Curious as to what causes this signal

[DamoRC]

I am building a inertial chassis dynamometer for my electric RC car.
It's a pretty basic setup, a couple of drums mounted on bearings.
To measure the RPM of the drum I am using a simple coil mounted on the frame and a small disk magnet (pulled from an old headphone speaker) mounted on the end of the drum. As the drum "rolls" the magnet passes the coil and creates some current flow etc etc.
To measure the RPM, the coil is wired to a headphone plug that I insert into the microphone jack of my laptop. I then wrote some software (in VB using the winmm.dll WAVEINXXX functions) to "listen" to the input from the microphone jack and count the pulses.

**broken link removed**[/IMG]

This all works just fine. But I am curious to know why the signal I get for each pulse looks the way it does, specifically the portion highlighted in the green ovals in the attached pic. Even though this is working, I was wondering if there is some simple component addition I could make to clean this up.



Thoughts / Comments appreciated.

 

[Hero999]

It's probably noise picked from another source being picked up by the coil. It could be from the motor's commutator, who knows?

 

[MikeMl]

A two or four pole low-pass filter with a cutoff just above the highest frequency you will measure would eliminate it, and leave just a sine-wave at the fundamental frequency.

 

[DamoRC]

Thanks for the filter suggestion Mike - I will read up about these.

Hero - I was less concerned about the noise in the region (which as you pointed out, is likely motor interference). I was more interested in why the signal has gone "high" in this region.
The signal moves to a max as the magnet approaches the coil and then moves to a min as the magnet moves away. But then it "bounces" back over the midline and drifts back down. I was wondering why it just doesn't return to the midline

 

[BrownOut]

It migh be 2nd order damped oscillations from your coil and stray capacitance, imposed on the signal and spurious niose. I'd like to know more about the VB program and any online information about sampling and displaying data you used for your project.

 

[DamoRC]

Thanks Brownout. Is there a way to eliminate this effect (or will the filters that Mike suggested deal with this)?

On the VB thing, the code I have is a little buggy right now but a google on scopes, winmm, waveinopen and VB should reveal a wealth of info. When I get the bugs worked out I can pm the source as it probably doesn't belong here.

 

[BrownOut]

Actually, after thinking about it, the midline signal is correct. Think about it, it swings high (positive) as the magnet approaches, then low (negative) as the magnet passes. At all other times, the magnet isn't affecting the coil, and thus the absence of a signal means a mid output. The overshoot and undeshoot might be explained by the weak effect of the magnet as it aligns with the coil around 180 degrees. The rest of the niose is probably inductive pickup caused by vibration or something. I'd just peak detect the signal, and call it a day.

 

[MikeMl]

What is the period of the signal in msec? i.e. how many samples per second.

 

[DamoRC]

I am collecting (in this example) 8000 samples (bytes) per second. Peak to peak is approx 110 bytes or 13.8 msec.

 

[RCinFLA]

Try putting a load resistor on the pickup coil.