Noir, it sounds like your expectations are higher than current technology allows.
I hope that in the near future Lexel will have his high powered boost drivers working with NarsilM. Then you can set your own levels and your own temperature limits.
I think his Bistro UI works on them, but I’m not very familiar with Bistro.
I think building your own light is the only way you’ll get it to your personal liking.
Hello maukka, your measurement indicates the max temperature of Mid is around 65°C, but with my new EC65, I got only 43°C (the max temperature measured around the side switch, during my whole Mid runtime test). My experiment is done under room temperature (25°C), with no cooling.
That is pretty weird, or maybe I should measure the temperature around the head instead?
Here’s the light at 25 minutes on mid. Already too hot to hold.
Thank you, maukka. I’m sure that my new EC65 can still be held after 30 mins of Mid. Its body tube was still under 42°C then, and I did hold it to feel if it’s too hot (it was not). So there are indeed some differences.
I don't think that is the case. A lower temperature limit is just a different value that is set in the firmware and has nothing to do with advancements in technology.
Technology is also not the limiting factor when it comes to improving the driver slightly. I am not talking about huge leaps like 5 A more in turbo or twice the efficiency, but small improvements that would allow the turbo to be sustained down to 3,3 V or maybe 3,0 V instead of just 3,5 V. Small things like for example double or shorted springs (lower voltage losses) would already help in this regard (for some reason Acebeam didn't bother doing this), or simply a slightly lower maximum (turbo) mode which can be sustained better.
As it is the EC65 (or rather its driver) is IMHO a bit too "on a knife's edge" (for lack of a better term).
Very true. If you want it done right, you have to do it yourself. One of the reasons I'm into modding 
.
Thanks for the great and exhaustive review!
Not sure if it was mentioned earlier, but the photo of the package you posted shows peak beam intensity of 10,600 cd.
That is for the Nichia version.
Good news, the Samsung 21700 30T not only didn’t break anything, it was able to cycle full turbo seven times before stepping down to the lower turbo level of 3150 lumens. Initial turbo lasts for 40 seconds before stepping down, which is quite a lot better than on one of my original Acebeam 21700s with 12 seconds.
The internal dc resistance of the 30T is only 13 mOhm.
I’m sure performance would be even better with that updated version like Toobadorz got. Have you talked to Acebeam about getting a replacement?
Mine is already a replacement after the first one broke on turbo, just like Toobadorz’s. I’m sure the difference in performance is due to the battery sample difference as that is already large between the two I have.
Okay, I must have forgot you mentioned it.
Is your turbo still at 3700 lm? Toobadorz said his was now closer to 4000 lm on the replacement light (for some reason).
I don’t have numbers for the first light. It could have well been lower than the 3750 on the new one.
edit: scratch that. The first sample was 3456 lm @ 0 sec.
I’m still thinking this light could benefit from spring bypasses.
I’m not an experienced tester and I’ve only ever tested one light that had a boost driver. I did however notice that as I was using a heavy wire across the tail cap and a clamp meter to measure the amp draw on turbo that it extended the turbo run time.
I further could watch the amp draw decrease as I compressed the driver spring.
Have you noticed any similar behavior?
Also, what battery do you think is best for this light.
The Acebeam doesn’t give much full power turbo runtime, but does give extra run time at lower levels.
The 30T gives a lot of full power turbo run time, but it’s 3,000mah capacity runs out pretty quick.
Maybe the Lishen/Liitokala 4000mah is a good compromise?
Also, that 20A protection circuit is the biggest I’ve heard of. To answer Noirs question, that’s probably why they set that 3.5v turbo stepdown limit. If they tried to set it to a lower voltage it would probably try to draw well over 20 amps from the battery and maybe burn up the driver in the process.
If there is a measurable difference using different batteries, then a spring bypass should lead to enhanced performance or not?
At 15 amps a spring bypass is always helpful.
The EC65 tailcap spring has a resistance of 17.76 mOhm, which translates to 0.266 volt drop at 15 A.
That’s quite a lot, as in significant.
And as the spring heats up, the resistance goes up. 10 amps for 30 seconds and the voltage drop is 0.276 V scaled to 15 A .
It’s using a phosphor bronze spring, right?
Then, either using a special alloy berrylium copper spring or a spring bypass would be the best way to go.
I think it should. When the driver pulls a heavy load from the battery, the voltage sags down. The more resistance, the lower the voltage. All of this extra resistance causes the driver to work harder.
As you can see with the 30T battery, when placed under a heavy load (15A) this battery can maintain a higher voltage level.
A lot of lesser load (10A) boost driver lights use double springs to reduce electrical resistance and keep the voltage high.
Since a boost driver is trying to put out a constant wattage, Ohms law states V x C = W. This is voltage times current (amperage) = wattage. So as a batteries voltage runs down, the current draw will go up (but only so far).
I suspect if the springs were bypassed, reducing resistance, then voltage going to the driver would be higher and it would run turbo for longer times.
Look at what the extra resistance of the 18650 adapter did to the performance. It hurt it a lot. Therefore, my theory is bypassing the springs would definitely help improve performance.
I’m surprised Acebeam did not improve the springs. This seems like an error on their part.