Skip to main content

I have a super streets taxi scene I've been thinking of automating.  It's a single lane dog bone using two super street "Y" for the reversing loops.  The roadbed is in place, so I'm not looking to modify the roadbed configuration.  One reversing loop is around Grand Central Terminal  (GCT) and the other around Columbus Circle.  The lane runs through the city.  I'd like to cue taxi's around GCT.  I'm thinking a taxi stops in the rear of GCT, a taxi in front of GCT then moves out, and the taxi in the rear advances to the front after which the scene repeats.

Mechanism would be roadbed embedded optical sensors to detect position, roadbed blocking, relay control logic and a solenoid to reset the "Y".

Any suggestions to improve the automated scene or mechanization would be appreciated.  Thanks in advance.

Original Post

Replies sorted oldest to newest

Just 2 taxis, right?  I'd think trying to do 3 or more can get tricky.

I'm not familiar with the super street "Y".  Does it have anti-derailing or must you pre-set (pulse a solenoid?) the branch direction when entering or leaving each loop?

What exactly are these optical sensors?  Can you mount them to detect direction of travel?  For example, if you are shooting a beam upward to reflect off the bottom of a taxi let's say you put a reflective bottom on just the right side of the taxi.  Then by placing the optics on the right side of the roadbed (relative to direction of travel) you can detect which direction the taxi is traveling on the straight.  This might simplify the relay logic.

Is soldering and/or working with small electronic components (resistors, capacitors, etc.) "allowed" or to be avoided at all cost?!

In any case, I'd think you can cobble together a handful of low-cost eBay relay modules (about $1 per relay) and get something running for, say, $10-20.

Sounds like a fun project! 

 

stan2004 posted:

Just 2 taxis, right?  I'd think trying to do 3 or more can get tricky.

I'm not familiar with the super street "Y".  Does it have anti-derailing or must you pre-set (pulse a solenoid?) the branch direction when entering or leaving each loop?

What exactly are these optical sensors?  Can you mount them to detect direction of travel?  For example, if you are shooting a beam upward to reflect off the bottom of a taxi let's say you put a reflective bottom on just the right side of the taxi.  Then by placing the optics on the right side of the roadbed (relative to direction of travel) you can detect which direction the taxi is traveling on the straight.  This might simplify the relay logic.

Is soldering and/or working with small electronic components (resistors, capacitors, etc.) "allowed" or to be avoided at all cost?!

In any case, I'd think you can cobble together a handful of low-cost eBay relay modules (about $1 per relay) and get something running for, say, $10-20.

Sounds like a fun project! 

 

Hi Stan,

I just want to say, you have ALWAYS been VERY helpful in the past.  Thanks again.  You are definitely an idea man.  To answer your questions:

I'm planning just 2 taxi's.  I'd like to have more, but given my constrains I don't see how I could accommodate more.

The super street "Y' is sort of like a manual switch.  There is no switch machine.  It works like this, given the top parts of the "Y" form the reversing loop.  The taxi enters the "Y" and travels in the direction depending on the "Y" position  (let's call it position A) and travels around the reversing loop.    As the taxi exits the "Y", the wheels of the taxi move the "Y" rails to position B.  When the taxi returns to the "Y", the taxi enters position B and exits position A.  Thus the "Y" not only returns the taxi in reverse direction but also reverses the taxi direction in the reversing loop.  The movable 'Y" rails are attached to a sliding bar which can be attached to an actuator type mechanism.  Since, in the subject taxi scene, the taxi's needs to move only in one direction in the reversing loop, the movable rails need to be reset after a taxi exits the loop, hence the idea of using a solenoid.

Think Azatrak detectors in the reflective mode.  I've mounted these in my complex trolley scene and they work very well.

Yes, I know how to solder but prefer to stay way from using discrete components and bed boards.  However, I am asking for new ideas.

My approach to the world sometimes is a little different than others.  I first consider functionality and cost second.  From my point of view, it makes no sense to spend money on something that doesn't do the job.  Of course, compromise is always part of the design process.

