Train Detection with Mutliple Sensors

[St. Simon]

Hi,

 

I'm part way through my project of interlocking my new Norwood Road Layout, and my attention has turned to how I will physically detect the trains.

 

The track was laid by the previous owner without insulated rail joints, apart from points, (its DCC), and as it has been done so nicely, I don't want to go and splice the rails to put them in. So I'm looking at infrared train detection.

 

My problem is that all the kits available on the market to do it are single sensor, this has the problem in that in a large section, if you have a light engine running through it, it will only see it once and then could get 'lost' within the section. Obviously this can be cured by putting in these single sensor boards in positions where at least one would also be covered if a train was in the section, however this has the disadvantage that these boards are quite expensive for the number I need.

 

Thus, my question is can I wire multiple sensors in such a way that only a single board is needed for each section?

 

Simon

[smokebox]

Could you arrange for the entry detector to operate a latching relay and then release (de-operate) the relay with another detector at the exit point of the section?

[Suzie]

You might like to get IRDOTs if you don't want to fit current sensors. They are available with a delay so that once triggered they stay 'occupied' for a while which for a moving train will be at least until it has got to the next sensor or leave the block.

 

The relay versions of IRDOTs can easily be paralleled.

 

Fitting current sensors will probably not be too hard. You just have to cut one rail at both ends of the block and feed it via the detector. Chances are that for most blocks you will already have an insulated joiner installed at one end where it meets a point, and there will be a dropper you can reuse.

[The Stationmaster]

Simon as we discussed last Saturday I do wonder if a 'strike in/strike out' arrangement might be the best answer?  Could be done with a latching relay circuit as suggested above or an electronic equivalent and no doubt through a micro computer.  Thus it would would like an axle counter but using only a single pulse to strike in and strike out of a section of track although that could leave part of a train or a divided train undetected.  

 

Then you could use either an optical detector or a reed switch.  

[Suzie]

If going to that level of sophistication an axle counter becomes quite simple!

[Colin_McLeod]

How does the axle counter work?

[St. Simon]

How does the axle counter work?

Hi,

 

Two types of 'Axle Counter', the Mechanical and Magnetic variety.

 

The mechanical variety is basically a treadle, which is physically activated by the flange of the wheel, a pair of these treadles can be used to provide direction detection.

 

The Magnetic variety is basically a magnet on one side of the rail and a detector on the other side. Installed, the magnet produces a magnetic field which induces an electrical current in the detector, if a wheel passes through the counter, it alters the magnetic field and thus alters the electric current, the change in thus current is detected and shows an axle passes through.

 

Both systems count the axles in and counter the axles out and if the net result of this is 0, then the line is clear, if the result is a positive or negative, it cannot be proved that the track is clear.

 

Simon

[St. Simon]

You might like to get IRDOTs if you don't want to fit current sensors. They are available with a delay so that once triggered they stay 'occupied' for a while which for a moving train will be at least until it has got to the next sensor or leave the block.

 

The relay versions of IRDOTs can easily be paralleled.

 

Fitting current sensors will probably not be too hard. You just have to cut one rail at both ends of the block and feed it via the detector. Chances are that for most blocks you will already have an insulated joiner installed at one end where it meets a point, and there will be a dropper you can reuse.

Hi Suzie,

 

The IRDOTS are what I'd like to use, but the boards are £10 or so, and I would need something in the order of 30 to cover my track circuits in full, so it really isn't financially viable. What I'm wondering is can multiple sensors be wired into one board.

 

I can't use the delayed feature as they assume that the train is moving and by the point at which it times out the train has left the section, where as it of come to a stand, thus there is no way of the interlocking proving the train has cleared a section.

 

The majority of my section limits are on plain line, so have no insulated joiners on. The problem for me in cutting the rail is that it is all painted, weathered and ballasted, so I could easily ruin the track, whilst I would also have to install an IBJ into the joint, which is a pain.

 

Simon

[Suzie]

Just cutting the rail with a cutting disc should not cause any damage, and fitting dummy fishplates should hide the join if that worries you. IRDOTs are the expensive way to go, so I think if you are budget concious you will have to bite the bullet and cut the gaps.

[crosti]

Hi there st simon

           I have a system of train detection using infrared detectors and a latch relay system. The way i have wired it is such that a train entering the section breaks the beam and latches up a relay and then when the last vehicle clears the section, the latch relay unlatches, this I am able to do by utilising a 470mf capacitor on the clearance  which charges up via the output relay on the detector when the the train has broke the exit infrared sensor and as soon as the last vehicle as passed the detector, and the output relay drops it discharges the capacitor into the latch relay thereby releasing it, the only snag with it is, it will only work one way at this present time !!!!

