Skip to main content

I have several Lionel Bump and Go Trolleys which I have sync'd to traffic lights at a 4 way intersection using Superstreet point to point runs.  I've built the circuits and installed on the layout.  I'm now testing.  It all works fine so far except I need the trolley to stop immediately when it crosses the IR detectors embedded in Superstreets just before the bumper.  The detector is designed to cut power which stops the trolley however there is a electronic delay.  The trolley crosses the detector, bumps, reverses, power cuts, coasts which all result in the trolley stopping about an inch just off the bumper/sensors.  The trolley restarts after 3 seconds when the detector times out.  I need the trolley to stop closer to the bumper and still on the sensor to avoid collision since the Superstreet run is just an inch or so longer than the trolley.  There are solutions, like relocating the reflective sensor which would be very very difficult at this point.  No space to make the Superstreet runs longer.

I'm wondering if I can delay/slow the trolley after it bumps.  I tried reducing trolley speed which helped but didn't solve the issue.   I think the trolley motor is DC rectified on-board.  The trolley bumps, reversing polarity which changes direction.  Any ideas to add an RC circuit across the motor to slow the trolley restart after the bump ?  Another idea is to add weight to the trolley.  Thanks in advance.

 

Original Post

Replies sorted oldest to newest

shorling posted:

... The detector is designed to cut power which stops the trolley however there is a electronic delay.  The trolley crosses the detector, bumps, reverses, power cuts, coasts which all result in the trolley stopping about an inch just off the bumper/sensors.  The trolley restarts after 3 seconds when the detector times out.

Untitled

In their description, it seems the MRD1 relay trips immediately when the detector is triggered.  The relay then releases after a delay (1/3 to 14 sec with Variable version, or 1 sec with Fixed version).  I don't understand how the trolley has time to reach the bumper and reverse if power is cut immediately upon crossing the detector (as it heads toward the bumper).  It appears  you have the MRD1-V Variable version if you are able to set the restart time to 3 seconds. 

Clearly I'm missing something. 

Attachments

Images (1)
  • Untitled
Last edited by stan2004

Hi Stan,

I'm using the MRD1 with "normally closed" contacts.  Contacts "open" when activated.  They are specially modified to "normally closed" MRD1 by azatrak.  The MRD1 does not open immediately on detection.  There is a 1/8 second delay.  As long as the trains remains over the sensors the MRD1 activates and opens the relay.  When the train moves off the sensors there is a fixed 3 second delay till the relay closes.  The trolley travels about 1.5 inches in the 1/8 second including the bump to reverse.  Unfortunately 1.5 inches is past the sensors on the reverse trip, so the trolley restarts in 3 seconds.

Obviously the main problem is the placement of the detector so close to the bumper. Relocating it or a new one, regardless of the additional work, would seem to be the best solution.

Given that, a non-electrical method - friction - may work. One way would be to pinch the wheels against the track with a small bar, adjustable for pressure. Just enough friction to slow the car, but allow it to resume travel when the power is turned back on. If the section of track where the trolley stops is not visible, one side of a piece of Velcro attached to the track, high enough to rub on its belly, may work. If located in a visible area, attach it to the underside of the trolley instead and place some smooth, prototypical track debris for it to rub on.

Dave

 

 

 

Dtrainmaster posted:

Obviously the main problem is the placement of the detector so close to the bumper. Relocating it or a new one, regardless of the additional work, would seem to be the best solution.

Given that, a non-electrical method - friction - may work. One way would be to pinch the wheels against the track with a small bar, adjustable for pressure. Just enough friction to slow the car, but allow it to resume travel when the power is turned back on. If the section of track where the trolley stops is not visible, one side of a piece of Velcro attached to the track, high enough to rub on its belly, may work. If located in a visible area, attach it to the underside of the trolley instead and place some smooth, prototypical track debris for it to rub on.

Dave

 

 

 

Thanks for the input Dave.  You are correct regarding sensor placement, however relocation at this point in construction is difficult and I'm searching for easier solutions.  Slowing the trolley down will work with friction being one method.  I was also wondering if added weight would allow the trolley to run a slower speeds ?

Additional weight would increase inertia.

About the IR sensing method used, the application seems intended for a triggered time delay, i.e. signals, switch throws, etc., not for a holding state depending on train position. That is usually achieved by separating the IR pair into a cross-track beam.

You could add another (longer) delay timer triggered by the one you have now. But it sounds like there is another issue with available distance before the adjacent collision zone anyway.

