Very true. 
Thanks for your input.
I'm interested.
Of course I’m interested. A DD driver with high PWM frequency medium and low modes is perfect.
Very interested!
The 3+ amps boost project, although with a lot of energy from various blf-members, never took off very well, it is perhaps too difficult. This, however, is within easy reach of any expert who wants to make it, and is very desirable too :-)
It could be a regulated one too, like the m6 has. Mcu, FET, shunt resistor. So the current gets regulated through Pwm. If someone doesn’t like the regulating he could short the shunt…
That’s what I thought about some times, maybe even some nanjg like stars to give some different mode choices for the mass of people….
You can’t regulate the current on a direct drive driver via PWM. Each pulse will carry the full current. The PWM only reduces the average current.
I did a direct drive driver for my 20,000 lumen salad bowl… An ultra high power direct drive LED driver
I also did a smaller one with surface mount parts and smaller FETs…
Hmm… would it be possible for the chip to calculate an average current passed through the LED and calculate a duty cycle to limit the current to a specified amount?
For example, if 100% duty cycle puts 4.0A to the emitter, how about 75% duty cycle to reduce the current to 3.0A average?
Of course, assuming no battery sag, but you get the idea.
I’ve been wanting to do a direct drive which regulates average current!
Hi,
How small was “smaller”? Also, what were the specs (input voltage, max current output, etc.)?
I would like to see the smaller version too…
I once made an attempt to make a dd driver for a triple XML sk98 but the space was so small I lost the joy on the project, but the tests were awesome. I got more than 8 A through 3 parallel LEDs with not optimal wiring from one cell and I could dim it…I even cut the metal plate from the FET and it seemed to fit, but I ripped of a leg of the tiny while installing, very frustrating…
so a nice layout board would be ideal.
Maybe if there is space we could manage to add any kind of over-voltage protection for the mcu so it could be used with 2 cells for the mtg guys here…this could be the perfect one driver fits all needs thing
It is around 50x50mm square and can handle around 40 volts, maybe 40 amps or so per channel. It really isn’t in a typical flashlight form factor. Has 4 big Deans connectors for connections, etc.
The problem with just reducing the average current to an LED with a direct drive driver is what current each actual pulse is. If your LED would fry connected directly to the power source, it would fry from the PWM pulses. Your power supply has to be close enough to the Vf of the LED that the peak currents are OK. It does not take much excess voltage to produce way too much current for the LED.
I think most aren't thinking of massive currents or voltage much over Vf. Just direct drive setups that leds already tolerate except with added intelligence.
At least thats the use for my version. AdaptiveDirect™℠③®៛❧ツ
Do you mean the spikes that would occur when you turn the FET (?) on/off/on to do the PWM might cause large current spikes and damage the emitter?
EDIT: Could those be “smoothed” using like a capacitor?
All a direct drive driver does is directly connect the battery to the LED. The LED can draw massive current if the battery voltage is even a little higher than the LED Vf. If you PWM the direct drive FET, you get pulses of that massive current.
A cap that could smooth out that waveform would be huge. You would need an inductor to smooth the current… but that makes your direct drive driver a boost or buck converter.
Hi,
Thanks for the explanation… sounds like a “catch-22” :(?
Off-topic: So with the 7135/current-regulated drivers, this (spike) problem is avoided because the current regulator just won’t produce the current rush when you switch the current on/off/on via the current regulator?
EDIT: With the direct drive/FET type driver, can you not cause the FET to just increase resistance instead of using PWM?
Yes, but that means the FET has to get rid of the current flowing through that resistance as heat. Also, FETs don’t have much of a linear operating range. They like to be digital full on or full off. Using them as analog devices can be tricky. You need a proper current sensor and some kind of feedback control loop.
Thanks again! No “easy” solution I guess, which is probably why these things don’t exist (yet)…
I’ve built a couple of them before. Mainly used as electronic variable current loads for power supplies. It actually would not be too difficult to build a driver using the technique… but ’7135 chips are cheap and easy to use.
With the copper mounted xml2 and xpg2 leds nowadays it makes more and more sense to direct drive them off a single Li-ion, most batteries are not capable of delivering the 8+ amps to blow the leds, even a 18650 IMR has enough voltage sag to keep the current through an xml2 low enough not to blow it , and with direct drive you get the maximum output possible for the set-up.