Welcome to the forum   
Welcome Guest! To enable all features please Login or Register.

Notification

Icon
Error

Share
Options
View
Go to last post in this topic Go to first unread post in this topic
Offline trainbuff  
#1 Posted : 18 April 2010 22:10:36(UTC)
trainbuff


Joined: 26/11/2006(UTC)
Posts: 507
Location: Atlanta, Georgia
Hi,

Is it possible or advisable to extend a contact track to include a switch in order to have a simple track occupancy circuit? For example if a wagon gets accidentally uncoupled in the ladder track or throat area leading into or out of the shadow station, can that be detected? I've looked at a c-track switch and it looks like a lot of cutting would have be done to make that work, essentially isolating the movable tracks and the tracks between the movable tracks and the frog. But I'm afraid this may make the switch considerably less reliable for delivering power to a passing locomotive.

Has anybody done some something similar, or have another method for occupancy detection over switches?

Essentially I want to know if the throat area, which is essentially 3 switches in a tunnel, is clear before I let a train in or out.
UserPostedImage

Thanks for any ideas or suggestions.

layout details: c-track, mobile station, and designing an old style analog control panel for switches, signals and occupancy detection. I have the books "The HO Signal Book 03402" and "Controlling, Switching, Running 07421" for references.
Offline Fredrik  
#2 Posted : 18 April 2010 22:57:20(UTC)
Fredrik

Sweden   
Joined: 13/07/2004(UTC)
Posts: 642
Hi,

in my opinion it's very advisable to have such a occupancy detection in the switches/points/turnouts. I have also made the cuts in C-track switches you mention. However now I use K-track (again) as this task is (IMHO) easier. The 3-way-switch (C) is impossible to "cut" - the X-switch (C & K) is very difficult, and afterwards somewhat unreliable... Haven't tried the large redius switches in C, but they can be used "out-of-the-box" with K.

If you use the "diode-trick" (if possible with this kind of feedback - I use it for S88-alike decoders) for the isolated rail it'll get more reliable.
Fredrik.
*ECoS 2 + ECoSDetector + SwitchPilot + ECoSTerminal; *Z21 + Loconet + Digikeijs + MGP; **CS3+ + CdB (** coming soon...)
WWW: MJ-fjärren
Offline trainbuff  
#3 Posted : 18 April 2010 23:08:15(UTC)
trainbuff


Joined: 26/11/2006(UTC)
Posts: 507
Location: Atlanta, Georgia
MJ-Teknik wrote:
Hi,

in my opinion it's very advisable to have such a occupancy detection in the switches/points/turnouts. I have also made the cuts in C-track switches you mention. However now I use K-track (again) as this task is (IMHO) easier. The 3-way-switch (C) is impossible to "cut" - the X-switch (C & K) is very difficult, and afterwards somewhat unreliable... Haven't tried the large redius switches in C, but they can be used "out-of-the-box" with K.

If you use the "diode-trick" (if possible with this kind of feedback - I use it for S88-alike decoders) for the isolated rail it'll get more reliable.


Thanks! So this is doable, good. Luckily I don't have hidden three-ways or X-switches (double slips?). But I do have, gaspOhMyGod, R1 curved switches. I think the connections in a curved switch are the same as a straight switch.

I will do some research on the "diode-trick" as I'm not sure what that is. If you point me in the right direction, it would be much appreciated.

Thanks
Offline Fredrik  
#4 Posted : 18 April 2010 23:39:19(UTC)
Fredrik

Sweden   
Joined: 13/07/2004(UTC)
Posts: 642
Ok,

the R1 curved switches are also "cuttable". There is a metal-plate in the middle of the switch, looks like a capital 'A', and begins at the end of the points. This metal plate needs to be divided (preferrably not removed as then the ground connection gets even worse). I cut this in 2 halves and glued it back down (won't stay otherwise). I also had to connect all 3 ground spade-connectors to each other with a wire (I think it was - check ground connections).

The diode-trick: It's simply to connect the isolated rail with the grunded one with a diode (don't remember which way to connect it right now...). This also allows the "isolated-side" to connect to ground, but not enoguh for a S88-detector to react - thus it get's more reliable as all wheels are in fact "grounded" Cool. I've read most about this in German forums - maybe someone else can explain it better.

Link: Diode-trick, 1:st picture. "Masse" = Ground (sorry, linked to German page).

To use this connection I today insert CAN-feedback-modules with integrated "diode-trick" (expansion-board), so I hardly need to think about it at all.