Here's is some more perspective.  My city scene is elevated and divided into two parts.  One part is done and installed on the layout. This installed part includes the trolley scene we chatted about a while back and more than half of the subject  taxi scene is located on this installed part.  I want to finish and install the top side of the second part.  In order to accomplish this  I must configure the blocks, install sensors and provide roadbed power to the blocks.   All wiring gets routed under the table where the control electronics will be located.  This allows me lots of flexibility to configure/modify control logic after the fact.  So my interests now are more like "preliminary design" options.  I can do the "detail design" later if my "preliminary design" is well thought out.

Thanks again

Steve

 

 

 

 

 

So if this is a case of "I'll show you mine if you show me yours" , here's a starting point.

shorling taxi idea 1

1. Two isolated blocks.  Block A and Block B.  A taxi will enter the main loop from Block A.  The other taxi will wait in Block B.  ONLY one of the two will be powered at a time - so either Block A is powered or Block B is powered.  The rest of the layout is always powered including the half circle on the far right.

2. Two sensors.  Sensor1 is on the straight.  It will generate a trigger when the taxi goes left OR right.  As it turns out, it doesn't matter that 2 separate triggers are generated.   Sensor2 detects a taxi has entered the waiting block B.

3. One solenoid.  The sole function is to set the "Y" to position B.  It cannot set the "Y" to position A.  The car entering the straight will set the "Y" to position A (if I understand your explanation).  Hence the solenoid only operates in one direction which saves a bit of electronics.

So here's the sequence.  Let's say the game starts with a taxi in Block A and a taxi in Block B with power applied to Block A. 

Step 1. Taxi in Block A starts moving into the straight.  There is no power in Block B so that taxi is stopped and waiting.

Step 2. Moving taxi hits Sensor1.  This fires the solenoid to switch "Y" to position B in anticipation of taxi returning after looping from left circle.  This also flips power from Block A to Block B.  So waiting taxi in Block B is powered and moves to Block A which is now unpowered.  Hence taxi stops in Block A and waits for next cycle.

Step 3. Moving taxi reaches left circle, makes loop, comes back to sensor1.  It passes sensor1 which again (redundantly) fires the solenoid to switch the "Y" to position B.  No problem.  Block A is still unpowered (taxi is waiting).  Block B is still powered but there's no taxi there so no problem.

Step 4. Moving taxi reaches Block B and hits Sensor2.  This toggles block power from Block B back to Block A.  So moving taxi loses power and stops in Block B.  The waiting taxi in Block A is now powered and starts moving.  We're back to Step 1.  Lather, rinse, repeat.

So components are:

1. A latching-relay which powers Block A or Block B - one at a time or the so-called SPDT (Single-Pole-Double-Throw) function.  Relay is "set" by a trigger from Sensor1 and "reset" by a trigger from or Sensor2.   You can buy a latching relay for a couple bucks but I'm not sure I've seen latching relays with convenient screw terminals (no soldering) for what I'd pay (we might have different philosophies on cost ).  I've shown in earlier OGR threads how to wire a few $1 eBay non-latching relays (no soldering required) to become a latching relay.

2. 2 sensors.  These can be your Azatrax pairs.  I've never used those but it appears they require 5V DC and are best used with an Azatrax detector board which runs $5-10 per sensor pair.  I may have that wrong.  You could also use a 10 cent Nd magnet disc glued to the bottom of each taxi to trips a 25 cent reed switch in the street roadbed.  It doesn't really matter what you use; it's the concept of a trigger signal when a taxi passes the sensor point.

3. Solenoid.  Again, I've never seen or used a Super Street "Y".  But the key is you only need to move the "Y" from A to B (as opposed to both A-to-B and B-to-A).  This means the solenoid only has to work in one direction which makes the electrics easier.  I figure a $1 solenoid should do the trick.  Pulse the solenoid for a fraction of a second (e.g. when Sensor1 is triggered).  For example something like one of these 12V solenoids I pulled from my parts stash would do the trick:

solenoids

A key is these solenoids only need a fraction of a second to do their thing.  A moving taxi over a sensor might be active for only a fraction of a second.  It appears that if you use an Azatrax detector board in conjunction with their sensors, you can dial in a delay so this would presumably activate the solenoid for a longer period.  This could allow use of another switch-machine mechanism like a geared-motor if the snap-bang solenoid plunging is too "violent"!

