Digitrax's SE8C Signaling -

[inobu]

Signaling is extremely difficult in the beginning. Trying to shift through the documents in order to get the a basic understanding is tasking in itself. I must admit the Digitrax product line is convoluted but it has to be in order to product the features we want in a layout. 

 

I had time to gather my thoughts into one document that may get you going. Even that was daunting. There most likely are typos and such but I wanted to get the foundation of the document completed. Hopefully I have time to clean it up.

 

It works in the newer (8 and above) Adobe readers as it is interactive to a degree where you mouse over the spreadsheet area. It takes the SE8C worksheet and puts the pieces together.

 

Hopes it help clear the mud.

 

Inobu

 

I compressed it.SE8C_D_DOC_1.pdf

 

Older Adobe reader will have a stacked mess

Edited March 13, 2016 by inobu

[inobu]

Second Part JMRI                                      - not sure if the PDF is working out -

 

 

The SE8C requires a driver of of some sort. Like a PC program to operate the signal logic. JMRI to be specific.

 

Here is the correlation between the Digitrax components and JMRI. The system is based on look up tables to setup the hardware and software. Here is an image that links the two.

 

There are 3 tables that represents the three components on the diagram.

• SE8C -------------  Signal Head
• BDL168   --------   Sensor or Blocks
• Switch   ----------   Switch/Turnout

 

You can make the correlation by matching the numbers to the tables. The individual LED's controls are located with the switches but the combination of the LED creates the signal head. Notice the sensor reflects states. These states are the drivers of the logic. In JMRI you can activate the sensors that will trigger responses via the signal heads. 

For example: If block LS161 was occupied its state would turn active. That active state would trigger the logic to issue the command sw266 t or thrown. The SE8C would then activate the 266 LED which is the yellow LED. Included in that same logic is command sw267 t. Because 267 controls the block (LS162) before LS161 it must stop the train in that block. So the red LED is activated by 267 in the thrown state and the train should stop in that block.

 

Referencing the controlling aspect sheet. In so many words this is a command reference sheet as well.

Note 267 thrown is the red LED

         266.T is a Yellow, Depending on the railroad Aspect the command 266 c would have made the yellow flash.

 

 

 

 Inobu

Edited November 22, 2016 by inobu

[inobu]

Putting Another Piece Together

 

This post is stepping closer to the signaling logic in JMRI. It is logic that drives signaling. The image below gives a visual representation of what need to occur in the logic definitions.

 

 

Before we can build the logic we need to understand the signals objective.

 

First off the reality concerning trains, As soon as a train starts rolling its potential to become a run away increases up drastically. The faster the train travels the higher the possibility. This holds true as some trains can take up to a mile to stop. With this in mine one can assume that the train engineer hopes that nothing is in the way more so planning on how to stop the train. So with this in mind we can understand the objective. Signals warns train of pending obstruction on the track as it needs to have enough distance or time to stop. 

 

Before we start take note that I used different addresses. I chose DRV 6-8 is change the numbering around. This help to better understand the address numbering and how to better manage the addresses.

 

So in the example you can see that block 2 has an issue. The logic fora occupied Block 2 triggers a yellow "prepare to stop" on HD8 in block 4 and a Stop on HD6 block 3 and HD2 block 2. In the case of block 1 it will have a red stop on HD1 in Block 1 on the east bound side but a green of the west bound as that block is after the trouble. Also the block before block 1 on the west side will have a yellow as it is a two block warning system.

 

As you look closer at the diagram you can see that the signal heads are label HD1-10. The tables reflect the lookup between the turnout table and the Signal Head table. Once again the switch commands are ties to the turnouts switch values. Managing these address is important as you don't want to get the two addresses confused.

 

Now the Logic

 

The logic is the coding within JMRI that looks up values and executes the actions created by you.

 

A simplified version of the logic statement is. 

