I try to make a sample first ,
What is the maximum current used for SBT90 on the forum? I recently made a drive sample of 18A current, which was originally used for 3 * SST40.
25A or so is roughly the max. As long as the driver has an output channel of a simple high performance FET, it will be fine for the SBT90.2.
PS: was really hoping for a e-switch though 
20A+ for SBT90.2...
Yeah, like TomE said, e-switch would do a magic in all setups.
Is the collar some kind of TIR optics, just turned around?
20A exactly would be my sweetspot, I would be hesitant to go for absolute max and rather go for 90%. Very bright anyway, I assume much better for led lifetime, and certainly more efficient.
I’m also quite interested in this new AA/14500/16340 driver.
Simon, could you please clarify a couple of things:
- Is the new AA/14500/1634 driver available with the Biscotti clone (Biscuit?) firmware, which means memory can be turned off?
- Does the new AA/14500/1634 driver with four modes (1–10–35–100) have memory, or does it always start in low?
I’d like to put this driver in a T2 host with gChart’s booster tail - it would probably become my default gift flashlight.
Sounds perfect to me! I would be more than happy to review a sample 
I dunno the details but basically it's a small reflector with a large center hole, but I don't think the reflector curve is parabolic - not sure, others know more. For mounting it, focus is the issue. Getting the focus, meaning distance off the LED, is important to get the max out of it. It's mounted upside down over the LED to capture the photons radiating outward, and bounces them back onto the LED surface. There's some good tech info on them here on BLF, this is just what I think I know offhand.
Thank you, Tom.
Actually this page shows how it's set up and mounted: https://meadowstar.com/technology/rlt-inside/, the diagram on the right entitled: RLT INSIDE BASIC ASSEMBLY
Nice! Like the sst40 or XM-l2. Hopefully sst40 and your 5 or 6A driver. Is it telescope or twist?
Stack an R050 on top of the R010 in the 6A driver to hit 7.2A, that’s almost the sweet spot of the SST40 (lumens/watt).
i will let manufacturer to produce AA/14500/1634 driver with biscotti clone,
it has memory,
The driver detects the temperature of the PCB board instead of the temperature of the host. In addition, because the heat transfer speed of the PCB is slow, the situation as described by you occured.
it’s sst40 with 6A driver ,it’s twist
Huh?
0min to 9min the LED is just heating up, getting less efficient and throwing fewer lemons. That’s perfectly normal. Especially only 3600 to 3400 or whatever.
At 9min it looks like a timed stepdown to 50% from 100%. Is that what you’re irked by?
That or “Oops, things are getting too hot, throttling down to 50%” (thermal) stepdown.
Ie, it may be “thermal” stepdown, but not running closed-loop. Ie, it doesn’t seem to be throttling down only enough to keep the temps in a safe range, but immediately kicks down to half-power to keep things from frying.
At least that’s what I’m seeing.
According to your standards, this is indeed a failure, or you can say that this temperature control is not humane.
If you think I’m saving costs, I don’t have to argue. After all, people have different perspectives on things, and the most ideal state is that everyone agrees.
If you think that I am deceptive, I object, and I may not explain the temperature control clearly enough.
A flashlight cannot satisfy everyone’s needs. For example, you prefer a flashlight with a longer brightness time, and the temperature represents the temperature of the flashlight head. I can design the corresponding flashlight according to these requirements.
Zabacha, is that you, Zero Air?
Your (temperature) measurements are inaccurate,
You might want to do some research before posting your prejudices and false accusations, since i have done those tests and wrote that review
That’s sad. I’m a big fan of push pull zoomies.