Attachments

Images (2)
  • solenoids
  • shorling taxi idea 1
stan2004 posted:

So if this is a case of "I'll show you mine if you show me yours" , here's a starting point.

shorling taxi idea 1

1. Two isolated blocks.  Block A and Block B.  A taxi will enter the main loop from Block A.  The other taxi will wait in Block B.  ONLY one of the two will be powered at a time - so either Block A is powered or Block B is powered.  The rest of the layout is always powered including the half circle on the far right.

2. Two sensors.  Sensor1 is on the straight.  It will generate a trigger when the taxi goes left OR right.  As it turns out, it doesn't matter that 2 separate triggers are generated.   Sensor2 detects a taxi has entered the waiting block B.

3. One solenoid.  The sole function is to set the "Y" to position B.  It cannot set the "Y" to position A.  The car entering the straight will set the "Y" to position A (if I understand your explanation).  Hence the solenoid only operates in one direction which saves a bit of electronics.

So here's the sequence.  Let's say the game starts with a taxi in Block A and a taxi in Block B with power applied to Block A. 

Step 1. Taxi in Block A starts moving into the straight.  There is no power in Block B so that taxi is stopped and waiting.

Step 2. Moving taxi hits Sensor1.  This fires the solenoid to switch "Y" to position B in anticipation of taxi returning after looping from left circle.  This also flips power from Block A to Block B.  So waiting taxi in Block B is powered and moves to Block A which is now unpowered.  Hence taxi stops in Block A and waits for next cycle.

Step 3. Moving taxi reaches left circle, makes loop, comes back to sensor1.  It passes sensor1 which again (redundantly) fires the solenoid to switch the "Y" to position B.  No problem.  Block A is still unpowered (taxi is waiting).  Block B is still powered but there's no taxi there so no problem.

Step 4. Moving taxi reaches Block B and hits Sensor2.  This toggles block power from Block B back to Block A.  So moving taxi loses power and stops in Block B.  The waiting taxi in Block A is now powered and starts moving.  We're back to Step 1.  Lather, rinse, repeat.

So components are:

1. A latching-relay which powers Block A or Block B - one at a time or the so-called SPDT (Single-Pole-Double-Throw) function.  Relay is "set" by a trigger from Sensor1 and "reset" by a trigger from or Sensor2.   You can buy a latching relay for a couple bucks but I'm not sure I've seen latching relays with convenient screw terminals (no soldering) for what I'd pay (we might have different philosophies on cost ).  I've shown in earlier OGR threads how to wire a few $1 eBay non-latching relays (no soldering required) to become a latching relay.

2. 2 sensors.  These can be your Azatrax pairs.  I've never used those but it appears they require 5V DC and are best used with an Azatrax detector board which runs $5-10 per sensor pair.  I may have that wrong.  You could also use a 10 cent Nd magnet disc glued to the bottom of each taxi to trips a 25 cent reed switch in the street roadbed.  It doesn't really matter what you use; it's the concept of a trigger signal when a taxi passes the sensor point.

3. Solenoid.  Again, I've never seen or used a Super Street "Y".  But the key is you only need to move the "Y" from A to B (as opposed to both A-to-B and B-to-A).  This means the solenoid only has to work in one direction which makes the electrics easier.  I figure a $1 solenoid should do the trick.  Pulse the solenoid for a fraction of a second (e.g. when Sensor1 is triggered).  For example something like one of these 12V solenoids I pulled from my parts stash would do the trick:

solenoids

A key is these solenoids only need a fraction of a second to do their thing.  A moving taxi over a sensor might be active for only a fraction of a second.  It appears that if you use an Azatrax detector board in conjunction with their sensors, you can dial in a delay so this would presumably activate the solenoid for a longer period.  This could allow use of another switch-machine mechanism like a geared-motor if the snap-bang solenoid plunging is too "violent"!