Edit: Inserted link.
Fredrik.
*ECoS 2 + ECoSDetector + SwitchPilot + ECoSTerminal; *Z21 + Loconet + Digikeijs + MGP; **CS3+ + CdB (** coming soon...)
WWW: MJ-fjärren
Offline trainbuff  
#5 Posted : 19 April 2010 00:40:30(UTC)
trainbuff


Joined: 26/11/2006(UTC)
Posts: 507
Location: Atlanta, Georgia
MJ-Teknik wrote:
Ok,

the R1 curved switches are also "cuttable". There is a metal-plate in the middle of the switch, looks like a capital 'A', and begins at the end of the points. This metal plate needs to be divided (preferrably not removed as then the ground connection gets even worse). I cut this in 2 halves and glued it back down (won't stay otherwise). I also had to connect all 3 ground spade-connectors to each other with a wire (I think it was - check ground connections).

The diode-trick: It's simply to connect the isolated rail with the grunded one with a diode (don't remember which way to connect it right now...). This also allows the "isolated-side" to connect to ground, but not enoguh for a S88-detector to react - thus it get's more reliable as all wheels are in fact "grounded" Cool. I've read most about this in German forums - maybe someone else can explain it better.

Link: Diode-trick, 1:st picture. "Masse" = Ground (sorry, linked to German page).

To use this connection I today insert CAN-feedback-modules with integrated "diode-trick" (expansion-board), so I hardly need to think about it at all.

Edit: Inserted link.


When you describe the 'A', you are looking from above?

I made a quick sketch of where I thought cuts had to be made for going straight and for diverging. Its easy to see what to cut when swtich is straight, but I'm not sure what to do when diverging.

When diverging, the cut has to be made between on both sides of the frog? I guess the frog can stay powered?
UserPostedImage

When going straight, only one cut has to be made, between the movable points and the straight track.
UserPostedImage

Offline Fredrik  
#6 Posted : 19 April 2010 00:48:46(UTC)
Fredrik

Sweden   
Joined: 13/07/2004(UTC)
Posts: 642
The top of the 'A' is located just at the frog (looking from above). the 'legs' lies to the right of the black rail leading to the frog, and to the left of the green one.

Also: I assume this is a R2 turnout - the R9 (or whatever they are) I don't know of, haven't used them. Also don't know if they (R9) have this 'A'. If they do (and for the other switches) the purpose of the plate is to interconnect ground (and for the wheel flanges to roll on it for connection) between both rails.

For the R1/R2 switches there was no need to cut anything in the rails - only tha 'A'-plate and the ground connection close to the connector spades (those cut on straight track).
Fredrik.
*ECoS 2 + ECoSDetector + SwitchPilot + ECoSTerminal; *Z21 + Loconet + Digikeijs + MGP; **CS3+ + CdB (** coming soon...)
WWW: MJ-fjärren
Offline trainbuff  
#7 Posted : 19 April 2010 00:56:48(UTC)
trainbuff


Joined: 26/11/2006(UTC)
Posts: 507
Location: Atlanta, Georgia
Thanks, yes its R2. I think I understand you. I will try it and see.

Thanks!
Offline trainbuff  
#8 Posted : 19 April 2010 07:10:46(UTC)
trainbuff


Joined: 26/11/2006(UTC)
Posts: 507
Location: Atlanta, Georgia
Ok, I'm still not understanding how the frog is isolated or not. Because I'm a visual learner, here is a picture:
UserPostedImage

In order to isolate the right rail, the right point and the right divergent rail for purpose of making contact track we should cut the 'A' to separate the points, where the red line is. When the 'A' is cut, is it also separated from the frog to point connection? If it is disconnected from the frog then it appears the frog will be un-powered. That means a locomotive going straight will experience a short piece of unpowered rail. Is that an issue?

There is a very small gap between the frog and left rail and right divergent rail at the frog to rail joint, so the frog is not getting power from those two rails.

My idea would be to cut the 'A' section in such a way that the left point is still connected to the frog. That way a locomotive going straight would have uninterrupted power.
UserPostedImage
But that seems like a complicated cut.

What am I missing?

Thanks for your help and patience!
Offline river6109  
#9 Posted : 19 April 2010 13:39:10(UTC)
river6109

Australia   
Joined: 22/01/2009(UTC)
Posts: 14,636
Location: On 1965 Märklin Boulevard just around from Roco Square
I think you are going over the top with this.
Having c-track what chances are there for your carriages to uncouple ?
Are you going bejond prototypical speeds ?
I would first look at your couplers and make sure they are compatible and secure.
I can understand your concern having a non detectable distance in your hidden layout area.

