Switch Point Detection continued

This forum is dedicated to Riding Scale Railroading with propulsion using other than steam (Hydraulics, diesel engines, gas engines, electric motors, hybrid etc.)

Moderators: Harold_V, WJH

User avatar
ChuckHackett-844
Posts: 60
Joined: Wed May 03, 2017 3:54 pm
Location: Tampa, Florida

Switch point position detection

Post by ChuckHackett-844 » Thu Oct 19, 2017 3:01 pm

We currently use turnout motors that are supposed to provide an output that indicates that the point is hard over one direction or the other but the micro switches in the turnout motors are unreliable for low current/voltage circuits. When they fail, people think that the signal system has failed (no green on double-head route choice signal) but it's actually the switch motor that has failed.

Does anyone have a design for a reed-switch (inherently waterproof) device that connects to the points to show "actual" as opposed to "commanded" point position? (i.e.: the reed-switch is attached to the points and would not be activated if there was a rock in the points, etc.)

I know I can design one but alas "so many projects, so little time".

A tested design will work, better if we can purchase some.

Regards,

Chuck Hackett (UP844)
Regards,

Chuck Hackett, UP Northern 844, Mich-Cal Shay #2
"By the work, One knows the workman"

Atkinson_Railroad
Posts: 137
Joined: Mon Jun 08, 2015 6:27 pm
Location: Michigan
Contact:

Re: Switch point position detection

Post by Atkinson_Railroad » Thu Oct 19, 2017 4:52 pm

Another new product idea for the hobby.

Envisioning a permanent magnet assembly attached securely to a point detector rod
and a weather protected adjustable, yet “fixed” reed switch combination mounted adjacent
to the turnout motor.

How many contacts normally open and normally closed would the reed switch arrangement require?
I’m guessing it would depend on the end user’s needs and that a single pole double throw
reed switch would be adequate in most designs.

Primitive illustration attached.

John
Attachments
SWITCH POINT DETECTION.jpg

User avatar
ccvstmr
Posts: 1469
Joined: Sat Jan 04, 2003 10:37 am
Location: New Lenox, IL

Re: Switch point position detection

Post by ccvstmr » Thu Oct 19, 2017 6:19 pm

Sounds like you're well on to something there. But you're right...detecting the position of the switch motor is NOT fail-safe. Shouldn't care where the motor is positioned...doesn't tell you where the points are. Instead, should detect the device being controlled...the switch points. Will add... detect the position of the hardware at a point furthest from the motor. Need to be sure all parts of the mechanical switch throw are moving.

Can understand the idea of using "sealed" components for weather protection. Would think there's low force activation switch of some kind that can be used. Question is, would those be reliable in the "dirty" railroad world? Aware your home RR does not use slap switches...but there are plenty of RR's that do use those for manual block control signals.

If you're going to incorporate a switch position control system with some type of automation and have (2) switch position detectors...it would be even MORE fail-safe if logic were added to prove one detector closed while the other detector is open. You'll can add a time delay to allow for point travel (when neither position detection switch is activated). In this manner, all the controls are being tested every time the switch points are moved. And...if there is ever a time when both position detection switches are activated...that should generate an immediate fault. An indication of tampering or foreign matter is preventing proper switch operation.

In short...there's more to a fool-proof, fail-safe, switch point detector system than the position detection switches themselves. Hope you'll keep this board informed of your progress and final solution. Carl B.
Life is like a sewer...what you get out of it depends on what you put into it!
I don't walk on water...I just learned where some of the stepping stones are!
I love mankind...it's some of the people I can't stand!

User avatar
ChuckHackett-844
Posts: 60
Joined: Wed May 03, 2017 3:54 pm
Location: Tampa, Florida

Re: Switch point position detection

Post by ChuckHackett-844 » Thu Oct 19, 2017 8:04 pm

John: I have lots of designs - just no time to implement them! :-) I was hoping someone would say "I have 10 I can send you!" :-)

I have experimented with many different reed switches most of them have way too much hysteresis (movement required to release after closure or vice-versa) to allow them to be used to detect (for example) if the switch point is within say 1/32" of the rail.

I have discovered that the push-in type window security alarm sensors (two white buttons, one a reed switch and one a magnet) do have a fairly small hysteresis and two sets could be used (one set for each point direction). I would arrange it so that they were field-adjustable (screw to adjust position). This is what I'll probably make use of some day with injection-molded mounts to easily attach one set to each point.

Another advantage of not using the micro-switches in the turnout motors is that they can be applied to manual-throw turnouts and still participate in the signal system.

Carl: It's very difficult to find a "micro switch" that is environmentally sealed, physically sturdy (derailments tear things up) ... and inexpensive. Magnetic reed switches are inherently dirt/sand/weatherproof (sealed in glass) and very cheap.

Agreed, must detect "point" position rather than motor position to give a "clear" signal but this is not done on a lot of railroads due to cost/convenience.

