Electric Motorcycle Drive & Hall Effect Sensor Ringing

[jegues]

Evening gents,

We've built the following circuit in the lab for our DC-DC converter for an electric motorcycle.

The blocks from left to right are,

• High speed optoisolator
• IR2110 Gate driver IC
• MOSFET Power buffer circuit
• MOSFET Power Module
• PMDC Motor
• ~48V Lead acid battery bank
• Hall effect sensor

Also note the circuit operates with two isolated grounds, one being the signal ground(i.e. ground for microcontroller), and the other the power ground. (i.e. negative of battery bank)

When operating this circuit, we measured the voltage at the gate of the bottom power MOSFET within our power module with respect to the power ground (yellow signal on the oscilloscope) as well as the output voltage of our hall effect sensor with respect to the signal ground.(blue signal on the oscilloscope)

We are using the hall effect sensor to monitor the current passing through motor and to relay this information back to our microcontroller which will take appropriate action if the current is deemed to be outside of acceptable ranges.

Please see the schematic above, as well as the following youtube video: http://youtu.be/9R0xHGUImaM
As seen in the video, we observe some ringing on both the yellow and blue waveforms that seems to be somewhat synchronized, as if the cause of the ringing was from the same source, or at least somehow related.

You'll also note some more erratic high frequency ringing fading in and out of both signals. I'm fairly certain this erratic high frequency ringing fading in and out is simply due to triggering issues within the oscilloscope when the video was taken, but can't say for sure.

Does anyone have any idea what the cause of this sort of ringing could be? Or any further tests we can do to narrow down the cause of the problem?

Our goal is to obtain a "acceptable" signal from the hall effect sensor which will then be filtered and passed to our microcontroller for analog to digital conversion. It must be "acceptable" in the sense that the filtered version of the signal that is passed to the microcontroller still contains all the information that the microcontroller may need. In other words, the microcontroller should be able to distinguish between a 100A current and 150A current within a couple peroids of the switching frequency. (which for us is 20kHz)

I look forward to any advice, suggestions, or comments you may have me.

Thanks again!

 

[berkeman]

I didn't watch the YouTube video yet (I'm at work), but you should try a "ground minus ground" test to see if the noise is truly differential, or if it is common-mode noise that your oscilloscope is not rejecting (o'scopes don't have very good CM rejection generally).

To do the ground-minus-ground test, take the measuring tip of your o'scope lead off of the signal wire and connect it to the point where its ground lead is connected (leave the ground lead still connected). The o'scope trace should be a flat line with no noise on it. If the ringing is still there, then the noise is common-mode noise and not strictly differential. Knowing that can help you to trace the source of the noise...