Loco Detection: What Is It and How Does It Work?

[hosh]

Don't worry though, because the "Community Moderators" (Censors) are working just fine! 😎

[ramuncho]

 Bonjour a tous

Je suis nouveau sur le forum plus je le suis regulierement (tous les jours) depuis un an

En ce qui concerne LD, mon avis personnel c'est qu'on ne verra jamais 

quand on veut sortir un produit ,il sort

[RB51]

Hi ramuncho - welcome to the forum.

 

Your post in English (I think)

Hello has all I is new on the forum more I am it regularly (every day) With regard to LD, my personal opinion it is that it will never be seen If it were to leave it would be it already left

--------------

Bonjour, merci de votre poteau. Vôtre n'est pas une vue rare, mais nous tous devons attendre et voir. Je contrsucted ceci utilisant une traduction APP, ainsi excuses si elle n'est pas lisible. R-

 

Hi, thank you for your post. Yours is not an uncommon view, but we all have to wait and see. I have constructed this using a translation app, so apologies if it is not readable. R-

[LCDR]

 Thinking about the practicalities of a Locomotive Detection system, and having read back through this thread I have to confess I am none the wiser about how in practical terms a loco detection system is supposed to work.

 

Some years ago I built a demonstration layout which was a single track that crossed a scenic section between two tunnels.. This operated all day long without human (or even my own) intervention because I had a shuttle module fitted. This was purely electronic operated by reed switches on the track and a magnet inder the train. As the train struck in at either end the module was triggered by an impulse from the reeds and after a brief time delay the current to the rails was reversed and the train moved off in the opposite direction. On reaching the other end, the whole sequence operated in reverse.  To make this work it needed insulated breaks in the rails, the reed switches and the magnets 'blu-tacked' under the train.

 

So my question is  - How would LD actually detect a train? Would there be a rail break that was energised by a wheel crossing it, a  reed switch operated by a magnet under the train, or a trip switch that closed as a wheel went over (like the old Tri-ang  train control kits) , or a passsive IR detector, or what? Would you need lots of them scattered round the layout associated with stations, points and signals, and would the electronic messages from the sensors be transmitted via the rails or some other method (radio perhaps?)  How would interference be prevented e.g. how would conflict between the power supply, command & control messages and train detection messeages be kept seperate?

 

Is there even a pre-existing system made by another manufacturer, in which case why does Hornby need to develop one?

 

[Howbi]

LC..........extract from May 2014 DCC handbook....."In it's simplest form Hornby's Loco Detection will be able to switch signals when a loco passes a sensor as well as stop or reduce loco speed, according to the signal state and is also capable of telling which loco is passing a sensor, what speed it is travelling at and which direction it is travelling in. The Loco Detection system is made up of three elements:- a loco detection module (LDM), track sensors & loco detector tags. Each LDM is designed to plug into a USB port on a PC and can read up to 48 track sensors." ............HB

[96RAF]

Expanding on what HBM has said:

The detectors on the track note only that something has occurred over its head, the labels under the loco/trains identify what this something is, and the modules connected to the sensors pass this data to RM, which has prior knowledge of what is moving, at what speed, and in which direction.

 

From this accumulated data some measure of control can be applied according to rules set up in accordance with the list of commands published in the RM forum sticky labelled LD command list. How this method of control will be applied is yet to be established, whether by program or some other yet to be advised  RM methodology.

 

Further than than that applied logic we cannot go, as Hornby has put the system on the back burner according to folk lore accumulated from various emails to HCC, etc.

[LCDR]

 Ah yes, but HOW does it do it? Is it mechanical, magnetic, electrical or magic?

 

Some kind of impulse will have to occur which says 'I am here and going in direction X'   There are 'axle counters' used on the real railway (and are replacing track circuits) which incorporate a 'Freddie' which detects the proximity of a passing wheel (I believe it is done with magnetic induction but I am not 100 % sure) , and by having two or more sensors adjacent and linked logically will calculate direction and speed digitally.  Or will there be a model railway version of a Balise, which receives a radio signal from a passing train?  If so will each loco require an additional bit of kit that sends signals to the detector? (ERTMS in miniature in fact. )  I am intrigued because all these devices will have to be very sensitive to detect the small movements involved in model form. How will it cope with a mix of plastic and metal wheels, I wonder?

 

We live in interesting times, indeed.

[St1ngr4y]

My understanding is that the "label" to which RAF refers, on the underside of a loco, is like a "bar code" which appear on the items you buy at the supermarket. The sensors on the track are bar-code readers, so when the loco passes over the sensor, it is just like swiping a can of peas at Tesco's  😀 This results in an electronic message being sent to the LD module containing the Id of the sensor and the DCC Id of the loco, and maybe an indication of the direction which I assume can be detected by the way the swiping occurred.

Ray

[Howbi]

LC..........further extract from May 2014 info.........The track sensors require a 1mm hole to be drilled through a sleeper for installation..Detector tags are stuck to the underside of vehicles and these are 'read' by the sensor they pass over. There are 106 encoded tags in the set presumably numbered from 001 to 106....there was much speculation as whether this is a bar-code type system but bar-code readers are much larger than the sensor used here..........HB

[hedley_reynolds]

Hi I have been looking at all of your inputs on this topic and I'm just wondering if Hornby are waiting to see how block signalling do with their infrared loco detection system as it controls 2 to 4 aspect signals as well as level crossing barriers and gates as the loco's pass over the sensor under the track also it's in n gauge as well as 00 gauge and I might be totally wrong about this but I'm just curious as it's been a long time since this topic arose in the forum

Hedley

[hosh]

Hornby have stated that the "sensors" will know train speed and direction but that is vague? Maybe they simply mean RM will know this since these parameters are set in and by RM. And "direction" might only mean forward or reverse, not north or south for example.

