Tram line plan

[gavino200]

I'd like to add a tramline. Below is sort of a mock up. The single viaduct doesn't even have track in it. I canalized it long ago. I'm planning to make a tram line between urban areas at both ends of the layout. I'm going to keep it DC, so I can run tiny trams that are a pain to convert. The tram in the picture is a non-working tomytec model. I have zero experience with trams. If someone could answer a few questions for me I'd really appreciate it.

 

 

1. I want to make the gradient very gradual. I have about four or five pillars for each change in elevation. I'm planning on filling these down so that I can decrease the Kato single viaduct gradient to one quarter or one fifth of its stock grade. Will trams have a problem with that?

 

2. I'm not sure whether to make a station at each end or a loop. If a loop, then what sort of circuit works best for a reversing loop? The rest of my layout is DCC with a Digitrax control system. 

 

3. If I were to automate the tram line, is it a problem that it's DC? Can I still switch to manual?

 

4. I'm guessing trams can turn on the tightest radii. Am I wrong?

 

Any other ideas, suggestions, or problems I don't seem to be aware of?

 

This is one end of the line. The ground level double track will go into a tunnel here and be hidden under the city which, will be elevated. The tram line climbs on single viaduct to the city where it will be at (elevated) ground level.

 

 

 

This is the incline between two kato piers of successive heights. I'm planning on filing down piers to reduce this gradient to approx one quarter of what you see.

 

 

Another view of the incline. I'll need to get some more straight viaduct track to keep the intervals even. Some of these intervals are with short straight segements.

 

 

I'm not sure if it's good to combine a corner with an incline so I'll keep the corner level.

 

 

Other end of the line. Nothing permanent here. I just laid some track as a mock up.

 

 

[railsquid]

20 minutes ago, gavino200 said:

I'd like to add a tramline. Below is sort of a mock up. The single viaduct doesn't even have track in it. I canalized it long ago. I'm planning to make a tram line between urban areas at both ends of the layout. I'm going to keep it DC, so I can run tiny trams that are a pain to convert. The tram in the picture is a non-working tomytec model. I have zero experience with trams. If someone could answer a few questions for me I'd really appreciate it.

 

 

1. I want to make the gradient very gradual. I have about four or five pillars for each change in elevation. I'm planning on filling these down so that I can decrease the Kato single viaduct gradient to one quarter or one fifth of its stock grade. Will trams have a problem with that?

 

 

I was originally planning to do this:

 

Tram crossing attempt by Rail Squid, on Flickr

 

but the trams (Modemo and Kato) didn't cope with the tight-curve-on-a-gradient at all well.

 

They seem happy going up the "mainline" gradient (about 1:30 incline, or 5cm rise over ca 1.5 metres) with gentler curves (280mm).

 

Quote

 

4. I'm guessing trams can turn on the tightest radii. Am I wrong?

 

 

Mine run round the Tomix 103mm curves fine, the Kato Portram is even rated for 90mm (!), Note I haven't tested the Tomytec 3-part tram mechanism, IIRC it might have a larger minimum radius.

 

Kato Portram going round 103mm curves:

 

 

Edited February 11, 2018 by railsquid

[kvp]

The safe minimum radius is 140 mm. Modemo trams don't like anything tighter than this. Tomytecs could usually go through a 103 mm, but not all and most hate S curves at this radius. At 140 mm, all the trams i have are reliable through S curves (like the Tomix mini turnouts), including the multi articulated ones.

[gavino200]

On 2/11/2018 at 2:41 AM, kvp said:

The safe minimum radius is 140 mm. Modemo trams don't like anything tighter than this. Tomytecs could usually go through a 103 mm, but not all and most hate S curves at this radius. At 140 mm, all the trams i have are reliable through S curves (like the Tomix mini turnouts), including the multi articulated ones.

 

I'm thinking of going with 103R assuming that I can file down the central plastic if I have trouble with non Tomix brands negotiating the corners.

[gavino200]

A little play with 'Anyrail' produced this decidedly ugly loop. But it's proof of concept. I'll easily be able to fit a loop at this end of the layout. So it looks like the Tram line design will be dumb-bell shaped. A single line with a tear-drop reversing loop at each end. 

 

[Kiha66]

Are you still interested in using DC for the tram line?  I've been playing around with some logic circuits and I may have found a solution that would allow both automatic running and manual control. 

[gavino200]

1 minute ago, Kiha66 said:

Are you still interested in using DC for the tram line?  I've been playing around with some logic circuits and I may have found a solution that would allow both automatic running and manual control. 

 

Yes, I am. The appeal of a DC track is that I could run tiny trains like a Stephenson's Rocket, and other novelties, that I wouldn't have a hope of getting a DCC decoder into. 

[gavino200]

It looks like it will also work with 140R track. But I like the effect of the 103R better. The 103R would be much easier to fit into a street grid.

