Trains.com

Subscriber & Member Login

Login, or register today to interact in our online community, comment on articles, receive our newsletter, manage your account online and more!

Atlas Signal System

10628 views
41 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    July 2009
  • From: lavale, md
  • 4,678 posts
Posted by gregc on Tuesday, January 24, 2023 6:35 AM

ATLANTIC CENTRAL
In most cases, both tracks of double track would/should be signaled in both directions

the layout has some double headed signals.  for now, the signal for the left track is just wired for STOP.

the block detection is availble for both tracks from PSX circuit breakers.  switch positions need to be wired into the processor to support use of the left signal.

processor firmware can easily be changed as functionality grows over time.

1617

greg - Philadelphia & Reading / Reading

  • Member since
    January 2009
  • From: Maryland
  • 12,897 posts
Posted by ATLANTIC CENTRAL on Tuesday, January 24, 2023 10:15 AM

gregc

 
ATLANTIC CENTRAL
In most cases, both tracks of double track would/should be signaled in both directions

the layout has some double headed signals.  for now, the signal for the left track is just wired for STOP.

the block detection is availble for both tracks from PSX circuit breakers.  switch positions need to be wired into the processor to support use of the left signal.

processor firmware can easily be changed as functionality grows over time.

1617

 

I understand, but I don't get why people build stuff "half way", or temporary in the first place?

Sheldon

    

  • Member since
    January 2017
  • From: Southern Florida Gulf Coast
  • 18,255 posts
Posted by SeeYou190 on Tuesday, January 24, 2023 11:01 AM

gmpullman
If you were so inclined, you could send me a couple of those signals and I could see about converting them to LED and give you a pattern to follow for you to do the rest of them, Kevin.

I am VERY inclined!

Unfortunately they are somewhere in the "wall of boxes" that is currertly surrounding the room where my wife and I are sleeping waiting for the master bedroom to be completed.

Can you give me a few months?

Big Smile

-Kevin

Living the dream.

  • Member since
    July 2009
  • From: lavale, md
  • 4,678 posts
Posted by gregc on Tuesday, January 24, 2023 2:22 PM

ATLANTIC CENTRAL
but I don't get why people build stuff "half way", or temporary in the first place?

what do you mean.  the club layout is a work in progress (going on 17 years)

greg - Philadelphia & Reading / Reading

  • Member since
    August 2003
  • From: Collinwood, Ohio, USA
  • 16,367 posts
Posted by gmpullman on Tuesday, January 24, 2023 3:00 PM

SeeYou190
Can you give me a few months?

    At your convenience, sir. 

I recall having some "bulb"-based signals way back when and on some of them the colored dye peeled right off so all three "colors" were clear.

It should be relatively simple to replace your incandescent lamps with LEDs. Sounds like a fun project Cool

You know how to find me Big Smile

Cheers, Ed

  • Member since
    January 2009
  • From: Maryland
  • 12,897 posts
Posted by ATLANTIC CENTRAL on Tuesday, January 24, 2023 3:13 PM

gregc

 

 
ATLANTIC CENTRAL
but I don't get why people build stuff "half way", or temporary in the first place?

 

what do you mean.  the club layout is a work in progress (going on 17 years)

 

 

So, the club I once belonged to reached a point where they wanted signals. The "signal committee" designed a complete working system, built the components, installed the system - with all the desired features, in about a years time.

40 years latter that system is still working.

Every layout is a work in progress, but my construction management background makes doing things "over", or half way, really difficult - especially when other projects have not been touched yet.

No offense intended to them or anyone, I just can't work like that. 

