I know just enough electronics to be dangerous
That said, I've never had the pleasure/need to work with MOSFET's till now. I need to dim an array of high powered LED's (not an uncommon challenge - I've read a few related threads on here already). I know a logic-level MOSFET is best to interface this LED array to one of the Arduino PWM pins.
My question: how to shop for a MOSFET that best fits the electrical parameters of both the Arduino's PWM output, yet properly drive the LED array? I'm very confused when looking at MOSFET datasheets on what are the most important characteristics to watch for, what they mean, and "issues" they might enter into the overall circuit. Like, will these parameters induce and unwanted voltage-drops, capacidence, inductance, etc... to the array meaning I'll need to re-consider the series load resistor, etc... ?
FWIW, my LED array runs roughly at 2.1 Amps @ 16VDC (about 35 watts) total to include all the LEDs and current protection resistors. It consists of three parallel circuits, each having 5 LEDs in series with a 1 ohm resistor. This totals to 15 3.2v 700ma LEDs and 3 1 ohm resistors.
Not asking for MOSFET part #'s, rather I'm looking for any tips on how to identify the right MOSFET for this circuit's needs would be greatly appreciated - thanks!
Okay, so sticking with "logic-level" MOSFET's should take care of the "turning on" (gate) requirement. And as far as the Source/Drain power handling requirements I'm looking at the following specifications:
Maximum Continuous Drain Current (Amps)
Drain to Source Breakdown Voltage (Volts)
Device Total Power Dissipation (watts)
BUT I see other specifications such as:
Static Drain to on Source Resistance (Ohms)
Input Capacitance (pf)
Forward Transconductance (mhos)
Would any of these be of concern in regards to additional resistance added to the driven circuit? Does the capacitance effect the max speed of switching or perhaps the need for additional protection circuits, build-up charges, etc... ? Sorry - although this is less an Arduino-interfacing question, I'm just trying to wrap my head around all this.
The most important thing is that it's a logic level mosfet which will fully saturate on with +5vdc applied to the gate, unlike standard mosfets that require +10vdc gate voltage to reach full rated drain current. This is a good candidate and it's datasheet should guide you if you select any other mosfet:
RDSon -- Drain->source resistance is the "on" resistance of the device, often in the dozen milliohms range. This together with the current and voltage you will run through it will determine how much power is dissipated in the FET when it is on.
Input capacitance determines how much current will rush into the gate when you first turn it on. The effective capacitance will take a little while to charge, and during that time the gate will suck current faster than your arduino logic pin would prefer -- thus you will need a resistor between the logic pin and the gate. Then the R-C timing of that resistor and the gate capacitance will affect how long it takes to turn on, how long it operates in non-saturation, and how much power is dissipated during this turn-on/off period.
Even for logic level FETs, a driver is likely a good idea if you will switch it on and off very often.
The most important thing is that it's a logic level mosfet which will fully saturate on with +5vdc applied to the gate, unlike standard mosfets that require +10vdc gate voltage to reach full rated drain current. This is a good candidate and it's datasheet should guide you if you select any other mosfet: N-Channel MOSFET 60V 30A - COM-10213 - SparkFun Electronics
Lefty
Thanks Lefty! Looking at that datasheet, I see the 'Gate to Threshold Voltage' maxes out at 2V. Is that simply saying anything 2 and above volts causes the S/D to switch on fully?
RDSon -- Drain->source resistance is the "on" resistance of the device, often in the dozen milliohms range. This together with the current and voltage you will run through it will determine how much power is dissipated in the FET when it is on.
Input capacitance determines how much current will rush into the gate when you first turn it on. The effective capacitance will take a little while to charge, and during that time the gate will suck current faster than your arduino logic pin would prefer -- thus you will need a resistor between the logic pin and the gate. Then the R-C timing of that resistor and the gate capacitance will affect how long it takes to turn on, how long it operates in non-saturation, and how much power is dissipated during this turn-on/off period.