Dave

Dtrainmaster posted:

Additional weight would increase inertia.

About the IR sensing method used, the application seems intended for a triggered time delay, i.e. signals, switch throws, etc., not for a holding state depending on train position. That is usually achieved by separating the IR pair into a cross-track beam.

You could add another (longer) delay timer triggered by the one you have now. But it sounds like there is another issue with available distance before the adjacent collision zone anyway.

Dave

Yes, the increased inertia would slow the trolley as it reverses, not sure how much ?  Also, if I could slow the motor upon the reverse a time constant or two with an RC circuit, maybe that would work.  The unexpected delay of 125 ms allows the trolley to travel an additional 1.5 inches.  If the delay was cut in half, the travel would be reduced to 0.75 inches which would be acceptable.

Let me see if I understand this.

a) Trolley approaches bumper and trips the MRD optical sensor.  MRD wait 1/8 second then removes track power (via relay).

When is track power removed relative to when it hits the bumper (which presumably flips an internal switch that reverse the motor polarity/direction)?

I think it will be frustrating to create a reliable/repeatable system based on timing.  It seems a wildcard is how much momentum or kinetic-energy the bumper absorbs on the reversal.  I'd think this would change from bump to bump.  So depending on the relative timing of when coasting starts vs. the reversing, the stopping point might move around.

I don't know how comfortable or interested you are in fiddling with electronic (vs. mechanical) solutions but ideas which might spawn others are:

1) lower the the speed of the trolley as it approaches the bumper.  In general, it always seems easier to stop something on a dime if it is going slower to begin with.  I assume you need some minimum speed to reliably flip the reversing switch so,

2) change the mechanical reversing switch to, say, a magnetically triggered reed or Hall-type sensor that flips a relay in the trolley.  The idea here is this would allow reversing at an exact point on the track and at an arbitrarily slow speed.

3) add a dynamic braking circuit to the DC motor in the trolley.  If you short the windings of a DC motor, it will stop instantly.  This is, for example, how the smoke units generate distinct puffs of smoke - the electronics shorts the motor windings which instantly brakes the smoke fan motor, airflow stops, and you get a puff.  So something similar could stop the trolley on a dime.

4) If the trolley is indeed "just" a DC motor with a bridge perhaps remove the bridge and drive the trolley motor with DC on the track.  To reverse direction you must now have an external relay (to flip DC polarity) but this means you don't need to "slam" into the bumper to reverse direction.  Of course the bumper itself can still be in place and you'd employ some combination of 1), 2), and 3) above.

Last edited by stan2004
stan2004 posted:

Let me see if I understand this.

a) Trolley approaches bumper and trips the MRD optical sensor.  MRD wait 1/8 second then removes track power (via relay).

When is track power removed relative to when it hits the bumper (which presumably flips an internal switch that reverse the motor polarity/direction)?

I think it will be frustrating to create a reliable/repeatable system based on timing.  It seems a wildcard is how much momentum or kinetic-energy the bumper absorbs on the reversal.  I'd think this would change from bump to bump.  So depending on the relative timing of when coasting starts vs. the reversing, the stopping point might move around.

I don't know how comfortable or interested you are in fiddling with electronic (vs. mechanical) solutions but ideas which might spawn others are:

1) lower the the speed of the trolley as it approaches the bumper.  In general, it always seems easier to stop something on a dime if it is going slower to begin with.  I assume you need some minimum speed to reliably flip the reversing switch so,

2) change the mechanical reversing switch to, say, a magnetically triggered reed or Hall-type sensor that flips a relay in the trolley.  The idea here is this would allow reversing at an exact point on the track and at an arbitrarily slow speed.

3) add a dynamic braking circuit to the DC motor in the trolley.  If you short the windings of a DC motor, it will stop instantly.  This is, for example, how the smoke units generate distinct puffs of smoke - the electronics shorts the motor windings which instantly brakes the smoke fan motor, airflow stops, and you get a puff.  So something similar could stop the trolley on a dime.

4) If the trolley is indeed "just" a DC motor with a bridge perhaps remove the bridge and drive the trolley motor with DC on the track.  To reverse direction you must now have an external relay (to flip DC polarity) but this means you don't need to "slam" into the bumper to reverse direction.  Of course the bumper itself can still be in place and you'd employ some combination of 1), 2), and 3) above.

