Yes, you are right.
@ Simon
Can you make bigger lighted switch for high output hosts like S21B, M21C etc?
The one in S21A (lighted as on S2+, C8...) is different than on S21B (non lighted).
Also, threads are different on A and B...
i can use adapters to apply 17mm 5A driver
ok ,i will produce 20mm lighted switches
Yeap but those are S2+ topics. This is bigger brother in question.
I was hoping to mix and match S21 family but not likely. :(
I was thinking maybe square threads are used on hosts which use pill and trapezoid where integrated shelf is instead of pill...(?)
If you swap the reflectors between them, how much will that affect the beam shape?
I like your 351D-S21A’s beam shape, but I’m tempted by the power of your CULPM1.TG-S21B (somehow I assume CULPM1.TG is much more powerful than 351D).
Will the OP reflector make the CULPM1.TG-S21B much less throwy and more floody?
I’ve noticed something seems rather strange (unique?) here.
Now in this thread I have first post on three-pages-in-a-row. Page 117, 118, and now 119.
I didn’t plan for that and I’m NOT a Hacker 
Is it only displayed like this on my web browser? Or is it real? Please check page 117, 118, 119.
.
.
Sorry for my off-topic trivia
Just received my S21B and (S2+ Sst40 5000k) 2 days ago. Here’s a quick review.
I love the S21B. Gonna buy another in grey.
Here’s a pic of the S2+ with orange lighted switch (S21B on low beam):
Did someone try 10 degree TIR optic for S2+ SST40 and does it throw more than reflector?
Unlikely, at least as far as TIR optics are built the same way as the bunch I've tried. On the one hand they do not recoup or redirect the radiation emitted at widest angles, closest to parallel with the MCPCB, or close to 90° with the normal vector of the emitter die. At those angles the emitted radiation is quite a bit:
As you can see in the above graph (it is the same or quite close for all led emitters), radiation between (let's say) 60° and 90° is relatively small… but don't get fooled: you must think tridimensionally as the led is. If you imagine a “half sphere” emitting light, now you can see that the area covered by these lower angles is a lot, if only the relative emission is less because you need to multiply that by the relative luminous intensity at those angles. But still, it is quite a bit of light which reflectors make use of, and TIR lenses… right now I think the cheap types don't, or to a lesser extent if anything. Or maybe I am wrong in this and it just is because of their angle. To really be comparable to a reflector throw wise, we'd need ultra tight angle TIR lenses. A 5°/10° TIR makes a incredibly big spot compared to any big throwy reflector, we need 2°/4° or 1°/2° TIR lenses to start speaking seriously. In any case, you get the idea: we need bigger and tighter angle TIR lenses. 
Can you please compare if the reflectors of both lights are thr same size?
I’m thinking of making a quad SST20 or LH351D using S21x as a host , and there is a thread of a guy on Reddit making one using S21A. If the size of the reflectors is the same then I might go for S21B. Is S21B a bit heavy? Thicker shell? From the picture, it looks a bit wider.
They are the same size. I asked just a few posts earlier 
Thank you very much.
Convoy L2 host question: can its long tube be “halved” or unscrew divided into two shorter pieces, and remove one of them to be able to make it shorter and single cell powered? 
It’s 1-cell but with an extension tube, so yeh.
I always used mine as a 1-cell.
What kind of modes this driver have? The 4 modes 0.1–3–30–100 would fit my needs the best.
YES PLEASE!!!