Remote sensor node needs to phone home to my rack with no grid and no cell. Is a 24/7 Starlink dish really the answer?
retained text
I have a remote site that only needs to send sensor readings and a heartbeat back to a server in my homelab. Every build thread I find solves it with a full Starlink dish running around the clock, and the power math bothers me. A Standard dish pulls 40 to 75 W, the Mini still 25 to 40 W. That is roughly 0.6 to 1.5 kWh a day, every day. Off-grid that means a few hundred watts of extra panel, a bigger battery bank to ride through a cloudy week, and a charge controller sized for all of it. All to move a few kilobytes into my lab. It feels like buying a truck to deliver a postcard. Duty-cycling the dish with a relay helps, but you are still waking a phased-array terminal to say "I'm alive, it's 4°C." Starlink is brilliant when you actually need bandwidth. For telemetry-class traffic it looks like the wrong tool. So two questions for people who have done this: 1. What are you using for the "few bytes, no grid, no cell" uplink in 2026? LoRa or Meshtastic back to a gateway, a sat IoT modem, NB-IoT, something else? 2. How does it land on the lab side? MQTT broker on the rack, a small VPS relay, a LoRaWAN gateway at home, or something I'm not thinking of? ***Edit:*** *Sorry, I should’ve mentioned the distance up front. The node’s about 30 miles out, behind a ridge with no line of sight to anything, no grid and no cell. Since it’s just a few bytes on a schedule, I’m going with a low-power satellite IoT setup rather than a 24/7 Starlink dish. Thanks all, the “what distance?” calls were fair.*
document fetched 2026-10-29 00:39 · canonical https://reddit.com/r/homelab/comments/1xhao2m/remote_sensor_node_needs_to_phone_home_to_my_rack