# How to self start a MOSFET multivibrator circuit

I have built a simple multivibrator circuit using IRF530s for the switchers. It does not start unless I introduce noise by touching one of the timing caps. I tried to include a circuit image but as a new user could not. It is a textbook multivibrator only using power MOSFET instead of Bipolar Transistors. How can I make this circuit start at power up?

simulate this circuit – Schematic created using CircuitLab

• You can see the circuit at: circuitlab.com/cs56fxj3 I have tried different timing values and voltage supplies. Both in the simulation and in the lab the circuit will not self start, but always starts with just a small touch of my finger. Commented Jan 10, 2013 at 17:52
• @user17689 How does one execute a small touch of the finger on the circuit sim? :-) Commented Jan 10, 2013 at 18:14
• no time to write an answer, but see virtual.cvut.cz/dyn/examples/examples/electronic/mosmulti/… Commented Jan 10, 2013 at 19:25
• That did it, Phil! It works in both the sim and the lab. I guess the 1p cap in your example provides just enough offset to make things start. Thanks for your help. Commented Jan 10, 2013 at 21:00
• The basic problem is that the threshold voltage of these FETs is about 4V (where they start to turn on) and they are guaranteed to be fully on at 10V according to the datasheet. As against 0.6 to 0.7 volts for a silicon bipolar transistor. So they either need a bigger kick to get started, or (as Phil's circuit does) positive grid cough gate bias to reduce the kick required. The 1pf imbalance probably helps simulation to start, though!
– user16324
Commented Jan 10, 2013 at 22:34

## 2 Answers

Instead of applying a static 20 V power supply or a slow start PSU, pick a voltage source that starts instantaneously. When simulating, pick a step voltage source that is 0V for the first second, then after 1 second make it step to 20 volt. This step introduces enough noise in the circuit to make it start.

Some observations:

1. Why are R1 and R2 connected form gate to GND and not from gate to V+? IRF530 is an N-Channel-MOSFET, and it need a continuous positive voltage on the gate to remain "on". Your circut can only provide positive spikes at the gates through the capactors, and maybe through Crss (Miller Capacitance), but no sustained positive gate voltage.

2. In your simulation, the circuit will have trouble starting up because it is exactly symmetrical. Disturb the symmetry (e.g. by making R3 = 1 k$\Omega$ and R4 = 1.001 k$\Omega$), and it may start. A real circut will have enough imperfections to start, no need to buy a 1.001 k$\Omega$ resistor in Schillicon Valley.