Yes, I always (even if using the turnout motor feedback) use two switches. Therefore there are three "states" of the turnout ... Left (left micro-switch activated), Right (right micro-switch activated), "in transition" (neither micro-switch activated), and "Error" (both activated ... supposedly can not happen, or transition timeout). These states are detected automatically by my signal system controllers and result in appropriate signal aspects (as defined by the user).

The switch motors are not really my responsibility but, when they fail (the points either don't move, or they do move but the micro-switches do not send the indication to the signal system) it LOOKS like the signal system has failed but it is just doing what it's told and passes that on to the engineer .. "Neither route is clear" and I get blamed because the engineer goes over and the points appear to be in the proper position - not fair but that's life.

Regards,

Chuck
Regards,

Chuck Hackett, UP Northern 844, Mich-Cal Shay #2
"By the work, One knows the workman"

User avatar
ChuckHackett-844
Posts: 60
Joined: Wed May 03, 2017 3:54 pm
Location: Tampa, Florida

Re: Switch point position detection

Post by ChuckHackett-844 » Thu Oct 19, 2017 8:08 pm

John,

Comment on your design: I think you need two separate switches since, due to construction variations, etc. the point gaps will not always be exactly the same. The ability to adjust the exact position of each point detector independently fixes this problem.

Regards,

Chuck
Regards,

Chuck Hackett, UP Northern 844, Mich-Cal Shay #2
"By the work, One knows the workman"

Atkinson_Railroad
Posts: 137
Joined: Mon Jun 08, 2015 6:27 pm
Location: Michigan
Contact:

Re: Switch point position detection

Post by Atkinson_Railroad » Thu Oct 19, 2017 10:16 pm

Chuck,

I knew you were going to point out the lack of a second switch! Ha Ha : )

Hence, the poor excuse disclaimer of the "primitive illustration" noted in my earlier reply.

Thank you for tossing out your Switch Point Detection question among the crowd.

Signaling has been a long, WAy over due neglected subject within the hobby.
You have taken it to a whole new level for the long time dwellers.

I'm still working on my "Searchlight Retrofit" and have made it this far: http://www.youtube.com/watch?v=Fez-nJA7ang

John

User avatar
ChuckHackett-844
Posts: 60
Joined: Wed May 03, 2017 3:54 pm
Location: Tampa, Florida

Re: Switch point position detection

Post by ChuckHackett-844 » Thu Oct 19, 2017 10:39 pm

In the video you mention a "Reverse Protection Diode" - unless you are attempting to protect the LED from voltages that exceed its reverse breakdown voltage I don't think you need this.

Even if you are, without seeing the circuit, it's not clear to me how the applied "reverse voltage" would be distributed among the three series (assumption) components.

i.e.: how much would be across the (reverse-biased) protection diode and how much would be across the (reverse-biased) LED and how much would be across the resistor (due to leakage).

I think, depending on exact characteristics (reverse-bias leakage, etc.), you could get more voltage than you expect across the LED.

Even though the LEDs in my signal heads have a Reverse Voltage of 5v, as far as I recall, I have never damaged one after accidentally installing it reversed and applying 12v (with current-limiting resistor) ... "your mileage may vary" :-)

Chuck
Regards,

Chuck Hackett, UP Northern 844, Mich-Cal Shay #2
"By the work, One knows the workman"

Atkinson_Railroad
Posts: 137
Joined: Mon Jun 08, 2015 6:27 pm
Location: Michigan
Contact:

Re: Switch point position detection

Post by Atkinson_Railroad » Fri Oct 20, 2017 12:57 pm

The discussion of Miniature Railroad Signaling deserves its own channel within the entire realm of the “love”,
regardless of scale or motive power.

Again, a place to discuss and politely exchange ideas related to outdoor Miniature Railway Signaling obstacles is needed on this forum.
The subject matter spans the entire history of railroading and obviously outdoor railroading.

Ya, I know… I may sound like Lew… pressing for standards.

Just making a re-attempt to nudge the powers to be… at considering the subject matter of signaling as an authentic endeavor
for the outdoor railroading hobbyist... here... on the forum.

John D. Atkinson

Harold_V
Posts: 16432
Joined: Fri Dec 20, 2002 11:02 pm
Location: Onalaska, WA USA

Re: Switch point position detection

Post by Harold_V » Sat Oct 21, 2017 1:53 am

Atkinson_Railroad wrote:Just making a re-attempt to nudge the powers to be… at considering the subject matter of signaling as an authentic endeavor
for the outdoor railroading hobbyist... here... on the forum.

John D. Atkinson
I'm not convinced there's need for an additional forum. The subject can readily be discussed in any of the venues that are currently available in the railroad related fora.

That said, if there's enough traffic to justify a sub-forum, it can be created, with related threads moved accordingly. I suggest the conversation continue, to determine need. If it's justified, it will come about.

H
Wise people talk because they have something to say. Fools talk because they have to say something.

Atkinson_Railroad
Posts: 137
Joined: Mon Jun 08, 2015 6:27 pm
Location: Michigan
Contact:

Re: Switch point position detection

Post by Atkinson_Railroad » Sat Oct 21, 2017 6:58 am

Acknowledging your “reality check” Harold.

