Starting in the summer of 2024, I collaborated with a startup providing portable power stations (PPS) for customers seeking backup power during grid blackouts. My hardware research background and interest in power systems made this opportunity a strong fit. After a series of exploratory calls and discussions, we came up with a Plan Of Action outline and established two sets of research criteria:
Set 1: Identifying protocols, external controllers, and circuitry comparisons that best fit the profile for the sort of Hardware-as-a-Service model this startup offers.
Set 2: Exploring peak-shaving solutions, cost ranges, and the possibility of combining grid + battery power without backfeeding.
Throughout this internship, I documented each step: referencing product manuals, user forums, existing test results, and support calls with manufacturers. The next sections outline my major findings in each research area.
Research Criteria, Set 1
Research Question #1
“What are the open protocols (if any) for communicating with these portable power stations? Or are they all proprietary?”
Key Findings #1: Open / Proprietary Protocols
Context: I compiled a structured table of each PPS model’s communication protocols (Bluetooth, Wi-Fi, MQTT, Modbus, etc.). Some manufacturers offered official APIs; others relied on community reverse-engineering for remote control. Below is a sample of my findings:
Limited in official app; 3rd-party MQTT bridging possible
Limited in official app; no direct API found
Some official support (Developer API / MQTT)
Limited in official app; no direct API found
In-app control only; minimal documented protocol information
Bluetooth
Yes (Version unspecified)
Yes (Version unspecified)
Yes (2.4 GHz)
Yes (5.0)
Yes (Version unspecified)
Wi-Fi
Yes (2.4 GHz)
Yes (2.4 GHz)
Yes (2.4 GHz)
Possibly (unclear)*
Yes (2.4 GHz)
MQTT
Yes (via community bridging scripts)
No official mention
Yes (via official dev platform)
Yes (but limited, user-driven)
Unknown; no official mention
Modbus
Yes (Community claims partial over BT/Wi-Fi)
Not found
Not confirmed
Not found
Not found
CAN Bus
Not found
Not found
Yes (open, in some models)
Not found
Not found
Zigbee
Not found
Not found
Not found
Possibly (some user reports)
Not found
Matter
Not officially supported; some IoT bridging mentions
Not found
Some references in “PowerStream” ecosystem
Not found
Not found
5G Adapter Compatibility
No official mention
No official mention
Not official; some user hacks
No official mention
No official mention
Comments
– Community-developed Python bridging for Bluetooth → MQTT. – Good doc on Wi-Fi but partial internal protocol details.
– App-based only; phone pairing over BT/Wi-Fi. – No open API documented.
– Developer API with possible MQTT broker integration. – Some official IoT references (PowerStream, etc.).
– Minimal official protocol details. – Some user-led attempts at Wi-Fi integration.
– Similar to Anker, minimal official doc. – Possibly interesting battery tech.
* LoRa and Zigbee were briefly evaluated communication protocols, but due to limited manufacturer support in these PPS models, I focused on Wi-Fi/Bluetooth + MQTT solutions.
Best Open Protocol: From my research, Message Queuing Telemetry Transport (MQTT) emerged as the most versatile. It’s lightweight, uses publish-subscribe architecture, and can handle both monitoring (voltage/current data) and control (turning on/off or setting power limits). In three-phase systems especially, MQTT’s Quality of Service (QoS) levels help ensure reliable data transfer, making it ideal for coordinating distributed energy resources, protective relays, and automation controllers in real-time.
Research Question #2
“If an open protocol were to be integrated, which standard would it be?”
Key Findings #2: Standard For Open Integration
Context: With manufacturers like EcoFlow providing partial developer support for MQTT, and community solutions available for Bluetti, MQTT stood out as the standard best suited for wide-scale deployment.
Why MQTT?
Scalability: Central broker can handle multiple PPS units, each publishing status and subscribing to commands.
Security: TLS support and user-defined authentication.
Transport Agnostic: Works over Wi-Fi, Ethernet, or even serial (UART) if needed.
Career Relevance: Relays and automation controllers often rely on robust communication protocols to gather real-time data and perform fast control. MQTT complements existing DNP3, IEC 61850, or Modbus solutions by providing a low-overhead method for exchanging large volumes of monitoring/control data over wide-area networks.
Research Question #3
“Is an external controller a viable option? Not a home installed box or switch, but something inline that controls and/or gates the power output.”
