Forums » Design Discussions »

Testing Hall Effect Sensors

Added by J. Simmons over 9 years ago

So, I am finally ready to start working on Holoseat again. I think the place to start is with . To that end, I have added several links about Hall Effect sensors to the Resources page. The main lessons I see from these links are:

  • There are 3 types of Hall Effect sensors (digital reed switch, latching, and linear). Watch the video on this page to see the difference. I think we should test out both the linear and the digital reed switch. The linear is what we have been discussing, but the digital reed switch may work just we well for our purposes and require fewer code changes (depending on sensitivity).
  • Many of the demos display the responses of the Hall Effect sensors on real time plots. This would be super helpful for testing out the behavior of the sensors before integrating into a new revision of the Holoseat.

Proposed next steps:

  1. Order a small selection of Hall Effect sensors to compare behavior in lab setting
  2. Set up lab tests for various combinations of Hall Effect sensors, magnets, and later pedal setups. Possible tests include:
    1. plotting sensor readings vs time
    2. plotting RPM vs time based on field strength and based on digital reed switches
  3. Extend desktop client to report debug messages using v0.2 hardware (this lays the ground work for both integration testing and for live communication between the Holoseat controller and the desktop app)
  4. Select single Hall Effect sensor design to replace reed switch and upgrade Holoseat

After this, we can come back around to looking at adding a second Hall Effect sensor to calculate the direction of pedaling and start implementing reverse walking.


Replies (7)

RE: Testing Hall Effect Sensors - Added by J. Simmons over 9 years ago

I am jotting down some notes on how I am thinking of implementing step 2 above. As I have been thinking about my last round of testing and how difficult selecting a decent sensor/magnet combination was, I realized what we need is a test rig that can spin a small wheel at a specified RPM so we have a known quantity we are measuring. This gives us a repeatable and reliable test framework for evaluating sensor/magnet parings. The following links/comments are my precursor to developing a block diagram and BOM for the test rig.

  • Basic plan
    • Wooden frame
    • Stepper motor
      • Small wheel mounted horizontally
        • Small metal plate on wheel to attach magnets to
    • Arduino
      • Running AccelStepper to enable stepper motor to be run at user specified constant speed (see also this forum post to see how steps can be converted to revolutions since this library uses steps/sec as input for speed setting)
      • Arduino also connected to the sensor under test
      • Arduino runs tests over various speeds commanded vs various speeds detected and reports them to control PC which plots of commanded and sensed speeds vs time
    • Use a bike computer with cadence to verify the stepper motor is running at commanded speeds
    • wooden mount for sensor under test with adjustable height so we can vary the distance between the sensor and the magnet

Plan to use old Arduino when I setup the actual test rig. Will likely repurpose the test rig when we get to v0.3 prototype to use it to test actual commands sent from Holoseat to computer.

RE: Testing Hall Effect Sensors - Added by Bryan Christian over 9 years ago

Sounds like a good plan. Nothing really to add to this.

RE: Testing Hall Effect Sensors - Added by J. Simmons over 9 years ago

Some additional notes.

  • Article covering how to select a hall effect sensor - one of the interesting points is a "sensitive" hall effect sensor is one that triggers at less than 60 Gauss, with an example high sensitive sensor triggering at 30 Gauss
  • This digital reed switch sensing range is 90G Trip, 65G Release, so not super sensitive, but might be worth testing to see what a more standard one is like
  • This digital reed switch is more sensitive: 60G Trip, 10G Release
  • This analog hall effect sensor is shown in the video linked above, unfortunately it comes from a different vendor than the first two sensors

A friend of Bryan's is using this sensor. Note, the spec sheet does not document its sensitivity, though comments indicate it is a digital reed switch style sensor. The built-in LED is an interest touch for debugging, but I think we might be better off going with bare sensors and using plots and other debug data to indicate sensing state. Worst case is we add an LED to the circuit like in the videos (this is probably what the sensor package is doing).

So, I think the next step is to do a BoM for the sensors (they need capacitors and resistors, etc) and then research about magnets. I have some 1/4" x 1/16" neodymium ones at the house I will include in the tests, but I want to see if we can include a wider range of strengths and sizes.

RE: Testing Hall Effect Sensors - Added by J. Simmons over 9 years ago

Bill of Materials (pt.1 - sensors)

Part # Description Vendor (linked) Unit Cost ($) Qty Total Cost ($)
TR 08 90G Trip, 65G Release Sensor Digi-Key 1.80 3 5.40
TR 09 60G Trip, 10G Release Sensor Digi-Key 1.83 3 5.49
TR 10 Analog Sensor Modern Device 2.75 3 8.25
TR 11 750 Ohm Resistors Pack Radio Shack ~3.00 1 3.00
-- Misc Hardware and Shipping -- 15.00 1 15.00
Total 37.14

RE: Testing Hall Effect Sensors - Added by J. Simmons over 9 years ago

On to magnets....

OK, so according to this page, very strong magnets will not damage that vendor's hall effect sensors. I am hoping that this is true of most vendor's sensors.

Found this great site for purchasing magnets (they include pull strength and surface gauss rating). I am thinking we could buy a set of different size/strength magnets to test. Possible examples include:

Ordering 5 of the first two, 3 of the next one, and 2 of the remaining ones, with shipping would cost $16.30 (ordering multiples in case some get lost, more of the smaller ones since they are easier to lose).

RE: Testing Hall Effect Sensors - Added by J. Simmons almost 9 years ago

Well, that's disappointing... The gaussboys site has a busted SSL cert. Will have to do some more shopping to find a magnet supplier.

(1-7/7)