1
\$\begingroup\$

I trying to find a MOSFET transistor that I can use to switch on/off a device that has a current of ~17 amps and a voltage of 12 V. I would like to be able to take the 3 V signal from Raspberry Pi to switch it on/off. Does anyone have any suggestions?

\$\endgroup\$

closed as off-topic by Wesley Lee, Brian Carlton, Voltage Spike, ThreePhaseEel, Dmitry Grigoryev Feb 28 '17 at 17:09

This question appears to be off-topic. The users who voted to close gave this specific reason:

If this question can be reworded to fit the rules in the help center, please edit the question.

  • 2
    \$\begingroup\$ Couldn't you find any from the usual suppliers or by searching for 3V switching MOSFETs? I'm really not seeing a question here, other than shopping. \$\endgroup\$ – Brian Drummond Feb 27 '17 at 17:17
  • 2
    \$\begingroup\$ There are probably hundreds of options. digikey.com/products/en/discrete-semiconductor-products/… First, assuming this is DC you are talking about, you most likely want an N-channel FET, and you should select something that is 'logic level', i.e. with Vgs(th) below 2.5V or so. Narrow down to the voltage and current range you want. The rest depends on essentially the type of package you are looking for and price. \$\endgroup\$ – AngeloQ Feb 27 '17 at 17:24
  • 2
    \$\begingroup\$ Turns out there are over 3000 on DigiKey meeting those criteria: digikey.com/short/32tr92 Also, be sure not to rate the current or voltage too close to your needs. If the actual expected current and voltage are 17A and 12V, get something more like 30A and 25V. Otherwise you are more likely to damage it. \$\endgroup\$ – AngeloQ Feb 27 '17 at 17:29
  • 2
    \$\begingroup\$ Threshold voltage is normally quoted against a drain current in the hundreds of microAmps; for a load of 17A you will need a lot more, probably more than the 3V output from the pi. You'll probably need to use a P-channel MOSFET driven by an NPN common-emitter. \$\endgroup\$ – Finbarr Feb 27 '17 at 17:48
  • 1
    \$\begingroup\$ 17 Amps?! Wowwie. What exactly are you trying to drive here? \$\endgroup\$ – KingDuken Feb 27 '17 at 17:50
1
\$\begingroup\$

High power MOSFETs can not be switched directly via 3.3V. Thus it's necessary to have a second stage which generates the control signal. I attached a suitable circuit.

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
  • 1
    \$\begingroup\$ The current is not so high that this is necessary. This MOSFET can do it directly from 3V. \$\endgroup\$ – AngeloQ Feb 27 '17 at 19:00
  • 2
    \$\begingroup\$ Thanks for your comment. You are right, there are actually some parts that fit the requirements. However it's bad practice to directly drive FETs from Raspberry outputs. \$\endgroup\$ – auoa Feb 27 '17 at 19:06
  • 1
    \$\begingroup\$ @AngeloQ In the linear region, at 3 V Ron will destroy the device at 17 A. \$\endgroup\$ – skvery Feb 27 '17 at 21:45
  • 1
    \$\begingroup\$ @skvery, why do you say that? Rdson is about 25mOhm at 3V (gs). Yes, higher than what could be achieved with higher Vgs, but it won't destroy it, unless the power dissipation is not dealt with properly. But that would apply at higher Vgs also. And we know nothing about the expected duty cycle of this application. Yes the Rdson curve is steep at 3V, and a different part might be more suitable if high duty cycle is expected, but my point was to show that a MOSFET can be driven directly without another transistor stage. \$\endgroup\$ – AngeloQ Feb 27 '17 at 23:27
  • \$\begingroup\$ @auoa. Why is it bad practice to connect FET directly to a Raspberry Pi GPIO??? If you 'isolate' with a BJT then under fault conditions it may still damage the 'Pi. \$\endgroup\$ – Jack Creasey Feb 28 '17 at 16:46
0
\$\begingroup\$

Mouser has a great table for finding parts like this, and it can be found here: MOSFET chart

-- edited based on a comment by @auoa - just use a transistor second stage to trigger the MOSFET

\$\endgroup\$
  • 2
    \$\begingroup\$ You should not use a relay or Boost converter to generate to control signal for the FET! Just use a second transistor stage. \$\endgroup\$ – auoa Feb 27 '17 at 18:49
  • \$\begingroup\$ I edited my answer to reflect this comment. Thank you, @auoa \$\endgroup\$ – SDsolar Feb 27 '17 at 23:23
0
\$\begingroup\$

Initially you should understand that many of the folks on Stack Exchange are professionals and simply won't respond to (or worse will seriously castigate you for) a badly formed question.

