Using PSX-AR reversing units as block detectors

[Airbuspilot]

Hi

I am using several DCCspecialities PSX-AR Units for reverse loop control in a G scale garden railway, they are all working well in that function. The layout is being wired for automatic control with iTrain as the controller and RR CirKits Watchman for block detection and current transformers as detectors.

In the initial phase I have 11 detected blocks. 7 are simple cable to insulated track with a current transformer feeding the Watchman. The additional two blocks are the reverse loops controlled by the PSX-AR and connected to the Watchman by current transformers. I am using a current transformer on one of the output wires from each of the PSX-AR's. The seven sensed blocks are working normally however the two reverser blocks are showing permanently occupied.

The DCC supply from the PSX-AR to the track uses a pair of 14 gage solid copper wires, the longest run is around 50 meters. The current transformers are wired with twisted cable to the Watchman, these cables are very short 15/30 cm. The current transformers are fairly close together but can be separated by sliding along the feed wire. The feed wires are parallel and separated by about 8 cm there is no change when the transformers are at max separation. The track insulators have been checked and the cables are fairly new and undamaged.

As there is no current being drawn from the PSX-AR until a train enters the section I don’t understand why the Watchman is showing occupied. Any suggestions appreciated.

Robin

 

[Gizzy]

Could it be that the reverse loops use a relay to change polarity?

The relays might be drawing current....

 

[Airbuspilot]

Could it be that the reverse loops use a relay to change polarity?

The relays might be drawinf current....

Hi Gizzy. The units are all electronic with no relay, the two output wires go to either side of the track so no train no current. As both units are working correctly in the reverse mode I don’t think I have a short or a current leak from bad wiring. I could understand one giving trouble but not both.
Robin

 

[AlanL]

As there is no current being drawn from the PSX-AR until a train enters the section

Are you assuming this or have you measured the zero current? (I wouldn't know how ro measure the DCC current). Any experts out there because I'm not?

I would suggest disconnecting the DCC feeds to the reverse loops and see if that clears the track occupied condition with the current transformers still in place.

My thinking is that perhaps there is some current leakage across the sleepers and tracks due to dirt on the reverse sections that is causing the issue.
Another less likely problem may be the long feeds (up to 50 metres) to your return loops that you described as fairly new and undamaged.
Could the length simply be a problem, capacitance and leakage? Again, any experts?

I'm not familiar with your system but can you increase the current threshold on your detection system?

AL

 

[Airbuspilot]

Are you assuming this or have you measured the zero current? (I wouldn't know how ro measure the DCC current). Any experts out there because I'm not?

I would suggest disconnecting the DCC feeds to the reverse loops and see if that clears the track occupied condition with the current transformers still in place.

My thinking is that perhaps there is some current leakage across the sleepers and tracks due to dirt on the reverse sections that is causing the issue.
Another less likely problem may be the long feeds (up to 50 metres) to your return loops that you described as fairly new and undamaged.
Could the length simply be a problem, capacitance and leakage? Again, any experts?

I'm not familiar with your system but can you increase the current threshold on your detection system?

AL

Hi Alan

I have had a the opportunity to trouble shoot this morning, it would appear that the long cables act as some kind of arial.

Disconnecting the output cables as expected show not occupied.
Connecting one output shows not occupied but with both connected back to occupied.
Disconnecting the track end with either one disconnect or both disconnected shows occupied.

I will run a single extra wire to replace one of the feeder wires, keeping it well clear of the existing cable to see if that works. Possibly a twisted pair of cables might solve the problem, not sure if i can find that locally.

Robin

 

[PhilP]

I was about to suggest a twisted pair..

DCC should be balanced, but it is carrying quite a bit of current (or can be) it would be quite easy to get one leg out of balance with the other.

When that happens, just like us, things begin to fall over.


PhilP

 

[AlanL]

Disconnecting the track end with either one disconnect or both disconnected shows occupied.

Doesn't that mean that it is detecting even without any current flow in either one wire or both wires to the track. How can there be current without both wires connected to the track?

The long lengths do have voltage on them but as they are not connected , there should (in theory) be no current flowing.

Does that indicate leakage to ground?

AL

 

[Airbuspilot]

Doesn't that mean that it is detecting even without any current flow in either one wire or both wires to the track. How can there be current without both wires connected to the track?

The long lengths do have voltage on them but as they are not connected , there should (in theory) be no current flowing.

Does that indicate leakage to ground?

AL

Hi Alan

That is exactly my problem, in theory there should be no current flowing.

I ran these wires myself, new rolls opened from new, cable pulled through flexible tube with a tracer, cables In place for around 6 months. There is an uninterrupted run from the base station (with one wire sensed) to a junction box where the wires run without further joints to the track.

The junction box allows transit through a wall. The three reverse units pass through the junction box but via separate tubes. Disconnecting at the junction box shows unoccupied, disconnecting at the track either one at a time or both together shows occupied.

My first thought was a damaged wire or a badly insulated section. As both units are displaying the same characteristics the probability of damaged cables leaking are unlikely, same for the track - no train no current flow. The reverser units are performing normally.

It seems the most probable cause is some sort of interference between the pair of wires in the tube. From the base station to the wall is around 10 m, from the wall to the first reverse unit is around 30 m.

