Train Detection Sensors and Methods

[C855B]

Like with everything else MRR, compression has to be considered. I see what you're saying, but you're not going to reasonably get 30-second pre-activation on an N-scale layout. I don't even know if you'd want it - think of all the N scale motorists you're inconveniencing. One speed detector pair on each approach should work fine @ ~4' or so.

Hmm... FWIW, yard limits here on BNSF - to my knowledge - are restricted speed, the FRA definition of "restricted" being 20 mph (plus a whole load of qualifiers). Something special about your location, or is it something historical?

[Mark W]

I always wanted to see a version that would vary based on speed.  I think proto signals are 30 seconds.  Using Arduino should make this very possible, you just may need several detectors.

Jason

This is actually my goal.  Warning period will be cut to about 5-8 seconds though. 

[wcfn100]

This is actually my goal.  Warning period will be cut to about 5-8 seconds though.

I'll get some popcorn. 

Jason

[wcfn100]

Hmm... FWIW, yard limits here on BNSF - to my knowledge - are restricted speed, the FRA definition of "restricted" being 20 mph (plus a whole load of qualifiers). Something special about your location, or is it something historical?

I should probably check that I'm not confusing it with the Cedar Falls Branch, but I'm still pretty sure it was 10mph.  As for anything special, besides the actual yard, within the 2 miles of track that went through Waterloo you had multiple grade crossings, a Train Order tower, an 8 span bridge crossing (on a slight grade) and 90 deg turn which crossed another railroad and a few blocks of street trackage* (not actually down the street in my era, but the street was divided by the tracks) and with that trackage were more crossings.

So...pick one. 

I think the info is in a post-passenger era TT but of course I can't seem to find that one.

Jason

[C855B]

Likely the biggest issue was that curve. We have something similar here - NS has a 10mph curve through five grade crossings including a divided US-numbered highway. Their 8-9000' container trains pretty much snarl the world.

[jagged ben]

The speed sensing is generally how the prototype works. This is why you frequently see a short-but-fast train approach a crossing, activate the gates at the speed the train entered the circuit, but the train stops short of the gates... and the gates go up after several seconds. This is a timeout to cover this situation. When the train moves into the "close" detection block, the gates lower. Sometimes if the train accelerates too fast from stop it is in the crossing before the gates are fully down - I've seen this multiple times. IOW, don't over-think this, the 1:1 situation isn't perfect, either.

There are also some rules about this, right?  i.e. the engineer is supposed to know that if he stops within a certain distance of a grade crossing he is then supposed to approach it very slowly when he starts moving again.  IOW the human factor is also at work. 

[SSW7771]

Mark - this is very interesting, I will be watching your progress closely.

[C855B]

There are also some rules about this, right?  i.e. the engineer is supposed to know that if he stops within a certain distance of a grade crossing he is then supposed to approach it very slowly when he starts moving again.  IOW the human factor is also at work.

Frequently there are special instructions covering specific locations where crossing signal activation during switching might confuse the gate/flasher timers, advising restricted speed, etc. Also, you will see rectangular "STOP" signs at certain low-traffic crossings where the train is to stop, wait for the gates to lower, then proceed across.

A strange workaround for this problem I encountered "back then" was the crossing next to the SP station in Bakersfield. The crossing circuit was setup to allow passenger trains to stop with the pilot almost-but-not-quite in the street, and the gates would raise after a timeout. It was a servicing stop and they'd be there 15-20 minutes. The weird part is there was a microphone with a big megaphone on a post just across the street. It was part of the crossing signal circuit - blow the horn, and the gates would lower.

[peteski]

I think that in N scale, where time and distances are compressed, having 6 sensors per track is a bit of an overkill. 4 sensors should be enough to detect the direction and the train entering or leaving the crossing area.

[C855B]

I might be getting what you're saying, that you can assume that trains are likely to be moving within a relatively narrow speed range and arbitrary timing becomes a "good enough" - when you hit a single approach sensor, you can then assume what direction it's coming from and that the train will be getting to the crossing within a few seconds +/- a few, and if it doesn't hit the crossing zone within a few + a few seconds, then timeout the gates.

[Mark W]

I'm designing this for best possible conditions.  In practice, I'm using 6 sensors when possible, 4 when restricted.  A pair will lie about 3' from the crossing, approx 6" apart.  This pair reads the speed of the train.  Once the speed is set, the sketch will then calculate the time it will take to reach the crossing at the recorded speed and delay the warning until time to reach <= 6.  This should be immediately for a train traveling at ~60smph.  The last sensor will be about 3-4 inches on the far side of the crossing to end the warning once clear.

Tonight I picked up a bunch more photocells and built a test loop of unitrack.  Instead of adding servos to the proof of concept, I've programmed the speed reader which is now working, but just barely.  I need to clean up the script quite a bit. 

[peteski]

Unless you have a huge-a$$ layout where the distance to the gates can be properly represented, the delay is also not needed.  With 4 sensors, simply activate the gates as soon as the 2nd sensor gets triggered (indicating speed and direction). If the sensors are far enough away from the gates, I suppose that you could introduce speed-dependent delay.  Then release the gates once the train departs the crossing (when it uncovers the last sensor - in either direction).

I realize that all this accurate emulating of the prototype action is cool, but is it really feasible or needed?  But have at it - it will be interesting to see how it works.

[railnerd]

I realize that all this accurate emulating of the prototype action is cool, but is it really feasible or needed?  But have at it - it will be interesting to see how it works.

Since Mark is doing Free-moN, the "layout" is pretty long.  What he's doing is probably overkill for an HCD, but might work great for a 8' module that is part of a longer, linear layout.

-Dave

[wcfn100]

Likely the biggest issue was that curve. We have something similar here - NS has a 10mph curve through five grade crossings including a divided US-numbered highway. Their 8-9000' container trains pretty much snarl the world.

I was wrong, it was 20mph.  The 10mph was for East Dubuque.


Jason

[jagged ben]

A strange workaround for this problem I encountered "back then" was the crossing next to the SP station in Bakersfield. The crossing circuit was setup to allow passenger trains to stop with the pilot almost-but-not-quite in the street, and the gates would raise after a timeout. It was a servicing stop and they'd be there 15-20 minutes. The weird part is there was a microphone with a big megaphone on a post just across the street. It was part of the crossing signal circuit - blow the horn, and the gates would lower.

That's awesome.  Try modeling that.