Hall Effect Sensors

[protoserge]

Oh no! I forgot to put the Gaussian blur effect on!


I feel your pain with the atrocious camera phone shots trying to clarify something.

[Katna.Grace]

We (3132) are using hall effect sensors with a buffering circuit this year. We developed them in the off-season for our 2012 shooter wheels.

We use Hall Effect sensors from RS (http://australia.rs-online.com/web/p...-ics/7384743/).

We are mounting the board flat underneath our shooter wheel, and 'hanging' the magnets down off the wheel. The magnets are epoxied into a piece of aluminium hex stock that was turned to a taper and had a hole drilled in it for the magnet. A thread was then tapped into the other end of the piece, to mount it on the end of the bolt holding the wheel on.

When mounting them, make sure that the sensor itself can't move, so that it can't be sheared off by whatever it's measuring or drift away from the magnet.

I'm attaching a picture of the board, and a picture of our magnet mounts (before the pcb is attached), as well as the schematic for the buffering circuit. Let me know if you have any questions.

[electroken]

A useful gear tooth sensor that wires directly into a digital input is the Allegro Microsystems ATS667LSG. We like to use them when we need speed information but not direction information and we don't need the highest resolution. We also like that they're $6.

In the photo below an ATS667LSG is mounted to a carrier board and installed in the side of a Cim-u-lator gearbox, with its face about 0.03" from the output gear. It gives us a nice, clean 27 pulses per output shaft revolution. The sensor on top of the right motor is just posing for the photo.

[Ether]

Originally Posted by electroken

It gives us a nice, clean 27 pulses per output shaft revolution

How are you decoding the signal?

[electroken]

Originally Posted by Ether

How are you decoding the signal?

That's a good question. I told the student writing the code what to expect for a signal, and he said "Yeah, I got this".

[Ether]

Originally Posted by electroken

That's a good question. I told the student writing the code what to expect for a signal, and he said "Yeah, I got this".

The decoding method makes a big difference in the quality of the signal, especially at high speeds (depending in the method used).

Could you ask him to join the discussion here?

[BitTwiddler]

Originally Posted by electroken

A useful gear tooth sensor that wires directly into a digital input is the Allegro Microsystems ATS667LSG. We like to use them when we need speed information but not direction information and we don't need the highest resolution. We also like that they're $6.

Nice find! A few questions.
1. Where did you purchase them?
2. Did you wire them with the bypass and lowpass capacitors?
3. Do you have any data on how it performs on chain sprocket teeth?

[MrForbes]

Originally Posted by BitTwiddler

3. Do you have any data on how it performs on chain sprocket teeth?

.....steel? aluminum?

[Kevin Sevcik]

Originally Posted by MrForbes

.....steel? aluminum?

Steel. Gear Tooth Sensors have an embedded rare-earth magnet along with a hall effect sensor. They detect a change in the magnetic field that you're not going to see with a non-ferrous target.
EDIT: Also, Digikey has 22,690 in stock. Better order quick before they run out.

[BitTwiddler]

Originally Posted by Kevin Sevcik

Steel. Gear Tooth Sensors have an embedded rare-earth magnet along with a hall effect sensor. They detect a change in the magnetic field that you're not going to see with a non-ferrous target.

I think Jim was yanking my chain (also non-ferrous) for not being more specific.
Thanks for the source info though. We'll get our order in ASAP!

[MrForbes]

Originally Posted by BitTwiddler

I think Jim was yanking my chain (also non-ferrous) for not being more specific.

It was a serious question, Dave, because the sprocket is aluminum and I thought they only worked on ferrous teeth.

[electroken]

Originally Posted by BitTwiddler

Nice find! A few questions.
1. Where did you purchase them?
2. Did you wire them with the bypass and lowpass capacitors?
3. Do you have any data on how it performs on chain sprocket teeth?

1. Digi-Key (hurry... only 22,000 left!)
2. The capacitor hiding in the epoxy is a 0.1uF 50V ceramic across power and ground.
3. We haven't tried reading sprocket teeth, but I would expect them to work well with steel sprockets and a small air gap.

[philso]

Originally Posted by electroken

A useful gear tooth sensor that wires directly into a digital input is the Allegro Microsystems ATS667LSG. We like to use them when we need speed information but not direction information and we don't need the highest resolution. We also like that they're $6.

In the photo below an ATS667LSG is mounted to a carrier board and installed in the side of a Cim-u-lator gearbox, with its face about 0.03" from the output gear. It gives us a nice, clean 27 pulses per output shaft revolution. The sensor on top of the right motor is just posing for the photo.

Electrotoken, am I right in thinking that the magnets in the hall effect sensor in the ATS667LSG detects the disturbances in the fields from the magnets mounted in the sensor package as the steel gears in the gearbox rotates through the field? Did you just connect it into one of the digital inputs on the Digital Sidecar? The datasheet does not seem to show any sort of calibration. Did you just stick it on, connect the wires and run?

[electroken]

Originally Posted by philso

Electrotoken, am I right in thinking that the magnets in the hall effect sensor in the ATS667LSG detects the disturbances in the fields from the magnets mounted in the sensor package as the steel gears in the gearbox rotates through the field? Did you just connect it into one of the digital inputs on the Digital Sidecar? The datasheet does not seem to show any sort of calibration. Did you just stick it on, connect the wires and run?

Your thinking is correct in that the magnet is buried in the sensor. I don't know if the 0.1uF cap I have between power and ground at the sensor end is strictly required, but it makes me feel better.

It really is as simple as wiring into the digital sidecar and counting pulses. We used two of these last year and will be using three this year.

[carneeki]

Originally Posted by Katna.Grace

When mounting them, make sure that the sensor itself can't move, so that it can't be sheared off by whatever it's measuring or drift away from the magnet.

Hi Katherine,

We (FRC #4774) are going to use the same sensor boards! Thanks for making them open for others

We (FRC #3132) *1 forgot to include the value of T1 on the schematic. It is a 2N2222 NPN transistor, a TO92 package. Part of me is wondering if we can make all components sit flat so the board could be laminated in a future revision?

BR,

Adam

*1 (Yes, I am a mentor for two teams)