Battery Solution for Satellite Ground Station Backup
I have spent enough nights in equipment shelters to know that a satellite ground station is one of the least forgiving places to lose power. A low Earth orbit pass lasts eight to fourteen minutes. If the pedestal stalls mid-track because a utility feeder dropped, the pass is gone, the data is gone, and no amount of apologizing in the morning report recovers it. A properly engineered battery solution for a satellite ground station is therefore not an IT accessory; it is mission assurance hardware. In this guide I will walk through how we size, select, and commission backup power for ground stations, from the pedestal drive to the waveguide dehydrator, using the same discipline we apply to any industrial lithium battery program.
Why a Ground Station Can Never Reschedule a Pass
Unlike a warehouse that can absorb a two-hour outage, a ground segment has hard, non-negotiable windows. A typical LEO tracking station sees four to twelve usable passes per day, each eight to fourteen minutes long, and constellation operators usually contract ground segment availability targets of 99.5 to 99.9 percent. Missed passes during launch and early orbit checkout are worse: orbit-raising burns and commissioning often depend on one or two contacts per day, and miss them and the spacecraft sits unattended for another revolution.
There is also a safety dimension. During a storm-related outage, the antenna must be driven to its stow position and held there, or a wind event can drive an unbraked dish into hard limits and destroy the positioner. That means the highest-torque load in the station must stay alive precisely when the grid is most likely to fail. Any backup power design that treats the pedestal as a non-critical load has misunderstood the site.
Mapping the True Critical Load
The first deliverable in every ground station battery project I run is a measured load table, not a nameplate sum. The categories that matter:
- Antenna pedestal drives: azimuth and elevation servo amplifiers draw 5 to 15 kW while tracking, with inrush two to three times rated current at each reversal. High-wind tracking is continuous and regenerative braking on elevation reversals pushes energy back onto the DC bus.
- High-power amplifiers: a 400 W to 3 kW traveling wave tube amplifier in an uplink chain consumes two to three times its RF output as heat, so a 1 kW TWTA is a 2 to 3 kW thermal load that must also be conditioned.
- Receive chain: LNAs, downconverters, and switching matrices are milliwatt loads individually, but they ride on -48 V DC plants with their own rectifiers.
- Baseband and modems: rack servers, modulators, and routers typically add 2 to 8 kW in a mid-size station.
- Waveguide dehydrator: 100 to 300 W continuous, trivially small but fatal to forget, because a depressurized waveguide admits moisture and takes the link down hours after power returns.
- Shelter HVAC: only a fraction is critical, but the amplifier rack fans must keep running or thermal shutdowns will follow the outage.
For a representative 7.3 m X/S-band station, the genuinely critical list lands at 15 to 25 kW. Deciding what rides through, and for how long, is an engineering judgment I make with the customer, and it usually splits into two tiers: keep-the-antenna-safe (pedestal, stow control, dehydrator, minimal network) and keep-the-pass-alive (full RF chain and baseband).
Sizing: Bridge Time, C-Rate, and the Stow Allowance
Utility outages at ground station sites typically last 10 minutes to 4 hours, and remote sites often run generator-first architectures where the battery only bridges the 10 to 30 seconds before the genset closes. For stations on unreliable feeders, I size for 30 minutes of full critical load, which covers the majority of transient events without forcing a generator into every tender.
Worked example: 18 kW critical load for 30 minutes equals 9.0 kWh of usable energy. Divide by depth of discharge (0.90), conversion efficiency (0.93), and end-of-life capacity (0.80) and the nameplate requirement is 13.4 kWh, which I round to a standard 51.2 V, 260 Ah LFP battery at 13.3 kWh. Then I check power: 18 kW over 13.3 kWh is roughly a 1.3C average discharge with short excursions above 1.5C during pedestal acceleration. That C-rate demands constant-power discharge curves from the cell datasheet at the minimum expected temperature, not a 0.5C nominal rating read off the first page.
I also reserve a stow allowance: two full stow operations plus holding park position add 0.5 to 1.0 kWh and a 2.5C peak, and I would rather see that peak served by the battery than by a voltage sag that trips the servo amplifier mid-storm. Where the customer wants multi-hour autonomy, every additional hour is pure energy sizing, but the peak power requirement does not change.
Architecture: -48 V DC Plant, AC UPS, or Hybrid
Most stations already have a -48 V DC power plant per ETSI EN 300 132-2 serving the receive chain and baseband, and legacy VRLA strings floating behind it. The RF and pedestal side runs on AC. My preferred architecture is a hybrid: a lithium battery on the -48 V plant for the DC loads, plus a separate AC UPS or inverter bay sized for the pedestal drive and amplifiers. This keeps each battery in its natural voltage domain and avoids a single large inverter carrying everything.