Hi Stan,

The IR sensors are located about 1/2 inch from the Superstreet bumper.   These are reflective mode sensors embedded in the roadway.  Once detected, the MRD1 takes 1/8 second to turn off power.  During the 1/8 second the trolley travels 1.5 inches and stops.  This includes the 1/2 inch to reach the bumper to reverse plus an additional one inch in the reverse direction.  Since the trolley travels one inch in the reverse direction the trolley uncovers the sensor which un-detects triggering a timeout which returns power after 3 seconds.

The entire trolley  setup is much more complicated than this one aspect and works great except for this glitch.  The system is not timing based.  It's based on trolley location detection.  Unfortunately, there is this unanticipated 1/8 second delay.  If the delay can be reduced, problem solved.  Moving the sensors or extending the roadbed would also solved the problem.

Thanks for your suggestions:

1)  I tried speed reduction.  Lower speed helps but trolley stall occurs before the delay is achieved.

2)  It would probably be easier to relocate the IR sensors than replace the bump switch with other sensor technology.

3)  I like the dynamic braking idea.  Maybe an RC circuit such as a resistor in series with the DC motor and a capacitor across the motor. The transient slows the motor enough in a couple of time constants when the DC source is reversed ?

4)  The trolley has a bridge rectifier and a DC motor.  With DC source still have to sync the trolleys to the traffic lights at the 4 way intersection.  Trolleys behave just like automobiles:  Go on green, stop at yellow light or run yellow depending on position, stop at the light on red. 

Azatrax has indicated they can reduce the delay to 1/16 second which should reduce travel from 1.5 inch to an acceptable 0.75 inch.  Still searching for alternatives as backups.

 

I read the words but still don't get it.  So the trolley travels about 1.5 inches in the 1/8 second.  Does that mean it is going, say, 1" and hitting the bumper, stopping, dynamically braking (since motor windings instantly see reverse polarity), accelerating in reverse, then going another, say, ~1/2" away from the bumper making for a total powered travel of 1.5".  Then power is removed and the trolley coasts a bit further (how much)?

Later he says 1.5" is past the sensors on the reverse trip which suggests (?) 1.5" refers to the distance from the bumper rather than the total distance travel in 1/8 second.  This may seem like minutiae but 1/4" or 1/2" is apparently critical to making this work.   Going 1.5" in 1/8" sec is about 33 scale MPH.  If during that 1/8" second there is a reversal during which the motor decelerates, stops, accelerates, that means the trolley is going much faster than 33 scale MPH which I'd think must be lowered to make stopping on a dime more attainable.

This may be a case of too much detail.  Ideally, the goal would be to have the trolley stop on a dime as Stan says.  The trolley can do that, there really is no coasting.  Cut power and it stops.  The electronic delay between sensor detection and the power cut relay opening is 1/8 second (125 ms) .   In 1/8 second the trolley travels an additional 1.5 inches. The issue here is the trolley moves off the sensors in the 1.5 inches of travel and stops.  With 1.5 inch travel the trolley moves into the collision zone.  If the trolley moved, let say only 0.75 inch and stopped, it would not enter the collision zone.  Like Stan says, moving a small distance is critical.

So, I'm searching for potential fixes.  The requirement is when the power is cut to the trolley, the trolley must be still be in the sensor detection zone.  So, the less movement the better,  If the trolley movement is limited to not more than 0.75 inch or 63 ms of travel time it should work.

I think my last response overlapped with your clarifications.  Anyway, if Azatrax shortens the sensor delay to 1/16th sec it appears this will solve the problem so now we're just having an after-action debriefing.

Some other thoughts.  If your sensors are embedded in the track bed aiming upward, is it practical to add a lip/skirt/fender to the leading edge of the trolley so the sensor "sees it" sooner?  This would only need to be a fraction of an inch long so not too obtrusive.  I suppose there could be mechanical interference issues depending on the location of the sensors and bumper geometry.

As to adding capacitance to aid stopping.  I may be misunderstanding your idea, but note that a typical DC can motor used in O gauge mechanisms will have a DC resistance of, say, 5 ohms.  So in terms of the electrical time-constant (R x C), you'd have to have a ginormous capacitor to have a meaningful impact.  For example, a 0.1 sec time constant when R is 5 ohms means a C of a whopping 20,000 uF.  Yes, you can find caps that big of course but if placed across a DC motor that reverses, it must also be of a non-polarized style which complicates matters. 

Not quite the same issue but I recall a recent thread where it was proposed to use a capacitor to provide DC motor backup energy to power through fraction-of-a-second losses of track power.  When the math was done, it turned out the capacitor was the size of a can of tennis balls... and it cost something like $100!   The take-away though was the flywheel (if it fits) is a much more efficient and inexpensive way to store energy to assist with coasting.  I bring this up because it stands to reason that the converse is true.  So in your case you want to stop instantly when power is lost.  I don't know what your motor is but if it has a flywheel on it perhaps you can remove it!

