Page 1 of 2

Train detector, welded rails - ideas???

Posted: Tue Feb 19, 2019 8:35 am
by DianneB
Our club has flat-iron rails with ties welded on every few inches. It is a good reliable system that stands up well over time but makes 'train detection' a challenge!

We have one blind spot just coming out of a corner where the forward view is limited by the clubhouse just after the curve. More than one engineer has had to make a "rapid speed change" when coming around the corner to find a stopped train hidden behind the clubhouse! With a major meet coming up this summer I would LOVE to be able to install a warning signal before the curve to indicate if there was a train behind the clubhouse!

I have experimented with brass contacts that make contact with the wheel flanges to detect a train but between getting stepped-on and corrosion, they didn't last very long. Industrial micro switches are too expensive.

We don't have a clear line of sight down the hidden track to use an optical beam-break detector. I could use a number of photo cells between the rails to detect the shadow of a passing or standing train but that would require a LOT of wiring and keeping the photo cells clean would be a problem.

Cutting the rails to make 'sections' is not an option.

I have scratched my head for years to find a solution but have nothing yet. :(

I may experiment with making a LONG coil (20 feet?) with outdoor data cable and placing it between the rails to see if I can detect any inductance change when a train passes over.

A mirror, if we had a place to install it, might give a view round the corner.

Any other ideas, ANY AT ALL???

Re: Train detector, welded rails - ideas???

Posted: Tue Feb 19, 2019 11:20 am
by rkcarguy
I like the inductance idea, didn't you use that before in another area of the RR?

Re: Train detector, welded rails - ideas???

Posted: Tue Feb 19, 2019 11:29 am
by cbrew
look into Ultrasonic Distance Sensors, they are used on cars, a good number of hobby kits on the market. just need to find a water proof version.

Re: Train detector, welded rails - ideas???

Posted: Tue Feb 19, 2019 11:46 am
by NP317
I like Chris's idea of the ultrasonic distance sensors for operating a signal.
Mirror is second choice for me. And simplest.
~RN

Re: Train detector, welded rails - ideas???

Posted: Tue Feb 19, 2019 11:51 am
by cbrew
yeah, the only issue with using anything but true track detection is how to tell of the train is still present.
can have a sensor counting in and out but nothing is fool proof.
I even looked into using RFid tech at one time. but it was just not cost effective.

Re: Train detector, welded rails - ideas???

Posted: Tue Feb 19, 2019 7:39 pm
by John Hasler
Looking for inductance change should work but I think you want the axis of your coil vertical. When a train is present part of the flux path that is normally entirely through air will be through iron. You might be able to modify a metal detector. A car detector of the sort used at parking lot entrances would also work and possibly only need recalibration.

Re: Train detector, welded rails - ideas???

Posted: Tue Feb 19, 2019 8:00 pm
by rkcarguy
How about a magnetic switch of some sort? One at the entrance to this section of track would "capture" it, the other would release it?

Re: Train detector, welded rails - ideas???

Posted: Tue Feb 19, 2019 9:04 pm
by Mark Landgraf
In Europe, the prototype signal systems use a coil that is attached to the side of the head of the rail. It is set up to count wheels as they pass. If 100 wheelsets enter the block, and a 100 wheelsets exit the block, the block is clear, otherwise the block is occupied. Traffic signal loop detectors would work too. If you use PLC's or other delicate electronics, it would be best to provide some sort of optical isolation so as to avoid damage from nearby lightening strikes. You might also consider LED emitter/detector units that can be place far enough apart that a wheel can pass thru the gap. Microswitches that can be depressed by the flange could also work. Personally, I like the electric eye light beam set at coupler height, with both units on the outside of the curved track, about 40-50 feet apart (so the beam passes thru the track centerline). Several of these pairs placed along the curve could provide adequate detection.

The new big rr PTC signal installations use a proximity unit for switch point detection. They contain two coils. One is constantly powered to set up a standing wave and the other coil is used to monitor changes in the standing wave as steel passes by or is in immediate proximity.

Mark Landgraf

Re: Train detector, welded rails - ideas???

Posted: Tue Feb 19, 2019 10:04 pm
by chiloquinruss
There is also the option of infrared detection. To warrant clearance make two circuits, Set them up as a trigger on, then at the other end set trigger off. Russ

https://www.adafruit.com/product/1568

Re: Train detector, welded rails - ideas???

Posted: Tue Feb 19, 2019 11:58 pm
by John Hasler
The "car detector" I mentioned is the same as a traffic signal loop detector.

BTW the PLCs I have had experience with are far from delicate. I think they are too pricey for this application, though.

Multiple LED-phototransistor pairs could be wired with all the LEDs in series and all the phototransistors in series, requiring only three wires. You would then get a "train present" signal as long as any beam was interrupted.

I don't have solution to the cleaning problem but you could get redundancy by using clusters. Have clusters of three LEDs with every LED in the cluster illuminating all three phototransistors in a matching cluster on the other side of the track. The phototransistors in the cluster would be in parallel (but the clusters would still be in series).

Re: Train detector, welded rails - ideas???

Posted: Wed Feb 20, 2019 6:25 am
by DianneB
Thanks for all the suggestions gang - gives me lots to think about.

Re: Train detector, welded rails - ideas???

Posted: Wed Feb 20, 2019 9:39 am
by kcameron
My understanding of wheel detector setups are the sensor is really a pair of them. This lets it compute which direction each wheel is passing. That gives you a shot at keeping the count right when a train enters, but then backs out of the block. The sensors are close enough so it starts to detect at the second detector while it is dropping from the first detector. This relation is to prevent any false actions. Think of the detectors showing two humps for each wheel. So depending on which rises first, that gives the direction. It then adds or subtracts from the wheel total. I've not tried to build one however, just read about the technical ideas. Our track uses the manual button system for our signals on 1", the 1.5" uses isolated rails and automatic detection.