
Will this work with the Biscotti firmware?
I just get a nail clipper and cut it, it breaks clean off.