• You must be a Supporting Member to participate in the Candle Power Forums Marketplace.

    You can become a Supporting Member.

4A output Driver. Fact or Fiction?

dat2zip

Flashlight Enthusiast
Joined
Jan 5, 2002
Messages
3,420
Location
Bay Area
Is it really possible now? That is the question.

With that quest in hand I determined there were two drivers I have that are capable of driving more than 1+ Amps to a low Vf (~3.5V) LED like the new Seoul P7 LED.

The SOB is actually capable of doing more than 1 Amp if the step down ratio is high enough. An input voltage twice that of the output would allow nearly double that of the internal switch current limit which is around 1.5A. A four to one ratio of input to output would allow driving 4A to the output with ~14V input limit.

I just received a new sense resistor that is 0.027 ohms. This is almost 1/4 the value of 0.1 ohms and nets a regulated output of 1.8A on the SOB. Of course that means minimum input voltage is 7V or higher to stay in regulation and to keep the SOB input at around 1A or less.

I've have wired up four Cree P2 in parallel to simulate a 4 die Seoul P7 and am using that as my load for my test configuration.

With one 0.027 resistor I got 1.8A to the LED and I had the supply at 7V. Lower input voltages would still work it would only make the driver work harder and generate more heat. My recommendation is to keep input current max at 1A or so even though the IC is good for 1.5A.

To get 4A you would use two of these. I will try two now and see how it goes and post the results as well.


As for the second driver. That's the new Iracongi or Irakongi. I actually like the latter version. Iracongi is the tiny box jellyfish that are super tiny and super deadly.

The driver is a buck driver and is on a large rectangular board with a rotary switch. Currently it is targeted as the driver for the Makita LXT box that sits on top of the Makita Dock base.
x800IMG_9979.jpg


Here's a picture inside the box to give you an idea of the size of the PCB.
IMG_9991.jpg


The board provides three regulated level control with the rotary switch and is a true white balanced driver meaning low doesn't turn yellow. There is a buzz from the PWM in the current design on medium and low. The driver IC on this board is a beefy buck converter and the board input voltage limit with the correct high voltage capacitors can run off 36V.

This board could drive the Seoul P7 at almost 1:1 input to output voltage ratio. I would suggest a 2:1 input to output or 6-9V input range for a 4A output configuration.

When I get a chance I will test this board as well in a 4A output configuration.

Since Buck converters input to output ratio step down you can take any buck converter and with the proper modifications possibly get the same results. It doesn't just apply to these two drivers.

Thermal management of the driver and LED of course are utmost important as input/output powers go beyond 10W or more.

Wayne
 
The SOB limited out around 1.8A and seems internally it thinks the output is shorted or something like that with internal circuitry kicking in preventing it from maintaining regulation at output levels higher than this. I need to look into this further.

The Irakongi board had issues as well. The driver definitely should not have any issues at these drive levels. It's a beefy driver. There was some stability issues and the driver was outputting 5.8A to the LED instead of the 4A I expected. The IC died shortly thereafter.

I believe I know why and have made changes to the board and ordered new boards to test. I should get the new boards later next week and will resume testing.

I think it would be nice to put one each P7 in the PRCAN vs 3 Rebels. That should up the amount of light each head produces. Depending on the power the PRCAN can dissipate will dictate how hard to drive the P7s. I would think 500mA to each die would be the sweet spot of good output and not too much heat generation. That puts the max output at 2A and wire both heads in series. Actually, it might be possible to wire 3 heads in series at this drive level on the Makita LXT battery.

Wayne
 
I replaced the IC on the Irakongi board and found I still had a problem. I finally chased down the problem to shorted diode.

I spent all evening chasing the final problem and found I had one pin on the IC not soldered. After I corrected that I was able to get the converter to buck down to the psuedo P7 setup and regulate 4A (3.91A) using a single low value sense resistor. I tested with another sense resistor for 2A and measured 1.96A to the LED.

I need to do more testing and determine if heatsinking of the IC, diode and inductor will be needed.

Other than that I am very pleased with my results.

My setup that I will implement will be two P7 in series with max current of 2A or so.

My power supply at 20V was pulling nearly 1A to drive the 4 parallel Cree setup which means input power was 20W. I did not measure the Vf when 4A was flowing.

Let's assume Vf was 3.5V @ 4A. That's 70% efficient. I doubt it's higher than that and it's probably lower than 70% efficient. That means between the IC, diode and inductor there is 6W of power that needs to be dissipated.

Wayne
 
Hi Wayne
Love your work.
Is four Crees in parallel the same as a P7 or should it be two parallel / two series? I ask from a position of ignorance and my box-door is still open if you want me to get back in it........
 
The P7 is 4 dies in parallel. To simulate the P7 I wired 4 crees in parallel to get something similar for a test load on the converter.

Wayne
 
Since the LED current target is 4A the average current in the inductor is therefore around 4A. The input current ramps to 4A and then some during the "charge" period. The duty cycle for the charge period is roughly Vout/Vin or appoximately 3.5V/18 or ~20% duty cycle for the charge on period. The power supply average current is ~1A or so, but, peak current is 4+ Amps.

The SOB peak input current is limited to 1.5A and will never regulate the desired output current.

Wayne
 
Gotcha thanks dat2zip. I hope poor current sharing doesn't bugger them up too fast then.
Good luck with the driver.

It might be worthwhile to make a board with multiple drivers on them. :huh2:

There are already other boards that have multiple drivers. Anyone know of one that might handle the task?

Wayne
 

Latest posts

Top