DIY simple AMC Driver for MTG2?

I am trying to figure out how to build a driver for the MT-G2…

I am thinking of doing something similiar to below… I was thinking of doing 2 groups of amc’s first group of 12 for approximately 4 amps… second group of 8 AMC’s off a switch that parrallel the first group.

So when the circuit is first turned on the MT-G2’s receive 4 amps…
When the switch is flipped and the second group of AMC’s is turned on the MT-G2’s receive 7 amps.

At 12.2 volts about 1000 lumens per led.

At 4 amps I will need a supply voltage of approximately 13.4 volts before the AMC’s start to regulate current and heat up.
At 7 amps I will need a supply voltage of approximately 14.2 volts before the AMC’s start to regulate current and heat up.

When the supply voltage drops below 10.5 volts or so, the leds wont light up… I am ok with this…

Does anyone see anything wrong with this idea? Also could I add a diode/Resistor as noted with the question marks to protect the system from random voltage spikes? If so what…

Also how many milliamps would 20 AMC’s require on their green leg to turn on?

Also I guess I could install a voltage boost circuit before this if I wanted to keep the incomming voltage at about 14.5 volts…

http://s15.postimg.org/ft9iip623/Driver.jpg

Thanks in advance for the feedback… I’m new at this…

I would keep the input voltage to two series cells (you can add more in parallel)* and use two Nanjg 105Cs, one with MCU and protected by a zener/resistor regulator acting as master, and another MCU-less as slave. Each set of 8 (or 12, if stacked) 7135s running one LED.

For MTG2/7135s you need a source that's around 2.5-3v higher (when fully charged at the start) than the Vf to get a decent regulated runtime out of it, before the input falls below Vf and current starts to plummet. A difference of 3-3.5-4v would be better but there aren't any commonly available cells that can get those numbers, with two cells at 8.4v we just have to make do with what we have, it's not ideal though. A little more would be better.

*that's assuming you're using individual li-ion cells, if you've got something like a 12v SLA you won't have enough voltage overhead... MTG2's Vf at 3A is 6.2-6.3v, only goes up from there.

Thanks comfychair… just ended up messaging you on the other forum. :slight_smile: Someone told me to talk to you… its for a motorized project, so there will be alternator voltage… so ideally around 14.4 volts… if there’s not enough overhead, I could put a boost converter ahead of the circuit and tune as required…

I’ll do some reading on modding the nanjg 105c’s … Which component is it that controls voltage to the atmel? and at 12v in i’d be close to the limit of what the amc7135’s can handle right?

I'm not on The Other Forum, so I don't know who the message went to... :D

7135s don't care about the input voltage, at all. They have no connection to the positive side of the circuit, so they can't be affected by it. They only go between the LED- and GND. The difference between the forward voltage of whatever LEDs are connected, and supply voltage, is the only differential voltage they are exposed to. Well, across their current-carrying pins, anyway. The voltage supplied to the MCU determines the voltage that will be applied to the Vdd control pin on the 7135s, so if you only give the MCU a max of, oh 5-5.5v, the voltage sent to the Vdd pins will be in the same safe range as well.

If you have lots of room, i.e. if this is fixed lighting and not a handheld portable light, just use a 7805 voltage regulator to power the MCU and be done with it. Let the voltage drop across whatever string of LEDs drop the supply voltage to a safe range for the OUT-to-GND pins on the 7135s.

Or, if this is fixed lighting and you have room, why not use a DC-DC converter board as a driver instead. There are boost, buck, and a few boost-buck models all over ebay for cheap. Most of them are adjustable for output voltage (which will determine the current thru the LEDs) or for both voltage & current. A converter will be much more efficient than 7135s.