Skip to main content

A Selection of Tools for EPICS Monday-Morning Integrations

Project description

The EPIC Monday-Morning Integration Toolkit

pipeline status coverage report Latest Release pylint

EMMItool is the "Swiss army knife" for rapid ad-hoc integration of scientific instrumentation into EPICS-based beamline control sytems.

Concepts and Components

EMMI is separated in different layers of which either can be used for specific types of integrations, ranging from rapid "Monday morning" hacks for attaching an existing non-EPICS capable device to a beamline, to a writing full-fledged EPICS-IOC support for your device:

  • The EMMI Device Architecture (EMA) offers abstraction layers for access to motors, signals and switches

  • EMMIdaemon is a stand-alone application that runs and presents an ad-hoc IOC on one side, with a REST-like HTTPS API support on the other side

EMMI Device Architecture

Controlling Motors

EMMI uses a simple yet extensible motor interface. The philosophy is that we're considering an ideal motor in our model. We don't concern ourselves with backlash compensation, acclereation ramps etc. We care about what the motor is supposed to do, as opposed to how it is going to do it.

For this, we consider the following properties:

  • position: R/W property which returns the current motor values, respectively moves to the specified absolute value when written to

  • position_relative: R/W property that facilitates a movement relative to the current position when written to; always returns 0.

  • state: a R/W property that indicates the current state of the (abstract) motor device; can be explicitly set within specific parameters to advance to specific states.

The state diagram of a motor looks like this in its most simple form:

          ┌──────────┐
          │   INIT   ├┄┄┄┄┄┄┄┄┄┄┄┄►┄┄┐
          └────┬─────┘               ┊
               ▼                     ┊
          ┌────┴─────┐               ┊
          │          ├┄┄┄┄┄┄┄┄┄┄┄┄►┄┄┤
  ┌───►───│   IDLE   ├┄┄┄┄┄►┄┄┐      ┊
  │       └────┬─────┘        ┊      ┊
  │            ▼              ┊      ┊
  │   ┌┈┈┈┈┈┈┈┈┴┈┈┈┈┈┈┈┈┈┐    ┊      ┊
  │   ┊                  ├┄┄┄┄┄┄┄┄►┄┄┤
  │   ┊     BUSY...      ├    ┤      ┊
  │   ┊                  ┊    ┊      ┊
  │   └┈┈┈┈┈┈┈┈┬┈┈┈┈┈┈┈┈┈┘    ┊      ┊
  │            ▼              ┊      ┊
  │       ┌────┴─────┐        ┊      ┊
  ├───◄───┤          ├┄┄┄┄┄◄┄┄┘      ┊
  │       │   STOP   ├┄┄┄┄┄┄┄┄┄┄┄┄►┄┄┤
  │       └────┬─────┘               ┊
  │            ▼                     ▼
  │       ┌────┴─────┐           ┌───┴────┐
  └───◄───┤  ERROR   ├┄┄┄┄┄┄┄┄►┄┄┤  FAIL  │
          └──────────┘           └────────┘

Which translates to:

  • INIT is the initial states after startup, device is undergoing custom configuration and is not ready yet.

  • IDLE device is ready to perform according to commands

  • BUSY is the state in which device is performing, most likely moving

  • STOP is the state in which the device is decelerating with the intention of coming to a standstill. This can be part of a regular IDLE-BUSY cycle (i.e. returning to IDLE once standstill is reached), or can be an intermediate state towards an ERROR state, ensuring that the device is stopped for handling of errors.

  • ERROR is a well-defined state which represents the device not peforming, but which is still part of the "well defined" behavior of the device. Such a state, for instance, is reaching hardware limits or impossibility to execute a command (e.g. because coordinates are outside of allowed range). The device is always in a standstill when in ERROR, which is ensured by the fact that ERROR is only entered through STOP. ERROR can be entered from all "regular" operational states (IDLE, BUSY or STOP), but not from INIT -- initialisation errors result in FAIL.

  • FAIL is the state of a fatal error, incompatible with "defined behavior" of the device. It is a terminal state, meaning that there is no system-supported from this state. A complete reinitialisation, typically encompanied by a power cycle or hardware reset is the action to be performed to advance from FAIL. It can be entered from any other state.

The BUSY state deserves extra explanation, as it's the main mechanism for extending the functionality of a motor.

