PDA

View Full Version : Fet Static Drain-tosource ON-Resistance



metron9
02-14-2007, 05:16 PM
I decided to try the LUXEON III (3 watt) LEDS in my drumlights. The traces are big enough to carry 1.4A on the board but I am down to a 1 OHM resistor at the 5V regulated supply. Here is my problem.

I am using the IRFL110 Fet. It can handle 1.5A continuous.

I tested the RED LED and a resistor without going through the fet and I get 1240mA When I go through the FET I get 660mA

Looking at the datasheet and this is something I have not had to deal with before as I was using 7.5 ohm resistors on the LUXEON I (1 watt) LEDS
The .54 Maximum ohms RDS(on) was a small in relation to the resistor I was using. It now comes into play as a 50% difference.

I do not see minimum and typical values on the datasheet.

Question is, how do I calculate the resistor I need if I don't know the typical and minimum values, or is this a situation where increasing the voltage and resistor will lower the percentage of current drop due to the RDS(on).

The forward voltage of the red Luxeon III os 3.51V MAX and 1540mA is MAX current.

Remember too these are not 100% duty cycle so I want to push the highest current allowed to get the brightest flash.

Darn now it's 1500hrs again, always happens when I am having fun that clock just spins and spins....

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!

Phil Pilgrim (PhiPi)
02-14-2007, 05:33 PM
Having such a high percentage of your excess power dissipated by the FET is not a good thing, since RDS(on) can vary so much from device to device and with temperature. Perhaps a different FET with an ON resistance in the milliohm range would be the way to go. That would ensure that the more predictable resistor carries most of the excess load. But variations in the LED's VFWD are also cause for concern with so little available headroom. Increasing the supply voltage would help, but at the cost of power efficiency.

The optimum solution would be to use a switching current regulator designed for driving high-power LEDs. Try a search on Google. You'll turn up a plethora of such devices.

-Phil

metron9
02-14-2007, 05:41 PM
I know a complete redesign should be done to do it right, but I have just tested the 2AMP 6V unregulated supply and that gives me the voltage I need. I don't need to conserve power, that's not a problem as the lights only turn on typically less than 50% duty cycle unless a snare roll holds one on for the duration of the roll but even then the maximum current is still pulsing. I will do some more testing but i think the higher voltage supply will work, I made the board traces 24 mil so I have plenty of capacity on the traces. I will look for a replacement SOT-223 with a lower on resistance as well.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!

Phil Pilgrim (PhiPi)
02-14-2007, 05:49 PM
If you can't find a lower-resistance FET in an SOT-223 package, you can always just solder another on top of the one you already have. FETs can be paralleled with near impunity, and doing so would halve your on resistance.

-Phil

metron9
02-14-2007, 05:51 PM
Just started looking at 17 pages of n-fets sot-223s and found digikey part 497-2519-1-ND

It shows Max 60 and 120 mohm I guess that is milliohm , perhaps I better sleep first before the sun comes up.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!

Phil Pilgrim (PhiPi)
02-14-2007, 06:07 PM
Be sure to check the gate voltage required to get the rated on resistance...

-Phil