Developer Guide

Tango device server

A MidWeatherMonitoring device has been developed to publish Weather Monitoring Station data automatically to interested clients.

Device properties

The following Tango device properties should be set:

  • Host: hostname of the Weather Station Modbus server

  • Port: port number to connect to

A MidWeatherMonitoring device will start in adminMode.OFFLINE and will need to be brought online e.g.:

dp = DeviceProxy("ska-mid/weather-monitoring/s1")
dp.adminMode = 0

Device attributes

The MidWeatherMonitoring Tango device uses the ska-mid-wms-interface package which publishes data for the following sensors:

  • windSpeed

  • windDirection

  • temperature

  • pressure

  • humidity

  • rainfall

The Modbus addresses for these sensors are sourced from the same register section (for an Acromag 961EN-4006 the addresses are 9-14).

For each of the default sensors the device server will create 4 attributes:

  • Engineering value - “<sensor_name>”. It is calculated from the ADC value using the following formula:

\[\begin{split}engineering\_value = (adc\_value - adc\_min) / (adc\_max - adc\_min) \\ * (engineering\_max - engineering\_min) + engineering\_min\end{split}\]

  • ADC value read directly from the Acromag device - “<sensor_name>ADC”.

  • The raw current reading from the sensor - “<sensor_name>Current”. It is calculated from the ADC value using the following formula:

\[\begin{split}current\_value = (adc\_value - adc\_min) / (adc\_max - adc\_min) \\ * (current\_max - current\_min) + current\_min\end{split}\]

  • The input signal status of the sensor read from the Acromag device - “<sensor_name>SignalStatus” of type InputSignalStatusEnum.

There is also an extra attribute called the rainfallRate, which calculates the difference in absolute rainfall every minute (mm/min).

Here are the current default attributes and their properties for the SKA-Mid Weather Monitoring Station Rev. B ICD issued 14 Feb 2025:

Attribute

Type

Unit

Min

Max

windSpeed

float

m/s

0.0

65.0

windDirection

float

°

0.0

360.0

temperature

float

°C

-50.0

50.0

pressure

float

mbar

500.0

1100.0

humidity

float

%

0.0

100.0

rainfall

float

mm

0.0

100.0

rainfallRate

float

mm/min

0.0

3.2

*ADC

int

counts

0

20000

*Current

float

mA

4.0

20.0

*SignalStatus

Enum (int)

0

2

Helm charts

The repository contains a Helm chart for deploying MidWeatherMonitoring Tango servers. Each individual device server pod configuration has been defined in charts/ska-mid-wms/weather_station_config.yaml. By default, 5 devicer server instances are deployed to a Kubernetes cluster but is configurable in charts/ska-mid-wms/values.yaml under deviceServers.wms.station_ids - this list corresponds to the simulator instances that will be created.

The MidWeatherMonitoring pods expect Host and Port properties that are used to connect the pod to a weather station server. To connect to hardware simply provide the IP address and port number to connect to in the values file on modbus_server_hostnames and modbus_server_ports respectively. If the WMS simulator is used, the default values are set in the Helm chart and do not need to be changed. The length of these lists must match, and the order of the entries in these lists must correspond to the order of the station IDs in station_ids. In the example below, the station ID is “1”, will be used to connect to the modbus server running on “1.1.1.1:8080”

deviceServers:
  wms:
    station_ids: ["1", "2"]
    modbus_server_hostnames:["1.1.1.1", "2.2.2.2"]
    modbus_server_ports: ["8080", "9000"]

Deploying

Helm charts describe a Kubernetes configuration. To run a Helm chart locally it is recommended to use Minikube. The Helm chart is available from artefact.skao.int. To check which remote repositories Helm knows of run:

helm repo list

If https://artefact.skao.int/repository/helm-internal is not in the output, it needs to be added e.g.:

helm repo add ska-internal https://artefact.skao.int/repository/helm-internal

Helm also needs to fetch the list of available packages:

helm repo update

A chart can then be pulled locally (to specify a version use the --version option):

helm pull ska-internal/ska-mid-wms

With Minikube started the chart can be deployed to the local cluster with:

helm install ska-mid-wms ska-internal/ska-mid-wms --create-namespace -n ska-mid-wms -f charts/ska-mid-wms/values.yaml

To redeploy first delete the installation and run the install command again.

helm delete ska-mid-wms -n ska-mid-wms