BTC — updated
Coin Web Mining
Account Cart

Power Meter and Monitoring for ASIC Miners

Power Meter and Monitoring for ASIC Miners

Knowing exactly how many watts each ASIC pulls — and how that draw changes hour by hour — is the difference between a farm that runs predictably and one that surprises the operator with bills, breaker trips, and silently degraded units. Per-circuit power monitoring catches firmware bugs, PSU failures, fan stalls, and overclocks that drift outside their lane. This guide covers power meter asic monitoring options across home and small-farm scale: Emporia, IoTaWatt, and Shelly hardware, what to log, what to alert on, and how to integrate with the broader monitoring stack.

Why per-circuit power monitoring matters

Whole-house utility meters measure aggregate consumption. That number lands on the monthly bill and tells the operator the total cost, but it does not say which circuit, which miner, or which time of day drove the draw. A misbehaving ASIC that has dropped to 60 percent hashrate but is still pulling full wattage looks identical on the utility meter to a healthy unit — only per-circuit monitoring catches the discrepancy.

Per-circuit data also unlocks accurate cost-per-hash calculations. Combining real wattage with pool-reported hashrate produces a J/TH efficiency figure that operators can compare to manufacturer specs and to neighbors of the same model. Drift in this figure over weeks or months is the earliest warning of fan-bearing wear, thermal interface degradation, or firmware regressions. The ASIC remote monitoring stack write-up covers the pool-side half of the same picture.

Hardware option 1: Emporia Vue

Emporia Vue is a residential energy monitor that clamps current transformers onto each branch circuit at the panel and reports per-circuit wattage at 1-second resolution. The 16-circuit version is the workhorse model and runs about $200; the 8-circuit version is cheaper and adequate for one or two miners. The Emporia app gives basic dashboards and the cloud service retains data for up to three years on the free tier.

The advantages: low cost, easy install (a competent DIYer can do it in two hours with the panel cover off), and broad smart-home integration via the Emporia API. The disadvantages: data lives in Emporia’s cloud, the API rate limits get tight for high-resolution polling, and the current transformers are sized for typical residential loads — operators with very heavy 30 A or 40 A circuits should verify CT range. Most home miners and small farms (under 20 units) start here.

Hardware option 2: IoTaWatt

IoTaWatt is the prosumer-to-light-commercial option. It’s an open-hardware energy monitor with 14 CT inputs, optional voltage reference for true power-factor measurement, and local data storage with no cloud dependency. It speaks InfluxDB, MQTT, and Emoncms out of the box, which makes it a natural fit for operators who want to land monitoring data in Grafana alongside pool and miner stats.

Cost is about $400 for the base unit plus $15–$30 per CT. Installation is similar to Emporia (CTs clamp around branch-circuit conductors at the panel) but the configuration is more involved — IoTaWatt requires assigning each CT to a circuit in the web UI and setting up the data destinations. The payoff is full local control of data, no cloud lock-in, and a community of operators sharing Grafana dashboard templates. Operators building a permanent monitoring stack typically prefer IoTaWatt over Emporia at the 5–20 ASIC scale.

Hardware option 3: Shelly EM and 3EM

Shelly EM is a single-phase energy monitor with built-in WiFi and a relay output, designed to monitor and optionally switch a single circuit. The 3EM is the three-phase version. Both report wattage, voltage, current, power factor, and accumulated energy via Shelly’s cloud or local MQTT. The hardware is small enough to fit inside a panel enclosure or behind a junction box.

The Shelly advantage is the per-circuit form factor with optional switching — operators can remotely cycle a single ASIC from the same device that monitors its draw. This is particularly useful for unattended deployments where a hung firmware needs a hard power cycle. Cost is about $50 per channel, which gets expensive at farm scale but is competitive at the 1–4 ASIC home setup. The 3EM at about $130 covers a three-phase rack PDU on a single device.

What to log and how often

For each circuit, the baseline data stream is wattage at 10-second resolution (averaged from higher-frequency samples). That cadence catches transient events without filling the disk. Voltage at 1-minute resolution flags brownouts and sag events that may stress PSUs. Power factor at 1-minute resolution flags PSU degradation — a healthy ASIC PSU runs near unity power factor, and a drop below 0.95 suggests the PSU’s input capacitors are aging.

Operators integrating monitoring with the broader stack typically store data in InfluxDB or Prometheus and visualize in Grafana. The ASIC remote monitoring stack overview covers the broader tool selection. For network-side monitoring, the mining network setup guide covers the switch and IP infrastructure that delivers miner telemetry alongside power data.