Sheldon

    

  • Member since
    July 2009
  • From: lavale, md
  • 4,678 posts
Posted by gregc on Tuesday, January 24, 2023 6:19 PM

ATLANTIC CENTRAL
Every layout is a work in progress, but my construction management background makes doing things "over", or half way, really difficult

i don't see reprogramming a board as doing things over.  Software testing begins as soon as basic features are available.   Reprogramming requires connecting a USB cable between a laptop and the processor under the layout and downloading the new code.

i see signaling on this layout evolving in stages

  • independent nodes, driving local signals depending on PSX block detection available on the panel (club layout had 3 panels)
  • linking nodes via WiFi to share block occupancy and add nodes without local PSXs
  • moving control logic from nodes to central controller and adding additional features (switch control) with CTC

this approach makes some capabilities available early before the system is more complete as well expectations to evolve 

there just one person that understands electronics in the club

1888

greg - Philadelphia & Reading / Reading

  • Member since
    January 2009
  • From: Maryland
  • 12,897 posts
Posted by ATLANTIC CENTRAL on Tuesday, January 24, 2023 9:16 PM

gregc

 

 
ATLANTIC CENTRAL
Every layout is a work in progress, but my construction management background makes doing things "over", or half way, really difficult

 

i don't see reprogramming a board as doing things over.  Software testing begins as soon as basic features are available.   Reprogramming requires connecting a USB cable between a laptop and the processor under the layout and downloading the new code.

 

i see signaling on this layout evolving in stages

  • independent nodes, driving local signals depending on PSX block detection available on the panel (club layout had 3 panels)
  • linking nodes via WiFi to share block occupancy and add nodes without local PSXs
  • moving control logic from nodes to central controller and adding additional features (switch control) with CTC

this approach makes some capabilities available early before the system is more complete as well expectations to evolve 

there just one person that understands electronics in the club

1888

 

Greg, I respect and even admire your dedication to the club setting. I enjoyed the club, and later a round robin group, that I belonged to in the past.

I will invite everyone I know to come operate my layout when it gets to a reasonable point of operation.

But I no longer have any interest in being part of the "team" layout building experiance. I'm not interested in the compromises or politics of working on a club layout, and I'm not intersted in have any help in any serious degree on my home layout.

I don't completely understand the approach you describe above? 

I had a much longer response here - but I thought better of it.

Sheldon

    

  • Member since
    July 2009
  • From: lavale, md
  • 4,678 posts
Posted by gregc on Wednesday, January 25, 2023 8:55 AM

ATLANTIC CENTRAL
I don't completely understand the approach you describe above? 

i'm not sure when you mentioned "contruction management" if you are describing building renovation or the construction of your relay based controllers.

i certainly understand that you don't want to start a renovation without having a fairly complete understanding of what is expected.   the customer can't change there mind midway thru contruction and isn't likely to say that's good enough and stop construction before it is complete

but it's not obvious that for a new unique situation that a relay based electrical signaling circuit (see diagram) doesn't undergo changes during bench testing; that the paper design isn't flawless

at least with the last project i worked on, a femtoCell, much of it was implemented on signal processors where the software can easily be changed rather than hardware which can't be changed

the hardware for a signaling system i'm discussing is nothing but a processor, I/O expansion chips and connections to signal LEDs, block occupancy detectors and switch position indicators.   All the smarts are implemented in software.

that software can start out as nothing but indicating STOP if the block is occupied.    the next feature might be to indicate APPROACH if the "next" block detector is available.   the feature after that may also force STOP if a particular switch is open, and that might be further improved by considering routes.   And moving all the smarts to a central element doesn't require any hardware changes either.

there were many "mistakes" in the original New Haven East Interlock code at a friends layout, where you pressed 2 buttons on either end of a panel to align all the switches between them.   Even after corrections, additional changes were made after using it.

and this has been my experience as an embedded firmware developer on data terminals, speakerphones, CDMA cell phones, optical transmission systems as well as the femtoCell radios i've worked on.

The software evolves as details, corner cases, become better understood during development and testing.    There are just too many details to completely identify during the paper design phase and this is acceptable because software can easily be changed or just replaced.  software changes are tested nightly.  experienced developers have a good feel for what to worry about.

physical construction -- building renovations or electronic circuitry -- cannot be changed as quickly or easily as software and that's why the  "approaches" are different

yes, model RR electronics are getting that complicated primarily because custom software can be developed for inexpensive generic processors used in household and industrial products

2051

 

 

greg - Philadelphia & Reading / Reading

  • Member since
    January 2009
  • From: Maryland
  • 12,897 posts
Posted by ATLANTIC CENTRAL on Wednesday, January 25, 2023 9:38 PM

Greg,

How do you build a partially designed signal system? Specifically how do you know what physical signal head configurations to put at each location, or even where those locations are, if you don't already know what aspects are required?

