5
\$\begingroup\$

I need to short-circuit a power source using a microcontroller to measure the short-circuit current capacity (less than 10A). So the impedance of the short circuit should be as low as possible.

I was using a 5V relay commanded by a BJT transistor and it works very well. But for space reasons I would like to test other options, like using a MOSFET transistor working as a switch. I tested with the logic level NMOS IRL540N as shown in diagram and it worked fine. Putting 5V at the gate closes the short circuit with about 36mΩ resistance.

But the problem is when the power supply is connected wrongly with reverse polarity the NMOS always conducts, even with a low microcontroller output. I don't understand why it conducts as VGS remains null even with reverse polarity connected.

How can I protect the MOSFET to be always off when reverse polarity is connected to drain-source?

The power supply could reach up to 25V. Is the MOSFET always secure? The absolute maximum rating for Gate-to-Source Voltage is 20V, but in this application VGS is always lower than 5V.

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
8
\$\begingroup\$

A glance at the IRL540N shows the reason for reverse conduction.

enter image description here

Figure 1. The MOSFET contains a built-in diode as part of the construction.

You can't add a diode because the voltage drop will be too high for your application.

schematic

simulate this circuit – Schematic created using CircuitLab

Figure 2. Modified circuit with built-in diodes shown.

You could add a second MOSFET in series but reversed. This would prevent either polarity passing unless the MOSFETS were turned on. The down-side is double the on resistance and the fact that the power circuit is not directly referenced to ground.

\$\endgroup\$
  • \$\begingroup\$ Go ahead, Chris, and please explain further. I used a similar circuit to this for an AC switch about 15 years ago and all was well when switched from a 12 V supply with a push-pull transistor driver. I'm working from memory. What have I missed? Ah - I think it's coming back to me. I had a transistor each side of common ... \$\endgroup\$ – Transistor Oct 10 '17 at 22:13
  • \$\begingroup\$ Actually, I'd mixed up which FET was flipped, so deleted that as maybe it's okay in normal polarity connection. But with reverse polarity, I think there is a VGS violation. \$\endgroup\$ – Chris Stratton Oct 10 '17 at 22:15
  • \$\begingroup\$ Does the edit look safer? \$\endgroup\$ – Transistor Oct 10 '17 at 22:17
3
\$\begingroup\$

The body diode causes your failure.

The preferred solution for high power, uses an Nch on the high side with a charge pump to boost Vgs while have a diode to gnd to protect the driver.

Such as 30V, 3.3mOhm MOSFET SPP100N03S2-03

enter image description here The load can be a 10 mohm current shunt for sensing or accurate control with a comparator and Vgs can be adjusted with a sweep signal from the charge pump rail to test the DUT.

\$\endgroup\$
  • 1
    \$\begingroup\$ You are right about the body diode. But your answer does not solve the problem. Battery could be connected upside down, and the body diode will conduct. \$\endgroup\$ – Chupacabras Oct 10 '17 at 22:16
2
\$\begingroup\$

MOSFETs have a built in diode between source and drain that will conduct if you reverse polarity. That is just how 3 terminal MOSFETs are made (the body is tied to the source). If you can find a 4 terminal MOSFET you might be able to pull this off, but more likely will destroy the MOSFET the first time voltage is applied backwards.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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