1. If block 2 is active (occupied)
2. initiate yellow light on HD8
3. initiate red light on HD6
4. initiate red light on HD4 
5. initiate red light on HD3
6. initiate red light on HD1

You will also enter the clearing logic. Replacing the red light for green.

 

1. If block 2 is inactive (Unoccupied)
2. initiate yellow light dark HD8
3. initiate green light on HD6
4. initiate green light on HD4 
5. initiate green light on HD3
6. initiate green light on HD1

The table are match with the diagram as you can see the correlation between the tables and components.

 

For example HD1 in block 1 is red. The signal head table state is red. The signal head is made up of the two switch addresses 297 and 298 from the Turnout table.

In the Turnout table it shows the command is thrown for 297 and closed for 298. Cross reference the spreadsheet it shows what command 297 t and 298 c represents. 

 

Components

 

SE8C is the Turnout table

 

 

. Block detection is the Sensor table

 

BD4 or BDL168

 

 

 Signal Head Table is the mast.

   

 

Hopefully this is helping to understand the signaling principles and the link between the electronics and software control in JMRI.

 

 

Inobu

 

Understand that this is not a tutorial but an overview that gives you insight of what a tutorial is doing. 

Edited March 17, 2016 by inobu

[inobu]

JMRI Logic

 

Here we will take a look at the logic tables and what happens. Once you have establish your Layout in Layout Editor and you can manually operate the blocks and signals via the table buttons. Next you must create the logic that drives the states and conditions that you want to occur.

 

In order to avoid confusion I left out something concerning the sensors. If you refer back in the last drawing you will find that there are types of sensors. One Designated as LS and IS. LS is a sensor that is derive from the Digitrax equipment. This is a fixed header type. The IS is internal sensor. This is a sensor created by you. This way you can chain or link multiple sensors and actions off of the physical activity from the layout. 

 

In the example here. The Logix table is triggered off by the active state of Dector 801 or Block 801. Because the states is changed to true JMRI will look up the Logix system name IX801. Within that look up it has 2 conditions or system names IX801C1 and C2 Occupied or Unoccupied. These labels are created when you add the conditions within JMRI.

 

In this case the LS801 state is Occupied so Block 801 is true so JMRI will look in the conditional user name "Detector 801 Occupied". This lookup will send JMRI to the consequential action which changes IS801 to active which activates the LED on Panel Editor and plays a bell wav file.

 

If the block was Unoccupied or should I say when the state changed to unoccupied in block 801 then the blue arrows would hold true and JMRI would that the actions of the blue arrow.

 

 

 

It is the logic in this table that drives the conditions and instruction. 

 

Here is where you will have tons of fun.

 

Adding a transponder and transponding to JMRI. You could add the WAV file of a Japanese station announcer stating the arrival of a train when it activated the block in front of the station block. Also stating the departure wave file upon another command.

 

That is what I'm working on next. I'm not sure if having JMRI play the sounds or program them into a decoder. I think having the ability to have JMRI play the wav or sound file to different channels would give you one central control point and announce all over the layout. 

 

I hope this gives you a general idea of what is going on in the world of JMRI. I know it was difficult for me as there was not a simple overview of what was what. That caused be to run back and forth picking up bits and pieces still wondering.

 

In any case hope it helps out.  

 

Inobu

 

 

Also, note in the unoccupied instruction set. Notice the 2 second delay. I placed that delay in order to emulate the delay reaction of the LED. It gives the visual feel of sending a command to a distant device and the time taken for the relays to trip and the board to update itself. 

Edited March 23, 2016 by inobu

[Ochanomizu]

Hello,

 

I use Digitrax and transponding, but I have not yet implemented signalling.  It is something I hope to do in coming years.  Thank you for this insightful introduction to the topic.  I shall bookmark it for reference for when the time comes.

[inobu]

Hello,

 

I use Digitrax and transponding, but I have not yet implemented signalling.  It is something I hope to do in coming years.  Thank you for this insightful introduction to the topic.  I shall bookmark it for reference for when the time comes.