Previously I suggested replacing the optical sensor with some kind of magnet in the track bed and then a sensor in the trolley with trolley electronics.  I realize this would not be a simple solution but I was pre-disposed to this thinking based from a bump-and-go modification to a handcar:

But in your case, perhaps you could simply put the 10 cent magnet on the trolley and the sensor in the roadbed.  This would mean negligible modification to the trolley.  The magnetic sensor (whether a reed switch or Hall sensor) would have negligible sensing delay and could trip one of those $1 (free shipping) eBay relay modules.

Last edited by stan2004

From my perspective, this problem is not inertia or rebound driven, but is due to controller lag within the tight parameters set by the owner of the trolley line. And said manager's refusal to alter the right-of-way.

Trolley Stop Delayed -RR

This could easily be fixed, assuming that the system has, and you have checked it for, the ability to override the sensor input when it is time to re-power the track. In other words, in its present configuration, will it restart when the trolley is stationary over the sensor?

My suggestion, change the IR detector to a cross-track type. This would allow earlier detection, and more importantly, state holding along more of the track section, regardless of where the car stops. Either split the current sensor's IR emitter/receiver pair, or obtain separate ones. This will also give you flexibility in placement, along with the ability to experiment, without having to alter the track.

Trolley Stop Delayed -E-R

Dave

Attachments

Images (2)
  • Trolley Stop Delayed -RR
  • Trolley Stop Delayed -E-R
stan2004 posted:

I think my last response overlapped with your clarifications.  Anyway, if Azatrax shortens the sensor delay to 1/16th sec it appears this will solve the problem so now we're just having an after-action debriefing.

Some other thoughts.  If your sensors are embedded in the track bed aiming upward, is it practical to add a lip/skirt/fender to the leading edge of the trolley so the sensor "sees it" sooner?  This would only need to be a fraction of an inch long so not too obtrusive.  I suppose there could be mechanical interference issues depending on the location of the sensors and bumper geometry.

As to adding capacitance to aid stopping.  I may be misunderstanding your idea, but note that a typical DC can motor used in O gauge mechanisms will have a DC resistance of, say, 5 ohms.  So in terms of the electrical time-constant (R x C), you'd have to have a ginormous capacitor to have a meaningful impact.  For example, a 0.1 sec time constant when R is 5 ohms means a C of a whopping 20,000 uF.  Yes, you can find caps that big of course but if placed across a DC motor that reverses, it must also be of a non-polarized style which complicates matters. 

Not quite the same issue but I recall a recent thread where it was proposed to use a capacitor to provide DC motor backup energy to power through fraction-of-a-second losses of track power.  When the math was done, it turned out the capacitor was the size of a can of tennis balls... and it cost something like $100!   The take-away though was the flywheel (if it fits) is a much more efficient and inexpensive way to store energy to assist with coasting.  I bring this up because it stands to reason that the converse is true.  So in your case you want to stop instantly when power is lost.  I don't know what your motor is but if it has a flywheel on it perhaps you can remove it!

Previously I suggested replacing the optical sensor with some kind of magnet in the track bed and then a sensor in the trolley with trolley electronics.  I realize this would not be a simple solution but I was pre-disposed to this thinking based from a bump-and-go modification to a handcar:

But in your case, perhaps you could simply put the 10 cent magnet on the trolley and the sensor in the roadbed.  This would mean negligible modification to the trolley.  The magnetic sensor (whether a reed switch or Hall sensor) would not have any neglible sensing delay and would trip one of those $1 (free shipping) eBay relay modules.

First Stan, I want to thank you for your thoughtful considerations.  Azatrax solution should work, but I'm a firm believer in Murphy.  The following are comments on your other suggestions:

1)  The skirt idea should work, but would not be attractive.  It would have to be higher than the bumper fence and extend past the trolley bumper.

2)  I still like the RC option.  I did some dabbling on this  a few days ago and came to the same conclusions you did:  large non-polarized capacitor.  I presume there is lots of open space inside the trolley.  I haven't taken the shell of yet.  Do you think this would work ?

3).  I get the flywheel inertia example.  That's an idea too.  I haven't seen the motor to determine if a flywheel can be mounted.

