Another Sensor

Here's the sensor from Neuhold mentioned on the ft:c forum. The breakboard is supplied without pins, so some soldering is required.
I soldered standard 0,1" HE14 headers and used a nice blue Berg connector (reference). The flat cable ends with 5 fischertechnik plugs.
To place the accelerometer on a model, I use duck-tape and a 15x15 red plate (reference 38246). 

Clip-on plate could be even better (reference 31506).
The connectors made the use of the accelerometer compatible with a breadboard. 

Comparison Between Sensor Implementations

The same accelerometer IC is used is the two break-boards:
  • left: Neuhold
  • right: Dimension Engineering
Apart from the voltage regulator and the accelerometer, the Dimension Engineering board includes a third IC, an operational amplifier. 

This avoid the nasty bar-code from appearing on the TX controller screen during boot-up.

When the Neuhold accelerometer is connected to the TX controller, the following screen may pop-up when turning the power on.

Here's the detailed explanation I received:
This bar code screen shows the serial number of the ROBO TX Controller. 
This screen shows up if the input I1 detects a resistance of 13 kOhm (+/- 1kOhm) during the start of the controller. 
This is far out of the specified measurement range (0 - 5kOhm). 
So normally this case shouldn't happen if you connect standard ft-sensors to the input I1.
By pressing one of the controller buttons you can return to the normal operation mode.
How to implement those boards into a fischertechnik model?

Neuhold Implementation

Dimension Engineering Implementation

Implementations Comparison Protocol

Both are positioned for the test! How do they perform comparatively? Are the measures consistent?

Here is the model I built to test the consistency between the Neuhold and the Dimension Engineering implementations of the accelerometer
The Robo Pro project is fairly classic, based on the CSV output file.
Picture

Sensor Implementations Results

To compare the two implementations of the same sensor, I use the linear model basic correlation coefficient.
  • The closer the coefficient to 0, the worse the correlation.
  • The closer the coefficient to +1 or -1, the better the correlation.
Picture
Results are excellent.

Both implementations of the sensor are consistent.

The y-axis (y1 and y2) wasn't tested on the model, so figures and non significant.

Previous

Next