Two integration details deserve attention. First, the pedestal regenerates: during elevation reversals the servo returns energy to the bus, and a lithium-backed rectifier absorbs it gracefully where a legacy thyristor charger will push bus voltage up until alarms trip. Confirm the rectifier tolerates regenerative energy or add a braking resistor. Second, ride-through coordination: the battery must hold the bus until the generator closes and stabilizes, typically 10 to 30 seconds, and the transfer logic must never let the UPS and genset fight over the same feeder.
Chemistry: LFP vs NMC vs Sodium-Ion vs VRLA
The chemistry decision for a ground station battery solution is rarely close. The station is a fixed asset with a 15 to 20 year life, space in the shelter is modest, and reliability per cycle matters more than mass.
| Chemistry | Gravimetric (Wh/kg) | Cycle life to 80% | ARC self-heating onset | Notes for ground stations |
|---|---|---|---|---|
| LFP | 150-180 | 4,000-6,000 | ~250°C | Default choice; best calendar stability at partial SOC |
| NMC 811 | 240-280 | 2,000-3,000 | 110-140°C | Only where floor loading forbids LFP; require 45°C cutoff |
| Sodium-ion | 100-160 | 2,000-4,000 | 200-230°C | Cold sites: 85-92% capacity at -20°C, but 1.5-4.0 V window breaks legacy -48 V plants and cabinet volume grows 25-40% |
| VRLA | 30-40 | 300-700 | n/a | Familiar but needs true float, dies young in cycling duty, 3-4 replacements per station life |
Lithium iron phosphate wins because a backup battery spends most of its life at partial state of charge, where LFP calendar aging of 1.5 to 2.5% per year at 25°C is the best in the industry. NMC saves floor space but its low thermal runaway onset is a poor match for unattended shelters. Sodium-ion is genuinely interesting for high-arctic or high-altitude sites that see weeks below -20°C and cannot afford preheating energy, provided the design accepts the wide voltage window from the start rather than retrofitting it onto a -48 V plant built for lead-acid.
Thermal and Environmental Design for Remote Sites
Ground stations live in hostile climates precisely because they need radio-quiet horizons: mountaintops, deserts, coastlines. Heat is the quieter killer: LFP cycle life falls from 4,000 to 6,000 cycles at 25°C to 3,500 to 4,000 at 35°C and under 2,000 at 45°C, so a dark outdoor cabinet in direct sun is a warranty claim waiting to happen. Reflective coatings, a sun shield, and filtered forced ventilation holding cell temperature near 30°C cost almost nothing in parasitic load.
Cold is the harsher constraint. Below 0°C a lithium battery must refuse charge or it will plate lithium metal permanently, so the BMS needs a hard charge lockout with an insulated heater pad sized for the site. At a 40°N mountain station the heater draws 3 to 8% of daily energy in winter, which is acceptable when fed from the -48 V plant but must be budgeted on solar-only sites. Specify IP55 minimum outdoors, an ePTFE pressure relief vent to stop condensation pumping, 316 stainless fasteners and coated busbars within 5 km of salt water, and shock ratings per IEC 60068-2-64 if the cabinet shares a foundation with the antenna pier.
Generator Integration and Recharge Discipline
At generator-backup sites the interaction rules are simple but frequently violated. The battery charger must not exceed 25 to 30% of the genset rating, or the generator either stalls on the rectifier inrush or runs for years at low load where wet stacking fouls the cylinders. Recharge at 0.2 to 0.3C: after a 30-minute bridge event the 13.4 kWh battery needs roughly two hours at 4 kW, comfortably inside a 15 kW genset budget alongside the station load.
Lithium chemistry also changes standby philosophy. VRLA wants a continuous 2.25 to 2.27 V per cell float; LFP wants to sit at 30 to 60% state of charge and receive a scheduled full recharge, monthly at minimum, so the BMS can rebalance and the capacity test can catch a failing cell. Delete the temperature-compensated float voltage from your rectifier settings; it is the wrong control law for this chemistry and quietly holds the battery at high SOC where calendar aging runs fastest.
Lightning, Grounding, and EMC at an RF Site
A tall antenna on a ridge is a lightning collector, and ground station shelters sit closer to that threat than almost any other battery application. Follow the site lightning protection zone concept: surge protection on every AC and DC conductor crossing a zone boundary, a single-point ground window, and battery cabinet bonding sized for the site’s ground ring rather than a generic water-pipe earth. I have seen surge remnants destroy battery string fuses that no datasheet transient rating anticipated.
Electromagnetic compatibility cuts both ways at a receive site. The station exists to hear signals near the noise floor, and a cheap BMS running switching converters at a few hundred kilohertz can raise that floor through conducted and radiated emissions, especially on receive bands where LNAs sit meters from the power cabinet. Require IEC 61000-6-2 and IEC 61000-6-4 compliance, ask for conducted emission data in the 150 kHz to 30 MHz range, physically separate the battery bay from the receive rack run, and use shielded, grounded power cabling. This one specification line has saved more receive sensitivities in my projects than any amplifier upgrade.