4)  The magnet idea look likes a keeper.  Easy and fast acting.   You can tell when I've been at this too long: how about a BIG electromagnet to clamp the trolley in position 

 

Dtrainmaster posted:

From my perspective, this problem is not inertia or rebound driven, but is due to controller lag within the tight parameters set by the owner of the trolley line. And said manager's refusal to alter the right-of-way.

Trolley Stop Delayed -RR

This could easily be fixed, assuming that the system has, and you have checked it for, the ability to override the sensor input when it is time to re-power the track. In other words, in its present configuration, will it restart when the trolley is stationary over the sensor?

My suggestion, change the IR detector to a cross-track type. This would allow earlier detection, and more importantly, state holding along more of the track section, regardless of where the car stops. Either split the current sensor's IR emitter/receiver pair, or obtain separate ones. This will also give you flexibility in placement, along with the ability to experiment, without having to alter the track.

Trolley Stop Delayed -E-R

Dave

Dave, thanks for all your inputs.  You clearly understand the issue.  Maybe I need to get the said Manager to be more flexible.

Regarding your comments.  Note sure I understand your first comment but I would like too.  I'm attracted to your fixed and easy statement.  Here is more information assuming there is no delay.  The sensors only operate when the traffic light is yellow.  A relay closes enabling the transmitter LED when the traffic light turns yellow.   If the traffic light is yellow and the trolley is over the sensors, power is cut to the roadbed section thereby stopping the trolley in position.   When the light turns red power remains cut but not by the sensors and the trolley remains in position. The trolley will not restart until the traffic light turns green.  When the light is red or green, the sensors are off.  If not, the sensors would prevent the trolley from restarting when the light turns green and the trolley would forever be locked in position.

I can't use across the track sensors.  There are actually 3 lanes in this Superstreet layout: two are parallel and one cross.  The two parallel lanes are butted with no clearance for mounting individual across the track lane sensors.

 

 

"The trolley will not restart until the traffic light turns green."

So the trolley will restart if it is sitting on the sensor. Good

"There are actually 3 lanes in this Superstreet layout: two are parallel and one cross.  The two parallel lanes are butted with no clearance for mounting individual across the track lane sensors."

I thought this problem area was a single stub track, crossing the main route (i.e., the interference point)?

FWIW, this method can also be done vertically, if you consider "E" as a street lamp and "R" being in the track on the previous drawing.

 

What ever you end up with, you owe Stan and I some pictures.

Dave

Dtrainmaster posted:

"The trolley will not restart until the traffic light turns green."

So the trolley will restart if it is sitting on the sensor. Good

Yes, the sensors  are off on red and green.  Power is reapplied when the light turns green and the off sensors all the power to route to the insulated section.

"There are actually 3 lanes in this Superstreet layout: two are parallel and one cross.  The two parallel lanes are butted with no clearance for mounting individual across the track lane sensors."

I thought this problem area was a single stub track, crossing the main route (i.e., the interference point)?

You are correct.  All three lanes are exactly the same configuration.  Each with a short stub.  Solving one solves all three.

FWIW, this method can also be done vertically, if you consider "E" as a street lamp and "R" being in the track on the previous drawing.

Not sure what you mean here: one sensor on the sidewalk and the other under the trolley ?

What ever you end up with, you owe Stan and I some pictures.

I was hoping a video of it actually working 

Dave

 

shorling posted:

2)  I still like the RC option.  I did some dabbling on this  a few days ago and came to the same conclusions you did:  large non-polarized capacitor.  I presume there is lots of open space inside the trolley.  I haven't taken the shell of yet.  Do you think this would work ?

3).  I get the flywheel inertia example.  That's an idea too.  I haven't seen the motor to determine if a flywheel can be mounted.

4)  The magnet idea look likes a keeper.  Easy and fast acting.   You can tell when I've been at this too long: how about a BIG electromagnet to clamp the trolley in position 

 

2) I think it can be made to do something but don't think it can be "engineered" without a lot of effort.  In other words you might come up with something by trial-and-error but doing the math to come up with a designed solution has too many unknowns.

3) Note that you don't want to mount a flywheel as that would INCREASE coasting distance when power is removed.  I think you want to decrease coasting distance!  My suggestion was to remove the flywheel (if it has one). 

