In case of PWM based drivers, Vf on lower modes is the same as on highest mode,voltage sag is the same on all flashlight parts as on highest mode! That's big disadvantage of PWM based drivers.Check HKJs graphs for 7135 based drivers,on lower modes,current is constant until battery voltage drops to ~3.4V,same as on high mode,no matter what mode you choose.Voltage drops at springs,switch,wires remain the same on every mode.
In case of non PWM (true constant current regulated) driver,on lower modes current will remain constant until battery voltage become lower than Vf of LED at that current,for example 3.0V@300mA. Also there will be less stress on battery,springs,etc.
But,there is even something worse in case of PWM drivers. LED efficiency is the same on low modes as on highest mode!That could mean ~100lm/W for 3Amp max. driver,~80lm/W for 5Amp max. driver and ~70lm/W for driver with direct drive on high.
In case of non PWM driver,LED efficiency is much higher at lower modes,because every LED is more efficient at lower currents.
In practice,that means for example ~150lm/W on medium mode,~200lm/W on low mode.That's 2-3 times better than for PWM drivers,so you could have 2-3X longer runtime on low mode,or 2-3x brighter low mode and the same runtime.
Thanx led4power - that all seems to make sense, and certainly is the downside to these FET based driver designs. The quick answer of "that's why there's low modes" is not quite accurate (I use that explanation myself!). Tivo532's Z8 based designed driver uses FET's for a "turbo" mode, and 7135's for regulated lower modes, which is a better solution. For example you can get the "wow" FET based 5A-6A level, than also get PWM based modes on a much lower valued amp limit of 1.5A for example, so you get the efficiency of the Vf for 1.5A in lower modes, not the Vf of 5A-6A -- basically best of both without having a full blown PWM-less design.
Um, Tom, didn’t he just say that 7135 based drivers keep the Vf at Hi levels even in low? So mixing the chips with the FET’s is not “the best of both worlds” but just more of the PWM world.
What driver is constant current regulated that allows the more efficient low modes? Anything in a 17mm? Anything that will deliver, say, 4.5A on High?
There is no budget one cell driver with this functionality.
…mixing things up, a Pwm pulse delivers every time the full current:
On a nanjg 2.8A
On a FET maybe 6A
So if you would use a FET with 3 additional 7135, you would be able to use all modes below 1 A more efficient and still have the option to use the FET for higher modes.
——
I am unsure if the LED efficient a is the same in a current peak and on average current of the same high….I doubt that. I am also unsure if a XML can achieve 200lm/W…
Yes, it's PWM's, but PWM's off of 1.5A and not 6A -- big difference. Also I'm saying 6A as an example of what you would get with a FET (direct drive), while that 1.5A is based on how many 7135's you populate. His board has 4 spots for 7135's, so you could use one or could use 8 (4 stacked), etc.
I don't know of anything in 17mm size that does regulated multi-levels - you are right, therefore, PWM's off a lower high mode is the best I know of right now in a 17mm size, that's Tivo532's board - it's a compromise solution. His design is in the public also, so anyone can change or improve on it -- of course I don't have that capability. Some guys are adamently against the use of the Z8 MCU for example, but it could probably be designed with an Atmel MCU as well - same concept of FET or 7135 controlled output based on program control.
Another interested aspect of Tivo's design is the ability to run the FET's in parallel, with the advantage of reducing resistance with each one added. Again, you can stack FET's. Now if the Tiny13A FET designed here in this thread could support a stacked FET, would that cut the resistance in half as well? Interesting....
I see how the 7135 chips in lower numbers would allow for a more efficient low, but how do they work together?
Not sure why I try. I’ll forget it all poste haste…
That is exactly what Tivo's East-092 build does, but his doesn't use the ATtiny13A chip, his uses an 8k flashable chip (he also has a 8*7135 driver as well)
It's ALOT more complicated that an ATtiny13A chip and a decent enough FET, plus the 7135's can only handle so much current..
What work together - the 7135's and FET's? You, from the firmware, control which output circuit you want to use - 7135 or FET's. It's two different circuits, throw a switch in the firmware (turbo ON or OFF), and the board operates as a FET output board, like comfy's here, or it operates like a regular Nanjg 7135 based board -- Best of Both, or actually is both...
Ahh, so if there’s a way to send it up in smoke I’d do it through my programming. I took programming in college. In 1980. Do you know how many times I’ve slept since then?
(Crimey but that makes me feel old! My wife was 4 when I was in college.)
[To clarify, she was 27 when we got married 11 years ago.]
Great info guys. Not sure if Comfy's approach will be harmful to the emitter still, but it's not going to stop me from implementing his approach on some of my lights. It's a great solution for existing drivers that don't do what I want them to do.
That is what the PWM function is for, control the overall power to the emitter, 4.2vdc won't instantly smoke a LED, but it will get HOT fast...if you didn't switch it off and on at varying levels then sure it more than likely would smoke...but this is why the PWM and lower modes come in...you can push the LED's at FULL BORE for a short time, then dial it back...the FET's just give you a HUGE pipe to push that current thru, the 7135's (that's why there is 8 gives you 8 little 1" waterhoses vs one big 8" main)
I started college in '75 in engineering, took a Fortran course and loved it. Eventually didn't finish my BS in EE, but went back for CS and got my first job in progamming in '78 at the age of 20. So, been working in software at every possible level for 35 years, more jobs/projects than I can recall. These MCU's are a throw-back to old times for me, but at least it's not assembly language...
While I've been trying to figure these things out I did read a few things on some of the AVR boards about guys writing for the attiny13a in assembly to squeeze more functionality in...that was WAY over my head.
Remember what they were saying back then, while the computer literally took up an entire room? “If the Automobile Industry were progressing at the rate of the Computer Industry a car would cost $20” Man, if they’d only known!
I dropped the programming when my Dad had cancer in ’83, got into the construction side of things and…well, became dumb.
He’s still free of cancer at 85 years old and I’m still dumb….
Been through the IBM deck of cards for writing software on (don't drop the deck! your lines of code will get scrambled), the paper tape thing (yes, I can read binary on paper tape), 8" floppy disks, even bought the original IBM PC with the 4.77 Mhz 8088 (close to state-of-the-art at the time) for $3,500 with a monochrome monitor, single 5 1/4" floppy, DOS 1.0, etc....
This shouldn't be any harder on the LED than any other direct drive driver, if there were a problem it would have been discovered and well-known by now. This is no different than an east-092, it just uses a different MCU that we can write custom firmware for.
, or actually is both...