All c-tracks rail current (left & right rail) is connected, by cutting the connection you will have power on side (outside) and the detection rail will be inside.
When your trains go over a curved setting turnout your inside rail is already set up for it.
When your turnout is set straight, your tunrout frog is touching the curved track.

Your above picture shows the intention to cut away this section.
This whole piece is connected on three spots.
a.) picture: underneath the turnout you have to bend the latch backwards, (visible)
b.) at the middle you have to bend 2 latches backwards. (visible)
c.) the third spot is connected on the underside of the moving frog.
How easy it is to seperate the frog pin from its connection (underneath) is unknown to me. there is a chance it breaks.

The only other option I could suggest is:

When both rails are seperated from the current,

1.) your turning frog (left hand turn) has no connection to the outside rail, so your detection rail (inside rail) does'nt need any alterations, except a cut in the rail on both sides of the turnout, adjoining tracks
2.) When your frog position is straight, your outside track becomes the detection rail.

What to do: Isolate the outside track of your turnout on both sides (a cut on both sides of the adjoining track), connect your detection wire to it and isolate the inside track detection by using a switch.

UserPostedImage
https://www.youtube.com/river6109
https://www.youtube.com/6109river
5 years in Destruction mode
50 years in Repairing mode
Offline Fredrik  
#10 Posted : 19 April 2010 22:16:28(UTC)
Fredrik

Sweden   
Joined: 13/07/2004(UTC)
Posts: 642
Hi,

@trainbuff: You're right about the cut on the last pic above - that's exactly how I made the cut!

@river6109: I haven't tried your solution, however it get's more expensive once an extra switch is involved...

There is no need to cut the rails in any case: Just isolate the connection between the switch and the next piece of track (on the '0' connection).

And another good reason to use this detection is the possibility that a train stops to early, or is to long for the selected track (might have entered the wrong track). The switches are the most dangerous areas and thus should be detected!! BigGrin (that's my opinion!)
Fredrik.
*ECoS 2 + ECoSDetector + SwitchPilot + ECoSTerminal; *Z21 + Loconet + Digikeijs + MGP; **CS3+ + CdB (** coming soon...)
WWW: MJ-fjärren
Offline trainbuff  
#11 Posted : 20 April 2010 02:19:18(UTC)
trainbuff


Joined: 26/11/2006(UTC)
Posts: 507
Location: Atlanta, Georgia
Hi River6109, I liked your idea pretty good, it eliminated what looks like a difficult cut. But there is one small caveat, when the switch is diverging, the straight track no longer becomes a contact track. On a curved switch that straight portion is pretty long. If I'm going through all this trouble I may as well maximize the contact area and use the other method. Recently a large flat car (non-Marklin) uncoupled in this area and had I not looked up at the last moment my BR-41 may have meet the floor. That's what started this for me. I agree about getting all couplers working, but I have so many different kinds, old and new, Marklin and others, some that are hard to bring into alignment, etc., that fixing them all is going to take a long time. I had a real laugh when you told me I was going over the top with this! I decided I was a goner soon after I started this hobby!

Hi Fredrick, thanks for the confirmation on how you did the cut. It looks like the frog is the only non-contact track area and that is only for wagons on the divergent track. Its so short, it should not affect the ability to detect occupancy on the switch. Just out of curiosity, what German key words would one use to search for this topic on a German forum?

Thanks to both of you for your time and help.
Chris
Offline river6109  
#12 Posted : 20 April 2010 11:57:19(UTC)
river6109

Australia   
Joined: 22/01/2009(UTC)
Posts: 14,636
Location: On 1965 Märklin Boulevard just around from Roco Square
trainbuff wrote:
Hi River6109, I liked your idea pretty good, it eliminated what looks like a difficult cut. But there is one small caveat, when the switch is diverging, the straight track no longer becomes a contact track. On a curved switch that straight portion is pretty long. If I'm going through all this trouble I may as well maximize the contact area and use the other method. Recently a large flat car (non-Marklin) uncoupled in this area and had I not looked up at the last moment my BR-41 may have meet the floor. That's what started this for me. I agree about getting all couplers working, but I have so many different kinds, old and new, Marklin and others, some that are hard to bring into alignment, etc., that fixing them all is going to take a long time. I had a real laugh when you told me I was going over the top with this! I decided I was a goner soon after I started this hobby!