 

[Kiha66]

The layout would use spring switches, and insulators (red) right after the switches.  As the trains will always go around the loops in the same direction, there is no need to reverse them.  The single track connecting the two would switch direction automatically when a train ran over a sensor (green).  The each sensor activates a relay to switch the polarity. As the speed and original direction is controlled by a throttle, you can stop, slow, speed up, or change direction as much as you like, although going around the curves in the opposite direction would require a lot more work to automate the switches.  If your interested I can start making a control diagram for how to wire the sensors and relay.

Each loop can be any direction, I just chose counter clockwise for the drawing.

[gavino200]

I'm not sure exactly what a spring switch is. But I was talking with KVP today. He pointed out that since Tomix switches are "power routing" flipping the switch would automatically flip the polarity in the straight stretch. 

The two switches could be wired together so that they flip at the same time. In that way all I would need is one sensor right next to the switch on the outgoing side of each loop rigged to trip the switch. Does that make sense. 

 

I have a fair bit to learn before I make a decision about this. Has your method been tried out before? Or would I be the guiney pig (I'm not necessarily opposed to that - just curious). What other hardware and software are you planning to use in the setup. I definitely am interested. 

[bill937ca]

Or you can just create a continuous dogbone and avoid polarity issues.

[gavino200]

1 minute ago, bill937ca said:

Or you can just create a continuous dogbone and avoid polarity issues.

 

Yes, I would. But I don't have the real estate, unless I sacrifice another line from my yard. :(

[Kiha66]

8 minutes ago, gavino200 said:

I'm not sure exactly what a spring switch is. But I was talking with KVP today. He pointed out that since Tomix switches are "power routing" flipping the switch would automatically flip the polarity in the straight stretch. 

The two switches could be wired together so that they flip at the same time. In that way all I would need is one sensor right next to the switch on the outgoing side of each loop rigged to trip the switch. Does that make sense. 

 

I have a fair bit to learn before I make a decision about this. Has your method been tried out before? Or would I be the guiney pig (I'm not necessarily opposed to that - just curious). What other hardware and software are you planning to use in the setup. I definitely am interested. 

 

My design wouldn't need a computer, just a latching relay in the simplest version.  The way a latching relay works is whenever it detects a train at one end it switches the polarity to be heading the opposite direction.  I actually found an old drawing of this arrangement, so it looks like I'm not the first to come up with it.  The LEDs let you know which direction the center track is currently going, and the Diode bridges are to keep the loops going in the same direction.  The reed switches could be substituted for a more modern sensor.  Here the power supply polarity would mater, but you can control the speed and stop and start manually, or let the tram run automatically.

A spring switch just means a train can run through it even when the switch is set to a different route, and the switch will automatically reset as the spring pushes the points back to the correct position.  Kato switches can do this, but you need to turn off live frog to keep the switch from shorting.

[gavino200]

Ok. I had to stare at that for a loooooong time. Thanks for the diagram.

 

Questions:

1. Would electromagnetic relays make a loud clicking sound?

2. Is it possible to use a solid state (quiet) solution instead, like an Arduino unit. 

3. What's the advantage of using relays over power routing switches. 

4. I don't remember weather running the wrong way through a kato switch ever caused a derailment. I don't think so, but it's been a long time since I've run anything in DC. I can't run through them now as it caused a short, but not necessarily a derailment.

5. I'm planning on using Tomix track for the end of the loops. Will Tomix switches work for this too?

6. What happens if you run the train very slowly and trip the reed switch before you get through the switch. Does the tram reverse?

 

Understanding this diagram takes me to the limits of my electrical understanding. But it looks like something I could easily build myself. 

Edited February 13, 2018 by gavino200

[Kiha66]

1) The smaller relays tend to be pretty quiet, you'll probably get more noise out of trains running over the joint between track sections.  Using the switches will make much more noise than a single relay.
2) You can certainly use a computer to do the same thing, and while it would give you more options it would be more complicated and at greater cost.  You could also use simple solid state chips and transistors instead of the relay.  A flip-flop chip and transistors would be completely silents and very reliable, plus would work at the same voltage as the sensor whereas using the sensor to throw the switches would require intermediate steps to change the power from 5V logic to 12v switch power.
3) From what I understand of the switch method, there runs a risk of short circuiting the whole track if they don't both change at exactly the same time.  Although now that I think about it if you add an insulator to separate the single track into two sections it would work fairly well if KVP has a simple plan for how to throw the switches via the sensor.  This plan has the advantage of already being finished design wise.

4) I haven't had a derailment yet with my switches, but I don't use them in this way often enough to be sure.
5) I only use kato track myself, so I am not sure whether tomix track can do this. 

6) As the loop is always powered to run in the same direction, the each sensor will only trip to reverse the center track once.  So even if you park the train over the sensor, or set it off multiple times, the center track will not reverse again until you set off the other sensor.