4) If the track manager is joined at the hip to the MRD1 sensor setup then so be it.  Without seeing a photo I'm assuming you have two holes drilled in the track bed (one for the Emitter, one for the Receiver).  But is said track manager wedded to the MRD1 or to optical sensing?  If you are willing to throw the MRD1's 1/8 sec (or 1/16 sec) delay under the bus you can keep the optical sensing concept using a module such as this:

Untitled

This has a Emitter and Receiver on the board so you'd probably remove and run wires to one to give flexibility in mounting to your existing holes.  I realize that one attraction of the Azatrax IR reflective sensors is they use modulated IR pulses which helps with ambient light/background noise rejection.  The module above also uses modulated IR technology.  You'd pair this with a $1 eBay relay module.  I'm not quite clear on your 3 sec delay but for another $1 or so you can get an eBay relay module that has a delay feature.  Yes, there's some wiring to connect various modules but the cost should be much less than the $15-20 which appears to be the MRD1 price. 

Or if you aren't wedded to optical reflective sensing, the magnet on trolley with magnetic sensor on track would actually be less obtrusive using the same low cost eBay relay modules.

Attachments

Images (1)
  • Untitled
Last edited by stan2004
stan2004 posted:
shorling posted:

2)  I still like the RC option.  I did some dabbling on this  a few days ago and came to the same conclusions you did:  large non-polarized capacitor.  I presume there is lots of open space inside the trolley.  I haven't taken the shell of yet.  Do you think this would work ?

3).  I get the flywheel inertia example.  That's an idea too.  I haven't seen the motor to determine if a flywheel can be mounted.

4)  The magnet idea look likes a keeper.  Easy and fast acting.   You can tell when I've been at this too long: how about a BIG electromagnet to clamp the trolley in position 

 

2) I think it can be made to do something but don't think it can be "engineered" without a lot of effort.  In other words you might come up with something by trial-and-error but doing the math to come up with a designed solution has too many unknowns.

3) Note that you don't want to mount a flywheel as that would INCREASE coasting distance when power is removed.  I think you want to decrease coasting distance!  My suggestion was to remove the flywheel (if it has one). 

4) If the track manager is joined at the hip to the MRD1 sensor setup then so be it.  Without seeing a photo I'm assuming you have two holes drilled in the track bed (one for the Emitter, one for the Receiver).  But is said track manager wedded to the MRD1 or to optical sensing?  If you are willing to throw the MRD1's 1/8 sec (or 1/16 sec) delay under the bus you can keep the optical sensing concept using a module such as this:

Untitled

This has a Emitter and Receiver on the board so you'd probably remove and run wires to one to give flexibility in mounting to your existing holes.  I realize that one attraction of the Azatrax IR reflective sensors is they use modulated IR pulses which helps with ambient light/background noise rejection.  The module above also uses modulated IR technology.  You'd pair this with a $1 eBay relay module.  I'm not quite clear on your 3 sec delay but for another $1 or so you can get an eBay relay module that has a delay feature.  Yes, there's some wiring to connect various modules but the cost should be much less than the $15-20 which appears to be the MRD1 price. 

Or if you are wedded to optical reflective sensing, the magnet on trolley with magnetic sensor on track would actually be less obtrusive using the same low cost eBay relay modules.

Again Stan, thanks for your thoughts.  All good information.  To understand my resistance to change, a further understanding of the layout scene is required.  This is a two tier New York City scene with Dept 56 sky scrapers, Superstreets, street lights, walkways, Lionel's full size Grand Central Terminal, a couple of accessories, people, etc on top.  Below are five stub tracks (one is Track 61 with FDR and his car) servicing the GCT terminal complete with platforms, people, etc.   Strip LED lighting is mounted on the underside of the top tier to illuminate the platforms.  The upper tier is about 5 x 6 feet plywood, separated into two pieces, one larger and one smaller.  The trolley's are located on the larger piece.  All of the wiring to support upper tier is mounted on the underside under the upper tier plywood.  After condensing, there are still 60 wires which make their way under the table routed through the hollow standoffs supporting the upper tier.  It took 3 people hanging from the ceiling (almost) to place the larger plywood section on the layout without damage.  A miracle in itself.  This larger section is installed and everything works except you know what.  There are 3 sets of IR sensors supporting the trolleys. Two sets are somewhat accessible if you can stand on your head.  Accessibility to the other set is much more difficult.  Sensor wire service loops is a question ?  The design concept is to never remove the plywood but provide wiring/controller flexibility under the table.

 Here's where I am at the moment:

1)  I'm working to get the modified MRD1 with the 1/16 second delay.  Requires only change out of MRD1 under the table. Try it and see if it works.  Will not really know unless it's tested.

2) Use the existing IR sensors with a different detector board.  Requires only changes under the table.  Similar to your idea above.

