Model Testing - Gyroscope

Tim Lovett Aug 2004 | Home | Menu

A Solid State Gyroscope used to track motion in a ship model

One option for comparing the effect of variations of hull geometry is to run parallel tests in open water. The differences between the models should yield meaningful data if the tests are run for long enough periods of time. Testing the models in a large lake or bay than a wave tank means that the tests can be run for extended periods of time - and in windy conditions where broaching risk is being investigated.

Solid State Gyroscope

Two important measurements are Roll and Pitch. They also happen to be pretty easy to measure, especially with the introduction of the solid state gyro. These units are popular with robotic experimenters and things like model helicopters. The gyro has no moving parts, sort of. Instead of a rotating flywheel the angular accelerations are picked up by a motion change in a vibrating cantilever. The gyro unit can be seen mounted in a 3D computer mouse below.

It uses Piezo gyros that are angular rate sensors, not linear references. There is no flywheel like a mechanical gyro mounted in gymbals, but are solid state and have almost no moving parts. The Murata ENC05-E sensors used in the older ones draw 2 mA. The MG100 used in the new ones draws about 20 mA. The basic concept of the piezo rate gyro is a focault pendulum the size of a pencil lead. A silicon beam vibrates and when it turns the vibration remains in the inertial frame of reference. Angular accelerations result in Coriolis effect - continuation of the original motion which alters the vibration mode.
Another type, the MEMS wineglass gyro based on making a 1 mm hoop of silicon, and setting up vibration similar to the mode you get when you run a moistened finger across the edge of a wineglass.
Once again, the vibration essentially stays in the inertial frame allowing you to sense rotation. The neat part is that the sensor is nickel plated silicon on a chip. This means small size (2mm^2) small power consumption and chip fab manufacture. As yet, they are not so easily sourced.
The gyration uses a clone of the ENC05-e. The MG100s are larger and use more power.

Gyroscopic Acceleration Sensing - the easy way...

A commercial gyroscopic computer mouse is capable of recording 2 orthogonal rotation axes simultaneously. The unit can be very simply fitted on board the model. Notice how the mouse is vertical so that mouse yaw measures the model roll, and mouse pitch measures model pitch.
(Only an engineer can put up with sticky tape when doing these things - I'll do it properly later. And if my students ever had sticky tape on their projects...)

That's about it for the hardware. The cordless mouse comes RF linked to a receiver plugged into the computer as a standard serial mouse. Can't get much easier than that, and with a range of 60-100 feet (according to the specifications), there should be no need to mount the laptop on board the ark model either. (Just as well, because it won't fit anyway.)

Recording the data

This really is easy stuff. Well, I suppose the only trick was to keep track of the mouse even if it moves off screen. To do this I just cheated. Reset the mouse back to center-screen if it goes too close to the edge, and counting each time. Of course, the acceleration setting of the mouse must be turned off in the Windows settings. The sensitivity had to be quite high, otherwise the mouse would not respond to very slow movements - allowing too much drift. It is accelerations we want anyway, so some drift is OK.

Visual Basic data logging

At first I thought I'd steer clear of VB running in Windows for mouse tracking, thinking the response would be too slow. It turns out to give 60 readings per second, including writing the data to a file. This is certainly adequate. So this grand high tech solution uses the good old MouseMove command, and about 2 screens worth of code.

(Visual Basic 6, Celeron 366 running Win98. The mouse is on PS2 port. Yes, that is the kitchen table, and a 6V bike battery...)


Here is the very first test data displayed in Excel. We still have to check that movement scales linearly and a few things like that.


Once the mouse gets properly mounted and I'm confident it won't be drowned, it's off to the nearest body of water with suitable waves.

The Plan

Make two model arks. One is modified and compared to the other to give a measure of the relative improvement of certain features. Here are some of the ideas that could be tested.


1. Gyroscopic cordless computer mouse. . The one I used is actually an outdated model, which is better because it comes with PS2 and RS232 ports which are rapidly going out of fashion.

2. VisualBasic 6, Excel and Win98 are Microsoft products. (Also a bit old now, but matching the laptop. Won't catch me out there in a row boat with my good laptop)