Hi Fredrick, thanks for the confirmation on how you did the cut. It looks like the frog is the only non-contact track area and that is only for wagons on the divergent track. Its so short, it should not affect the ability to detect occupancy on the switch. Just out of curiosity, what German key words would one use to search for this topic on a German forum?

Thanks to both of you for your time and help.
Chris


Chris,
One thing you have to weigh up is: are you going to do this on a permanent basis, coupling uncoupled carriages together.
My example doesn't leave out any section whether it is a ground section of track or a detection section of track.

I can't see your point, when the switch is diverging, the straight track is no longer a detection track.
Why would you need a straight detection track when your carriages are diverging ? and visa versa.
Are you saying when the turnout is set to turn your carriages are going straight ahead and therefore you've got no detection ?
You are making it more complicated as it is in real modeltrain time.

When your turnout is straight, why would you need a detection section on the curved track ?
You have one common ground conection from one side of the turnout to the other.

Your trains are parked in a shadow station and should be far enough from any turnout to allow the next train to enter into a different track.
All you have to do with your shadow station is to make sure your blocksystem is set up properly.
If for instance your carriage(s) gets uncoupled while sitting on the turnout section, your blocksystem has to be 2 section apart, otherwise if a train has already entered the last blocksystem section, it is to late to halt the train (crash).
If everything is alright your signals will be green.
Solong that turnout switch is straight you will have permanent detection on the outer track section of your turnout(s).
If the track is occupied (shadow station track) the memeory will diverge it to another track and if your turnout has to diverge to a curved section this section will be detected if any carriages have been left behind.

One more question: are the trains entering or leaving the section (sketch).
If they are leaving this section by entering into the turnout, things are bit different.
Regardless which track you are on, while there are carriages sitting on any section of the turnout (straight or curved), the S88 will tell the memory the track is occupied.
The same principal applies. you will not have carriages on the curved track sitting when your train just left somme carriages on your straight track and visa versa.
The difference between the 2 examples: train leaving or entering the shadow station is:

When trains are leaving your left behind carriages on the turnout regardless straight or curvced section will not allow any other train to leave the shadow station.
If your trains are entering the shadow station you need 2 Blocksystem sections to prevent any train entering the last block section (trainlenght) before the turnout.

Chris life ment to be easy.

Good luck.

You can duplicate this option for every turnout and it does'nt matter which way the frog is switched.

John








https://www.youtube.com/river6109
https://www.youtube.com/6109river
5 years in Destruction mode
50 years in Repairing mode
Offline trainbuff  
#13 Posted : 21 April 2010 04:12:33(UTC)
trainbuff


Joined: 26/11/2006(UTC)
Posts: 507
Location: Atlanta, Georgia
river6109 wrote:
trainbuff wrote:
Hi River6109, I liked your idea pretty good, it eliminated what looks like a difficult cut. But there is one small caveat, when the switch is diverging, the straight track no longer becomes a contact track. On a curved switch that straight portion is pretty long. If I'm going through all this trouble I may as well maximize the contact area and use the other method. Recently a large flat car (non-Marklin) uncoupled in this area and had I not looked up at the last moment my BR-41 may have meet the floor. That's what started this for me. I agree about getting all couplers working, but I have so many different kinds, old and new, Marklin and others, some that are hard to bring into alignment, etc., that fixing them all is going to take a long time. I had a real laugh when you told me I was going over the top with this! I decided I was a goner soon after I started this hobby!

Hi Fredrick, thanks for the confirmation on how you did the cut. It looks like the frog is the only non-contact track area and that is only for wagons on the divergent track. Its so short, it should not affect the ability to detect occupancy on the switch. Just out of curiosity, what German key words would one use to search for this topic on a German forum?

Thanks to both of you for your time and help.
Chris


Chris,
One thing you have to weigh up is: are you going to do this on a permanent basis, coupling uncoupled carriages together.
My example doesn't leave out any section whether it is a ground section of track or a detection section of track.

I can't see your point, when the switch is diverging, the straight track is no longer a detection track.
Why would you need a straight detection track when your carriages are diverging ? and visa versa.
Are you saying when the turnout is set to turn your carriages are going straight ahead and therefore you've got no detection ?
You are making it more complicated as it is in real modeltrain time.

When your turnout is straight, why would you need a detection section on the curved track ?
You have one common ground conection from one side of the turnout to the other.

