Can not get more than 3 amps from Nanjg 105D driver master & slave

I am currently attempting to upgrade my Cateye EL-700RC Triple Shot Rechargeable LED Cycling Bike Light.

I started by swapping out the three LEDs to Cree XM-L2 U2’s from here:

They were wired in series to support the original 700mA driver. The improvement was very noticeable.
Next, I bought three Nanjg 105D drivers from here: and a 1s4p 18650 3.7v cell holder from here:

I started by changing the wiring of the LED’s to parallel. Next, I wired up one Nanjg 105D. It worked great, pulling a total of 2.85 amps from the battery on full beam and measuring 2.9v across the LED’s.

Then I removed the Attiny 32 chip from the second Nanjg 105D and wired it by following the master-slave circuit diagram here:

There was a very slight improvement in light output and the current draw increased to 3amps. However, the LEDs are rated to pull 3amps, so 3x 3amps = 9 amps in total. I would have expected ~6amps in total being drawn by the two drivers.

To see if it is the batteries that are the bottleneck, I swapped them out with a desktop PSU. Agin, by suppling 4v I was seeing 3amps being drawn, not ~6amps.

Ive checked and double checked the wiring without finding any issue. Am I missing something? Does anybody know how I might increase the current to the LED’s?


It’s been a long time since I did one of these. I seem to remember having the battery wire going to the master board, and then continuing to the emitter. Another wire would then connect the positive leg on the AMC chip of the master to a positive leg of an AMC on the slave. The neg. wires of the two boards would splice together and connect to LED neg.
It’s been a while.
Old Lumens can do this in his sleep.

How are you testing you amps ?

as far as a cheap DMM with thin leads?

everyone else was seeing “around 2 amps” frmo a group buy light… i was seeing just under 1 amp…

later on? when i got THICK bare copper wire that just fit into the leads connection on the DMM?

all of a sudden, all of “my” amp measurements suddenly jibed with “everyone elses”…

just saying…

What batteries are you using? Ultrafire?
Really nice bike light BTW. :slight_smile:

You don’t need power to the slave board, just pwm, LED~~, and ground. There’s no mcu to power up and the 7135’s only need the other connections. The master supplies the pwm signal that controls the 7135’s. The pics aren’t quite clear enough to see problems so you might try checking for continuity errors, mainly that led~~ and pwm are continuous on both boards.

The 7135s need some power or they’ll just be sitting there, dead. A wire should go, at least, from a pos. leg of a 7135 on the master to a pos. leg on the slave. That way they’re all connected in a series.

Nope, not at all. 7135’s run on the negative side of the circuit. Current flows through led- pin to the ground tab with on/off controlled by the pwm signal which is the only power it requires.
Edit- you might be thinking of single mode drivers with no mcu to supply that signal in which case power does go through a reverse polarity diode to those pins.

If you run a wire to from B+ to Vdd(pwm pin) you get one mode only, high. The 7135’s are wired in parallel not series so every chip gets locked on. Easy way to get a single mod driver but it loses low voltage indication since the mcu no longer has control of output.

Hi Folks!
Thanks for all the responses.
@Ouchyfoot, to answer your question, I am using the master-slave circuit from here: Ive updated the link in my main post.
@chadvone and @sedstar, to measure the amps drawn from my 4 parallel Li-ion cells, I used a Fluke DMM 17B. I also used a Tektronix PS281 DC power supply instead of the li-ion cells and unfortunately found the same reading: ~3amps.
@jhalb, yes, the Cateye EL-700RC is a great light. At five years of age, its showing its age as far as LED technologies go. However, it means that I was able to pick one up on ebay for ~25euro ($28). :slight_smile:
@Rufusbduck, thanks for the info! I have checked for contuinity of B- and PWM on both boards but I will check again to make sure. I have removed the MCU from the slave board so I dont think it matters that I have B+ connected to it. However, I will remove it to make sure.
Could the wires im using be a bottle-neck?
If it helps, ive attached some more photos.
You can see the whole set-up below. There are 2.1amps being pulled as shown in the photo on a full charge.
Imgur: The magic of the Internet
Bottom view of driver:
Imgur: The magic of the Internet
Top view of driver:
Imgur: The magic of the Internet

I still say to check the continuity of the legs on 7135s between the master and slave. Those chips are what is going to increase your amps, and somehow it doesn’t register 8 extra chips.

Too thin and too long wires… is the bottle-neck.

I know there is a bunch of ways to add a slave, but this is how I did mine. The above picture is from relics triple. He had a set of wires going to each emitter (still parallel) but you don’t need that because yours are wired in a home run.
positive from battery to master.
Positive from master to LED +.
Small wire (blue in diagram) running soldered to the Positive leg of 7135 on the master, and other end soldered to positive leg of 7135 on the slave board. (I remove the MCU from the slave)
Neg. to LED wire on both boards spliced together into a single wire to LED neg.
That’s it.
That’s how I used to do it.

Oops. I almost forgot. Slave board ground connected to master ground.