I’d go for Samsung 30q button tops.
Hopefully they’ll be available over at nkon.nl soon.
Yes at beginning of OP is link to Yoms cell test results
Grandpa Yom, I like it! 
Only my mother can call me Yom, or Yommy 
Hahaha Yom , how #*## autocorrect suggests Tom You and Ton
So my previous post Yoms = Toms
I am a gambler so i went with Sony VTC6 or whatever they are called, unprotected and flat tops, so i am in the process of evaluating if i should just do a solder blob, or actually solder on a little 2 - 3 mm cobber disk for a more pro look.
But i am a lazy old bastard and i am not sure i can find my little cobber disks in the mess that still prevail after i sold my house 6 years ago and moved into this temporary apartment.
Only opened about half of my moving boxes to get by, the rest are still stacked up and unopened.
At least my little cheap soldering station are unpacked years ago, cuz as you know a real man cant go by without using a solder iron at least every 2 month. 
we still use 7135 on this big light
Why we dont use a driver constant current fet like the one already in this forum(lq-2, i guess)
I am thinking about using MCU, control the DAC (for example, mcp4921 12bit DAC), go to op-amp, control the fet
No-pwm
If only my programming skill is high enough to create some driver :person_facepalming:
Are you talking bout LD-2? It's not open source, great driver, but he chose to keep it commercial, not in the open source community - the cost would be much higher. I don't think it can do smooth ramping based on the design - it's got limitations.
We contribute the board design and firmware and leverage China manufacturers to build them, best of both, keeps the cost down.
The Texas commander is the same priceable and open source. The issue there is that it can’t regulate more then about 2-3A before it would overheat. There are ways around this but with such a large driver it would be better to simply use a buck driver and setup the cells in series. Then you could regulate the power and not use PWM all at the same time.
Buck drivers are hard when space is tight like in 17mm drivers. Just not enough room to use the components you need. The inductor alone takes up the whole 17mm pcb.
The pwm-ed FET driver of the Q8 is simple and cheap to build, incredibly robust and hardly produces any heat at the used 20+ amps. With the almost optimal voltage difference between batteries and the used XP-L2 leds it is also rather efficient.
Using a non-PWMed FET driver a la LD2 greatly complicates the design and makes it very expensive.
(try find an affordable FET that can take the enormous heat of limiting the current to 10A 
)
As far as i know, buck boost driver uses PWM to change the voltage
Yeah, too bad it’s not open source
We have a lot of space for driver, maybe we can try something new instead adding a lot of 7135 IC
The is only internally in the buck circuit. The output current is constant current + a small ripple voltage that is determined by price and size of the components mostly.
In the Texas buck for example it should have less then a 3% ripple IIRC.
The Q8 is finished and is coming with the driver that is already in it.
Any changes would be for a possible change in the production version after the group buy or for a later “S” model. In which case a buck driver like the one in the GT would be perfect along with switching to 4S cells setup.
A fet that can handle that heat is not hard to find
Even with 1V drop on fet, 12A will generate 12W of heat, pretty low, and if we parallel 2 or 3 fet, the heat that each fet must handle is not high, and since the Rds drop, i’m pretty sure heat is not an issue
I think when i have time i will try to make an CC driver(with my elementary coding :person_facepalming:
12W is a TON of heat for an enclosed FET with no way to dissipate that heat. I was desoldering FET’s at a mere ~3-4W of heat dissipation.
Plus a buck driver would be far more efficient, particularly in the low modes.
Yeah, a few post before i mentioned about using buck driver and 2s2p battery configured, but looklike that all things are fixed so i guess this will be discussed in next version
Btw, example before i use 1V vold drop, it’s too high i guess, so in real life, the heat is about 5w i guess
Hey The Miller add 1 more for me for a total of 4 now.