Your trains are parked in a shadow station and should be far enough from any turnout to allow the next train to enter into a different track.
All you have to do with your shadow station is to make sure your blocksystem is set up properly.
If for instance your carriage(s) gets uncoupled while sitting on the turnout section, your blocksystem has to be 2 section apart, otherwise if a train has already entered the last blocksystem section, it is to late to halt the train (crash).
If everything is alright your signals will be green.
Solong that turnout switch is straight you will have permanent detection on the outer track section of your turnout(s).
If the track is occupied (shadow station track) the memeory will diverge it to another track and if your turnout has to diverge to a curved section this section will be detected if any carriages have been left behind.

One more question: are the trains entering or leaving the section (sketch).
If they are leaving this section by entering into the turnout, things are bit different.
Regardless which track you are on, while there are carriages sitting on any section of the turnout (straight or curved), the S88 will tell the memory the track is occupied.
The same principal applies. you will not have carriages on the curved track sitting when your train just left somme carriages on your straight track and visa versa.
The difference between the 2 examples: train leaving or entering the shadow station is:

When trains are leaving your left behind carriages on the turnout regardless straight or curvced section will not allow any other train to leave the shadow station.
If your trains are entering the shadow station you need 2 Blocksystem sections to prevent any train entering the last block section (trainlenght) before the turnout.

Chris life ment to be easy.

Good luck.

You can duplicate this option for every turnout and it does'nt matter which way the frog is switched.

John



Hi John,

Totally unrelated, I was browsing a Model Railroader Magazine (May 2007) and was surprised to see a switch wiring diagram very similar to yours. Interesting coincidence. Evidentally your method is used else where too.

By your response I can tell you are way ahead of me in using blocks, shadow stations and computer (memory) control. At this point, with my limited experience, all I think I need is a red light on the control panel when anything is on a switch regardless of switch position. But I'm going to take what you said in consideration.

The shadow station is a two way affair and I don't know what level of signal and automation I will eventually use. Besides my layout is small. Right now I envision a simple control panel where I can see what track is clear and set the signals, switches and run a train accordingly. In some sections the signals will control the track (shadow station) and in others, the operator will have to mind the signals.

Thanks for your response,
Chris

Offline river6109  
#14 Posted : 21 April 2010 13:04:40(UTC)
river6109

Australia   
Joined: 22/01/2009(UTC)
Posts: 14,636
Location: On 1965 Märklin Boulevard just around from Roco Square
Chris,
Having been there and done this many times over, your topic, seeing the sketch, I came up with the idea.
In your case, find out the simplest way of dealing with trains coming and going by controlling them via signals.
As you've mentioned your layout is'nt that big, your tasks of managing your trains in the shadow station should't be to difficult.
I do not see cutting up turnouts is the best solution.

John
https://www.youtube.com/river6109
https://www.youtube.com/6109river
5 years in Destruction mode
50 years in Repairing mode
Offline Olle3770  
#15 Posted : 03 May 2010 02:51:10(UTC)
Olle3770

Sweden   
Joined: 29/01/2009(UTC)
Posts: 75
Location: Eskilstuna, Sweden
This is the way I did it. Works for me.

UserPostedImage

Cut the connectors on the underside of the switch on the outgoing side (on the right in the picture). Cut the connectors on the upper side between the rails and the tongue (red strokes) using a Dremel or something similar.

Now you should almost have the X shaped (including the tongue) part inside the turnout isolated. The outermost rails are connected and serves as "0". The tongue will connect the X part with "0" so you need to isolate it on its "outsides". I use some sturdy lacquer (polyurethane or whatever on some) and a piece of silverish sticky tape on some to test what works best (blue strokes). The important thing is to make sure the tongue doesn't connect the X-part to the outer rails.

Now you need to isolate stuff. Take the brownish isolation gadgets and put one on the "right" side "0" connection. This way it won't short the incoming tracks detection. Then put the same kind of brownish isolation gadgets on one of the two outgoing (right side) "0"-connections. Which one depends on the turnout, but you have to do some thinking by yourself BigGrin. Now (on the underside) connect the "0"-section of the isolated V-part of the switch with the track occupancy detection "0" part of the rail going into the switch from the left side. This makes the switch a part of the incoming track. I use some wire and two 74995 with one of the 74995 being connected to two wires, one going to the X part and one going to the S88.

The green portion in the picture is the only place unseen by my scheme. To incorporate even this you have to cut one rail and do some wiring. Haven't tested this but an idea is shown below:

UserPostedImage

Since most of my rolling stock is longer than the 1.5 cm's we're talking about I don't care about it. The section before the switch will signal occupancy and when a train reach the tongue it will connect the X-part with one of the outer rails thus "grounding it".