Thanks,

 

Yes, do it in phases and you will get a lot of enjoyment out of it.

 

Inobu

[gavino200]

This is definitely just over the horizon for me. The question for me is to transpond or not transpond. What does transponding add that can't be provided by other mechanisms? Doesn't using transponding mean your are limited to only using Digitrax decoders? Is that worth it.

 

Also, I like the idea of having working LED signals changing colors around the layout. What are some of the things that you have them do? Is it just that as train moves around the layout it's followed by a trail of red lights that would serve as a warning for a following train to slow down? With automation the trains would get a red signal to stop at a station. I'm guessing you can have a station platform show a red signal when occupied, and that crossings will close and turn red as a train approaches. What are some other cool things you can do with this setup?

 

Where do you get your station announcement wav files? Is there an online library? Or do you record them yourself?

 

On a (slightly) related note, do you use the DS64 for turnout control? I watched this YouTube video about it. Apparently there is a problem whereby it sort of goes bonkers while the loconet system boots up. This guy plans on solving the problem by adding a small delay circuit. Have you had this issue? How did you solve it?

 

Edited February 7, 2017 by gavino200

[kvp]

Yes, all of the above are possible with signals. I prefer manual exit and entry signalling from a control panel. Block signals and crossings can be automated easily.

 

Transponding is good if you want to know which train is in which block without inputting the starting positions manually into the computer. Yes, it's mostly a digitrax only feature but kato decoders are also made by digitrax. I think that without computer control it's not too useful.

[inobu]

Gavino,

 

Now here is the payoff of wiring this way. You just have to disconnect the BDL168 and connect the 

rx-4 harness. The PM42 would be to the right as the board is labeled. 

 

The RX-4 handles transponding.

 

The RX-4 are like "wire taps" that listen to the information passed on the wires. Depending on where you want to listen is where you wire

the "taps".  The example here is up stream which gives general zone identification A B C D. They work in groups of 4.

 

 

lets say you wanted more detailed reporting then you would place the RX-4 on the down stream side or the BDL168's output.

 

There are some limitations. The BDL168 can support 2 RX-4's which means it can report 8 segments.

 

So if yard 1 was connected to Zone D you could run all 4 wires in one RX-4 for reporting. 

You could then use the other 7 RX-4 on single wires or however you wanted.

 

 

In this diagram you can see the my example is upstream and down stream would be wires below the BDL168. 

 

Inobu

Edited February 7, 2017 by inobu

[kvp]

Apparently there is a problem whereby it sort of goes bonkers while the loconet system boots up. This guy plans on solving the problem by adding a small delay circuit. Have you had this issue? How did you solve it?

Apparently it's a RTFM situation:

The manual: http://www.digitrax.com/tsd/KB804/ds64-set-up-using-option-switches-opsw/

 

 

 

OpSw 06 determines whether the DS64’s outputs power up automatically

 

Thrown (factory default) at power on - all DS64s and their 4 outputs power up automatically to their last state.

 

Closed at power on - the DS64 powers up and the 4 outputs do not power up until they receive a command.

Solution = set OpSw06 to closed (active)

[inobu]

Apparently it's a RTFM situation:

The manual: http://www.digitrax.com/tsd/KB804/ds64-set-up-using-option-switches-opsw/

 

Solution = set OpSw06 to closed (active)

kvp,

 

Why are you commenting on a problem you do not know about? The issue is in the way the command station starts up. When a layout is powered up the DS64 starts its boot sequence. Before it gets a chance to complete the boot sequence the command station will establish control over loconet causing it to stop its reset action. At that point some DS64 will not complete its reset and some switches can be thrown either way. This is the very issue that changes some decoders if left in programing mode on power up. It has nothing to do with reading the manual and all about having, using and understanding the product.

 

Inobu

 

 

. 

