Switched voltage across an inductor

[greanie]

I am on my first power electronics project but I am getting significant noise from my MCU's pwm signal on other lines. I am also having trouble validating my output voltage profile.

I have a simple circuit:



Across J1 is an inductor.

I have implemented this on a dev kit that includes a microprocessor and some spare board area. The linear regulator and FET are implemented on the spare board area.

The PWM that switches the N-type MOSFET at ~ 100kHz is generated by a microcontroller. The microcontroller is programmed with a PWM profile and then commanded serially to play that profile. Below is an example:



I am looking to verify that the resulting voltage across the inductor is as expected. I am using a PC-based oscilloscope(PicoScope by PicoTech). The resulting signals are as follows:



Channel A: PWM signal after a low-pass filter (roll-off ~ 2kHz, 1k-0.1uF, 200us settling time)
Channel C: Anode of diode, when the FET is on, it is ground and when the FET is off it is the low side of the solenoid which should be floating back to 12V (seems to be nonsense)
Channel B&D: Sensors that have been disconnected for now

(Channels B&D: have a digital low pass filter with a 10kHz roll off applied, otherwise, the noise from the PWM makes the scope signals unreadable.)

I have also a second scope running off a laptop (to avoid a ground loop that closes the FET) looking at the voltage across the inductor.

I am finding that when I set pwm to 100%, I get 100% out on both the PWM (5V) and across the inductor (13.7V). When I set the PWM to 50% (in firmware), I am getting 45% on the PWM (2.2V) and 30% across the inductor (4.1V). At 30% set in firmware, I am at 20% at the output of the low-pass filter on the PWM signal and at 10% of the full voltage across the inductor (again, low-pass filter in between). Can anyone explain the difference between the PWM signal going into the FET and the voltage across the inductor? Is it real or is it due to measurement error induced by the low pass filters?

Is there a way to measure the voltage across the inductor without using a second scope?

As already stated, I am getting a lot of noise from the PWM signal onto the other signals. I have had to put software low-pass filters on those channels even though they are essentially DC signals. Also, for example, on channel B of the plot provided, I am getting significant ripple even though the scope probe isn't connected. How can I better isolate my circuit from the PWM signal? Also, how can I measure the voltage across the actuator without creating a ground loop that causes the FET to close?

Any suggestions or ideas greatly appreciated!

 

[BradtheRad]

Not sure if this will cure the problem, but here is a technique to improve supply filtering to a sensitive device.

Add a resistor to isolate it from noisy circuitry. Add a smoothing capacitor.