It is indeed a [fringe] within a fringe activity.

(An analogy of using snooker here in the United States comes to mind. Unless there’s a table to play the game on,
there’s no way to learn and practice it.)

As time marches on, and research into the past takes place, there is at least a record of me putting the idea forth.

Thank you.

John

kcameron
Posts: 105
Joined: Wed Aug 17, 2011 4:23 pm
Location: Syracuse New York

Re: Switch point position detection

Post by kcameron » Sat Oct 21, 2017 8:12 am

On how to mount or protect the sensors:

My thought is mounting the parts below the rails and between the ties. Any derailment will fall onto the rail heads and ties and not put much force further downward. I'm thinking that a form of whisker or finger is the better way to feel the movement of the point. Depending on the scale of the track, we are looking at travel of 1/4" to 1/2" I think (I'm in 1" scale) yet the critical part, point contacting stock rail, has to be about zero regardless of scale. A gap of 1/64" or more is likely enough to drop the signal.

One thought I'm not sure about is how close would the sensor need to be to the point?? I'm leaning to back a bit from the point. Mainly to stay away from the throw bar and the point closure, just to make sure we don't interfere with either. For mounting I think something that would clip to the outside of the stock rails and span the width of the track under it. Along that line would be the two sensors feeling for when the point rail has moved far enough, one for each point. The two sensors would attach to that bar, likely setscrews or some threaded movement.

Now I had said a whisker to feel the movement of the point, another idea would be a plunger that gets pushed down (or released) by the point rail. I've seen enough of these types of sensors in industrial applications.

So that's some more ideas for consideration. Let's hear some more concepts.
-ken cameron
Syracuse Model Railroad Club http://www.SyracuseModelRr.org/
CNY Modelers http://www.cnymod.com/
Finger Lakes Live Steamers http://www.fingerlakeslivesteamers.org/
Member JMRI Developer Team http://www.jmri.org/
mailto: kcameron@twcny.rr.com
In the Upstate New York US area of the world

User avatar
ChuckHackett-844
Posts: 60
Joined: Wed May 03, 2017 3:54 pm
Location: Tampa, Florida

Re: Switch point position detection

Post by ChuckHackett-844 » Sat Oct 21, 2017 9:24 am

kcameron:

Totally agree that any device needs to be installed below the ties. 90% of the derailments I see are at switches.

Mounting issues:

To be most effective, I see it requiring one device per point - attached directly to the point. If one attaches to any kind of tie rod I fear that any connection slack/hysteresis/loose connection would diminish the ability to detect precise point position. Yes, in theory this would work with careful construction but, lets face it, when using volunteers construction is not always precise and things are not always well maintained.

A "clip" would seem to have to rely on a rail size/shape and clearances since it's most effective closer to the tip of the point where clearances are minimal. At the moment I'm thinking drilled and tapped holes in the web of the rail. Not easy to do in the field but can be done from the outside faces and does not require a nut on the inside faces. Maybe some kind of adhesive? But that would make maint/replacement difficult.

Detection:

I don't think you will find a whisker/lever/micro switch that is rugged, weatherproof (below the ties it will be submerged in heavy rain/flood), inexpensive and have a hysteresis (on-to-off, off-to-on movement) required to reliably detect a point gap below 1/32".

I had considered optical (photo-sensor slit) but problematic in dust, dirt and sand.

An inexpensive hall-effect detector (kind of like a variable reed-switch) is one approach that would work well and could be easily adjusted in the field but it requires solid-state logic which brings up the issue of protecting it from lightning, etc. It could be handled by my signal controllers (which have analog inputs) but this device can't rely on that for general application.

Thinking about it more, I can imagine a device that incorporates an encapsulated hall-effect detector and single chip micro-controller ($1 - $2 in reasonable quantity) that would be "teachable" in the field (i.e.: trained to the actual closed/open positions of a given turnout after install) and provide "contact-closure" output to any indication device. ... but, I fear the hobby/demand would be too small to drive the cost down enough to get it into the "commodity" price range. What would someone be willing to pay for a) an encapsulated device as described to incorporate into one's own mounting design or b) ready to install device?

I think the alarm system buttons I described earlier can detect down to at least 1/16" and hopefully down to 1/32" but I have not had time yet to actually test them. 1/32" is higher than desirable but might be the sweet spot in the cost/benefit curve.

On the subject of reed-switches (in fact there is one sitting on my keyboard): (warning: getting technical here) They can be made more precise (lower hysteresis) if one were to be able to create a steeper magnetic field gradient than that generated by an open magnet alone. Here I'm thinking of some kind of shaped poll pieces attached to the magnet that concentrate the field. My engineering background is lacking in this area (or I've been out of college too long :-) ) to figure that one out.

Chuck
Regards,

Chuck Hackett, UP Northern 844, Mich-Cal Shay #2
"By the work, One knows the workman"

Post Reply