comfychair-inspired quick+cheap+lazy single-sided 17DD FET-driver (poor man's nanjg92)

I made that wrong conclusion myself :slight_smile: just some months ago and thankfully wight explained it to me, so now i wont forget it :wink:

I don’t know why, but i like being wrong on an incorrect assumptions, when someone takes the time and explain how and why to me.

I always remember those times as some of my most treasured learning experiences :slight_smile:

On some boards the stars are dummies and mode group changes are done just after a blink in the lowest mode of a group.

I came late to the party?

can i just think “FET” anytime my first electronics idea is to think “transistor” ? (because thats what happens…lol)

in my “pea brain”, when FETs (and MOSFETs) started getting popular for circuits, i kinda “see” a transistor in the “breadboard of the mind”

I initially had trouble understanding transistors and old-fashioned power tubes other than knowing they amplified signals. And had trouble understanding the difference between them.

my “mentor” (or elmer if you will, lol…) broke my head ice by explaining the transistor as a traffic cop? for example a 10mw signal might come across the collector-emitter legs a 50mw signal… i remember there were “graphs” for selecting a “linear” chunk of the graph, so as to accurately represent the waveform.

=

so… in my pea brain? we are taking a 1A nanjg driver… and applying the 1A output to the base of the FETsistor…. so 5A is flowing across the other 2 legs? and the output of the driver is driving the base of the FETsistor… so our new “line out” is the output of the FETsistor? meaning its drawing its power from the battery?

everyone is talking about problems with PWM output for lower modes? that makes sense to me, as the PWM is just a sequence of pulse train, which constitutes a signal. Transistors had among other specs, a Mhz rating for what speed signal it would handle before destroying the signal (and the rating was a sick joke), you had to look at the graph and select a portion of the graph as an operating region, so as to preserve linearity of the source signal. Source signal here being the PWM pulse train.

Rufus,

You are not kidding. Keeping up with all of who is who is a job in itself. Pretty soon BLF will end up needing a directory of sorts, with credits :) I do not hold a candle to these people and their talents but it sure is nice to able to watch and learn from all of them and you!

Sometimes it’s interesting to see how something came about. That was just my perspective on the events so I’m sure it seemed different to someone not privy to the pm conversations. This forum has been really great about people stepping forward and adding their skills unselfishly. It’s fun to watch. :slight_smile:

BLF is more like an advanced light forum than it is a budget light forum. I have seen on other forums, not all flashlight forums, where it is purchase and show (similar to CPF) but here it is what the manufacturers do but out in the open. I truly wonder how many actually watch and read THIS forum for ideas, etc.

Totally agree with bugsy36. My entire idea for modding and building flashlight is inspired from BLF. I’ve seen on other forums but always come back here. I’d love to watch someone with dedication working on some stuff that result in perfect just as expected and share it with anyone else with honor. Sure thats a nice thing we get only on BLF.

I tried to build one of these Nanjg92 drivers by following djozz’s way just now, and I got very good results. 8)


^The drivers were ordered from Banggood. All the 7135 chips have been harvested for the other use next time.


^Soldering work is much easier than stacking the annoying 7135 chips! The FET was bought from ebay as introduced by our fellow members - but wait, now they were sold out already?!


^22awg wires and spring were added to the driver. The spring was reviewed by djozz here and its resistance was found fairly low, so I didn’t bother to attach copper braid on it.


^Assembled everything in my HD2010.

The modes are all working correctly and the current draw on high mode is 4.9A - 5.0A measured at tail with my INR 20R, and it was not even fully charged at that time.

I am very satisfied with the outcome considering all the parts are relatively cheap, easy to build and produce amazing result. The only ‘problem’ is that the low mode is still quite bright (as highlighted by the others before) and there is a little visible PWM pattern. I am not sure if the MCU can be flashed because its PWM output pin is seemingly different from the Attiny13A on the common Nanjg 105C. The PWM output pin of the Attiny13A is on Pin 6, while for this AK-47 it is on Pin 7?

Thank you djozz and comfychair for sharing such an excellent method with us. :beer:

if you use Nanjg 105C (D)
You can change the firmware of the microcontroller.
Use Pin6 for Turbo mode and Pin5 for Low Modes.
Use on pin5 one or more AMC7135 chips.
You can use this firmware

#define MAX_MODES 2

#ifdef REVERSE_SW //from Lo to Hi
#define SetMod0() Mod0075 //Pin5 PWM 8kHz
#define SetMod1() Mod0350 //Pin5
#define SetMod2() Mod2100 //Pin5+6
#define SetMod3() Mod2100
#define SetMod4() Mod2100 // Mode 5

#else //from Hi to Lo
#define SetMod0() Mod2100
#define SetMod1() Mod0350
#define SetMod2() Mod0075
#define SetMod3() Mod0075
#define SetMod4() Mod0075 // Mode 5
#endif

EBay seller E-best_trade has the ph4030al mosfets listed. As the original seller has sold out, I just ordered 20, the lowest they would go on best offer was 65c each. Will report back when they arrive. mosfet

Crap, now only I see the reply. Thanks for the information, but I am not gonna flash the driver at the moment, mostly because I feel lazy to remove it from my HD2010. :smiley:

I have a silly question here.

Let’s say I have a poor FET with high internal resistance and it doesn’t deliver more than 3A to the emitter, can I just leave another one or two 7135 chips on the board, to sort of lower the overall resistance?

Thanks.

Yes. in parallel

So that would help to reduce the overall resistance (higher current to the LED)?

+350 мА

that is only true for an ideal powersource. In reality the extra current via the 7135 will cause a slight voltage drop over the driver (which in total is the good thing) which will lower the current through the FET a bit. ...or not quite like this but someway like this

Hmm, how would that be considered a good thing?

I still think that a poor FET + 1 or 2*7135 is still better than 12*7135?

[quote=bibihang]

You can connect two FET in parallel

Good idea, thanks!

Update! And a bump!!!
As the original FET seller sold out I ordered 20 from ebay seller E-best_trade Link they arrived in 8 days from Hong Kong. And test good, genuine original.
.
I decided to build a Fet+1 driver. This turned out to be really easy with just a wire link from pin 5 to FET gate.
.

.
.
I used Dr jones basic MiniDrv firmware and added a couple of lines to make it duel PWM. Sorry about the screen shot don’t have any other way of showing.
The output is low(2ma), medium (350ma), high (1.2A), turbo (4.5A), with a 20 second turbo step down and no memory always starts on low.

.