[kvp]

Why are you commenting on a problem you do not know about? The issue is in the way the command station starts up. When a layout is powered up the DS64 starts its boot sequence. Before it gets a chance to complete the boot sequence the command station will establish control over loconet causing it to stop its reset action. At that point some DS64 will not complete its reset and some switches can be thrown either way. This is the very issue that changes some decoders if left in programing mode on power up. It has nothing to do with reading the manual and all about having, using and understanding the product.. 

The DS64 is not responding to loconet commands until it's operational. If it's set to only power up the outputs when it receives its first valid command packet, then it will have the proper shutdown state in memory before it starts to listen to commands but only outputs power to the turnouts when it has its first command from the fully operational command station. This requires the DS64 to be set to resume operation where it left off, instead of resetting the turnouts to their default position as this reset could get aborted if the first command arrives too early. This is what the video was about (turnouts moving before full command station bootup) and apparently this seems to be the right solution for it. If you don't belive me, just contact Digitrax and ask them. No thinkering with the power should be necessary.

 

ps: Also if any turnout moves more than once during startup and not to its default or saved position while the command station boots up, then the physical key input is probably enabled on the DS64 and left floating.

[inobu]

kvp, check you pm.

 

Inobu

[gavino200]

Gavino,

 

Now here is the payoff of wiring this way. You just have to disconnect the BDL168 and connect the 

rx-4 harness. The PM42 would be to the right as the board is labeled. 

 

The RX-4 handles transponding.

 

The RX-4 are like "wire taps" that listen to the information passed on the wires. Depending on where you want to listen is where you wire

the "taps".  The example here is up stream which gives general zone identification A B C D. They work in groups of 4.

 

 

 

lets say you wanted more detailed reporting then you would place the RX-4 on the down stream side or the BDL168's output.

 

There are some limitations. The BDL168 can support 2 RX-4's which means it can report 8 segments.

 

So if yard 1 was connected to Zone D you could run all 4 wires in one RX-4 for reporting. 

You could then use the other 7 RX-4 on single wires or however you wanted.

 

 

 

In this diagram you can see the my example is upstream and down stream would be wires below the BDL168. 

 

Inobu

 

Thanks Inobu. That was very well explained. It makes perfect sense.

 

PS where do you get the wav files? BTW I'm going to Tokyo in a few weeks. Are there any sounds you need? I'd be happy to make some recordings.

Edited February 8, 2017 by gavino200

[gavino200]

It sounds like both of you guys have found a different workaround/solution for the problem. As always thanks for your input and knowledge.

 

Gavin

[inobu]

Thanks Inobu. That was very well explained. It makes perfect sense.

 

PS where do you get the wav files? BTW I'm going to Tokyo in a few weeks. Are there any sounds you need? I'd be happy to make some recordings.

I get the waves from you tube and use Adobe Audition to edit the wav. file. At one point I was trying to put them in sound decoders but found it was a waste of time and a decoder. Just let JMRI play the file and place the speakers where you want.

 

After toying with the speakers I found this PAM8403.

 

This is a 3watt amp. So you can get decent tiny speaker in place and drive the layout with surround sound. I haven't tested it yet as I have been busy,

I was working on this a few months ago. I took a month for the amps to get in so I had to shelf the project.  Its on the project bin as we speak.

 

I paid a whopping $10 for 20 of them including shipping.

 

 

It works on 5 volts drives 3 watts which is enough for us.

 

You can then build the wav files to channels which is mapped to the speakers.

 

Inobu

[inobu]

It sounds like both of you guys have found a different workaround/solution for the problem. As always thanks for your input and knowledge.

 

Gavin

No, the issue is still there even with those switches activated.

 

Inobu

[kvp]

 

 

No, the issue is still there even with those switches activated.

Have you checked with a board that has up to date software and the right opswitch set? (also, connecting the unused mechanical switch inputs to a fixed state may help eliminate the spurious throw events)