I find this quite simple, so simple that I might miss something, but what? Hope this makes sense.
Offline trainbuff  
#16 Posted : 03 May 2010 05:54:51(UTC)
trainbuff


Joined: 26/11/2006(UTC)
Posts: 507
Location: Atlanta, Georgia
Hi Olle

That's a good solution and good explanation. I agree the 1.5cm are negligible. Thanks for sharing.
Offline xxup  
#17 Posted : 03 May 2010 07:22:09(UTC)
xxup

Australia   
Joined: 15/03/2003(UTC)
Posts: 9,457
Location: Australia
MJ-Teknik wrote:
Link: Diode-trick, 1:st picture. "Masse" = Ground


Oh this looks good. Can someone please translate the text and post it as a sticky entitled "The S-88 diode trick"?

It looks like it might work with a converted Hamo lok that currently won't run on my layout due to the DC wheels not working on contact tracks..
Adrian
UserPostedImage
Australia flag by abFlags.com
Offline efel  
#18 Posted : 03 May 2010 22:46:41(UTC)
efel

France   
Joined: 23/02/2005(UTC)
Posts: 800
Hi,

The diode trick is well known but does not work 100% with few decoders (including marklin).
It should be replaced by a zener (as I have used for years), or by Perz's trick: see
https://www.marklin-users.net/fo...aspx?g=posts&t=12351

Fred
Offline Tdl  
#19 Posted : 06 May 2010 12:07:38(UTC)
Tdl

Netherlands   
Joined: 30/03/2006(UTC)
Posts: 71
Location: Amsterdam
Train detection on a turnout is a necessity when the tower man cannot see the turnout. This will be the case for turnouts in hidden parts of the layout. Most computer programs also require train detection on turnouts.

On my K-track layout I have implemented axle detection. This requires cuts in the rail.
For a reliable axle detection the isolated rail segment should be somewhat longer than the longest axle gap in a train. For a 27,5 m couch in HO scale the axle gap is approx. 22 cm.

However cuts in the rail weaken the construction. This is particularly disavantagous in turnouts.

In turnout complexes in which I cannot make axle detection sections of the minimum length, I use 'darkness detection'.
After some experiments I have choosen for infrared. The receiver is in the track and is connected to a S88 decoder contact. In the visible part of the layout the sender is at the ceiling of the room. In hidden parts of the layout the sender is placed over the track.
I have made PCB's on which there are 8 infrared receivers.

The light openings between the wagons and in certain wagon types (for example open frame container wagon) cause
erronous detection. To minimize the effect I have taken three measures:
- in the receiver filter short duration changes in the train detetection;
- on critical places e.g. a crossing place two recievers in series;
- in the computer program short duration changes in track occupancy are ingnored.

Modifications to the trains, such as applying reflective surfaces, are not neccessary (this was a prerequisite).
Currently only one of my trains is not detected reliably. It is the PBKA Thalys trainset: this train has a light grey underside. I assume that a simple solution would be to darken the underside of this trainset.

Bottom line: For me darkness detection is the better choice for train detection on K-track turnout complexes.
Offline Olle3770  
#20 Posted : 09 May 2010 02:33:22(UTC)
Olle3770

Sweden   
Joined: 29/01/2009(UTC)
Posts: 75
Location: Eskilstuna, Sweden
trainbuff wrote:
Hi Olle

That's a good solution and good explanation. I agree the 1.5cm are negligible. Thanks for sharing.


Thanks, you're welcome. Happy to share with what little I can contribute. I owe you people so much. If you (or anybody else) tries it you're all welcome with feedback and ideas for improvement (P.M is OK if you like). Especially a foolproof way of isolating the tongue from the outer rail. The "right" tape seems to do the trick, but for how long will it stick? The lacquer seems to work as well, but will it brittle and fall of in time? It looks to me as if a train is negotiating the switch from the "wrong" side and the switch is in the "wrong" position the wheels will push the tongue aside and there will be very little, if any, mechanical contact with the "isolation material" part of the tongue (which initially worried me). But it is hard to know using "Mk I Eye Ball" technology.

I will include a supervisory module in my computer program which will alarm me if any switch operation as a side effect results in a track occupancy change related to the section the switch is in. This way I might spot problems as they arise.

