High power LED Cont...
 
Notifications
Clear all

High power LED Controller

8 Posts
5 Users
0 Reactions
3,466 Views
Posts: 393
Topic starter
(@soakes94)
Honorable Member
Joined: 18 years ago
[#18165]

Hi, I'm wanting to control some high power RGB LED's not sure what watts though. Either 9W (3W per colour) or 3W (1W per colour) I had also considered getting a 40W LED from mouser. (10W per channel, RGB+W).

Anyway, I have found this Constant Current Driver, http://www.instructables.com/id/Power-LED-s---simplest-light-with-constant-current/

If I was to use this to drive me LED's, how would I connect a pic16f688 to it? Do I connect a transistor to the constant current OUT of the driver or do I adjust in the input voltage to the driver?

I'm really just after a efficient and cheap way to control high power LED's using a PIC16F688.

Thanks
Steve


7 Replies
Posts: 2541
(@lightup)
Famed Member
Joined: 17 years ago

You could use an existing DC controller, say Renard, DMX like RPM's or the equivalent.
Like a cooking recipe, there are many ways to create a tasty dish using the same 10 ingredients.

Are you "scratch" building your own?


Reply
Posts: 6728
(@p-short)
Illustrious Member
Joined: 19 years ago

One way is to use an NPN small-signal transistor (e.g. 2N2222 or 2N3904) with collector to the MOSFET gate, emitter to ground, and base through a resistor to the output of the PIC in a Renard controller.


Reply
Posts: 2541
(@lightup)
Famed Member
Joined: 17 years ago

I took a look at the circuit used. Since the voltage source was selected to be 4VDC to 6VDC, you could use a 5V power supply to power your PIC and the LEDs, including the driver.

I would modify the circuit a bit so that your PIC drives the current source directly. Here's how I would try it.
Take R1 and remove the resistor from the power supply side and connect R1 to a PIC's output. Since R1 is 100K, there is practically no load for the PIC to drive. Leave the rest of the circuit as is. Now the PIC will provide a high output or a low. During a high output the driver circuit behaves as a normal current source. During a low the current source should be off. Using a PIC as a PWM you should be able to dim it. One possible issue is that a low on R1 may not completely shut off the current source. One would first breadboard the driver circuit and test it while either grounding R1 or connect it to the + side of the power supply.

With regard to efficiency, the less power used the better - until the circuit doesn't work properly anymore. So, the power supply should go as low as possible while still providing the desired LED current. This will require no heatsink. The higher the power supply voltage goes, while still providing the desired constant current, the more voltage the FET will get as a left-over. Since the current is constant but now the voltage across the FET is higher the hotter the FET will get until it requires a heatsink to keep the FET in a safe zone.

Hope I didn't confuse you.


Reply
Posts: 2541
(@lightup)
Famed Member
Joined: 17 years ago

One other thing.
The simple circuit I started with may have problems with dimming at high frequency rep rate due to the gate charge capacitance. Meaning that the value of R1 is too high. It can be lowered to provide more gate current to charge its capacitance, but you must stay within the limits of what a PIC's output can provide. If the PIC can't provide the neccesary current, you will need to provide another means to get more current into the gate. Now your at PShort's suggestion to improve that aspect. Depending how you use the additional transistor may lead to a PIC low output to provide a current source and a high PIC output turning the current source off. Breadboarding will resolve the design before production phase starts. lol

Goes to show that the simplest is not always the best solution.


Reply
Page 1 / 2
Share: