For organizations that rely on machinery for their day-to-day operations, an outage can be costly. Organizations with expensive machinery who run on thin margins will often optimize their costs by running at or close to capacity, or by minimizing redundancy between machines of a similar type. When a machine breaks, begins operating out-of-phase, or goes offline, the manufacturer loses money. Monitoring these machines for outages or out-of-phase behavior is preferred, but many of these machines do not come with built-in remote monitoring systems, nor do they contain logic-controllers that may enable a retrofit to add connectivity and remote monitoring and control.
This app provides a simple to construct power monitoring device that can be placed in-line with any AC-based machine. Using off-the-shelf hardware and modular components, this solution can be used to add monitoring to a machine regardless of its age or on-board monitoring capability. The hardware and firmware monitor AC energy data (RMS Current, RMS Voltage, Power Factor, Line Frequency, Active/Reactive/Apparent Power) and send regular readings, changes, and configurable alerts to the Notecard and on to Notehub.
For the Dr. Wattson energy monitor build you will also need
Male-to-female grounded extension cable or suitable cables to wire an IEC or NEMA AC inlet and outlet to Dr. Wattson. 16 gauge is recommended as the minimum dimension. Please select suitable wiring gauge for the maximum load used by your machine.
(optional but recommended) Corded female NEMA socket for USB power, 18-gauge diameter minimum, such as a spliced male-to-female 18-gauge extension cable
2 18-gauge color-coded insulated wires to power the Dr. Wattson board. (This can be taken from the extension cable used to build the USB power outlet)
Soldering iron (to melt and bridge solder jumpers on the Dr. Wattson board for I2C address configuration)
Wire cutter and stripper
Spade connectors or solder
Crimper to crimp AC wires to spade connectors (if used)
Heat shrink tubing and a source of heat, such as a heat gun
The Dr. Wattson energy monitoring board monitors power by being looped into the mains AC wiring that powers the machine being monitored. Additionally, solder jumpers are configured to select the I2C address of the Dr. Wattson board.
This solution can be used to monitor a single machine, or monitor multiple machines at a facility. When monitoring multiple machines at a facility, it can be useful to group the monitors at a facility into a Fleet. For more details, see The Fleet Administrator's Guide.
The application firmware found under the firmware folder can be built using these development environments:
PlatformIO extension for Visual Studio Code
Arduino extension for Visual Studio Code
We recommend using one of the VS Code extensions, since they are easier to set up and use, and provide a comprehensive development experience. However, if you're familiar with the Arduino IDE, that can be used as well but requires a little more setup.
There is no special setup required for the project beyond what is normally required to configure a PlatformIO project in VSCode.
This tutorial explains how to install and use the PlatformIO.
The PlatformIO project is located in the firmware folder, where you'll find platformio.ini that configures the project, including libraries required, location of the sources and compile-time definitions required.
The source code for the Arduino project is under firmware/notepower/ in this repository. We have included the correct configuration in .vscode/arduino.json which selects the Swan board as the build target and configures the required compiler options.
Before building the project, you will need to install the required libraries listed below.
Before compiling and uploading the sketch, be sure to install the STM32Duino board support package. The tutorial Using the Arduino IDE in the Swan documentation shows how to install support for Swan in Arduino IDE and how to compile and upload firmware.
You will also need to install the required libraries, and increase the serial receive buffer size, detailed below.
The Arduino framework by default provides a very small Serial input buffer, which means that if a developer wishes to use the serial port in a way that receives a large volume of data quickly, the data will be truncated and missed.
The workaround, which is required by this sketch, is to add a compiler flag that increases the serial buffer size.
Close the Arduino IDE if it is currently open.
Find the location of the platform.txt file for the board that you are building for. When building for Swan, which is an STM32 board supported by STM32Duino, this is located at
Create a file in that directory called platform.local.txt containing this line:
There are two ways to configure the ProductUID created in the Notehub setup above - either using the in-browser terminal to send a request to the Notecard, or by editing the firmware source code. For more details on what the ProductUID is and how it set it please see this guide.
With the Dr. Wattson board looped into the flow of power, the machine to be monitored is connected to the Load outlet and power is supplied to Dr. Wattson and the machine via the Line inlet.
During development and testing, you will typically power the Notecarrier and Swan via USB cables from your computer. When the application is deployed, you can use a USB power adapter plugged into the 18-gauge wired outlet.
To ensure the setup is working as expected, it's a good idea to test the application before deploying it in a real-life setting. For this, you will need some kind of load, ideally with variable power. In our testing, we used a desk lamp with OFF, LOW and HIGH settings.
The app is configured using a number of environment variables. Configuration includes how often regular power monitoring events are sent, and the thresholds for anomalous behavior that trigger an alert.