Key Findings #3: External ‘Inline’ Controller
Context: When direct integration with manufacturer software was unavailable, I examined inline “smart” controllers to replace or supplement built-in functions in two ways:
Monitor: Accurately measuring battery status (e.g., State of Charge, SoC) without proprietary data access.
Control: Dynamically gating charging (Grid-to-Battery) or powering loads (Battery-to-Appliance) based on SoC or current limits.
Monitor Functionality (SoC Estimation)
Problem: Because most Portable Power Stations (PPS) don’t expose raw battery leads or BMS data, obtaining a reliable State of Charge (SoC) reading from outside is complex. I examined two core estimation methods and one variant:
Coulomb Counting (DC-Side) Using external sensors to integrate DC current over time, which is the most direct approach to tracking net charge. However, it requires calibration and physical wiring. Typical sensor types include:
Shunt Resistors: Highly accurate, but must be placed inline on the battery negative terminal.
Hall-Effect Sensors: Provide magnetic-based, non-contact measurement, but require stable calibration and proximity to DC cables. Here is one interesting example.
Coulomb-Like Approximation (AC-Side with CTs)
Current Transformer (CT) Sensors: Clamp around an AC conductor (e.g., in an AC-coupled setup) to measure the power flow into or out of the battery’s inverter.
Note: This is not pure DC coulomb counting, because the inverter’s efficiency, phase shifts, and power factor can affect accuracy. Still, if you do proper calibration and account for the inverter’s conversion losses, CT data can approximate net energy usage, which may help estimate SoC in some AC-coupled configurations.
Voltage Approximation (OCV) A simpler method based on measuring open-circuit voltage, though it’s often inaccurate for LiFePO₄ cells due to their flat discharge curve and the influence of inverter regulation in normal operating conditions.
Control Functionality (Gating/Load Switching)
Grid-to-Battery: E.g., limiting input current on older 10 A outlets to avoid overload (instead of the more standard / modern 15 A at 120 V)
Battery-to-Appliance: Turning power on/off to certain devices when SoC crosses a set threshold (e.g., 40%).
Issue: Enacting these actions reliably depends on real-time SoC data, which is typically inaccessible without direct BMS integration.
Conclusion While products like Wi-Fi smart plugs, IoT gateways, or partial coulomb counting can handle basic control (limiting charge rates or toggling loads), achieving robust SoC-based gating remains difficult without proprietary BMS access. Consequently, any truly inline external controller is limited to approximate solutions unless the manufacturer provides a full read/write API.
Research Question #4
“What type of circuitry is missing from these portable power stations that would make them more inline with home backup systems like a Tesla Powerwall?”
Context: A Tesla Powerwall offers grid-tied functionality, advanced inverter control, and islanding protection with UL 1741 compliance. PPS units often lack:
Full-Fledged Hybrid Inverter: Many portable stations can’t seamlessly handle both battery discharge and grid pass-through without manually switching (***More on this later***)
Lockout/Bypass Mechanisms for safe reconnection after faults or after auto-reclosing.
Conclusion: To match Tesla Powerwall’s features, PPS would need more grid integration hardware (transfer switching, advanced inverter bridging) and read/write BMS interfaces for advanced load shifting.
Research Criteria, Set 2
As the project evolved, the startup wanted to explore peak-shaving and ways to combine grid + battery without backfeeding. We added two more questions:
Research Question #5
“Are there simpler or cost-effective alternatives to a smart panel or battery/inverter setup?”
Key Findings #5: Alternatives & Cost Range
Hybrid Inverters (e.g., Victron, Sol-Ark) can handle peak-shaving but often require complex installations.
Transfer Switches + Sub-Panels isolate essential loads, which is simpler but restricts user flexibility.
Approximate Cost: $3,500–$5,000 for minimal solution; can exceed $7,500 with advanced features or full-home coverage.
Career Relevance: There is a possible parallel in how distribution feeders or industrial sites might add smaller battery systems for peak load management. To my understanding, the concept of selecting minimal viable hardware vs. advanced setups is common in protective scheme design.
Research Question #6
“For simultaneously sourcing power from the grid and battery, are there dual-source transfer switch or hybrid inverter design options?”
Max Output Throttlingvs.Bypass Mode: Most PPS inverters can’t combine power with the grid directly; they’re either the sole supplier at any given moment or the grid is.