[meil]

Hi,

 

I'm part way through my project of interlocking my new Norwood Road Layout, and my attention has turned to how I will physically detect the trains.

 

The track was laid by the previous owner without insulated rail joints, apart from points, (its DCC), and as it has been done so nicely, I don't want to go and splice the rails to put them in. So I'm looking at infrared train detection.

 

My problem is that all the kits available on the market to do it are single sensor, this has the problem in that in a large section, if you have a light engine running through it, it will only see it once and then could get 'lost' within the section. Obviously this can be cured by putting in these single sensor boards in positions where at least one would also be covered if a train was in the section, however this has the disadvantage that these boards are quite expensive for the number I need.

 

Thus, my question is can I wire multiple sensors in such a way that only a single board is needed for each section?

 

Simon

Yes and if you care to join MERG ( http://www.merg.org.uk/ ) you would fine numerous ways and circuits of doing it there,

[Gordon H]

Yes and if you care to join MERG ( http://www.merg.org.uk/ ) you would fine numerous ways and circuits of doing it there,

 

Including a dual IR detector - known as the 'Dual Hector', kit 72A.

[Theakerr]

Over here we can get miniature reed switches that will fit both under and at the side of the track (Miniatronics Corp. Part #32-058-05.  5 for US$12.95 this summer in Seattle).  Suggest a simple latching on activated by the loco and and off activated by the brake van.  FYI, certain Hornby locos (L1 and 01 for sure) will activate a reed switch under the track.

[St. Simon]

Yes and if you care to join MERG ( http://www.merg.org.uk/ ) you would fine numerous ways and circuits of doing it there,

 

Hi Meil,

 

I'm already a member of MERG, but I thought that there kits were much the same as those offered by Heathcote etc, but I'll have a closer look now!

 

 

Including a dual IR detector - known as the 'Dual Hector', kit 72A.

 

I'll have a look at that, thanks Gordon!

 

Simon

[Junctionmad]

Train detection can be simple or complex, 

 

firstly you have to decide if you want track circuit occupancy , a la the prototype , i.e. right down to a single wagon 

 

the alternative is train presence detection, i.e., the assumption is the train is not detached within the occupancy section. 

 

magnetic sensors , witrh a rare earth button magnet fitted to each wagon and end , can be used as a form of axle counting , but you'll need some electronics to do the counting, all you need is a hall sensor on the entrance and exit of each section . There are several articles on MERGs site, my own experience is it can be hard to  prevent adjacent track triggering , unless the magnet is close to the track base, and this prevents visual and mounting difficulties 

 

Optical sensing is discussed extensively on the MERG site too, and there are kist for it ( HECTOR), personally Ive found IR to be either overly sensitive  ( i.e. when you put you r hands onto the layout you trigger detectors ) , or too insensitive and you miss detections 

 

The best for me , is to create section breaks and do current detection, I just slit the track with a mini drill , MERG has several kits for TOTI ( train on track indication ).  Furthermore if you want complete realism , you can fit resistors to wagon wheel sets and they will trigger the detector . Im awaiting some resistance inks/paint from the USA to see if I can just " paint " them on.  

[DCB]

For prototypical operation what you need is a notebook and pen, and two reed switches connected to a bell.  Put a magnet under the last vehicle, brake van? and the reed switches at the end of the section.  When the train leaves the section the bell rings twice and you can reset the points and signals and write it in the book.   Just like the real thing.   Best not to forget you left 907 on the up main before accepting an up express though.

[Junctionmad]

For prototypical operation what you need is a notebook and pen, and two reed switches connected to a bell.  Put a magnet under the last vehicle, brake van? and the reed switches at the end of the section.  When the train leaves the section the bell rings twice and you can reset the points and signals and write it in the book.   Just like the real thing.   Best not to forget you left 907 on the up main before accepting an up express though.

I think you could as easy hire a " bobby", he could then flag trains etc 

[Theakerr]

Not completely sure what your overall objective is (you did mention interlocking), but as an out of the box thought, would some sort of camera system (CCT) work.  Prices have dropped dramatically over the last few years.

