Discussion thread for the Knucklehead 4A Buck Driver by Mattaus.
The Knucklehead is a switch mode buck driver capable of up to 4A continuous drive current. It can support 3-18V input and will drive a LED/string of LEDs provided Vout is less than Vin. The driver is based around the LED2001PHR IC from ST Micro. It has been designed to use leaded parts as much as possible to make assembly at home easy.
As usual the design may not be optimal and better, more suitable, or cheaper parts may be available. That's the point of this thread.
Also note that the voltage divider resistors used for cell voltage detection (R2 and R3) are 10k and 3k and will not function properly with standard NANJG or similar firmwares. These values are only really suitable for 1S battery packs. The ideal resistor ratio is 100:7. With this ratio at 16.8V (4 fresh 4.2V batteries) this gives you 1.1V to the MCU, down to 0.2V at the MCU for a depleted 1S pack (3V - this driver will not function below 3V). Possible resistor values are for example (R1/R2) 10K/0.7K, 100K/7K etc.
Thank you for creating a DIY buck solution. This seems like it will be very handy driver with all the higher Vf emitters around these days.
A high current buck that has recently impressed me is the HX-1175b that is in the UF-T90 (The SST-90 light) and the Lustfire 3U2. Mine is currently delivering at least 11 amps to an MT-G2 and behaving nicely. I have a few on order from ebay. If you would like me to take some macros or anything, please let me know.
Can this driver be configured to output more than 4A? I’m waiting on the boards and inductors, everything else is in hand.
Would an R010 at R1 give closer to 5A? Or what about stacking a second R025? I’d like somewhat more than 4A but less than 5. Trying to figure out if this will cause any issues…
I have no idea how this all works, but it seems like the buck controller IC itself is limited to 4A maximum. Matt, do you think that we could stack two of these ICs on top of each other, then stack the necessary input capacitors as well to help smooth out the increased current spike?
I asked Matt to start this thread partly due to my own selfish interest and partly to help him outsource some of the research that might prove necessary to get his driver dialed in and find it’s limits so by all means let’s get some parts and try it out. Since he was generous enough to open source a project he invested a significant amount of time in I think the least we can do is share that investment and help it along. Shipping from digikey is somewhat atrocious for a small cart. Anyone interested in doing a gb for 10-20 sets for us to try out? Also, Dbcstm, where did you source the coil craft inductor?
I got the inductors off ebay from the UK. Not budge parts as far as that goes but what the heck, I wanted to build 3 of these and have an immediate need for one to rebuild a Crelant 7G5CS for a friend, the driver is out on that one.
The LED2001 IC is limited to 4A, and you couldn't just stack them together I'm sorry. You could possibly parallel them but I haven't even thought to try that.
Dont use the female ones. Thats my experience anyway. I dont mean all females are like that, just the ones in my life. I think I'm only getting in deeper.
The PWM is a part of the UI flashed to the MCU, is it not? So that will depend on us when we put the driver together.
The first one I build will certainly have the STAR firmware, with an electronic switch keying off pin 3 of the ATiny13A and probably running the same set-up as my M6 with the reversing capability so the electronic switch can take you up or down through the modes, but probably 5 modes instead of 7. This for the Crelant 7G5CS. Or that’s the plan…
200mV. So if you want to run an LED at 3.4V you need 3.6V minimum input. Vout is 0.1V to VIN. So to drive an XML at 4A you'd really want a 2S battery pack otherwise it won't run for very long.
Also note the PWM frequency of this IC is not fantastic. If you want to DIM to 2%, you can only do it with a PWM frequency of 200Hz. 1kHz will give you 5% dimming ratio. The faster the PWM, the higher the posible dimming ratio. That's how I'm reading the datasheet anyway :)