hi everyone, im fairly new here (post # 2), but i thought it would be good to get into a good online community for electronics as it is my livelihood (well, will be once i graduate... I am currently a student)

basically i am designing a microcontroller-controlled (sic) LED adjustable-color room light. basically put a few strings of 1W or 3W Red, Green, and Blue high-power LEDs (probably luxeons or something like that) together, and use a microcontroller to PWM each channel (RGB) to adjust brightness and color. the input will probably be something like a potentiometer through the ADC so there is a dial that just adjusts color. also thought about doing a capacitive touch-sense pad where X changes color and Y adjusts brightness... BUT I'm still in the design stages and little extras like that can be ironed out later.

what I need now is a way to drive high power LEDs from a PIC. Obviously I cannot drive them directly; current requirements are too high. I would need some sort of amplification or at least switching, where the PIC controls the gate.

I picked this project because I don't forsee the amplifier stage being particularly tricky. I am about to take a formal class in FET operation and amplifier design (followed by a class in BJT operation and amp. design). the project is a good mix of things I already know (microcontrollers, basic circuit design/analysis) and things i havent formally learned yet (transistors/amplifiers). so I'm coming here for help with the bit i dont know and wish to learn about.

thanks


EDIT: also, if anyones got suggestions for a microcontroller, please feel free to weigh in. most of my work has been with the MSP430 which is too much for what I need here. thinking about a PIC as those are fairly industry standard; also considering picking up an arduino dev package because of the open source software and low cost, and pre-built dev board and programmer. dont really know much about pics yet...



**EDIT**: said "LED controls gate"; meant "PIC controls gate", took me a while to notice

 

High power LEDs are driven by a Mosfet that is switched on and off at a high frequency by Pulse-Width-Modulation to control the brightness. If you use a linear amplifier to control the brightness then the amplifier will waste power by getting hot.

The microcontroller makes the PWM signal. The Mosfet must have a logic-level input rating.

 

such FETs seem scarce to me (at least, google searches arent turning much up) I need FETs that can handle about an amp (the higher the better, but ill probably be around 800-1200mA), and be triggered by TTL or something similar, about 5V and 20mA.

what would these be called? i need some good terms to search for.

 

IRLZ24's will work for high current (17A) on logic levels (0-5v). They come in a TO-220 package.
2N7000's can also be driven by logic level voltages. (TO-92 package)
2N7000's have an Rdson of around 6 Ohms. They can sink up to around 200mA continuous.

 

ah, excellent. thanks. perhaps I will try those IRLZ24s, even though 17A is a bit much, only need about 2. less heat then

 

how about the STP2NK60Z
( http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=497-4377-5-ND )
looks like it might have the right characteristics. 1.4A, thru hole, N channel.

however the Vdss is 600V (i think its a max)... my actual Vdss will be something like 3-6V, depending on what works. thoughts?

 

hi everyone, im fairly new here (post # 2), but i thought it would be good to get into a good online community for electronics as it is my livelihood (well, will be once i graduate... I am currently a student)

basically i am designing a microcontroller-controlled (sic) LED adjustable-color room light. basically put a few strings of 1W or 3W Red, Green, and Blue high-power LEDs (probably luxeons or something like that) together, and use a microcontroller to PWM each channel (RGB) to adjust brightness and color. the input will probably be something like a potentiometer through the ADC so there is a dial that just adjusts color. also thought about doing a capacitive touch-sense pad where X changes color and Y adjusts brightness... BUT I'm still in the design stages and little extras like that can be ironed out later.

what I need now is a way to drive high power LEDs from a PIC. Obviously I cannot drive them directly; current requirements are too high. I would need some sort of amplification or at least switching, where the LED controls the gate.

I picked this project because I don't forsee the amplifier stage being particularly tricky. I am about to take a formal class in FET operation and amplifier design (followed by a class in BJT operation and amp. design). the project is a good mix of things I already know (microcontrollers, basic circuit design/analysis) and things i havent formally learned yet (transistors/amplifiers). so I'm coming here for help with the bit i dont know and wish to learn about.

thanks


EDIT: also, if anyones got suggestions for a microcontroller, please feel free to weigh in. most of my work has been with the MSP430 which is too much for what I need here. thinking about a PIC as those are fairly industry standard; also considering picking up an arduino dev package because of the open source software and low cost, and pre-built dev board and programmer. dont really know much about pics yet...

How about SCRs or even Triacs? That's how we built color organs back in the dark ages.

Eric

 

How about SCRs or even Triacs? That's how we built color organs back in the dark ages.

SCRs and Triacs work great if you're dealing with AC. But LEDs really need DC to operate properly - and if the current isn't being switched on and off, then an SCR or TRIAC would be "stuck on" once turned on until the current through them fell to zero.

 

how about the STP2NK60Z
( http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=497-4377-5-ND )
looks like it might have the right characteristics. 1.4A, thru hole, N channel.

however the Vdss is 600V (i think its a max)... my actual Vdss will be something like 3-6V, depending on what works. thoughts?

It has a higher Rdson than the 2N7000; therefore in the TO-92 package, it wouldn't be able to carry the current. Also, since it isn't specifically a logic level gate, your mileage may vary on the threshold voltage.

I've used 2N7000's for logic level control switches, and they work.

The IRLZ24's may be overkill, but they are specified for logic level gate control.

If you want something physically smaller, try these:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=IRLU024NPBF-ND
$1.08/ea, logic level, I-PAK case (sort of like a small TO-220) 17A.
They'll work just fine.