I've been looking at the double-slip crossing and I think I can do the same to that one, but I'm not sure yet. I don't own any three-way switch so I can't tell if the "same scheme" can be applied to it. In my world, three-way switches and double-slips mostly belongs to the yard, and the yard is always occupied by something. Automating a yard is a later challenge for me. I will be happy If I can get the main line(s) swinging and then manually bring train-sets in and out of circulation.
Offline jfl_207  
#21 Posted : 17 May 2012 11:27:43(UTC)
jfl_207

Belgium   
Joined: 16/05/2012(UTC)
Posts: 6
Location: Pont-de-Loup
Hi,
greetings from Belgium.

I know this thread is one year old, but i wanted to share with you a solution we are using to get detection on a switch with C track.

I'm new to this forum and yesterday while searching on Google i've found some idea on how to do it.

the best one come from this forum.

The problem is that either you have to make a not so easy cut, or you have to use some isolating stuf, or you have to use cable to make link between different part of the switch.

so we have mixed the idea and we have found a solution.

Detection track c swith marklin

we just need to:
  • Cut the two rails just after the frog (the red line)

  • Remove the "A" piece (shown by the arrow) to remove it, no need to cut anything. just pull it gently with some Pliers.

  • Cut the 3 small metal link under the switch (the 3 blue lines)


After this modification, you get detection on the switch with just a dead zone of the length of the frog.
We have run an engine a the smallest speed and no trouble at all to run trough the switch.

sorry if this has already be posted.

and sorry for any mistake as english is not my native language.
thanks 3 users liked this useful post by jfl_207
Offline French_Fabrice  
#22 Posted : 17 May 2012 16:47:58(UTC)
French_Fabrice

France   
Joined: 16/05/2011(UTC)
Posts: 1,475
Location: Lyon, France
HI,
I've just discovered this thread today, so I'm quite late ....

Just a suggestion, don't know if it's applicable in your case. What about Viessmann 5233 "feedback decoder with track occupancy detector" ?
If a loco stays on a turnout and if the light is on, then you know the track is occupied.
If it's a wagon, then the documentation says you must use resistor paint on the axles (never experimented myself).

link to the 5233 documentation : http://www.viessmann-mod.../5233_92126_03_DE-EN.pdf

If this device can fit your needs, the no need to cut anything.

HTH
Cheers
Fabrice
Offline jfl_207  
#23 Posted : 17 May 2012 17:17:25(UTC)
jfl_207

Belgium   
Joined: 16/05/2012(UTC)
Posts: 6
Location: Pont-de-Loup
Hi,

it's also a solution.

But with this, the resistive axle only work with two rails systems.

on three rails system you need a contact between center rails et external rails.
so, if you don't put a skid under the wagon, you can't detect it.

Jean-Francois

Offline French_Fabrice  
#24 Posted : 17 May 2012 18:04:48(UTC)
French_Fabrice

France   
Joined: 16/05/2011(UTC)
Posts: 1,475
Location: Lyon, France
Originally Posted by: jfl_207 Go to Quoted Post


But with this, the resistive axle only work with two rails systems.

on three rails system you need a contact between center rails et external rails.
so, if you don't put a skid under the wagon, you can't detect it.

Jean-Francois



oh yes, you're right! sorry for having read the doc too quickly Blushing
cheers
fabrice
Offline perz  
#25 Posted : 30 May 2012 23:27:25(UTC)
perz

Sweden   
Joined: 12/01/2002(UTC)
Posts: 2,578
Location: Sweden
Just one comment from my own experience: Contact tracks are sufficiently reliable for detecting locos, which usually are heavy enough to give enough contact pressure on the wheels. They are also sufficiently reliable to detect moving cars, if you have some kind of de-bounce circuit (which is built into the S88 and equivalents). But for cars dropped behind, standing still, contact tracks often fail.

Maybe you can say that it is better to have an unreliable detection than none at all. But personally I would rather invest in better couplings, if dropping cars is a real problem.

Offline Pierobon  
#26 Posted : 05 July 2012 03:42:42(UTC)
Pierobon

Brazil   
Joined: 03/12/2011(UTC)
Posts: 16
Location: Here
Originally Posted by: perz Go to Quoted Post
Just one comment from my own experience: Contact tracks are sufficiently reliable for detecting locos, which usually are heavy enough to give enough contact pressure on the wheels. They are also sufficiently reliable to detect moving cars, if you have some kind of de-bounce circuit (which is built into the S88 and equivalents). But for cars dropped behind, standing still, contact tracks often fail.

Maybe you can say that it is better to have an unreliable detection than none at all. But personally I would rather invest in better couplings, if dropping cars is a real problem.