3) Abandon the existing IR sensors and replace with ?  Choices are different IR sensors, magnetic switch, limit switch, proximity sensor, etc.  Interface sensor with a relay to control power.  Maybe the existing sensor holes/wiring could be reused.  Requires both under top tier and under table modifications.  At this point IR sensors may not be a good choice since this same technology may exhibit delays.  Essentially no delays with switch/relay technology.  Also, your suggestion.

shorling posted:

3) Abandon the existing IR sensors and replace with ?  Choices are different IR sensors, magnetic switch, limit switch, proximity sensor, etc.  Interface sensor with a relay to control power.  Maybe the existing sensor holes/wiring could be reused.  Requires both under top tier and under table modifications.  At this point IR sensors may not be a good choice since this same technology may exhibit delays.  Essentially no delays with switch/relay technology.  Also, your suggestion.

Well, I feel I must defend the honor of IR technology!  I don't know how Azatrax does it but there is no reason for IR to be the limiting delay factor in your application.  Here's a timing diagram from a typical IR sensor using modulated/pulsed IR technology.

Untitled

As the diagram indicates a typical modulated IR sensor will trip after it sees between 7 and 15 cycles of the correct IR pulsing frequency.  So with a typical pulsing rate of 40 kHz, that's less than 1 millisecond.  The electro-mechanical delay of the relay armature moving might be longer than that! 

Attachments

Images (1)
  • Untitled
stan2004 posted:
shorling posted:

3) Abandon the existing IR sensors and replace with ?  Choices are different IR sensors, magnetic switch, limit switch, proximity sensor, etc.  Interface sensor with a relay to control power.  Maybe the existing sensor holes/wiring could be reused.  Requires both under top tier and under table modifications.  At this point IR sensors may not be a good choice since this same technology may exhibit delays.  Essentially no delays with switch/relay technology.  Also, your suggestion.

Well, I feel I must defend the honor of IR technology!  I don't know how Azatrax does it but there is no reason for IR to be the limiting delay factor in your application.  Here's a timing diagram from a typical IR sensor using modulated/pulsed IR technology.

Untitled

As the diagram indicates a typical modulated IR sensor will trip after it sees between 7 and 15 cycles of the correct IR pulsing frequency.  So with a typical pulsing rate of 40 kHz, that's less than 1 millisecond.  The electro-mechanical delay of the relay armature moving might be longer than that! 

Stan, I like the way you think     I'll  keep your timing chart in mind.  I'm getting the 1/16 delay MRD1.  We'll see how they play.

I also had another thought.  I could add in a series resistor to slow the trolley when it enters the stub roadbed on a yellow light.  This is a little tricky since the trolley stalls at the bumper in response to low inertia bumps.  When the reverse switch transfer is only partial, the trolley loses power and is dead in the water.

Also, some earlier Lionel Birney Trolleys use Pullmor motors.  I was thinking the Pullmor has better low speed control than Lionel's current Birney.  However, it appears these old Pullmors are based on the old Gang Car.  My Gang Car does have better low speed control but needs a mini-crash to reverse ?

gunrunnerjohn posted:

Very few Pullmor motors have better low speed control then almost any can motor.  For one thing, the Pullmor is a three pole motor, that "feature" kind of limits it's low speed potential.

Hi gunrunnerjohn,

Thanks for the input.  I'm in basic agreement with you relative to the limits of the Pullmor motors low speed operation.  I'm wondering if the subject trolley has a gear ratio favoring low speed in combination with a reverse mechanism that only requires a light touch.  This might allow lower speed operation which I'm looking for to solve another technical issue.  I suspect this is not the case and that the trolley operation is more like the gang car but I'm willing to be surprised.

gunrunnerjohn posted:

I know my gang car and the Lionel #60 trolley need a pretty solid whack to reverse, no pussy-footing around with those at the bumpers.   The MTH trolley with the electronic reverse and electrical switch bumpers reverses with a much lighter touch.

You are correct as usual.  I'm looking for a short Trolley (the MTH is an inch too long) that can run at slow speed and still reverse.   My thought is that maybe the subject trolley with the Pullmor motor and  reduction gearbox is capable of lower speed with higher torque than the current DC units  ?  If so, there next question is how to eliminate the need for the whack.

I'm guessing the mechanism that swivels the gang car people is similar to the mechanism that reverses the trolley pole.  I don't care if the trolley pole reverses.  If there is a way to defeat the pole reversing mechanism, maybe the drag on mechanism along with the need for a whack can be reduced substantially.  The reversing mechanism throw on the trolley is also large which contributes I suspect to the oversized bumpers.  There  may also be a detente in the mechanism. 