It will have to wait until after Christmas but I will try running a twisted pair and see if that helps.

Thanks for your comments

Robin

 

[AlanL]

The mysterious world of electrons ........

I've had a similar problem with my automatic signals. The signals are triggered with the passing of a train. I have 11 signals and all worked 100% of the time except for 1 signal that seemed to be permernantly triggered. The difference with this 1 signal was the length of the single core wire back to the controlling circuit. I solved the issue by modifying the control circuit. Originally the circuit used an optical isolator to trigger. The optical isolator is a neat solution, solid state, but it works on very low currents. I replaced the optical isolator with a miniature relay and now it has worked 100% for many years since.

As you are using Watchman detectors, I looked up their manual http://rrcirkits.com/manuals/WatchMan-manual-a.pdf and the detection current can be adjusted from 0.5ma tp 20ma.

Perhaps have a look at changing the sensitivity.

AL

 

[Airbuspilot]

Hi Alan

During the first lock down I discovered Arduino micro controllers, I looked at several detection systems. For me the best is current sensing, next Hall sensors and magnets.

I did look at changing the sensitivity but I think I will wait to see if twisted wires work first.

Roin

 

[Cobalt6700]

I have been playing with CT's for block detection - and you are right - what you are seeing is likely induction from other wires around the wires from your CT to the Watchman.

From the Watchman manual:

Simple lengths of the commonly available CAT-5 or similar cables may be used to connect up to 4 CT coils, one coil for each pair. Each coil should be connected to its input terminals with its own twisted pair wire. Do not connect any CT coil or its wiring to any other coils or wiring. Do not run the twisted pairs in close proximity to severe electrical noise sources such as unshielded florescent lamps with electronic ballasts. Do not use straight wiring. The twisted pair wiring serves to cancel out most electrical noise in the sensitive circuits.


There are two possible issues: Induction from the Watchman to the CT, and Induction / wire capacitance of your track connecting wires.

For the first one, running it in shielded ethernet cable with one end put to a ground may help to protect your signals from unwanted noise. DCC is a very 'noisy' signal and will easily be 'picked up' by cables around it.

The second one - the track connection wire diameter and length will affect the magnetic field of the wire (that is how a CT works). Really long wires have a higher capacitance and can trigger a block occupied alarm.

I'm not going to get into the whole 'twisted / untwisted' DCC feeder wire debate - I don't have enough knowledge to back that up. I didn't really see it making a difference though - however I was only testing over 5M.

My system's CT's are located on a PCB so I don't have issues with the first, however the second has caused me some head scratching. In the end, I added a trim pot across the CT which allowed me to adjust the 'sensitivity' of the CT.

It seems RRC also did that - they describe a setting on the Watchman which allows you to set the sensitivity of each CT line - they even have an 'auto' function to do it for you

The other thing you might see (if it is DCC feeder induction) is that a train on the other block will trigger both - again due to the induction. I don't think there is much you can do about that, other than shield the cables or physically move them apart

 

[Airbuspilot]

I have been playing with CT's for block detection - and you are right - what you are seeing is likely induction from other wires around the wires from your CT to the Watchman.

From the Watchman manual:

Simple lengths of the commonly available CAT-5 or similar cables may be used to connect up to 4 CT coils, one coil for each pair. Each coil should be connected to its input terminals with its own twisted pair wire. Do not connect any CT coil or its wiring to any other coils or wiring. Do not run the twisted pairs in close proximity to severe electrical noise sources such as unshielded florescent lamps with electronic ballasts. Do not use straight wiring. The twisted pair wiring serves to cancel out most electrical noise in the sensitive circuits.


There are two possible issues: Induction from the Watchman to the CT, and Induction / wire capacitance of your track connecting wires.

For the first one, running it in shielded ethernet cable with one end put to a ground may help to protect your signals from unwanted noise. DCC is a very 'noisy' signal and will easily be 'picked up' by cables around it.

The second one - the track connection wire diameter and length will affect the magnetic field of the wire (that is how a CT works). Really long wires have a higher capacitance and can trigger a block occupied alarm.

I'm not going to get into the whole 'twisted / untwisted' DCC feeder wire debate - I don't have enough knowledge to back that up. I didn't really see it making a difference though - however I was only testing over 5M.

My system's CT's are located on a PCB so I don't have issues with the first, however the second has caused me some head scratching. In the end, I added a trim pot across the CT which allowed me to adjust the 'sensitivity' of the CT.

It seems RRC also did that - they describe a setting on the Watchman which allows you to set the sensitivity of each CT line - they even have an 'auto' function to do it for you

The other thing you might see (if it is DCC feeder induction) is that a train on the other block will trigger both - again due to the induction. I don't think there is much you can do about that, other than shield the cables or physically move them apart

Thanks for your reply. I agree that it's almost certain to be a problem of induction between the wires. For the wires from the CT to the Watchman I'm using twisted pairs and they are quite short, I have 7 CT's which are working normally and only the 2 connected to the watchman are giving a problems so I have discounted these wires as the problem for now.

I will try running a new twisted pair from the terminal box to the track to see if that helps then try to adjust the sensitivity of the individual CT's on the watchman if necessary. I have no idea how a twisted pair would solve the problem but there seems to be a consensus that it helps.

Robin