My choice of device would be BSC019N02KS-G.

But not really a helpful answer to a not very well thought out or formulated question of course. Why is it a good device choice and what parameters make it viable for you?

You could start by being more helpful and enlarge the relevant information you supply in your question:

  1. At least describe the load, is it resistive, inductive, a bit of both? Does it have surge current (like a motor) when it turns on?
  2. If you are a hobbyist you probably have physical constraints such as the device package to deal with. If you are building on a breadboard, surface mount (most of the most high performance and newer devices) devices can be a challenge. In this case we know that you won't be using a breadboard since the contacts are not rated for 17 A ...right?
  3. Do you need a low side or high side switch? This will impact the switch architecture used. A low side switch could be a single device, a high side switch would require more components driven from a 3.3 V system.

The only real information you supplied is A) 3.3 V driven and B) 17 A @ 12 V

Let's create some requirements and find a device (there will not be thousands of choices). Assumptions:

  1. 0-3.3 V drive voltage
  2. Low side switch (aim for a single N-Channel FET)
  3. 17 A @ 12 V resistive load (no inductive snubbing required)
  4. Needs pins (not surface mount)

In searching for a suitable device we would never search for devices that only supported Idss(cont) of 17 A. I'd suggest you would allow for at least 10-20% more current capability ...so let's search for N-Channel FETs with say 19-20 A continuous capability, later I'll show this is way too close to the requirements, and you should select much higher Id.

Here's my starting point on Digikey (you could do the same on Mouser).
Notice that I've ordered the column of Id(continuous)
Scanning through the devices in the 19-20 A range you can notice the devices with the required Vdss, the lowest VGS(threshold) and drive voltage (which impacts RDS(on) to suit your needs.
Remember that the FET is an analog device, so though many (including me) talk about logic level FETs, what is really being described is the VGS at a particular current being useful for logic driven circuits.

In this case I spotted a device that supports Idss=19 A, VGS(threshold)=2.5 V, Drive voltage=4 V and in a TO-220 package. ....the ONSemi FDP8870-F085. Its datasheet is here.

From the datasheet we can find the saturation characteristics: enter image description here

This shows that we can expect that with 3.3 V drive from the Raspberry Pi you can saturate the device with a 17 A load. If the load has surge characteristics, for example if you were driving a motor that might stall and draw 3-4 times the running current then you'd be in a lot of trouble here. The device would leave saturation and the dissipation would soar and probably destroy your device.

But technically this device fits the bill if you never exceed 17 A @ 12 V.

Now what might be a better selection?
If we now select a device with Id(cont) of say 5* the 17 A, we'd search for something around 85 A ....say we look for 80-100 A.

Scanning the tables you can find an 80 A device, the OnSemi FDP8860, its datasheet is here.

The saturation characteristics are similar to the FDP8870 (called the On region characteristics this time), but it's capable of much higher current at the same Vgs.

enter image description here

So with our 3.3 V drive from the Raspberry Pi could expect this devices to stay saturated through to an Id of at least 60 A plus. Much better.

Both these devices are legacy parts (some years old), but there are new device families that have extremely low VGS(thresh) called Super Logic FETs. These devices have VGS(thresh) < 2 V and current capability into the hundreds of Amps.

If you look up the Infineon BSC019N02KS-G and compare it's datasheet you'll see this is capable of switching 80 A DC and pulses of potentially hundreds of Amps. The downside is that most of these ultra modern devices are exclusively SMT packages and not very hobbyist friendly.

Finally, what would the schematic look like for any of these devices:

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
  • \$\begingroup\$ You should really add a series resistor between GPIO and Gate. Otherwise you are charging a capacitor without current limitation! \$\endgroup\$ – auoa Mar 1 '17 at 0:04
  • \$\begingroup\$ Given the size of the gate capacitance, IMO you do not need current limiting. The GPIO's have inherent series resistance which you can measure by simply loading a GPOI set hi and measuring the voltage drop with load. But if you wanted to add a series resistor of say 200 Ohm, it would do no harm. \$\endgroup\$ – Jack Creasey Mar 1 '17 at 0:09

Not the answer you're looking for? Browse other questions tagged or ask your own question.