crankshaft simulation

[jaydee1]

Has anyone ever used a Hall Effect sensor to measure engine speed and crank angle? I'm trying to figure out how to have it measure crank angle of a single-cylinder engine using any DAQ system. I just need to know what parameters or VI's are used to translate the data output from a Hall Effect sensor into crank angle in degrees. I have a 35-teeth wheel attached to the crankshaft and the sensor installed right in front of it so i'm getting 35pulses/ revolution. I guess my question is how do i manipulate the pulses i'm getting into crank angle in degrees? Do i go about it like an incremental encoder or do i use tone extraction, etc. Any suggestion will be appreciated. Thanks.

[Robbob]

Jaydee,

The hall effect sensor will output a digital signal...so any digital board would work for this.  The problem here is that there is no index.  Encoders on motors often have what is known as a Z-Index, an extra pulse that shows where the reference location is.  Hall effect sensors are not doing any phase measurements(and they cant sense direction because of this).

I have four recommendations for you...from easiest to hardest.

1) Put the crank into a known position and zero out your system before running it. From here on, you know that you started at position X, so every digital pulse you get, you moved (360/35) degrees.  Awkward and prone to user error.

2)  Add another hall effect sensor, and get a gear with one tooth or an additional mark on your existing gear(maybe a hole in it?) Every time this one position pulses, you reset your other counter/position sensor.  This is a common implementation.

3)Implement a real encoder with a Z index.

4) Shave off one of the teeth of your gear. In software, detect when you have that extended low pulse and you can get the zero position here. Make sure the gear missing a tooth doesnt effect your mechanical performance(dont do this on a drive gear!) and you have to program this well to implement it, as your system may be tricked when your engine slows down rapidly

You dont need tone extraction or anything like that...just get a board with a counter. Use the count edges example in LabVIEW. Every digital rising edges is (360/35 = ~10.3) degrees.  Once you go over a full revolution, either use a modulus function to keep your degrees under 360 or reset the counter every time it hits 35.

[jaydee1 · Accepted solution]

Thanks Rob, but I already figured this out a while ago; i guess i should have deleted the post. I have a 36-teeth gear, removed one tooth to act as index (aligning it with the TDC marker on my engine flywheel) so that everytime, my sensor outputs a long pulse corresponding to 3 valleys (valley-missing tooth-valley), that indicates my TDC in every revolution. That, together with my engine speed data, makes it easy to calculate crank angle interval (which varies based on engine speed). Also, the hall-effect sensor doesn't output a digital signal. It's a proximity sensor outputting an analog signal that corresponds to teeth and valleys on the gear, so the signal essentially takes the form of a digital square wave, but is not actually digital. Depending on your supply voltage, you don't exactly get 0's and 5's, but you get 0.1ish and 4.7ish values which still simulate a square waveform, so an analog input module on the compactrio (which is what i'm using) would work just fine. All the same, thanks for your response. 

[Merc_Bert]

jaydee1, I'm looking at doing something similar with a 60 tooth wheel missing two teeth. How did you implement the missing tooth detection without messing up your RPM calculation? I realize this is a long shot since this is a old post but figure it would be worth a try.