[St. Simon]

Hi,

 

My overall objective is a way of detecting the presence of a train over the full length of a block section without cutting the rail and inserting IBJ's and at minimum cost. This will activate the coil of a Track Repeating Relay with the contacts being an input into the interlocking and to change the indications on the control panel.

 

Hope this makes it a little clearer!

 

Simon

[Junctionmad]

Hi,

 

My overall objective is a way of detecting the presence of a train over the full length of a block section without cutting the rail and inserting IBJ's and at minimum cost. This will activate the coil of a Track Repeating Relay with the contacts being an input into the interlocking and to change the indications on the control panel.

 

Hope this makes it a little clearer!

 

Simon

I want to make a concert flute , by drilling a piece of copper pipe, too.  The reality is a bit more complex. 

 

Im not being facetious

 

constructing a non-invasive detection system is not straightforward, all the current mechanisms have drawbacks as have been discussed , IR  can be setup to have  entrance/exit detection, hence only requiring two sensors. but IR has its limits  and spurious triggering is one of them and can cause unwanted  detections that have to be manually reset 

 

Axle counting is equally problematic in miniature form, 

 

Multiple sensors to effectively remove detection black spots are expensive and tedious to install etc 

 

There is no magic bullet here, the best bet is to buy a couple of test systems and see what you can live with 

 

The only semi-practical system that delivers near prototypical operation , is current monitoring and resistor wheel sets  ( for DC layouts  also AC frequencies can be used for detection  ) , effective but tedious to implement and has limitations ( requires  track segmentation ) 

[Chris Turnbull]

I don't know if this is any use to you but I have used IR detectors to interlock my layout and activate the SPAD indicators.  See this article:-

 

http://www.gauge0guild.com/gazette/article16_10.html

 

Chris Turnbull

[dasatcopthorne]

Simon.

 

I needed a very basic detection for a hidden fiddle yard.

 

I settled on the MERG toti. A light sensing LED switches on an

LED when a train blocks out light. The thing is, they are cheap.

 

Best of luck.

 

Dave

[Theduke71000]

Hi,

 

I'm part way through my project of interlocking my new Norwood Road Layout, and my attention has turned to how I will physically detect the trains.

 

The track was laid by the previous owner without insulated rail joints, apart from points, (its DCC), and as it has been done so nicely, I don't want to go and splice the rails to put them in. So I'm looking at infrared train detection.

 

My problem is that all the kits available on the market to do it are single sensor, this has the problem in that in a large section, if you have a light engine running through it, it will only see it once and then could get 'lost' within the section. Obviously this can be cured by putting in these single sensor boards in positions where at least one would also be covered if a train was in the section, however this has the disadvantage that these boards are quite expensive for the number I need.

 

Thus, my question is can I wire multiple sensors in such a way that only a single board is needed for each section?

 

Simon

Simon,

[Gordon H]

If the layout is DCC you are committed to a lot of expense with regard to train detection.

 

 

Detection with DCC is not necessarily expensive.

It partly depends what you want to do with the detection information.

If you just want to light an LED on a panel there are lots of simple circuits that will do this, e.g. MERG Pocket Money Kit PMP7 (807), costs GBP 1.50.

Unfortunately Dave didn't say which MERG TOTI he settled on, as there are several types available. It sounds like the LDR train detector, Pocket Money Kit PMP2 (802), which is even cheaper.

[St. Simon]

Simon,

Track already laid, if glued in place does NOT need plastic railjoiners. You simply use a mini-drill with a carbonundrum disc to cut a slot through each rail were needed ! Saves money !

 

If the layout is DCC you are committed to a lot of expense with regard to train detection.

 

 

Hi,

 

I know I only need to cut the track with a disc, but knowing my luck, I'll create a gash in the board. I'll have to install some sort of insulation where I've cut, otherwise the rail will expand and close the joint.

 

Detection with DCC is not necessarily expensive.

It partly depends what you want to do with the detection information.

If you just want to light an LED on a panel there are lots of simple circuits that will do this, e.g. MERG Pocket Money Kit PMP7 (807), costs GBP 1.50.

Unfortunately Dave didn't say which MERG TOTI he settled on, as there are several types available. It sounds like the LDR train detector, Pocket Money Kit PMP2 (802), which is even cheaper.

All I want it to do is energise a relay coil, that's all.

 

At the moment it's looking like I'll have to cut the track, it's just going to have to be careful work to avoid ruining the lovely track. Thinking about it, IR emitters and detectors might be unslightly and hard to disguise.

 

Simon