Hall Effect Sensor on DC motor case vs EMF -> Tachometer?

squidx

I'm interested in using a Hall Effect sensor, (I've ordered the Melexis from Parallax for a start), to get the speed of a DC motor. Usually, I think you'd glue a magnet or something on the shaft, but that would off-balance my small motor, and I'd rather go for something a bit more subtle.

Does anyone know if these sensors are sensitive enough to be triggered by the magnets of the motor passing by as it runs? Would gluing it to the case of the motor be close enough to sense the mag field? Or maybe carefully cutting a hole in the case would do it? (Obviously, I'm thinking of a motor with moving magnets, not moving coils).

Alternatively, has anyone measured back EMF on a DC motor? I've hooked up a transistor between the negative wire and ground on mine and running the output through a speaker I have around, I am getting some interesting clicking noises, but I don't have an o-scope to confirm the validity. Also, I'm getting a steady hum, so I probably need to put some resistance on the transistor trigger.

Any thoughts on this would be helpful.

Alex

LilDi

If you are using metal gears, the magnet can be glued on the back of the Melexis sensor and by placing the front of the sensor near the gear teeth.

http://www.melexis.com/prodmain.aspx?sPN=MLX90217

Here is a schematic for a back-emf motor control PCB.

bulkhead

It might be just as easy, but with less mechanical/magnetic affect to use optical encoders. A simple QRD1114 and a dull/shiny patch on your motor's shaft/gear/moving part could do the trick. To take it even further, you could use two of them and create a "quadrature encoder" so you can know the direction too. Use resistors to adjust sensitivity.

As for back EMF, this idea has been explored by some. www.acroname.com/robotics/info/articles/back-emf/back-emf.html. I think it can be done several ways, one involving a quick ADC measurement of the back EMF voltage to determine angular velocity (this method could be prone to "drift" from integrating the angular position approximations over extended periods of time), and another that somehow "counts" the passing of the brushes over the gaps in the commutator (try googling for exact details, I can't remember). However, the first method could be dependant upon temperature (is it?) and the contact of the brushes on the commutator, which degrades over time, and I haven't seen much about the second method. I think it's because of those factors that could cause inconsistency in the readings that back EMF techniques haven't been widely used.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I'm new to the propeller!

LoopyByteloose

Some Hall sensors recognize each pole. If one were to use a magnet that had a hole in the middle, it would be balanced. In somecases, you might have to divide by two to get actual RPMs. And with some Hall sensors, a one pole system would not work as it expects to be reversed on th next pass.

I have doubts that you could easily use the magnetic field within the motor to measure RPM.· Sure, you have a rotating magnetic field, but it is adjacent to a strong and opposing stationary field.

Optical sensors are easily available and multiple holes can balance the weight. The only drawback is that they may require a cleaner environment.

If you want to use your computer as an inexpensive O-scope, their is free software availabe that works from the sound card.· While you are roughly limited to audio frequency range, that is certainly enough for RS-232 work and motor rotation. I've build a little two channel opto-isolated input to my computer.· If I really connect something wrong, the isolators take the damage, not the computer.

I've considered measuring back-Emf by supplying a PWM drive to the motor and reading the back-emf inbetween on pulses with an ADC.· From what I understand, back-emf is the opposite polarity as the voltage that is driving the motor, so blocking diodes might help limit the input to the ADC when the motor is operating in a driven forward mode.· I've know idea it this could be worked out with an h-bridge in place.


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
PLEASE CONSIDER the following:

Do you want a quickly operational black box solution or the knowledge included therein?······

···················· Tropically,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan

Post Edited (Kramer) : 3/26/2008 9:14:30 AM GMT