The circuit below is used as a relay in order to cut-off the common ground C- and Motor-.

In my system, the m otor driver (Motor+,Motor-) is connected in parallel with the supercapacitors (C+,C-.) With the Safety_Sig_3.3V, I want to cut-off the motor driver's supply from the supercapacitors.

As far as the circuit, 5V_BB is the power supply in one side of the isolated_gate driver ADUM3123CRZ and the 12V is the power supply on the other side of gate driver. BB_GND and FC_GND are the two grounds respectively. C- and FC_GND is the same ground. The Safety_Sig_3.3V is the input of gate driver. The Gate_In gate driver's output which is connected with the MOSFET's gate.The disable pin is connected with the ΒΒ_GND (ground). The MOSFET is an NTMTS0D7N06CLTXG. R1 is a pulldown resistor at the gate of the MOSFET.

The problem is that when I turn off the switch and after some time turn it on again, the motor makes a spike. As a result the, wheel turns for 0,5sec and then is ok again.

The motor driver on the other side supplies a 1KW DC motor (48V) which is connected to a vehicle's wheel.

How is this spike explained?


enter image description here


1 Answer 1


The motor has considerable inductance (probably some ~mH), so its current cannot be set to zero instantly. An opposite voltage must be applied to cause the current to fall towards zero. This is known as flyback. When the voltage is set by the MOSFET's breakdown voltage, the MOSFET absorbs some energy, heating up rapidly.

This particular MOSFET is rated for 1.75J (quite a lot for a little SMT device!) at 40A. I would guess the problem is either:

  1. You are switching off at much higher current than the MOSFET can handle in avalanche (i.e., >40A).
  2. The motor has more inductance and therefore energy than the rating (about 2.2mH).
  3. The MOSFET cannot handle repeated avalanche events, and functions for some number of cycles (dozens?), but eventually fails.

AFAIK, most MOSFETs are not rated for repetitive avalanche, or if they are, it's for some given lifetime (millions, billions of cycles?) and very low avalanche energy. There is a wear mechanism which leads to eventual failure.

Best practice is twofold, then:

  1. Connect a TVS in anti-parallel with the MOSFET. Probably a SMCJ48A or something like that will do. Exact size requirements can be calculated from peak motor current (presumably, up to LRA (locked rotor amperes)) and motor inductance (measure it if necessary).
  2. The MOSFET Vds(max) must be rated for equal or greater than the peak clamping (surge) voltage of the TVS. Typically this will be ~30% above the nominal, so, a 60V transistor is marginal and I would recommend 80V. This also allows some flexibility in supply voltage (what is this, a golf cart sort of thing, so it's not actually 48V, but as high as 57V under charge (lead acid?), maybe even more in certain conditions (load dump?)?), so a 56-60V TVS would be nice in that case.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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