Thru-hole logic-level MOSFETS are getting a bit hard to find. Most manufacturers are using SMT/SMD nowadays. It makes the assembled PCBs smaller and lighter in weight. They can get away with that by using newer MOSFETS with lower Rds(on) values; since less heat needs dissipating, they don't have to spend money on big heatsinks - and the people necessary to install them. Besides, if you look - there isn't much of a price difference between a MOSFET rated for 1A than there is for one rated 20A.

You're generally better off to stay with a Vdss about double, perhaps triple what you're expecting to see in the circuit. In broad terms, the higher the Vdss, the higher the Rds(on).

 

You're generally better off to stay with a Vdss about double, perhaps triple what you're expecting to see in the circuit. In broad terms, the higher the Vdss, the higher the Rds(on).

so what youre saying is that, if i expect a source-drain difference of 5V, then i should get a FET with a Vdss rating of 10-15V? then I will also have to find the Rds(on) and consider that when choosing my current-limiting resistor up top. and then a bleeder resistor on the gate so it actually turns on/off again... and im only operating at 200Hz so this can be a high value, given capacitances are sub-100pF.

heres what I'm envisioning (Sorry the drawing isnt great, the software sucks) I have my FET tied to ground on the drain, and the source is at the bottom of the chain as seen in the pic. basically voltage level of 5V on the gate should turn it on or off, right? unless im badly mistaken. there will be 800mA to 1400mA going through that series.

the LEDs have a fwd. drop of about 2.2, but im just using round numbers here, the voltage up top is whatever i need it to be, and ill have as many LEDs in series as i can fit with the voltage i have.

 

so what youre saying is that, if i expect a source-drain difference of 5V, then i should get a FET with a Vdss rating of 10-15V?

If you could find one with a Vdss rating that low, it would work. Seems like the minimum Vdss ratings I've seen lately are around 20v. The Rds(on) and total gate charge are of more concern when dealing with low voltage circuits. The higher the total gate charge, the more current you'll need to drive the gate with to turn it on and off. The higher the Rds(on), the more power will be dissipated in the MOSFET itself.

Then I will also have to find the Rds(on) and consider that when choosing my current-limiting resistor up top. and then a bleeder resistor on the gate so it actually turns on/off again... and im only operating at 200Hz so this can be a high value, given capacitances are sub-100pF.

Well, the Rds(on) of even puny MOSFETs like the 2N7000 is quite low; 5-6 Ohms. The Rds(on) of the IRFZ24 and the like are so low that it's practically like a short piece of wire when it's turned on.

Heres what I'm envisioning (Sorry the drawing isnt great, the software sucks) I have my FET tied to ground on the drain, and the source is at the bottom of the chain as seen in the pic. basically voltage level of 5V on the gate should turn it on or off, right? unless im badly mistaken. there will be 800mA to 1400mA going through that series.

OK. Your voltage drop across the MOSFET is not correct.
If the MOSFET is ON, Vds will be Id x Rds(on).
If the MOSFET is OFF, the Vds will be close to whatever the "voltage up top" is.

The LEDs have a fwd. drop of about 2.2, but im just using round numbers here, the voltage up top is whatever i need it to be, and ill have as many LEDs in series as i can fit with the voltage i have.

LEDs are rated for Vf @ a given current.
Usually, there is a typical and maximum Vf specified for a given current.
For example:
Vf=2.2v @ 25mA (typ), 2.6v @ 25mA (max)

You'd use the typical Vf @ current for your resistor calculations. Out of a batch of LEDs, around 80% will fall within a few percent of the typical Vf. The remainder can be off by significantly more. The distribution of Vf at a given current, with a large enough sample of LEDs, will look like the traditional bell (Gaussian) curve.

To find out how many LEDs you can drive in a series "string" for a given supply voltage:
MaxLEDs = Integer( (SupplyVoltage -0.5) / VfLED)
The "-0.5" is headroom for your current limiting resistor(s). The larger that number is, the more tolerant your strings will be of voltage fluctuations and variations in individual LED Vf's - but the more power you'll dissipate in the current limiting resistors.

Let's say your supply voltage is 12.
MaxLEDs = Integer( (SupplyVoltage - 0.5) / VfLED)
MaxLEDs = Integer( 11.5 / 2.2)
MaxLEDs = Integer( 5.227...)
MaxLEDs = 5.
Then to calculate your limiting resistor, assuming your current rating is 25mA:
Rlimit >= (Vsupply - VfLED(total))/LEDCurrent
Rlimit >= (12- (5 x 2.2)) / 25mA
Rlimit >= (12 - 11) / 25mA
Rlimit >= 1 / 0.025A
Rlimit >= 40 Ohms.
The closest standard value is 43 Ohms. Alternatively, you could use two 20 Ohm resistors in series.

 

ah, thanks. got my LED maths done already though, that I understand. mine are either the 1W or 3W variety, hence my need for a 2A or so FET.

I thought Vgs had to be higher than Vds to turn it on, i.e. if source is gnd and drain is +25V ("what it is up top") then to turn it on my Vgs would have to be 25V or more

 

Most Mosfets turn on completely when their gate to source voltage is 10V.
Logic-level Mosfets turn on completely when their gate to source voltage is 5V.
The max allowed gate to source voltage for most Mosfets is 20V.

The gate voltage has nothing to do with the max rated Vds voltage.

If you use a Mosfet as a source-follower then the gate voltage must be higher than the supply voltage to turn on the Mosfet.