In its most simple form, a motor has only one task: to move an (abstract) axis to a specified value. However, slightly more complex real-world applications may differentiate more strictly on the type of movement to be performed:

  • a slewing movement is performed autonomously with maximum speed within parameters, towards a specific target;

  • a jogging movement, e.g. triggered by a joystick, is performed with a predefined speed as long as a condition (button press) actively persists;

  • a tracking movement is a movement that is bound to time constraints, e.g. hitting specific coordinates at specific times;

  • a homing movement may be used to define a slewing towards a hard-coded parking position;

  • a dialing or tweaking movement may be a manual correction on top of a predefined tracking path, etc.

Even more complex moves require several stacked types of movements (e.g. a tracking requires a slewing into position first, and accept tweaking input while performing the actual tracking).

As far as EMMI is concerned, we don't care about the complexity of the movement itself, we only care about representing high-level states of operation at an EPICS interface level -- roughly speaking, to us, the motor "does" or "doesn't do" anything. To model this, we allow splitting the BUSY state into sub-states, hiearchically denoted (e.g. BUSY.SLEW or BUSY.TRACK...). The restriction is that they all must either end in STOP, before returning to IDLE or entering ERROR, or must definitively fail directly into FAIL.

These supplementary states my be controlled by variables and properties which EMMI will happily manage and pass through to the EPICS interface, but will not understand or touch -- e.g. speed limits, accelerations, homing coordinates etc.

However, as it is customary with EPICS, EMMI will manage designated boolean PVs to trigger these states and indicate the successful performance of the action. This allows to a certain degree easy implementation of the "HOMF/D", "TWF/D" or "JOG" class of commands of an EPICS motor record.

This results in a 4-layer architecture that leads from the hardware controls to the EPICS variables:

  • The Axis Control is the layer (within Python) which directly serves the hardware interface API, e.g. typically a lass wrapped around a pySerial interface.

  • The Engine is a layer which enforces an API compatible the state diagram above, with the properties position, position_relative and state as described.

  • The Connector is a translator between the motor engine and a generic EPICS IOC generator, e.g. as provided by pythonSoftIOC.

  • The IOC Generator is a library that does the actual EPICS work.

In the spirit of "integration", we acknowledge that typically the first layer has already been written, and there is legitimate concern to reuse it. The only restriction EMMI imposes is for the Axis Control to not block. For the last layer, EMMI makes heavy use of pythonSoftIOC.

What remains is the Engine and Connector, which EMMI implements in the classes MotorEngine and MotorConnector within emmi.eda.

Typically, MotorConnector needs to specifically dock to user-supplied Axis Control code. There are three ways to do this:

  1. Write from scratch or re-write your Axis Control to be Engine compatible. This is, of course, the preferred method, but not always available.

  2. Use the supplied MotorConnector, which is a highly configurable template that makes heavy use of Python's "duck typing" to attach to a more-or-less compatible Axis Control. This has the advantage that it doesn't require tampering with "tried and true" hardware control code of the Axis Control layer.

  3. Write your own Motor Controller from scratch, paying attention to reflect the Engine API, as described above.

As a side note, this architecture also gives a natural layer at which to attach useful, yet hardware-independent, unit testing: by replacing the MotorConnector with a mock-up class that behaves as it's supposed to, all the layers between there and the EPICS interface can be tested in a suitable, automated CI/CD environment.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

emmi-0.1.3.tar.gz (56.8 kB view details)

Uploaded Source

Built Distribution

emmi-0.1.3-py3-none-any.whl (48.8 kB view details)

Uploaded Python 3

File details

Details for the file emmi-0.1.3.tar.gz.

File metadata

  • Download URL: emmi-0.1.3.tar.gz
  • Upload date:
  • Size: 56.8 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.11.3

File hashes

Hashes for emmi-0.1.3.tar.gz
Algorithm Hash digest
SHA256 4aae2a0913a450f98bb825a8fd4fae1642a1ca6cbd380a0fe205e7d04f8801f4
MD5 6ef0e7593b5c116dda64dcbf05b4606f
BLAKE2b-256 00b9d99df531922d6ab68c6f2bdcaa8d52d91d3d98856ba79da5c2e867a38e22

See more details on using hashes here.

File details

Details for the file emmi-0.1.3-py3-none-any.whl.

File metadata

  • Download URL: emmi-0.1.3-py3-none-any.whl
  • Upload date:
  • Size: 48.8 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.11.3

File hashes

Hashes for emmi-0.1.3-py3-none-any.whl
Algorithm Hash digest
SHA256 9506969d470c51c0dbba33410d27539fb699757141204a9f7f39e83e6147e34f
MD5 b54a94ea6d0956d1c693d6872867a187
BLAKE2b-256 0f878d633d0b1250d732f1e82d50f23d80ddc935f8bb56be5be9d9e3b8f3b7f9

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page