Alerts are generated when the current, voltage, or power use is outside the configured range or if a change greater than a given percent is detected.
These are the environment variables that you can configure according your use case:
heartbeat_mins: how many minutes between sending power notifications. The default is0 which means do not send regular power monitoring events, only send alerts. Sending a heartbeat event periodically provides regular monitoring of the source supply and equipment supply.
alert_under_voltage, alert_over_voltage: send an alert when the measured voltage is above or below the specified values in Volts. The default setting is 0 where no alerts are sent regardless of the measured voltage.
alert_change_voltage_percent: send an alert when the voltage changes by more than the given percent. The default value is 15, which sends an alert when a 15% or greater change is detected. Set to 0 to disable percentage change alerts.
alert_under_current_amps, alert_over_current_amps: send an alert when the measured current is above or below the specified values in Amps. The default setting is 0 where no alerts are sent regardless of the measured current.
alert_change_current_percent: send an alert when the measured current changes by more than the given percent. The default value is 15, which sends an alert when a 15% or greater change is detected. Set to 0 to disable percentage change alerts.
alert_under_power_watts, alert_over_power_watts: send an alert when the measured power is above or below the specified values in Watts. The default setting is 0 where no alerts are sent regardless of the measured power.
alert_change_power_percent: send an alert when the measured power changes by more than the given percent. The default value is 15, which sends an alert when a 15% or greater change is detected. Set to 0 to disable percentage change alerts.
These environment variables are set in Notehub, either per-device, per-fleet or per-project. For example, if you want all machines to send power monitoring events every 5 minutes, you would set heartbeat_mins to 5 at the project level in Notehub.
A power monitoring event is sent every heartbeat_mins minutes or when an alert is sent.
Events are sent to the notefile power.qo, and have this structure in the event body:
"current": 0.2846, // Line RMS current (A)"frequency": 59.8125, // Line AC frequency (Hz)"power": 7.9, // Line Power (Watts)"voltage": 118.6, // Line RMS voltage (V)}
Regular monitoring events are not immediately synched to Notehub, but are sent once per hour, as given by the outbound property in the hub.set request. You can change this behavior by setting the preprocessor symbol SYNC_POWER_MONITORING_NOTES in app.h.
The event body also includes these fields:
reactivePower: The measured reactive power (in VAR).
apparentPower: The measured apparent power (in VA).
powerFactor: The power factor - active power divided by apparent power.
When the device detects over or under voltage, current or power, or detects a change in these greater than the configured percentage, a power monitoring event is sent as above, with an additional property alert that lists the comma-separated reason(s) for the alert. Depending upon the cause of the alert, you may see one or more of these values present:
undervoltage, overvoltage: the measured RMS voltage is not within the bounds set by the environment variables alert_under_voltage and alert_over_voltage.
voltage: the measured voltage changed by more than the percent specified in the environment variable alert_change_voltage_percent.
undercurrent, overcurrent: the measured RMS current is not within the bounds set by the environment variables alert_under_current_amps, alert_over_current_amps.
current: the measured current changed by more than the percent specified in the environment variable alert_change_current_percent.
underpower, overpower: the measured active power is not within the bounds set by the environment variables alert_under_power_watts and alert_over_power_watts.
power: the measured active power changed by more than the percent specified in the environment variable alert_change_power_percent.
When an alert is triggered, it is immediately synched to Notehub.
Now that we have power monitoring events and alerts available, routing them from Notehub is our next step.
Notehub supports forwarding data to a wide range of API endpoints by using the Route feature. This can be used to forward your power monitoring data to external dashboards and alerts to a realtime notification service. Here, we will use Twilio SMS API to send a notification of an alert to a phone number.
Fill out the required fields for the Twilio SMS route, including "from" and "to" phone numbers, where "from" is your virtual Twilio number, and "to" is the number of the phone that receives the alerts. We will not be using placeholders for these numbers, but will use a placeholder for the message, so set the message field to [.body.customMessage].
Under "Notefiles", choose "Selected Notefiles", and check power.qo from "Include These Notefiles".
Under "Data", select "JSONata Expression" and copy and paste the contents of jsonata/route.jsonata into the text field "Insert your JSONata expression here".
Power alert from machine-12: overcurrent,power. 120V, 25A, 3000W.
These are the parts of the message:
The first part of the message indicates which monitored machine the alert pertains to by its serial number, here "machine-12".
The alerts are next. Here, "overcurrent" and "power" alerts indicate that the measured current was higher than the alert_over_current_amps environment variable, and that power changed by more than alert_change_power_percent.
Finally, we have the power information, showing the measured voltage, current and power at the time of the alert.