Even for logic level FETs, a driver is likely a good idea if you will switch it on and off very often.
Gardner - thanks for that crash-course, and the details about the inherent R-C behavior, makes complete sense. Yes, I was planning to place a resistor between the logic pin and the gate. I'm also inclined to put another resistor between the gate and ground so that the gate "sees" absolute 0v on logic lows - or do you think that would be over-kill because most MOSFET's already incorporate internal zdiodes, etc... ?
Thanks Lefty! Looking at that datasheet, I see the 'Gate to Threshold Voltage' maxes out at 2V. Is that simply saying anything 2 and above volts causes the S/D to switch on fully?
No exactly, not fully on. The threshold voltage is that which causes the mosfet to just start to conduct drain current, could just be milliamps with +2vdc applied. One has to actually look at the graphs in the datasheet to see the gate voltage Vs drain current, there will be lines showing current flow for various gate voltages, but +5v should show fully saturated current.
No exactly, not fully on. The threshold voltage is that which causes the mosfet to just start to conduct drain current, could just be milliamps with +2vdc applied. One has to actually look at the graphs in the datasheet to see the gate voltage Vs drain current, there will be lines showing current flow for various gate voltages, but +5v should show fully saturated current.
Lefty
Got it! Looks like the saturation graph in the part you gave above is on page 4 (PDF). I now see that anything above 2v starts allowing current to flow D/S. Guessing the right numbers are amps in %, at around 5v gate I'm seeing 82% at 25 degrees C. Are they saying 85% of the devices total rated amp handing - in which this case is 30 amps?
So what will happen when we place a small resistor on the logic pin to prevent the gate from initially sucking on it too hard, eventually the current draw will slow to a trickle allowing the gate-side of the resistor to reach 5V (somewhere there of) allowing the fet fully saturate thus we get the full on state. The Arduino's PWM pin is well protected, while we are able to fully saturate the FET - fantastic!
One last follow-up (and this is more Arduino-related) : since the R-C phenomenon of the gate limits how fast one can hammer a given mosfet on/off, and taking into account that the PWM pin hums at a mere 490Hz, regardless of duty cycle does anyone feel I need worry about these R-C datasheet characteristic when selecting a mosfet - especially among the more popular logic-level mosfets on the market?
One last follow-up (and this is more Arduino-related) : since the R-C phenomenon of the gate limits how fast one can hammer a given mosfet on/off, and taking into account that the PWM pin hums at a mere 490Hz, regardless of duty cycle does anyone feel I need worry about these R-C datasheet characteristic when selecting a mosfet - especially among the more popular logic-level mosfets on the market?
Well there are two issues here.
The lack of a series gate resistor just means there will be higher initial gate current flowing as it charges the gate's capacitance. The Arduino output pins are pretty rugged so I've never been too concerned, however the 'correct' answers is one should do the gate charge max current calculations and if they feel that value would exceed the avr pin current recommendations, then add the needed series resistance to correct that state. Most seem to just throw in a 100-330 resistor and call it good, I've mostly ignored the resistor and so far have had no problems.
The 'speed' of turning on and off the mosfet is effected by the gate drive current capacity, any series gate resistor, the capacitance of the gate/source, and the switching frequency of the gate signal. This all determines how often and how long the mosfet spends in the transition between being fully off and fully on. During that time the mosfet is dissipating heat energy because of it's Ron is higher then it's rated full current value, thus real power is being dissipated across the device. So under these worst case situations a mosfet can destroy itself very quickly with say a 20 amp load even though the device may be rated for 60 amps max. Heat sinks can help deal with this situations, as well as purpose designed gate driver chips that can supply very high momentary gate charging/discharging currents. Also using parallel mosfets can spread the heat across two devices, but of course then the gate capacitance is doubled also. That make sense?
The very best reference for this sort of thing I have found is in International Rectifier's application notes. AN-1084, Power MOSFET Basics (http://www.irf.com/technical-info/appnotes/an-1084.pdf) is an excellent place to start using power mosfets.
