Ummm… That’s pretty much the definition of the amc7135 linear regulator, isn’t it? We typically run them around 16 kHz. Almost all of our DIY-friendly drivers include one or more of them.
I don’t think a 7135 can sustainably sink that much wattage. Above like 5.5V in or something I remember them failing pretty fast, at least not delivering fully 350mA.
They are cheap enough to try out though. Some major potting and heatsinking is going to be needed though.
I could be misunderstanding the question. I was taking the 7.5 volts not as a requirement, but more of “I can’t get to 350 mA unless I push 7.5 volts, and I don’t want to go that high as it’s too wasteful.”
That chip is meant for powering strings of LEDs from a much higher voltage, so higher “overhead” is expected. Think of it as a ballast resistor, only one that isn’t linear resistance.
Use a single 7135. You can run them at higher freqs, but output won’t be linear. Eg, if you want an effective 20, you might have to hit them with 25 or so.
Ahh. You may be right. I guess they’ll have to clarify what parameters are fixed. If a lower input voltage is possible, then heck yea 7135 all day. Just make sure it’s a ‘raptor claw’ not a ‘crappy sailboat’. 
Oh. And here is a source of some 7135 drivers made for flashlights. Most run 16.5kHz or so PWM as is. May or may not work off the shelf for you. Good luck!
Thanks a lot for the info. Any tips on where I can buy the 7135 IC and be sure it is a “raptor claw” version. Is it available in North America? I’m in Canada . Or do I need to order from China? I’m new to the forum and am sure this has been mentioned elsewhere but I can’t seem to find it right now. And no, I’m not married to 7.5 volts. This is just what I was using with the other driver IC. I see from a spec sheet that the 7135 has a low dropout voltage of 120 mV. I guess the ideal situation would be to use a voltage 120 mV above the Vf of the LED . I’m using the R G and B of a color XM-L RGBW so maybe I’ll aim for around 4 volts to have some headroom.
Thanks again!
I linked to MTN electronics website. They are located in NW USA. Here is the bare 7135 chip >> https://www.mtnelectronics.com/index.php?route=product/product&path=25_76&product_id=60
I won’t say I guarantee it, but I would bet highly on it. Like ‘almost guarantee’. They have great CS so you can ask I’m sure 
Thats wrong with step down modules on ebay for 2$?
I assume you are suggesting a drop down converter to get the supply voltage closer to the Vf of the LED?
I wouldn’t think this technique of using the Vdd of the 7135 as an enable pin would be the best way to drive it? I would drive it with a FET in series with the LED, correct? Curious if anyone has tried using shunt dimming with the 7135, putting the FET in parallel with the LED. I know on some switching drivers this allows higher PWM frequencies but I suppose since there is no feedback resistor in the the case of the 7135 it would see a short circuit.
PWM’ing the Vdd of a 7135 works perfectly well. I would bet there are literally millions of flashlights out there that PWM the Vdd pin of a 7135. Some 7135 brands respond to this better than others. As said before, the “raptor claw” (from MicrOne / Nanjing Micro One) seem to be higher quality and respond to PWM signals faster. Others (like the “crappy sailboat”) have a slower turn on time, as so they do a bit worse at low PWM values at higher speeds.
Thanks so much for this. So in other words if you are only using 1 7135 is it peoples opinion on the forum that there is no benefit in using a FET with the 7135?
Thanks again for all the great advice.
That’s all what I said way up-top. It’ll work, but may be non-linear. So if you want “20” brightness, to accommodate slowness turning on and settling in, you might have to hit it with a 25 pwm instead.
Thanks @Lightbringer. Sorry to belabor this but just trying to understand. Are you saying that if I use a FET in series with the LED when using the 7135 I won’t have the same nonlinear issues I would if I switch the 7135 using PWM at the Vdd pin?
If you absolutely must have it 100% linear, then you may want to just put a FET in series. But if 5% (or so) accuracy is good enough, I wouldn’t worry about that at all. For LEDs, I don’t see this having any impact whatsoever. It’s not like you’re designing an exacting-standards medical device, right?
No I’m sure it will be fine. I just wanted to understand where the non-linearity was coming from. I wasn’t sure if there were also non-linearities when the current through the device was switched on and off via the FET. Thanks for having clarified. In terms of switching Vdd I can use use a microntroller pin, right? Probably using a PIC running at 5 volts.
Sorry for the hint of impatience, but what is the full story? What is the application and your goals? Is this a fun project just to learn and build a circuit yourself? Or do you just need a working solution? There are many, many drivers available off the shelf that will likely do exactly what you want with as little as reprogramming the mcu, which there are many firmwares here at BLF to help get you started.
I’m fine with programming the mcu’s. I do that all the time. It’s a pretty specific application so it will need pretty specific firmware . It’s really just the driver I have been a bit stuck on. I’m driving 18 XM-L color LEDs, so 54 7135’s which will be sequenced very quickly. It’s for a hologram application. I’m a bit surprised there aren’t more IC’s like the 7135 by the usual subjects ie. TI, Analog Devices, Maxim etc. Apparently the 7135’s at Mountain Electronics isn’t the raptor claw version but maybe that’s all that’s available out there for now. Maybe using a series FET inline with the LED would work ok at high pwm frequencies with the sailboat version 7135.