Cycle control from SFA nodes (and UI) is possible only by calling pre-made LM PLC macros, which can use macro control functions to manage cycles.
For the heavy industry processes we strongly recommend to use dedicated hardware production cycle controllers and control/monitor them using EVA ICS as a supervisor only.
To configure cycle you may specify action and interval, or you may set cycle parameters one-by-one after the cycle is created.
To specify action and interval during cycle creation, use the following format:
eva -I lm cycle create <cycle_id> [action] [interval <seconds>]
Example, start unit unit:lamp/lamp1 (call start macro function) with interval 0.5 seconds.
cycle create cycle1 @start('unit:lamp/lamp1') interval 0.5
Another example, run macro macro1 every 100 milliseconds:
cycle create cycle2 macro1(1, 2, x=test) interval 0.1
Cycle ID must be unique within a Logic Manager.
Each lcycle object has the following parameters:
autostart cycle will start automatically as soon as LM PLC server is started
ict interval notify counter. Every ict iterations notification is being performed.
iterations (read-only) cycle iterations since the last start/stats reset
interval cycle interval, in seconds
macro macro, which is called as a worker
on_error macro, which is called if cycle error has been occurred
status (read-only) cycle status, changed when start/stop commands are executed. Can be:
- 0 cycle is stopped
- 1 cycle is running
- 2 cycle got “stop” command and will stop as soon as current iteration finish
When cycle is running, attempts to change parameters ict, interval or macro will return an error.
Macro, defined in on_error cycle property is called when:
- exception worker macro raised an exception, on_error macro args
- _1 “exception” word
- _2 exception object
- timeout/exec error macro execution took more time than cycle loop
interval is set to, or worker macro exited with non-zero code. on_error
macro args contain:
- _1 “timeout” or “exec_error” word
- _2 serialized worker macro execution result
the macro can e.g. stop cycle execution, send critical event to controller core or just log an error and let cycle continue.
Theoretically, cycle interval can be up to 1ms (1kHz worker frequency), but don’t expect stable cycle loops on a slow/busy hardware. In real life, software controllers handle well production loops up to 200/300ms (3-5Hz), with lower values (~100ms = 10Hz) users may expect 2-3% of iteration loss. Stable logic-rich cycles with the interval, lower than 20ms (50Hz), are nearly impossible.
If worker macro perform a calls between EVA ICS nodes, don’t forget about network timeouts.
To let cycle run with a maximum precision and avoid timeout errors, it is strongly recommended for the low-interval cycles:
- set up dedicated LM PLC instance
- turn off controller logging
- turn off controller action history (set keep_action_history to 0)
- if worker macro performs calls to Universal Controller, make sure controlled unit state isn’t replicated to other nodes in real-time via notifiers or stored in any state history databases (make dedicated group for such units and don’t subscribe notifiers to it)
- read common recommendations about using EVA ICS in high-load environments.
Cycles vs jobs
Both cycles and Scheduled jobs do similar functions: start macros with the specified interval.
When you should use cycles
- you need to have complete control and monitoring on the interval loop
- you need to periodically stop / start interval loop manually
- interval is in seconds or milliseconds
usage example: automation loops.
When jobs are better
- you need a simple instrument to run the task periodically
- you need to start a task with the specified time schedule, e.g. every day at 9:00
- interval is in seconds, minutes or hours, precision is not important
usage examples: statistic tasks, cleaners, system checkers.