I have a microcontroller that outputs a 50% duty cycle PWM signal at 500kHz. That is then connected to the gate of a MOSFET. Now, I want to measure the PWM signal at the gate of the MOSFET. If I connect the probe, to measure it, I measure 15Hz and the signal is no longer a square, but a "sawtooth like". If I don't connect the probe, the MOSFET is switching at 500kHz.

I have a SIGLENT SDS1102DL oscilloscope, its channels are rated 1MOhm and 16pF. Is this the reason the frequency decreases?

How can I overcome this problem?

  • \$\begingroup\$ Is there a x10 switch on your probe? If so, try using it. \$\endgroup\$
    – Majenko
    Commented Dec 21, 2014 at 14:18
  • \$\begingroup\$ @Majenko It gets better, but still.... I've noticed, that if I connect the probe to the microcontrollers PWM pin insted of MOSFETs gate, the frequency doesnt change. I guess there is a loss connection or something somewhere. Thanks anyway. \$\endgroup\$
    – Golaž
    Commented Dec 21, 2014 at 14:48
  • 6
    \$\begingroup\$ If you're seeing 15Hz (!), I would expect you're seeing aliasing. Try starting with a high sweep frequency and reducing it. It's unlikely the micro can drive the MOSFET gate directly at 500kHz, most MOSFETs have a LOT of input charge and your micro is capable of tens of mA or less- you probably need amperes of capability to drive the MOSFET smartly at 500kHz. That's a much worse problem than the breadboard. \$\endgroup\$ Commented Dec 21, 2014 at 14:59
  • 3
    \$\begingroup\$ @Majenko: No, that is very unlikely the problem. The issues with breadboards are overrated, especially at such a low frequency as 500 kHz. Maybe there is a extra 10-20 pF capacitance on that line. That shouldn't bother a micronctroller output producing a 500 kHz square wave. I think Spehro is on the right track. Even high capacitive loading shouldn't change the frequency since that net is not part of the oscillator itself. This is almost certainly a measuring artifact. \$\endgroup\$ Commented Dec 21, 2014 at 15:14
  • 1
    \$\begingroup\$ Combining Spehro Pefhany and Majenko anwsers worked. I soldered the circuit on a PCB and used a BJT push-pull MOSFET gate driver. Thanks! \$\endgroup\$
    – Golaž
    Commented Dec 21, 2014 at 16:35

1 Answer 1


Combining Spehro Pefhany and Majenko solutions worked.

I soldered the circuit on a PCB to get rid of stray capacitances and used a BJT push-pull MOSFET gate driver, to supply more than a mere 20mA to quickly turn it on and off.


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.