OK, your New Haven East project is a passenger yard throat, correct?, I will guess that all the signals are dwarf signals. And I understand you were solving both route selection and signaling. And it is very complex. 

Personally I would have solved the route selection issues BEFORE deciding on signal indications. That is what I do with my relay based system.

On most prototype systems, for the greater part of the history of North American railroading, Interlocking signals (speed or route based) have consisted of multiple stacked heads, two or three heads per incoming route is typical.

Before worring about how to control them, it would seem to me knowing what the desired aspect for each condition needs to be should be the first information gathered.

Once you know that, then you can develope a simple truth table for each condition.

Then, software or hard wired logic is pretty easy to work out - but I will admit, I have not written any code in modern times. The last code I wrote was for a 1982 vintage Cutler Hammer PLC.

Now, in the case of simple block signals between interlockings, the red, green, yellow thing is pretty easy. And open non CTC turnouts put that block red, that is not a difficult wiring problem by any method.

I am not challenging the use of solid state or processor based systems here, I'm just suggesting it makes more sense to know everything you want it to do first - then design and build it.

That requires understanding the prototype system - AND, deciding if you need every feature of the prototype, AND, then making reasonable practical compromises for our model worlds.

DISCLAIMER - I know a fair amount about real trains, real train operation, real signal systems. I have no interest in modeling some very limited stretch of a real railroad down to the last inch (or complex signal aspect). 

My system is full of short cuts and compromises, and few people would know that just to observe it in operation.

On the prototype, the default aspect is RED. All signals are told to be RED by Normally Closed (no power to the coil) relay contacts until they are told otherwise - that way broken rail, damaged wiring, stuck turnouts, burnt up detection relays and a host of other failures stop or slow the trains.

And again, I will make the point that most model layouts, even pretty large ones, have very little need for the typical permissive block signal used on the prototype - the distances between our Interlockings, or Control Points, are too short. How many signals do you need in 20' or even 50'? That is the typical length of a control/signal block on my new layout. And I use a signal mid way thru each block as well, so that makes it every 10' to 25'. My average train is 15'-20' long. That is actually a pretty realistic selective compression of typical practice in my 1954 era.

I have a folder full of logic diagrams for different track configurations, with signal head configurations based on typical prototype practice from my era. 

My logic processor is my detectors, the relays that drive the turnouts, and "permission" from the Dispatcher. The turnout relays have multiple spare contacts to provide the necessary interlocking logic in both directions of travel thru each Interlocking or Control Point. You have already seen the turnout control schematic - there is a relay that repeats the postion of each turnout - some contacts are used for track power routing, some for the button logic of the turnout selection, and rest are used for the signal logic thru the interlocking. All the relays are 4PDT ice cubes.

It's not that complicated - it is a building block system. In a rare complex situation, I need a repeater or two.

Sheldon

 

    

  • Member since
    July 2009
  • From: lavale, md
  • 4,678 posts
Posted by gregc on Thursday, January 26, 2023 7:02 AM

ATLANTIC CENTRAL
How do you build a partially designed signal system? Specifically how do you know what physical signal head configurations to put at each location, or even where those locations are, if you don't already know what aspects are required?

Sounds like by "signalling system" you mean having the blocks defined on a completed layout along with knowing the type of signal (multiple heads) at each block boundary and what aspects are needed for various conditions including route and block occupancy.

The club certainly doesn't have that, like most layouts

It started after seeing that the PSX has a block occupancy output and using one to control LEDs mounted on top of a building to alert members that a train was ahead of them.   Then I was asked to make the signal in the photo work.    We then moved a 2nd signal facing the tunnel to the other side of the tunnel and made that work after rewiring some blocks.   Then we combined the outputs from an AR and 2 PSXs to signal a hidden reverse loop track and then added APPROCH because the next block was from a PSX on the same panel.

So "the plan" is evolving.  signals are being added that aid members much like i imagine signals were first implemented on the railroads.   Of course it will never be prototypical.