Monitoring, BMS, and Remote Telemetry
Unattended stations mean the battery must report, not just respond. I require cell voltage sampling at 1 Hz with ±5 mV accuracy, pack current and temperature at 10 Hz, and a trip chain that opens the main contactor in under 5 ms on a cell overvoltage. The management interface needs SNMPv3 and Modbus TCP so the station NMS treats the battery like any other rack element, plus four dry contacts for the legacy alarm panel: summary alarm, battery on discharge, low capacity, and high temperature.
Trend the four numbers that predict trouble: resting cell voltage spread, temperature rise per discharge, 1 kHz internal resistance, and cumulative amp-hour throughput. Internal resistance rising 25 to 30% flags end of life 300 to 500 cycles before the capacity cliff, and the failure you care about is not gradual fade but a battery that cannot deliver 1.5C on the night of a storm pass.
Commissioning and Acceptance Tests
Do not accept a ground station battery on paperwork alone. The five-test sequence I write into every tender:
- Insulation resistance at 500 V DC greater than 100 MΩ before first energization; below 10 MΩ, find the moisture path before power flows.
- After a full charge and a 2-hour rest, cell voltage spread under 30 mV. Spread that only appears under load points to busbar connection resistance, not the cells.
- Delivered capacity at 0.2C of at least 95% of nameplate on the acceptance discharge.
- A step-load test from 10% to 100% of critical load with the bus dip held under 10% and recovery inside 50 ms, proving the servo amplifiers will not reset mid-pass.
- A full-load thermal scan after 30 minutes: any busbar joint more than 15 K hotter than its neighbors gets retorqued and retested.
Repeat the step-load and capacity tests annually, and archive the results against the baseline. A battery solution for a satellite ground station is a 15-year asset; the trend line is worth more than any single reading.
Frequently Asked Questions
How long should a satellite ground station ride through a utility outage?
For grid-tied stations I design 30 minutes of full critical load, which covers the majority of feeder events and generator start delays. Generator-backed sites can drop to a 5 to 15 minute bridge, while solar-only remote stations usually justify 8 to 24 hours of autonomy sized against their worst weather month.
Can a lithium battery replace the existing -48 V VRLA plant without changing rectifiers?
Usually yes, with two changes: disable the temperature-compensated float and set the rectifier output to the LFP charge voltage window, typically 54 to 56 V for a 16S string, and confirm the plant tolerates regeneration from the pedestal. Rectifiers with adjustable voltage limits and no float-dependent logic need no hardware change.
What battery chemistry works best at remote cold-climate stations?
Sodium-ion holds 85 to 92% of capacity at -20°C where LFP drops to 70 to 80%, and it accepts low-rate charge below freezing without plating. The trade is a 1.5 to 4.0 V cell window that requires purpose-built electronics and 25 to 40% more cabinet volume. For moderately cold sites, LFP with a heater pad remains simpler.
Does the pedestal drive need backup power during a storm outage?
Yes, and it is the hardest load on the bus. The antenna must reach and hold stow during high wind, and the drive draws 5 to 15 kW with peaks above 2.5C during the move. Treating the pedestal as switchable load risks losing the antenna to the very storm that caused the outage.
How do I keep BMS switching noise from interfering with sensitive receivers?
Require IEC 61000-6-2 and 61000-6-4 compliance, request conducted emission plots for 150 kHz to 30 MHz, separate the battery cabinet from receive racks, and use shielded, grounded power cable. On receive-critical sites I have measured noise-floor improvements of several decibels from cable rerouting alone.
What discharge rate should the battery support for short-notice passes?
Size for the worst simultaneous case: full tracking load plus amplifier draw plus a stow operation, which in a mid-size station lands at 1.3 to 1.5C continuous with 2 to 2.5C peaks for tens of seconds. Demand constant-power discharge curves at your coldest expected cell temperature, not a nominal 0.5C rating.
How often should a standby lithium battery be recharged at low state of charge?
Hold standby between 30 and 60% SOC and schedule a full recharge at least monthly. The full charge lets the BMS rebalance cells and gives you a natural capacity checkpoint. A lithium battery held at 100% for float, as VRLA was, ages two to four times faster through calendar aging.
Which certifications should I require in a ground station battery tender?
Require UN38.3 for transport, IEC 62133-2 at cell level, IEC 62619 or UL 1973 at system level, UL 9540 with a 9540A test report for fixed installation, NFPA 855 compliance for the shelter layout, and IEC 61000-6-2 and -6-4 for EMC. For outdoor cabinets add the relevant environmental class per ETSI EN 300 019.
Further Reading
References