 

[gavino200]

6 minutes ago, Kiha66 said:

1) The smaller relays tend to be pretty quiet, you'll probably get more noise out of trains running over the joint between track sections.  Using the switches will make much more noise than a single relay.
2) You can certainly use a computer to do the same thing, and while it would give you more options it would be more complicated and at greater cost.  You could also use simple solid state chips and transistors instead of the relay.  A flip-flop chip and transistors would be completely silents and very reliable, plus would work at the same voltage as the sensor whereas using the sensor to throw the switches would require intermediate steps to change the power from 5V logic to 12v switch power.
3) From what I understand of the switch method, there runs a risk of short circuiting the whole track if they don't both change at exactly the same time.  Although now that I think about it if you add an insulator to separate the single track into two sections it would work fairly well if KVP has a simple plan for how to throw the switches via the sensor.  This plan has the advantage of already being finished design wise.

4) I haven't had a derailment yet with my switches, but I don't use them in this way often enough to be sure.
5) I only use kato track myself, so I am not sure whether tomix track can do this. 

6) As the loop is always powered to run in the same direction, the each sensor will only trip to reverse the center track once.  So even if you park the train over the sensor, or set it off multiple times, the center track will not reverse again until you set off the other sensor.

 

 

I like it. I'd have a preference for using a chip or transistors. 

 

I have a Tomytec mechanism on the way from Hobbysearch. No idea when it will get here. I never check. When It gets here I'll set up a test track with a simple DC loop and a junction for it to go backwards through (in spring switch fashion) to see if there are any problems.

 

I'd probably repeat the experiment with the Tomix switch when I get it to see if it works. That will be a while as I need to play around with Anyrail before I'll know what to buy. Then order it and wait for regular SAL.

 

Are your building one of these for yourself too? It seems like a fun project.

[Kiha66]

I don't need one myself, but I've found it fun to design and actually apply all the automation engineering classes I've taken.  If you're interested I can keep working on a finished design with solid state parts, and maybe a board mount for them.  

[gavino200]

1 minute ago, Kiha66 said:

I don't need one myself, but I've found it fun to design and actually apply all the automation engineering classes I've taken.  If you're interested I can keep working on a finished design with solid state parts, and maybe a board mount for them.  

 

Yes-sir-ee-bob!  :)

[gavino200]

6 minutes ago, Kiha66 said:

I don't need one myself, but I've found it fun to design and actually apply all the automation engineering classes I've taken.  If you're interested I can keep working on a finished design with solid state parts, and maybe a board mount for them.  

 

If you're building a prototype, don't buy components. Message me or email me a link for what you need. I'll buy them and have them sent to you.

[Kiha66]

2 minutes ago, gavino200 said:

If you're building a prototype, don't buy components. Message me or email me a link for what you need. I'll buy them and have them sent to you.

 

Thanks!  I'll try to have a finished design first, and have one of my professors take a look at it before buying any actual components.  One of my teachers has a circuit board making machine I can borrow to make a board for the components. 

[gavino200]

Just now, Kiha66 said:

 

Thanks!  I'll try to have a finished design first, and have one of my professors take a look at it before buying any actual components.  One of my teachers has a circuit board making machine I can borrow to make a board for the components. 

 

Cool. Btw there is absolutely no rush with this.

[Kiha66]

While looking at sensor options, I found Azatrax already makes a circuit for this track plan!  Seems they beat me to the idea.

http://www.azatrax.com/reverse-loop-system-2.html

[gavino200]

Is it possible to use as a power source, the output from a DCC decoder? The input to the decoder could come from the (now completely unused) power bus that used to power my small yard (now abandoned). That way the line would have it's own designated short protection. And the DC trams could be controlled from the Digitrax cab, just like the DCC trains.

 

[kvp]

Actually you don't need a relay for polarity as the Tomix turnouts will handle exactly that and then the light trams won't have to cut the blades mechanically. No isolation needed either. (=less derailments and stalls)

 

The arduino could be replaced with and edge triggered set/reset flip flop or a relay (if you add a TTL input transistor driver to the TCS sensor output) and by using a series capacitor to ground the turnouts could be controlled by switching 12V or gnd to them. The arduino just makes the construction less solder intensive as everything could be bought assembled, just add 12V to the arduino. (TCS sensors are 5V TTL and can only drive digital logic circuit inputs without overloading)

 

ps: The 140mm minimum is needed by all modemo, some dauphin and longer tomytec trams to not bind in the curves as you can't file down the curvature of the rails. 103 mm will cut the rolling stock options roughly in half.

[gavino200]

22 minutes ago, Kiha66 said:

While looking at sensor options, I found Azatrax already makes a circuit for this track plan!  Seems they beat me to the idea.

http://www.azatrax.com/reverse-loop-system-2.html

 

This looks perfect (although less fun than a custom designed circuit). This setup uses powered switches rather than swing switches, right?