ATLANTIC CENTRAL
Personally I would have solved the route selection issues BEFORE deciding on signal indications. That is what I do with my relay based system.

Of course, you don't want to rewire you hardware especially after it's installed

if it's not already clear, software changes are made at home, tested there by grounding inputs simulating block occupancy outputs.   Once satisfied, it just takes a couple minutes on the layout to reprogram a node.   Of course it takes longer if new connections to signals, PSXs or switch machines are needed.

ATLANTIC CENTRAL
Before worring about how to control them, it would seem to me knowing what the desired aspect for each condition needs to be should be the first information gathered.

Once you know that, then you can develope a simple truth table for each condition.

As you know, the logic for a 3 aspect signal is trivial.   

More complicated aspects can be implemented in software by matching a set of input conditions to a particular aspect.  it can be table driven, not explicit logic and easy to correct

ATLANTIC CENTRAL
I'm just suggesting it makes more sense to know everything you want it to do first - then design and build it.

of course it does.   but how do you know it's complete and correct?   How much double checking do you want to do before building it?   I spent a career doing this.  Once all the challenging requirements are understood it's often faster to test and modify software.

2293

greg - Philadelphia & Reading / Reading

  • Member since
    January 2009
  • From: Maryland
  • 12,897 posts
Posted by ATLANTIC CENTRAL on Thursday, January 26, 2023 8:56 PM

Greg, I'm not really good at all this multi quote editing so I hope you can just follow my responses.

Do you mean there is no portion of the layout that is relatively complete to allow that area to be properly blocked and signaled? It does not have to be the whole layout.

I don't know how you build layouts, or how many you have built or been involved with, but many of us "old timers" tend to build benchwork, track, controls then scenery, in that order - and depending on layout size, break that down into moderate sized sections.

I know, and have known, lots of people with layouts that are/were "relatively" complete - that is completed trackwork, working control systems, all benchwork covered with some degree of scenery, etc.

The original control system used by Ed Ravenscroft, documented in MR, that my system is largely based on, was completely modular allowing each new section of the layout to be built, wired, signaled, operated and then intergrated into previous and following modules seamlessly.

But all that is different from just going off willy-nilly with 100 projects started and none complete.

"how do I know it's complete or correct" - not sure I understand that statement. I know its complete and correct because I know what I am modeling, I know what I want, I know how real trains work, I know what compromises I am willing to accept or not accept in my modeling - and I am not involved in "management by committee".

AND, I am using artistic license to create a believeable selectively compressed fictional model world, I am not building an inch by inch replica of a real life place, nor would I want to.

So it is complete and correct when I say it is.

I'm not downplaying the flexibility of using software, I get that. But the learning curve is too high for me after all these years being away from it. And I am not modivated to relearn new versions of it for any other purpose. Believe me, I considered a lot of options 15 or so years ago - DCC, computers, etc. I looked at everything that was on the market, and the available DIY solutions like Bruce Chubb's CMRI - THEN.

But I'm not into the planned obsolescence thing, I'm not replacing things that work for my needs just because something else has come along. 

That goes for model trains and their controls, garden tractors, automobiles, vinyl records (making a strong comeback), paper books, and of course relays.

I think everyone should enjoy this hobby however it "happens" for them.

But for me that is not some willy-nilly constantly changing series of experiments. That was the first five or six years from age 10-16. After that I pretty much knew what I wanted from this hobby.

Several small sections of benchwork are up, the ceiling is ready in 1/2 of the whole space, my workshop is 90% operational, several major sections of benchwork will be next.

And, related to this conversation, another modeler asked me if I expected to make changes to the layout as it gets built. I explained that there will surely be minor adjustments as it goes along. But that this whole concept has been in use on previous layouts, and has been carefully refined. Major changes are VERY unlikely.

That is what planning is for......

Sheldon  

    

    

Subscriber & Member Login

Login, or register today to interact in our online community, comment on articles, receive our newsletter, manage your account online and more!

Users Online

There are no community member online

Search the Community

ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT
Model Railroader Newsletter See all
Sign up for our FREE e-newsletter and get model railroad news in your inbox!