To measure the nominal V gs threshold (starting-to-turn-on voltage) of the FET, short point 2 and point 3 and then measure the voltage between point 2 and point 4 as before. Then leakage current in milliamps is approximately = (multimeter reading in millivolts)/(10 4). You can determine if any residual leakage between Drain and Source by shorting point 3 and point 4, then measuring the voltage across the 100KΩ resistance feeding point 1 from the battery. You should see the voltage vary from low in first test to the actual battery voltage applied (usually 9 V). With points 1 and 2 shorted together, measure the voltage between point 2 and point 4, then short point 3 to point 4. Measurement of the device’s R dsON begins by removing the link between points 1 and 2, then measuring between point 2 and point 4 to obtain an approximate reading of resistance on the multimeter. The multimeter should read a voltage between points 2 and 4 that is quite low. Logic-Level FETs can be switched on at voltages of between 0.3 and 1.5 V.Ī simple circuit shown here for an N-channel FET helps detect whether the device works correctly as a switch. The reason is most power FETs require a gate-source voltage bias of at least 4-5 V to switch on. The problem is that testing a FET with a modern multimeter alone becomes problematic. Unfortunately, modern multimeters use a low excitation to measure resistance (1–2 V) to ensure simple active probing of circuit elements doesn’t damage them. The body diode will show cathode at the drain for an N-channel device and at the source for P-channel device. The diode junction measured is the FET body diode. The other direction should measure infinite resistance – or at least a few megaohms. If the drain-source path is healthy, applying the ohmmeter probes in one direction should read a short. Connect the gate of the device to the source terminal. They tend to open (like a fuse) when overloaded, disconnecting the MOSFET gate.Īnother frequent failure mode for a FET is a shorted drain-to-source. Another tactic is to add subminiature gate resistors. To avoid overloads, some designers add a zener diode between the source and gate–zeners fail short so they can limit damage in the event of a MOSFET failure. Thus it’s best to check the drive circuits of dead MOSFETs. The overload will also affect any other paralleled MOSFET gates. This can put the drain voltage back onto the gate where it feeds (via the gate resistors) into the drive circuitry, possibly causing voltage and current levels exceeding the limits of components in that section. MOSFETS that fail often go short-circuit drain-to-gate. The gate will discharge and the meter reading should go high, indicating a non-conductive device. With the meter positive lead still connected to the drain, short the source and gate. The MOSFET’s internal capacitance on the gate has now been charged up by the meter and the device is ‘turned-on’. Now move the positive probe to the ‘Drain’. (Hold the MOSFET by the case or the tab but don’t touch the metal parts of the test probes with any of the other MOSFET terminals until needed.) Touch the meter positive lead onto the MOSFET Gate. Connect the Source of the MOSFET to the meter’s negative lead. where the device is non-conductive with 0 V applied to the gate, operating like a switch). First consider the procedure for testing an enhancement-mode MOSFET (i.e. The usual testing procedure is for use with a digital multimeter in the diode test-range with a minimum of 3.3 V over d.u.t. A common multimeter can quickly do in-circuit tests, which are not totally definitive but generally provide acceptable go/no-go information, using either the meter’s diode-check or ohms mode. The sensible alternative is to test the transistor. Also, it is not reliable because an outside defective component can instantly destroy the replacement with no visible evidence. One approach to troubleshooting is to substitute a known good component, but that is a costly way to go. A bad transistor can sometimes be detected by its partly burned or distorted appearance, but more often there is no visible indication.