02-05-2018, 03:56 PM
(02-02-2018, 08:51 PM)Canard Wrote:(02-02-2018, 02:41 PM)KevinT Wrote: The signals at Bearinger Rd worked perfectly but the ones at the Quiet Pl and Old Albert St pedestrian crossings still clearly needed some tuning.
I generally very much appreciate that "things are hard", and am usually the first to hold up my hand and stop people from being negative toward crews and so on. I work in automation, and design complex, million+ dollar systems for cleanrooms with 20+ servos and hundreds of sensors.
...but I absolutely cannot fathom why it is so hard to get railway signals to work properly. Why is this so hard?!!?! IT'S ONE INPUT.
I suspect it's at least 6 inputs, 4 of which are analog.
For each track (northbound and southbound) that crosses a road there would be a basic 'are the rails shorted together' circuit in the immediate area of the road and sidewalks to tell the system when the crossing is occupied (digital input), and an analog approach circuit on either side for a predictor to detect not only when a train is coming, but how fast it's closing in. Each of the analog sections would need to be tuned, tested, then tuned and tested some more to get everything dialed in just right.
Even the 'digital' inputs are actually analog, because they have to use AC frequencies for detection because of the DC traction power return in the rails. Each of these AC signalling currents must be isolated from the DC currents, and detuned from the AC noise generated by the motor controllers in the Ion vehicles. I could easily see that taking a lot of tweaking, multiplied by X number of crossings, plus integration with the signalling and detection systems that direct the train movements themselves.
If anything, it's fiendishly complex (although should be at least somewhat familiar to the partners that make up GrandLinq as they've already done this in other parts of the world).
...K