Alerts that actually matter

Three alert categories are worth setting up from day one. First, wattage out of expected range: if a 3,500 W miner draws 2,800 W or 4,100 W for more than 5 minutes, something is wrong. Second, sudden draw changes: a step change of more than 200 W in under 30 seconds usually indicates a fan stalling, a hashboard dropping out, or firmware crashing into a reduced state. Third, sustained voltage sag: if voltage drops below 220 V for more than 60 seconds, the circuit is overloaded or the utility is delivering low voltage.

Avoid the trap of over-alerting. Anything that fires more than once a week without action quickly trains the operator to ignore notifications. Alerts should map to a specific operator response: “go check the unit in rack 3” or “open a ticket with the firmware vendor.” Background noise alerts (slight wattage drift, momentary network blips) belong in dashboards, not in alert channels. Detailed hashrate-correlation alerts are covered in the hashrate dropout troubleshooting guide.

Anomaly detection beyond simple thresholds

Once a few months of baseline data accumulate, simple statistical anomaly detection becomes practical. Compare each miner’s current wattage to its 7-day rolling average — a 5 percent persistent deviation is worth investigating, even if it stays within the alert threshold. The same approach catches J/TH efficiency drift earlier than raw-wattage alerts, because the comparison normalizes for legitimate hashrate variation.

For larger farms, time-series anomaly detection in Prometheus or InfluxDB Cloud catches correlated failures — a sudden drop in three units on the same circuit suggests a PDU or breaker problem rather than three independent failures. This kind of pattern-matching is hard to do from a spreadsheet but cheap to set up once the data is flowing into a time-series database.

Cost vs benefit at home and small-farm scale

For a single-ASIC home setup, a $50 Shelly EM or a $200 Emporia Vue pays back almost immediately by catching the first PSU degradation event before it kills the unit. For a 5–10 unit setup, an IoTaWatt at $500 installed gives full per-circuit data and integrates with whatever dashboards the operator runs. For 20+ units, the IoTaWatt scales with additional CTs or operators move to commercial-grade three-phase meters from Shelly, Carlo Gavazzi, or Schneider Electric.

The total monitoring spend rarely exceeds 1 percent of the hardware capital, and the catches it produces — early PSU failures, drifting overclocks, miscounted circuits in the panel layout — easily justify the investment. Operators planning farms at the $100,000 starter scale should budget monitoring as a line item, not an afterthought.

Integration with utility billing and TOU rates

In states with time-of-use (TOU) electricity rates, per-circuit monitoring data is the basis for shifting load to lower-rate windows. Operators can correlate measured kWh against the utility’s TOU schedule to calculate actual cost-per-hash hour by hour, and to evaluate whether participating in demand-response programs is profitable for their specific load profile. EIA publishes residential and commercial TOU rate data by state.

For multi-meter installations or for separating mining load from household load (which matters for tax accounting and sometimes for utility rate eligibility), a dedicated sub-meter on the mining sub-panel is the cleanest answer. That ties back to the sub-panel installation guide and the load-calculation work that comes earlier in the planning process. Operators selecting hardware to fit the monitored circuits can browse the Coin Web Mining catalog for current-generation ASICs with published wall-power figures.

For operators with multiple sites, the per-site power data is also the input to a longer-term decision: which site to expand and which to consolidate. A unit running at the lowest per-kWh delivered cost is the unit that should grow; a site whose effective rate has crept up over the past year is the one to consider relocating. That comparison is only honest when the per-circuit data is captured consistently across sites, which is the underrated benefit of standardized metering.

References

What's the cheapest way to monitor ASIC power at home?
A Shelly EM at about $50 per channel or an Emporia Vue at about $200 for 16 circuits. Both clamp current transformers onto branch-circuit conductors at the panel and report wattage to a phone app or cloud dashboard.
Do I need cloud-free monitoring?
Operators who want full data control without vendor lock-in typically prefer IoTaWatt, which stores data locally and pushes to InfluxDB or MQTT. Emporia and Shelly rely on cloud services by default, which is fine for most home setups but limiting for larger or more privacy-conscious operations.
What alerts should I set up on a power meter?
Three categories: wattage out of expected range (more than 5 minutes), sudden draw step-changes above 200 W in 30 seconds, and sustained voltage sag below 220 V for more than a minute. Avoid alerts that fire more than once a week without action.

The miners that justify the monitoring spend are listed in the Coin Web Mining catalog, with published wall-power and hashrate figures that feed into the J/TH efficiency dashboards described above.