gunrunnerjohn posted:I like the train direction indicator, I'd like to figure out an easy and bulletproof way of determining which way a train car is moving. Ideally, this could be done without wheel sensors, I'm thinking some sort of accelerometer, if out phones can do it, we should be able to as well.
For less than $2 shipped you can get an eBay 3-axis accelerometer module with analog outputs based on the Analog Devices ADXL335B sensor chip which puts out 300 mV/g. I'd think this could be the basis for an inertial direction indicator (no wheel connection) and doesn't require a microcontroller or programming.
But in train applications there could be the problem that you really want to indicate the "next" direction of motion even before it happens. Like the backup light on a automobile (or warning beeper on a truck), it activates when you put the vehicle into reverse even before actual movement. This makes sense in that the purpose of a backup light is to indicate/warn. In a previous application, the objective was to turn on the reverse headlight on the last car in a subway consist. Using a wheel indicator, the reverse headlight would only come on once the consist started backing up. The alternative was to slave the reverse headlight to the lead car's reverse light via a wireless link (no tether). So when the lead car with the motor drive was put into reverse (whether from conventional or command control) the trailing car's reverse headlight immediately slaved to the "next" direction of motion even before it started moving. This is what I think you'd want.
Now, as for installing a direction sensor in a piece of rolling stock, what is the application? That is, I assume it's not for directional headlight indication.
Of course to steer this on-topic, as we are talking about optical sensing, I've wondered if anyone has taken the time to unravel the mystery of the optical mouse which uses image processing to detect motion. It's amazing that for $10 or whatever you're getting a device that is continuously taking what amounts to a video of the surface and comparing pictures to detect which way the mouse is moving...no moving parts! So in principle such a device could be focused on the track bed to detect movement. Unfortunately all the one's I've seen use red LED to illuminate the surface so you'd see a red spot on the track bed; but in principle the same optical processing should apply to illuminating the track bed with IR or something invisible to the eye. Again, this method can only detect actual motion as opposed to the "next" direction of motion.