I like the idea of using dynamic braking to stop a DC motor equipped unit. For some reason I think real trolleys and traction units may do this as well. Depending upon how much room is in the trolley you could rig up a normally closed relay. In the closed position it can be wired to short the motor causing the braking. When track power is applied the relay could open unshorting it and providing normal operation. A small (9v?) battery may be needed to provided the shorting power when track power is cut and the relay is closed.

WOW, where's this thread going ?  I didn't say anything about a battery unless were talking about a capacitor ?  I'll restate the issue as follows:  As Stan has correctly said, I need to stop the trolley on a dime, immediately,  when it's position is sensed to avoid the collision zone.   The trolley is modern Lionel, AC powered, rectified to a DC motor. bump and go.  Trolley is running on Superstreets and the problem is on a stub track about an inch longer than the trolley.  The stub ends in a bumper fence.  The Azatrax MRD1 IR sensors embedded in the road are about 1/2 inch from the fence.  As the trolley approaches the bumper it is detected, hits the bumper, reverses and travels about an inch and stops.  Since the trolley is now off the sensors, it restarts after a time out.  So the trolley travels about 1.5 inches after it is detected before it stops.  As Stan correctly calculated the trolley is going about 34 scale MPH when it hits the fence.  Doing the math, that translates to about 1.5 inches in 125 ms all determined with my calibrated yard stick.  If the trolley could the slowed enough upon reversal, the 1.5 inch travel would be reduced and it would stop over the sensors out of the collision zone.  So ruffly, the total travel should be limited to an inch or less.  Azatrax has cut the 125 ms  in half to 63 ms and the updated MRD1 are in transit to me.  This thread is about Plan B in case Plan A doesn't work.  The RC idea is for the power feed to spend time recharging the capacitor starting when the trolley is reversed instead of sending that energy to the trolley motor, thus slowing the trolley. When power is cut to the trolley, the trolley stops immediately.   The trolley must stop over the IR sensors or the timeout will automatically restart the trolley.

gunrunnerjohn posted:

The reversing in the #60 trolley is a mechanical switch that has some friction, hence the large bumpers.  I don't know that there's an easy way to substantially reduce the friction in that assembly, I never gave it any thought.  I like to solve these issues with electronics, not mechanics.

I did take a look at the exploded parts view for the 6-18404 Pullmor trolley.  The bumper slide bar has a teethed slot in which a pinion rides to reverse the trolley pole.  This has to be a high source of slide bar friction.  Looks like the pinion can easily be remove.  So the question is will the Pullmor trolley with its gear reduction gearbox and low friction slide operate reliability a lower speed than a modern lionel DC motor Brill trolley ?

If this is an AC motor, dynamic braking won't work without field current.  OTOH, if it's a can motor, just shorting the armature will indeed dynamically brake the motor, that's exactly what I do in my Super-Chuffer to define the smoke chuffs.  With a can motor when it's unpowered it acts like a generator, shorting the armature puts a load on the generated current and adds drag to the rotation.

gunrunnerjohn posted:

If this is an AC motor, dynamic braking won't work without field current.  OTOH, if it's a can motor, just shorting the armature will indeed dynamically brake the motor, that's exactly what I do in my Super-Chuffer to define the smoke chuffs.  With a can motor when it's unpowered it acts like a generator, shorting the armature puts a load on the generated current and adds drag to the rotation.

It's a DC motor and when the power stops, it stops.  The concern here is how to lower the trolley velocity upon reverse restart to limit its travel to less than 3/4 inch in 63 ms. The trolley probable takes 6 volts or so before it even considers moving.

Last edited by shorling

Another idea: delaying the trolleys restart after the bump.  Remember the trolley motor is DC powered by full wave rectified AC.  When the trolley bumps the DC polarity is reversed, so there is a potential signal change of state and also an edge.  Suppose there is a relay in the trolley which shuts off power at the bump and restarts lets say X number of seconds later.  This process is triggered by the state change or edge.  The trolley bumps, power in the trolley is cut off for X seconds.  It's like a trolley stop.   When the traffic light turns yellow and the trolley is still stopped over the IR sensors, power is cut to the tracks so when the trolley restart "times out" the trolley is still stopped by the IR sensors at the yellow light thereby compensating for the IR sensor delay.  Seems like it would work.  First thoughts are a latching relay with a timer.  Any ideas on how to mechanize this or maybe there is something on market that does this or can be adapted ?

Add Reply

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