mirror of
https://github.com/DCC-EX/CommandStation-EX.git
synced 2024-11-26 17:46:14 +01:00
72 lines
4.1 KiB
Markdown
72 lines
4.1 KiB
Markdown
Railcom implementation notes, Chris Harlow Oct 2024
|
|
|
|
Railcom support is in 3 parts
|
|
1. Generation of the DCC waveform with a Railcom cutout.
|
|
2. Accessing the railcom feedback from a loco using hardware detectors
|
|
3. Utilising the feedback to do something useful.
|
|
|
|
DCC Waveform Railcom cutout depends on using suitable motor shields (EX8874 primarily) as the standard Arduino shield is not suitable. (Too high resistance during cutout)
|
|
The choice of track management also depends on wiring all the MAIN tracks to use the same signal and brake pins. This allows separate track power management but prevents switching a single track from MAIN to PROG or DC...
|
|
Some CPUs require very specific choice of brake pins etc to match their internal timer register architecture.
|
|
|
|
- MEGA.. The default shield setting for an EX8874 is suitable for Railcom on Channel A (MAIN)
|
|
- ESP32 .. not yet supported.
|
|
- Nucleo ... TBA
|
|
|
|
Enabling the Railcom Cutout requires a `<C RAILCOM ON>` command. This can be added to myAutomation using `PARSE("<C RAILCOM ON>")`
|
|
Code to calculate the cutout position and provide synchronization for the sampling is in `DCCWaveform.cpp` (not ESP32)
|
|
and in general a global search for "railcom" will show all code changes that have been made to support this.
|
|
|
|
Code to actually implement the timing of the cutout is hihjly cpu dependent and can be found in gthe various implementations of `DCCTimer.h`. At this time only `DCCTimerAVR.cpp`has implemented this.
|
|
|
|
|
|
Reading Railcom data:
|
|
A new HAL handler (`IO_I2CRailcom.h`)has been added to process input from a 2-block railcom reader (Refer Henk) which operates as a 2 channel UART accessible over I2C. The reader(s) sit between the CS and the track and collect railcom data from locos during the cutout.
|
|
After the cutout the HAL driver reads the UARTs over I2C and passes the raw data to the CS logic (`Railcom.cpp`)for analysis.
|
|
|
|
Each 2-block reader is described in myAutomation like `HAL(I2CRailcom,10000,2,0x48)` which will assign 2 channels on i2c address 0x48 with vpin numbers 10000 and 10001. If you only use the first channel in the reader, just asign one pin instead of two.
|
|
(Implementation notes.. potentially other readers are possible with suitable HAL drivers. There are however several touch-points with the code DCC Waveform code which helps the HAL driver to understand when the data is safe to sample, and how to interpret responses when the sender is unknown. )
|
|
|
|
Making use of Railcom data
|
|
|
|
Exrail has two additional event handlers which can capture locos entering and exiting blocks. These handlers are started with the loco information already set, so for example:
|
|
```
|
|
ONBLOCKENTER(10000)
|
|
// a loco has entered block 10000
|
|
FON(0) // turn the light on
|
|
FON(1) // make a lot of noise
|
|
SPEED(20) // slow down
|
|
DONE
|
|
|
|
ONBLOCKEXIT(10000)
|
|
// a loco has left block 10000
|
|
FOFF(0) // turn the light off
|
|
FOFF(1) // stop the noise
|
|
SPEED(50) // speed up again
|
|
DONE
|
|
```
|
|
|
|
Note that the Railcom interpretation code is capable of detecting multiple locos in the same block at the same time and will create separate exrail tasks for each one.
|
|
There is however one minor loophole in the block exit logic...
|
|
If THREE or more locos are in the same block and ONE of them leaves, then ONBLOCKEXIT will not fire until
|
|
EITHER - The leaving loco enters another railcom block
|
|
OR - only ONE loco remains in the block just left.
|
|
|
|
To further support block management in railcom, two additional serial commands are available
|
|
|
|
`<K block loco >` to simulate a loco entering a block, and trigger any ONBLOCKENTER
|
|
`<k block loco >` to simulate a loco leaving a block, and trigger and ONBLOCKEXIT
|
|
|
|
|
|
Reading CV values on MAIN.
|
|
|
|
Railcom allows for the facility to read loco cv values while on the main track. This is considerably faster than PROG track access but depends on the loco being in a Railcom monitored block.
|
|
|
|
To read from prog Track we use `<R cv>` response is `<r value>`
|
|
|
|
To read from main track use `<r loco cv>`
|
|
response is `<r loco cv value>`
|
|
|
|
|
|
|