Great reply.  Thanks for writing back.  Exactly what I'm looking for.  You implementation is slightly different than mine, actually better.

I was thinking 3 sensor sets each embedded in the road bed.  One set in each block and a third set before the "Y".   Each taxi would stop in blocks A and B at the sensor location rather than at the block entry point.  Provides more custom stopping point in each block.

Second, I was thinking of a delay between when the cue taxi stops and the station taxi starts.  So, at some both taxi's would be stopped at the station for the delay period.

Blocks A and B are controlled by latching relays.  Detection in a block stops the taxi in the block.  Solenoid is spring loaded type.

Here's my idea of operation.  Cue taxi enters station loop and triggers detector which sets latching relay stopping the cue taxi.  After a delay, the sensor drops out and resets another latching relay which powers the taxi at the station.  The station taxi moves out and exits the "Y' passing over the "Y" sensor.  The "Y" sensor triggers (momentary) the solenoid and resets the cue latch.  The cue taxi advances to the station position and stops, setting the station latch.  The taxi returns and passes over the "Y' sensor again (does no harm) but an  interlock could prevent this unnecessary action.  The taxi enters the cue position and stops.  Cycle repeats.

 

Last edited by shorling

Yep.  I can see your point about sensors in each controlled block for more precise stopping.  I showed sensor2 in Block B which would do that...but the taxi would simply stop as soon as it entered unpowered Block A.  So while I believe the stopping point would be somewhat repeatable, if you've already "cut" your blocks then you'd have to re-layout track sections to change the stopping point.  Clearly just moving the stopping sensor would be preferable in such a case!

I like the time-delay idea so that both taxis are stopped for a while.  Adds suspense and drama!  To that end, if using low-cost relays, I'd suggest using a so-called Delay-ON relay module ($1-2 on eBay).  So in this application a Delay-ON relay would delay application of power to Block A.   That is, when taxi enters Block B and hits sensor2, a relay swaps power from Block B to Block A.  But by inserting a Delay-ON relay, the power to Block A will be delayed by, say, 10 seconds...so both taxis will be stopped for an interval.

And that brings up another idea.  If you use a Delay-OFF relay module (also $1-2 on eBay), you can protect the solenoid in the unusual case where the taxi stalls/derails/whatever right on top of sensor1.  In normal operation, I figure sensor1 will trigger for a second or so as the taxi passes over it.  If this controls the power to the "Y" solenoid, then the solenoid gets power for 1 second which should be plenty of time to flip the "Y".  But if it stalls, the solenoid would get continuous power which is generally a bad thing.  A Delay-OFF relay could be set to turn off power to the solenoid after 1 second (or whatever).

Addendum:

I dug up this thread where another member had an application somewhat similar to yours.  I bring it up in case you're not familiar with these attractively-priced eBay relay modules:

alternating%20out-back%20trolley%20Rev%202

In that thread, I showed how to hook up 2 regular (non-latching) relays on the right module to "make" a latching relay.  Last time I looked, a latching relay module with "set" input and "reset" input and screw-terminal connections (so no soldering) is about $10 each on eBay.

latching relay module 10 bucks

 

Attachments

Images (2)
  • alternating%20out-back%20trolley%20Rev%202
  • latching relay module 10 bucks
Last edited by stan2004
stan2004 posted:

Yep.  I can see your point about sensors in each controlled block for more precise stopping.  I showed sensor2 in Block B which would do that...but the taxi would simply stop as soon as it entered unpowered Block A.  So while I believe the stopping point would be somewhat repeatable, if you've already "cut" your blocks then you'd have to re-layout track sections to change the stopping point.  Clearly just moving the stopping sensor would be preferable in such a case!

I like the time-delay idea so that both taxis are stopped for a while.  Adds suspense and drama!  To that end, if using low-cost relays, I'd suggest using a so-called Delay-ON relay module ($1-2 on eBay).  So in this application a Delay-ON relay would delay application of power to Block A.   That is, when taxi enters Block B and hits sensor2, a relay swaps power from Block B to Block A.  But by inserting a Delay-ON relay, the power to Block A will be delayed by, say, 10 seconds...so both taxis will be stopped for an interval.

