Pwm signal on analog oscilloscope.

Thread Starter


Joined Oct 3, 2012
Hello, can I analize a pwm signal with a analog oscilloscope? I have a 150mhz scope, I want to check the gate signal of the final stage in a Class D amp. The result is a sequence of dirty signal like this. ( image from internet, but what i see on the scope is the same).
Dirty signal.JPG

Regards, Renato



Joined Oct 2, 2009
What you observe is normal.

You are seeing repetitive PWM pulses with different pulse widths. Adjust your HOR sweep to a longer time scale and you will get a better picture.

Also set your oscilloscope to trigger once and then stop. You will be able to pick out each PWM pulse.


Joined Jan 30, 2016
If your 'scope has a persistence (or digital phosphor) function, adjust it so you can see both the rising edge, the most extreme (RH) falling edge and the following rising edge. What you are then seeing is called an eye diagram. It shows you two key pieces of PWM information.

#1 the time period of your PWM
t = (rising edge 2 - rising edge 1) - which for true PWM should be constant; and

#2 the minimum and maximum duty cycle
Dmin = (rising edge to shortest falling edge)/period,
Dmax = (rising edge to longest falling edge)/period.

To get the specific duty cycle for a given input voltage, you need a static input. With a typical AC coupled amplifier the only useful figure is with the input grounded for which the output PWM duty cycle should normally be 50%. The eye diagram will show the total variation either side of 50%, which represents your system noise level. There may be a bias adjustment for the amp to set 50% for no input.

Edit: If this is the gate signal. you can see Vgs-max & min, and the rise time + damping, and by triggering on the falling edge, the fall time + damping. What else did you want to see?
Last edited:

Thread Starter


Joined Oct 3, 2012
Yesterday i did a bit of test, with the right configuration, as you say, i was able to see the correct wave (or just part of it) and with the two channel that them are correctly opposed. So the mosfets are correctly driven. Many thanks for help.