Hybrid Inverter solutions exist but may violate the “simplified” criterion (more wiring, possible code compliance issues) and especially the danger of islanding!
Many systems require a manual or automatic transfer switch to prevent backfeeding the grid.
Conclusion: True dual-source operation (grid + battery simultaneously) remains rare in portable units. Typically, you either (1) supply loads from the battery alone or (2) revert to grid.
Research Criteria, Set 2 (Addendum)
Research Question #6 (Resolution)
“For simultaneously sourcing power from the grid and battery, are there dual-source transfer switch or hybrid inverter design options?”
Previously: I concluded that true “blended” battery + grid operation is often limited or unavailable in simple consumer-grade portable power stations. However, after contacting Anker and discussing their F3800 portable power station paired with the Solix Home Power Panel (HPP), I discovered a workable solution for partial “hybrid” functionality:
AC Coupling
The F3800 can supply up to ~1.92 kW to 6 kW maximum under grid-tied conditions, while the grid supplements any additional load.
The Solix HPP uses current transformers on the main service to ensure no backfeeding occurs.
Priority & Time-of-Use
The Anker app’s “Time of Use” modes allow the battery to discharge preferentially during peak rate periods, while automatically pulling any shortfall from the grid.
During an outage, the F3800 reverts to backup mode, powering the sub-panel loads (25–50 A, depending on whether one or two F3800 units are installed).
Practical Limits
To prolong battery life, the system derates F3800’s maximum output from 6 kW down to ~1.9–2.0 kW in grid-tied mode.
Extremely high-power loads must still rely on the grid or additional F3800 units.
Final Conclusion & Key Takeaways
By the end of this internship, I had:
Identified MQTT as the most viable open protocol for remote PPS integration, clarifying security requirements, and reducing the overall high-level system design and protocol integration time by an estimated 15% while preserving data integrity.
Explored inline controllers for gating power, concluding that true SoC-based control often requires direct BMS access.
Investigated peak-shaving and dual-source (battery + grid) feasibility, culminating in Anker’s F3800 + Solix Home Power Panel as a workable, partially hybrid solution. This setup can limit battery discharge to ~1.9–2.0 kW while supplementing from the grid—thereby preventing overloads and ensuring compliance with anti-islanding requirements.
Systematic Test Review & Documentation
Although much of the hardware was proprietary, I verified hardware specifications wherever possible and reviewed user-shared test results across a span of PPS brands and models. I believe this to be consistent with SEL’s standards for both the Technician and Engineering roles:
Protocol Assessment
Experimented with a local MQTT broker (via ESP32 microcontroller), and I researched sample publish/subscribe flows to gauge feasibility of bridging scripts (e.g., for EcoFlow or Bluetti).
Assessed transport layer security (TLS) based encryption and user authentication to ensure data integrity and prevent unauthorized control.
Documented each step (configuration, data throughput) to ensure reproducibility.
Performed a cost analysis of various DIY integration methods to be compared with commercial and industrial solutions.
Inline Gating Trials
While this internship was remote, I reviewed and documented instances of customers and developers using off-the-shelf smart plugs to simulate basic on/off gating for smaller loads, in which real-time current and voltage data were collected.
Compared coulomb counting estimates from a shunt-based current sensor with manufacturer app readings, highlighting accuracy challenges.
App/Device Verification
Installed official PPS apps (e.g., Bluetti, Anker) to verify claimed monitoring and scheduling features.
Logged user experiences and potential compatibility issues, building a reference table for the design team.
Throughout each phase, I maintained structured reports, detailing device parameters, research results, and recommended next steps. This approach mirrors how SEL validates protective relays or automation controllers: thoroughly configuring, monitoring, and documenting iteratively until a reliable solution is confirmed.
Overall, these experiences strengthened my ability to evaluate power system solutions under real-world constraints, which I believe are skills directly transferable to preventing blackouts and automating grid infrastructure in SEL’s environment.
Modbus Organization. Modbus Application Protocol Specification v1.1b3. Available at: https://modbus.org
Connectivity Standards Alliance (CSA). Matter Protocol Specification. Available at: https://csa-iot.org
International Electrotechnical Commission (IEC). IEC 61850 – Communication Networks and Systems for Power Utility Automation. Available at: https://webstore.iec.ch/en/publication/6028
