0

I would like to turn on a thin film heating element. I will be controlling the on/off with a 5VDC pin, possibly PWM, using an ATTINY85. The thin film heater votage is 24 VDC at 2A (max. constant). I currently have an IRF540N MOSFET to do this but I imagine there's something better suited for the task. How do I detemine which MOSFET would be best, subject to the contraints. I am very new to this but I'm here to learn. I also would like to avoid a heat sink, if possible.

pgpilot326
  • 1
  • 1
  • 1
    You're saying you're controlling with a 5VDC to the gate and also that you're just below the 20V Vgs max, so which is it? With 5V to the gate that mosfet turns on, but not very on. Still at 2A, probably more than on enough for you. See the gate voltage chart on page 3. You haven't mentioned what you're using to drive the gate (output of 555 timer or a mosfet driver or a logic circuit...?) or what frequency you're switching at. There's almost always something better. This mosfet has 44mOhm resistance *with the gate at 10v* so your resistance will be higher. – K H Jan 28 '21 at 03:06
  • You need to estimate your switching and conduction losses and based on that you can guess if your mosfet is an adequate size package and if you need a heat sink or a better mosfet. Your Rds on also goes up as temperature goes up so you may need to take that into account. For your conduction losses, P=I^2*R so if you double current you quadruple your power dissipation. Since you probably don't need to switch a heater at particularly high frequency, I'd suggest you compare mosfets looking for ones with low RdsOn *at the gate voltage you plan to use* or with RdsON low enough that 2x is fine. – K H Jan 28 '21 at 03:14
  • Thanks for your response! I though Vgs is the difference between the two - gate and source. The Vgs th is 2.5 to 4V which I thought was the "turn on" voltage needed at the gate. I'm using attiny85 (possibly pwm) pin to switch it on. – pgpilot326 Jan 28 '21 at 03:14
  • 4V is the voltage at which they decided the mosfet conducts in a useful way, and according to the chart on page 3, it's good enough for more than 10A, but that number may require significant cooling. You're well below 10A, but you'll still have extra resistance and in an ideal situation it's nice to keep losses so low that the mosfet doesn't need a heat sink, so to make it under that threshold, every bit of wasted power in the mosfet matters. – K H Jan 28 '21 at 03:16
  • 1
    Add the extra information to your question rather than putting it in the comments. Depending on how fast you need to switch and how much current your ATtiny pin can source(probably not much) and your mosfets gate capacitance, you might want to add a mosfet driver to switch the mosfet fast enough. So look up the voltage and current specs on your ATTiny output and add that to your question as well. – K H Jan 28 '21 at 03:25
  • 1
    @pgpilot -- Gate voltage: check the max Vgs(th) ... it's 4V ... defined at 250uA. This is much closer to being an "off" gate voltage than an "on" gate voltage. So a 5V drive will produce the 2A, but with little headroom and possibly degraded speed on the way up. Other mosfets might give more flexibility here, or it might not matter // standard shpiel: always check the "Safe Operating Area" chart (should be ok provided adequate heat sink) -- Figure 5 in this datasheet. – Pete W Jan 28 '21 at 03:25
  • I'm a hobby guy so I was leaving this in case someone authoritative came along to help you, but at 2A and a reasonably pessimistic 0.1 ohms, your I^2R losses would be (4)*0.1=0.4 W, so provided it isn't getting heated up by the heater and it has a thermal path for a little bit of cooling, my gut feeling is that you're fine with this mosfet, but that's based on the assumption you keep your switching losses very low. – K H Jan 28 '21 at 04:33
  • K H, thanks for all of your input! I found a posting on the arduino forum (https://forum.arduino.cc/index.php?topic=70635.0) that talks a bit more about all of this. I don't have the background to know how to check some of these things but I know enough that I don't want to waste power. I'm guessing a logic level device is more appropriate. I also read a post here (https://electronics.stackexchange.com/questions/157065/explain-in-laymans-terms-vgs-and-vgsth-of-mosfets) and think I'm closer on the whole Vgs thing... it's not the difference in Vds and Vg as I previously thought (24-5=19). – pgpilot326 Jan 28 '21 at 06:20

1 Answers1

1

If you want to control the FET with the output of an ATTiny, you want a logic-level MOSFET, which are designed to have their gates driven at lower voltages that microcontrollers would operate at (e.g. 5V).

The logic-level equivalent to your IRF540 is the IRL540. The IRL540 is well within your parameters for current and voltage capabilities, and it has a fairly low Rds(on) of 0.07ohms at 5V Vgs and 0.11ohms 4V Vgs. This means you should have no trouble running it without a heatsink even if you only drive it at 4 volts—\$P = I^2*R\$ is less than half a watt of heat dissipation, which shouldn't be too much of a problem for the FET, but it might get a bit toasty...a small heatsink certainly couldn't hurt and is very cheap.

You could look around at other logic level FETs to see if you can find one with a lower Rds(on) if you determine that's too hot to run without a heatsink.

Edit: Even better would be the IRLZ44, with an even lower Rds(on). Driving it at 5V, power dissipated is \$2^2*0.028 = .112W\$. RthJA is 62 C/W, so \$0.112*62 \approx 7℃\$ increase above ambient. If your room temperature is a toasty 25C, your FET will heat up to 32C. Quite hot, but well within the operating parameters, and certainly better than the 50C+ that the IRL540 would reach.

flashbang
  • 79
  • 7
  • Thanks for the info! Quick question: would R change according to Ohm's since the .028 Ohms is at 31 A? This would give $$\frac{31}{2} \times .028 \Omega = .434 \Omega \Rightarrow .434 \times 62 = 26.908^{\circ} C$$, right? – pgpilot326 Jan 28 '21 at 07:14
  • No, that's just the value they tested it at. If you drive the gate at the advertised Vgs your R(ds) on will be the same no matter how much current you're passing. There are two gotchas: temperature—there should be a graph showing the relationship between Rds(on) and temperature, and ensuring your Vds is higher than Vgs - Vth (aka the "triode region" of operation). At 12V Vds and 5V Vgs you're well into the saturation region though. – flashbang Jan 28 '21 at 07:46