 

The troubling thing about the entire system is that the train ID will be sent to RM, implying that RM doesn't know where all trains are at all times. There can only ever be 1 train headed toward 1 sensor at any given time (else collisions), so RM "should" always know what train just triggered a sensor, but it appears it wont.

 

And this is where it gets very intriguing - if RM has no memory of where trains are and what they're doing, then how can it even restart a train stopped at a signal? It would seem all new commands come from a triggered sensor - stopped trains do not trigger sensors! LOL

 

My money says that's why this product is on ice! :)

[Fishmanoz]

Good point about direction hosh.  On speed, RM in general would only know it if it can calculate how long a loco has been progressing on accel/decel between one speed step and another. 

 

As for knowing where anything is, RM doesn't, except it does know if a loco is at a sensor, but it doesn't know where the sensor is.  It has no concept of something approaching something else.

 

And if something is stopped, it can only be restarted by a program or a conditional command given by something else triggering a sensor. 

[96RAF]

RM has no idea of distance either. As far as it is concerned the sensors are dots in space, as are points. It has no concept of straights or bends or distances twixt points and/or sensors, so speed is of no real importance either, but maybe time is between sensing events.

[hosh]

And if something is stopped, it can only be restarted by a program or a conditional command given by something else triggering a sensor. 

 

The only way I see a stopped train being restarted is by a green signal. That requires 2 things - RM remembering where a stopped train is ("Resume loco on signal" ??) and some sort of synchronising of the signals/sensors.

 

Start changing points and have trains crossing over tracks heading in different directions though and this system simply cannot work. You need to know where stuff is and where it's headed. You cannot do that with signalling for opposite directions, and even if you could, it would be extremely end user unfriendly and unintuitive.

[hosh]

Just had a bit of a re-read over this thread.

 

What are people's thoughts on the copyright claims? Anyone believe this?

 

I find it very amusing actually - do they really think someone might copy this? 😆

[96RAF]

Maybe its the other way round Hosh.

Hornby may have encroached on someone else's patent, and have to prove otherwise, or maybe they are waiting for it to lapse.

Rob

[Fishmanoz]

Good thought from Rob, although the info some time ago on having had a light bulb moment might suggest otherwise.

 

To clarify copyright versus patent too.  Copyright is ownership of rights in written works, including code for programs in electronic equipment, and it exists automatically from the time it is written without any other action needing to be taken.  So I own copyright in this post when you see it.  I would suggest there is copyright in RM but unlikely to be in LD.  Patent on the other hand is to protect inventions in material form and must be applied for.  It must be new and not subject to already existing patent, and also must be unpublished.  Consequently, Hornby cannot tell us details of LD if they are applying for patents as that would publish it.  

 

It's an interesting conundrum whether to patent or not.  In fast moving technology, it's better to go to market early without patent and make your profits while others catch up.  Here, if you wait for a patent, others will pass you. Also once patented, your invention is published and others can figure how to get around it.

[96RAF]

Agreed Fishy.

A case in point was the original Black and Decker Workmate, which was cloned almost before it came to the market and patents did little to protect it. There was a great hoo and haa about it at the time.

 

Lots of stuff in the old days was marketed with the caveat Patent Pending presumably in hope of gaining some protection.

 

With the advent of cheap knockoffs these days a patent seems to be a stable door event anyhow and apart from being (often badly spelled) named lookalikes I don't suppose these cloners could even be prosecuted as their internals and other materials are so shoddy they could say it is a different product.

 

Rob

[hosh]

Not sure what I find more funny - the implication that someone might want to copy Hornby's broken system, or some peoples belief that the actual reason for the delay is copyright related!

 

😆

[LCDR]

 Why not have a device that emits and reads a unique passively generated electro magnetic code, much like contactless credit cards do?  The chip attached to the loco, the reader in the track as already suggested. The chip could actually link to the DCC address. (Or perhaps that IS what is proposed?)

[Michael_A]

The likes of digitrax and Uhlenbrock already make detection systems. The trick is to integrate it with RM......

[hosh]

Exactly. And if you know how to do the software properly, there is no need to get fancy with the hardware.

 

The alarm bells went off for me as soon as I saw Hornby's proposed detectors.

 

Tbh, if I see any system using other than current detection I would have zero interest. It is as simple and as adequate as anything else.

[96RAF]

I've said it before and I'll say it again...

....Rocrail has been using every detection system known to man for years, including but not limited to:

current detection,

Ir sensing,

bar codes,

RFID,

a man with a flag and a bloke with a megaphone.

Why Hornby is still faffing about defies logic.

[Fishmanoz]

LC, the tag on the loco doesn't seem to be as sophisticated as a chip, but the tag ID does get entered into the loco setup information in RM, along with all the other setup info like DCC address, functions and CVs. So the system knows which loco has arrived at which sensor.  Locos can also be configured in groups and commands which can be issued on arrival at a sensor can apply to individual locos (the one that's arrived at the sensor, or any other particular loco), or to a group of locos, or to all locos.

[96RAF]

It all seems fairly simple ...

...so why can't RM /LD do this yet:

11trains under control...