And that brings up another idea.  If you use a Delay-OFF relay module (also $1-2 on eBay), you can protect the solenoid in the unusual case where the taxi stalls/derails/whatever right on top of sensor1.  In normal operation, I figure sensor1 will trigger for a second or so as the taxi passes over it.  If this controls the power to the "Y" solenoid, then the solenoid gets power for 1 second which should be plenty of time to flip the "Y".  But if it stalls, the solenoid would get continuous power which is generally a bad thing.  A Delay-OFF relay could be set to turn off power to the solenoid after 1 second (or whatever).

Addendum:

I dug up this thread where another member had an application somewhat similar to yours.  I bring it up in case you're not familiar with these attractively-priced eBay relay modules:

alternating%20out-back%20trolley%20Rev%202

In that thread, I showed how to hook up 2 regular (non-latching) relays on the right module to "make" a latching relay.  Last time I looked, a latching relay module with "set" input and "reset" input and screw-terminal connections (so no soldering) is about $10 each on eBay.

latching relay module 10 bucks

 

Stan, Lots of good ideas as usual.  Thanks !!  I'll have to look into the delay on and delay off relays on ebay.  I use the ebay relay modules already that do not have a delay such as the one shown in your schematic.  I'm planing on adding an on/off toggle switch to power all blocks.  Then I can run just run one taxi through the reversing loops or have the scene control.

I still like the idea of having several, maybe 3, taxi's cued up at the station rear in addition to the taxi in the front.  I could envision how that would work: When the front taxi moves out, the first cue taxi advances to the front.  The remaining two cue taxi's advance in position, one taxi at a time.  When  the taxi returns from the loop, it stops at the back of the rear cue.  Delays would still be used, however all the taxi moves must be completed before the loop taxi returns.  Also, this would only work if the cue taxi's were manually placed in their starting position.  Seems like this scene should work if cue size is varied to 1, 2, or 3 cue taxi's on the layout.

shorling posted:
...I still like the idea of having several, maybe 3, taxi's cued up at the station rear in addition to the taxi in the front.  I could envision how that would work: When the front taxi moves out, the first cue taxi advances to the front.  The remaining two cue taxi's advance in position, one taxi at a time.  When  the taxi returns from the loop, it stops at the back of the rear cue.  Delays would still be used, however all the taxi moves must be completed before the loop taxi returns.  Also, this would only work if the cue taxi's were manually placed in their starting position.  Seems like this scene should work if cue size is varied to 1, 2, or 3 cue taxi's on the layout.

Right.  It seems you could just use traditional "red-green" block signaling with any occupied block disabling power to the next block.  The key would be to have the taxi stop ON the sensor so that you don't need to mess with latches - whether this latching is done with relays or digital logic chips.

I'm not familiar with Super Street taxis but if going slow enough, I'd think they might be able to stop ON a sensor even with coasting momentum.  To insure this, perhaps run the voltage to the queue blocks a couple of Volts lower using 50 cents of 3 Amp diodes (using diode dropping).  If starting Voltage is an issue, simply use a $1-2 delay OFF relay module to temporarily boost the voltage to a block (for 1 second or whatever) to get the taxi started.

More great ideas !!!  Coasting is not an issue.  For all practice purposes, when the power is off, the taxi stops.  The issue is detector latency.  This was a major problem with my trolley scene since I needed the trolley to stop instantaneously.  This is not the case here.  Super Street straight roadbed sections are 10 inches in length.  If each section is a block, plenty of room to stop.

Thanks for all your help.   I'm having spine surgery next Thursday, so with my back issues I've been relegated to planning rather than doing for the last 6 months.  Probably after the first of the year when I have recovered, I can begin the "doing" part.