I would rather go to the fishing shop and buy some lead, or to the beach and get a bucket of sand, this way you can have a heavy ballasted and reliable cargo or passenger car detection... LOL

[EDIT: following lines added]
You can also make real-weight cars...
How much weight do you have to put in a car, so it get proportionally heavy as it's 87 times bigger version? Hmmm, that depends on the coach, let's say for a typical ICN2000, ICE, Eurostar, TGV, or any other modern recent train that weights around 50 tons per coach, you would have to make a 575kg model coach. Drool
And if you think about those heavy cargo with 100ton+, would be a single model car weighting more than a ton... Blink is that possible?
Offline river6109  
#27 Posted : 05 July 2012 06:10:09(UTC)
river6109

Australia   
Joined: 22/01/2009(UTC)
Posts: 14,636
Location: On 1965 Märklin Boulevard just around from Roco Square
Hi everyone,

problem: detection or secure couplings.

Couplings from different manufacturers come in different shapes and forms.

Märklin: reliable to a certain extend but if you mix old static couplings (the most reliable for heavy trains) with new KKK couplings you run into problems (different height on some)
Fleischmann: These are the most reliable couplers for heavy weight and long trains but they do uncouple in curves.
Roco: 2 couplers come into my mind and again under heavy load the train carriages uncouple in curves but the latest coupler is one of the safest coupler each coupling latches onto the other but again with none Roco carriages this coupling can interfere with the buffers and can cause derailments.

John
https://www.youtube.com/river6109
https://www.youtube.com/6109river
5 years in Destruction mode
50 years in Repairing mode
Offline jfl_207  
#28 Posted : 05 July 2012 08:37:59(UTC)
jfl_207

Belgium   
Joined: 16/05/2012(UTC)
Posts: 6
Location: Pont-de-Loup
Originally Posted by: Pierobon;344312
You can also make real-weight cars...
How much weight do you have to put in a car, so it get proportionally heavy as it's 87 times bigger version? Hmmm, that depends on the coach, let' Go to Quoted Post

And if you think about those heavy cargo with 100ton+, would be a single model car weighting more than a ton... Blink is that possible?


Hello,

Sorry, but i have to disagree with your way of getting scale weight of a model.
If a real coach is let say 50Tons, your model would be 50.000.000g divided 3 times by 87 (one time per dimension) so your car would be +/- 76g.

Nmra recommendation are 1 oz + 1/2 oz per inch of lenght (+/- 28g + +/- 14g per inch).
so let say we have a car of 25meters. in ho this would be 28 cm. (for ease of calculation, i will round it to 10 inch (25.4 cm)
nmra recommended weight would be 28g + (10*14g), so 168 g.

reference : see this link Nmra directive for weight of a model


thanks 2 users liked this useful post by jfl_207
Offline Pierobon  
#29 Posted : 18 July 2012 06:38:49(UTC)
Pierobon

Brazil   
Joined: 03/12/2011(UTC)
Posts: 16
Location: Here
Originally Posted by: jfl_207 Go to Quoted Post
(...) 50.000.000g divided 3 times by 87 (one time per dimension) so your car would be +/- 76g.


Sorry about my miscalculations.

I was just starting to write things like $*#(&!@%#!( because of you disagreeing with me, but after just a little tiny bit of thinking I came to the conclusion that you are probably smarter than I am, and that your maths are 3 times more precise than mine. LOL

Now seriously, thanks for real.
Offline jfl_207  
#30 Posted : 18 July 2012 10:31:56(UTC)
jfl_207

Belgium   
Joined: 16/05/2012(UTC)
Posts: 6
Location: Pont-de-Loup
Originally Posted by: Pierobon Go to Quoted Post


I was just starting to write things like $*#(&!@%#! LOL

Now seriously, thanks for real.


As long as your $*#(&!@%#! are in H0 scale, it's not a problem BigGrin Flapper



Offline Pierobon  
#31 Posted : 29 July 2012 16:28:51(UTC)
Pierobon

Brazil   
Joined: 03/12/2011(UTC)
Posts: 16
Location: Here
Problem solved for me:

Buy this: http://www.ajckids.com/products/Trix/62611
Operate all lokos via cantenary.
If loko is a diesel or steam, use a battery (you have even more realism because you will have to stop the loko to refuel - change batteries). Perhaps you can use a rechargeable battery and a yard section with center rail to recharge...
Or you can leave just the turnout without center rail and use the battery for the train to go just that couple of inches.

Anyways, i'm doing it with just cantenary lokos. LOL
Users browsing this topic
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.

| Powered by YAF.NET | YAF.NET © 2003-2024, Yet Another Forum.NET
This page was generated in 1.392 seconds.