2

I am using this BS170 MOSFET to drive a fan using a pwm signal generated by a 3.3V MCU according to the following schematics:

enter image description here

(I know about the wrong symbol it's just for the packaging on the PCB)

So my gate gets connected to ground using the pulldown resistor and to the PWM signal.

I have two questions:

  1. Should i still use the D1 diode for back EMF protection or i can relay on the internal one in my transistor.
  2. Is my transistor is suitable for a fan using 0.32A. It should handle 0.5A so it should work right?
Ahmad Elbadri
  • 211
  • 3
  • 9

4 Answers4

2

1) Depends on the fan, if it has inductance (A DC brushed motor) then you may want some kind of over voltage protection like D1. If its a brushless DC motor (like a PC fan) then it already has a circuit built in for overvoltage protection or it doesn't need it.

2) It says 500mA right in the datasheet, your fan can't be more than this. A 320mA fan will be fine. The mosfet also has a 5Ω RdsOn which means its like a 5Ω resistor when its on. At 320mV this will cause the voltage of the fan to be 1.6V, a fan with a controller may or may not be able to tolerate this.

Voltage Spike
  • 75,799
  • 36
  • 80
  • 208
  • 5 ohm Rdson would give a 1.6V drop @ 320mA, which would leave 10.4V for the motor. –  Jan 17 '17 at 23:11
  • should that be ok for a PC fan or the 1.6V drop won't let it on ? considering that am using it in PWM ?? – Ahmad Elbadri Jan 17 '17 at 23:15
  • A controller in BLDC fans may not like to run 1.6V above ground. If the current is ramping up\fluctuating in a starting condition, that 1.6V will also change, the controller may not like that. Better to go with a switch with lower resistance or at least be aware of it, depends on if this is for hobbyist work or a product. – Voltage Spike Jan 17 '17 at 23:16
  • Try it, see if it works. – Voltage Spike Jan 17 '17 at 23:19
  • 1
    How would the fan know it's 1.6V 'above ground'? What would be its reference for this 'ground'? –  Jan 17 '17 at 23:21
  • Put a meter across the FETS source and drain and measure the difference in voltage – Voltage Spike Jan 18 '17 at 04:28
  • I've downvoted this - the bit about the mosfet rating is wrong (it's not in-rating at 320mA), and the bit about the fan worrying about an offset from 'ground' is also wrong. –  Jan 18 '17 at 09:04
  • How is it wrong? If you put a 5 ohm resistor on the ground of a microprocessor, bad things will happen. – Voltage Spike Jan 18 '17 at 17:10
2

For your first question, yes keep the diode - if it isn't needed because of something special about the fan, it will do no harm. If it is needed then leaving it out will damage the reliability of your circuit. The diode in your MOSFET is not useful to do the same thing.

For the second question, personally, I wouldn't use this MOSFET for this job - it's a 'small signal' MOSFET, and you would almost certainly be much better with a power MOSFET of some kind:

  • The Rdson is too high - at 0.32A, you'd be dissipating 0.32*0.32*5 = 512mW which is more than the device is rated to dissipate. And that's before you start to think about switching losses caused by your PWM switching.

  • There are no characteristic graphs for Vgs as low as 3.3V - although you're above the threshold at 3.3V, you're clearly not operating the device in the way the designers intended.

Additionally you should almost always add a resistor in series with the gate of a MOSFET, to control the turn on speed/gate current.

Here's a post about choosing a FET: Selecting a MOSFET for driving load from logic

Community
  • 1
  • using a heat sink won't improve the situation ? – Ahmad Elbadri Jan 17 '17 at 23:30
  • I'm confused why you would want to put a resistor in series with the gate of a MOSFET. – Bradman175 Jan 17 '17 at 23:30
  • it's a pulldown resistor to ensure that if the gate is unconnected it's pulled to 0 V – Ahmad Elbadri Jan 17 '17 at 23:34
  • @AhmadElbadri Sorry I wasn't referring to you. That 10k pulldown is correct. – Bradman175 Jan 17 '17 at 23:36
  • @AhmadElbadri Using a heatsink won't make 512mW less than 500mW, which is the abs-max for the device. It's not sensible to use an unsuitable part and then try and keep it going with a heatsink. Just use something more suitable. –  Jan 18 '17 at 08:36
  • @Bradman175 - a series gate resistor is hardly necessary with such a tiny (unsuitable) FET, but in general with a power MOSFET you want to be able to control the switching speed, in order to trade-off peak current in the drive circuit (the FET's gate looks like a short circuit at the instant you turn it on), switching noise (ringing caused by very fast switching) and switching-time dissipation (caused by slow switching). If you don't explicitly include a resistor you just get something which is function of the gate capacitance and the drive impedance, and you can't adjust it. –  Jan 18 '17 at 08:42
0

For the first question, I believe depending if your using a threshold voltage to manage the power to the fans, a diode will be useful.

For the second question, If I understand the question correctly, you want to manage the current to the fan to prevent overloading to the first Fan or the second actually. Seemingly, if the circuit is trying to maintain two fans that require different amounts of power, I suggest testing that Fan 1 can operate until a certain threshold is met and when the Fan 2 is needed to start.

Overall, according to the specifications, the maximum threshold voltage is 3.0V, so I would work around the voltage to determine the amount of current to each fan. Consider adding/removing resistors based on the 3.0v threshold.

JDavila
  • 36
  • 3
-1

Normally a low Vce(sat) saturated BJT needs 3% to 10% base current. Considering the fan is a BLDC motor the current is commutating and aliasing effects can be had with PWM. A 1 Ohm MOSTFET also works but costs more.

These things need to be tested to verify the assumptions.

When I tried this approach I was not satisfied with the results and went with a 3 terminal ADJ regulator with a thermistor controlled voltage to adjust the fan speed for smooth quiet operation with pot selected setpoint and 5'C range from 0 to 100% speed. ( all for under $2)

This Infinion BSS806N H6327 would be a better Nch FET choice for low 57mΩ RdsOn and low cost.

Tony Stewart EE75
  • 1
  • 3
  • 54
  • 182