You can have as many as you like, ive done up to 4, 3 sitting, each advances and takes a loop on his own,  with 4 activation sections, and with the throw of one switch you have a nonstop loop, now 4 wouldnt work with the reversing Y, but 2, 1 traveling, 1 stopped, alternating trips, wouldnt be hard at all, and No cost, no relays, no sensors, just pull out the street pins and add some wire...

the Y would be lightly sprung, allowing taxi to leave then points close back....just a thought...

Last edited by Squirrelstrains
Squirrelstrains posted:

You can have as many as you like, ive done up to 4, 3 sitting, each advances and takes a loop on his own,  with 4 activation sections, and with the throw of one switch you have a nonstop loop, now 4 wouldnt work with the reversing Y, but 2, 1 traveling, 1 stopped, alternating trips, wouldnt be hard at all, and No cost, no relays, no sensors, just pull out the street pins and add some wire...

the Y would be lightly sprung, allowing taxi to leave then points close back....just a thought...

Hi Squirrelstrains:  Thanks for the ideas !!  I'm trying to automate the scene since the workload of running the trains themselves is pretty high.  I do like the idea of a lightly spring loaded  "Y" points.  Have you tried this ?  The taxi's are so light in weight but maybe with the right spring rate it would work.

My goal was to have different cars travel main street one at a time, but i had to settle  for 4,

Using the isolated sections it runs by itself, stops starts advances all hands off.

You can also return to a normal  loop simply by two DPDT switches  grounding the isolated  sections

I can email you a wiring diagram if you like i have a hard time posting it on here

Good luck its a fun project, i enjoyed mine and get lots of comments.

Last edited by Squirrelstrains
Squirrelstrains posted:

My goal was to have different cars travel main street one at a time, but i had to settle  for 4,

Using the isolated sections it runs by itself, stops starts advances all hands off.

You can also return to a normal  loop simply by two DPDT switches  grounding the isolated  sections

I can email you a wiring diagram if you like i have a hard time posting it on here

Good luck its a fun project, i enjoyed mine and get lots of comments.

Thanks, my email is in my profile.

I prefer tripping a solinoid via #1 sensor, #2 switch or #3 short outer rail isolation rail vs longer sectioned isolated  especially with low wheel count, traction tires, and/or light weight vehicles just because their track contact is already weaker comparied to a normal loco. I.e. the long single rail power feed tends to the most unreliable of those options for power loss & stall reasons imo. I keep to short isolating sections and relays & maybe capping for a linear coil .

I like reed switches as well. If for nothing else the ease of moving magnets or reeds appeals to me. Opto sets can be a pita at times too, morning sunshine to camera flash to I R remotes to overhead lightbulb conditions can drive you nuts ....or they may work reliably for years on end. Thats a players call on prefered use  

You will likely not miss continuous looping unless you run your SS alone often.  A bypass for the stand's stop but leaving a pickup stop in play for a single (private sector vehicle? ) would fit MY sensibilities at least. A toggle and/or a relay are all it takes for looping as you noted.

Adding another stand/or stop at the other end just requires stacking the circuits, though each added length you can manage to fit seperating the stops makes timing/latching easier.

A company is mentioned in a currently running thread on anti collision without relays that produces a variabley timed "cascade" board system aimed at ladder siding control. They are tandemly stackable modules. I dont recall the co. name, but it may be a plug and play option or closer to it. Last I looked there  were some additional automation modules in the works but ive no idea how that has progressed since I haven't  been to the site in near a decade. (might be DC & G scale based, but no taxi/mow set up is likely to be above the amp range offered)  Its simple stuff  to repoduce sure, but premade and stackable is refreshingly simple too,

I think you could get an easier, better adjusting, lighter give rate from a stretched type coil spring vs finding balance on a torsional  one (that looks offhand made to just maybe snap lock points left or right too).  Im thinking a lawnmower's butterfly valve/governor spring would be about right. You only need a slight deflection of points and a return; these are delicate springs on most càrbs anymore and long or short hooks both so hiding  the coil could be as versitile as a "hollow log" , stray pipe , tool/electric boxbox  up to about 3" away . The hook lead/ rod hid in brush etc. 

 

Add Reply

Post
×
×
×
×
Link copied to your clipboard.
×
×