RM/LD dealing with malfunctions

[hosh]

Haven't thought too much about this recently so decided to start a thread on the subject to see what peoples thoughts are on the issue. By malfunction I mainly mean a sensor not being tripped for whatever reason.

 

It would seem that a system would be needed whereby RM would need to know what sensor a train is supposed to trigger next. That way, depending on how the train left the last sensor (from a dead stop or from whatever speed) we could measure the time the various scenarios take to get to the next sensor, add a little buffer time and then instruct all the trains to stop if this condition isn't met i.e. that the train made it to that sensor by the prescribed time. And RM could have a message like "Train T failed to arive at Sensor A" and then we inspect.

[Fishmanoz]

My first thoughts hosh - this is a NASA project.

 

May I ask that you tell us where you see we stand with current or hypothesised RM capability towards achieving this, and what extra capability would be needed to actually achieve this outcome?  And for me, to give this some structure, it needs to end up with specifics in terms of program commands, setups for sensors, points and signals, sensor instruction sets and use of arrows on track.

 

i have some ideas in my head, but currently there are lots of wholes I can see, so am interested in how you see it might happen.

 

Anyone else have ideas?

[hosh]

Firstly, an automated system without something like this is a dangerous system. Someone running $500+ locos won't want to see them smashing into things.

 

As mentioned in the "RM/LD seeing ahead/behind" thread, RM is going to need to see ahead. There will need to be some way to instruct/notify sensors/signals ahead.

 

So assuming RM can indeed handle non straight signalling, being able to communicate with sensors/signals ahead will be necessary.

 

So for a train on it's way from sensor A to B, A could issue an instruction like "Notify sensor B to expect Train T within 45 seconds". If the train doesn't arrive at B within 45 sec then RM shuts down all trains and could simply display that instruction as the error message.

 

Really NASA stuff?

[Chrissaf]

So far, I've kept well clear of contributing to this topic, in the guise of the many threads it has consumed, some now locked by admin, but.....

.

It's a TOY.......not an automated 'Docklands Light Railway'......I think you all expect too much.

[Fishmanoz]

Good comment Chris.  We'll look forward to another next year?

 

hosh, yes still NASA.  Communicating to points and signals - no problem using current sensor instructions.

 

Communicating to sensors - currently no facility to do so directly, the one and only thing sensors do is communicate detections back to RM along with the known info inherent in that (sensor and loco IDs, speed and direction).

 

So each sensor would have to have  instructions in it to say "if loco detected at A is not detected at C within S seconds, stop all locos and sound warning" and such instructions would be needed for all sensors that can be adjacent to the original sensor A.

 

I think this is clearly far too complex for a sensor setup instruction or instructions, it would have to be in a program run by the sensor.  And that program would have to contain a variable to identify which loc ID to look for at C etc, which would be the one detected at A.  And all of that is way beyond current programming capability which has been established only to run trains, set signals and points etc.

 

So doable, but NASA project to achieve it.

[hosh]

Yes, it would be a bit of initial setup work for the user. Up to them though - not necessary, just a nice option for the carful. If I had a layout full of locos that cost north of $500 each then i think the effort would be worth it.

[96RAF]

This is all getting a bit complicated for me, even as a professional engineer, but have you looked at Rocrail and JMRI, which are both open source hence the code is freely viewable, to see how they handle automatic operations and fail safe.

 

I feel you may be trying to reinvent the wheel when these 2 applications have been running successfully for ages and could be examined for clues.

[PJ_model_trains]

What do you want for Christmas in a few years time, Pro-Ho-Ho

 

Joking aside, Hornby do not jump in at the deep end, they will bring the system out, see how it goes, attend to teething problems and then listen to futher advancement comments, requests and theories.

 

My thoughts are, don't expect everything in version one, you won't get it, some ideas may seem complicated now but there is possibility to grow into things so, although hosh wants to put a space ship on the moon, one day it happened so who knows.

 

Hornby will no doubt bring out a version they can build on, first brings in revenue for work done to date, the Pro version brings in more revenue for advancements and continued support. It also gives us something to look forward to, to talk about (as if we need this), and to make sure any advancements are what is needed for users rather than what Hornby think we need. So I think we will see Pro-Ho-Ho. Time will tell.  ;o)

 

[St1ngr4y]

How many more threads are we going to see about LD. Maybe there should be a new section in the Forum devoted to LD .....

Ray

[RB51]

,Good idea Ray. Then those that wanted to could have one place to go to "read all about it" and those that didn't could aviod it. Although I am interested in the subject in a general way I am more than a little bamboozled by the various LD discussions most of which seem to be conjecture. I dip into the various LD threads once or twice a day, and then leave no further forward in terms of understanding. R-

[St1ngr4y]

,Good idea Ray. Then those that wanted to could have one place to go to "read all about it" and those that didn't could aviod it. Although I am interested in the subject in a general way I am more than a little bamboozled by the various LD discussions most of which seem to be conjecture. I dip into the various LD threads once or twice a day, and then leave no further forward in terms of understanding. R-

I'm as keen as anyone to see what Hornby will release, but the discussions about how it will work seem to be going round and round in circles. As I said months ago, I think a lot of people are going to be extremely disappointed with its capability if and when it does arrive.

Ray

[RB51]

There will always be folk who are disappointed, that is just a fact of life. I have moved on to another LD type product which although a little daunting to begin with is:-

 

1. Already out there and has been for a few years

2. Very cheap and easily expandable

3. Has a support forum - not great but it is available

4. Integrates with many DCC controllers

5. Via other software will work with the Elite

 

R-

[PJ_model_trains]

How many more threads are we going to see about LD. Maybe there should be a new section in the Forum devoted to LD .....

Ray

 

A very good suggestion Ray.  Have you contacted Adam?

[PJ_model_trains]

,Good idea Ray. Then those that wanted to could have one place to go to "read all about it" and those that didn't could aviod it. Although I am interested in the subject in a general way I am more than a little bamboozled by the various LD discussions most of which seem to be conjecture. I dip into the various LD threads once or twice a day, and then leave no further forward in terms of understanding. R-

I'm as keen as anyone to see what Hornby will release, but the discussions about how it will work seem to be going round and round in circles. As I said months ago, I think a lot of people are going to be extremely disappointed with its capability if and when it does arrive.

Ray

 

I agree fully Ray

 

I placed a message on another post earlier saying that I feel LD will be a good starter version, they will leave room for a Pro version. HRMS will, whilst dealing with any teething issues (new product teething problems will need ironing out) but they will continue to listen to our comments and suggestions and build in advancements that the end users want, that they feel are suitable. This is how they built RM I personally cannot see them working any differently with LD.

 

I agree, anyone thinking NASA type thoughts will no doubt be disappointed and have a lot more to moan about. LD, I think, will be a good entry level package with lots of room to grow. ;o)

 

[Sloop]

From what I have read I do not think it would be of much use on my small layouts, apart from stopping a possible derailment occasionally and making it a little easier to swap trains in the station.

 

It would be good for me if they could release just enough for auto- profiling locos though.

 

John b.