// Decision Engine 07

Size your module power budget

Estimate total node power draw for NVIDIA Jetson, Google Coral, and Hailo accelerators. Select module, operating mode, workload utilization, and peripherals to get system power totals with safety margin, PSU specification, current draw, cable gauge, and battery capacity. Supports single-node deployments from fanless edge sensors to powered enclosures.

Supports: NVIDIA Jetson (5 modules) · Google Coral (2 accelerators) · Hailo (2 processors)
AI Module Selection
// Select a family first
// Select a module first
▸ Advanced: Utilization modifier

Duty cycle: how often the module runs at full power (0.5× = 50% active, 1.3× = over-subscribed concurrent workloads)

▸ Advanced: Manual power override

Override all calculations with a known value (e.g., from datasheet or real measurement)

Use manual watts
// System Power
ℹ Selected value updates as you change module settings.
System Budget
Select all that apply
Fanless (passive heatsink)
Small 40mm fan (1W)
Medium 60mm fan (1.5W)
Large 80mm fan (2W)
Liquid cooling pump (3.5W)
Active thermal solution (4W)
Custom cooling
Select all that apply
× 0.4W
× 0.2W
%
%
▸ Step 13: Battery sizing

Leave runtime at 0 to skip battery calculation

hours
%
%
%
// Select module and power mode to continue
Quick Results
Module draw
Planning peak
Typical range
Confidence
Multiplier
1.0×
Power Breakdown
Component Watts
Operating total
Current draw
PSU recommendation
Battery capacity
Planning Notes
Configure the system to see planning recommendations.
{ }
Assumptions
Configure the system to see detailed assumptions.
FAQ

What is module power draw?

Module power draw is the electrical consumption of an AI accelerator (Jetson, Coral, Hailo) under a given workload. It varies by operating mode, workload type, and utilization.

How do I choose between operating modes?

Operating modes trade power for performance. For edge inference with single streams, use the lowest power mode. For multi-stream or compute-heavy loads, use higher modes.

What is utilization modifier?

Utilization modifier scales the module power based on actual workload duty cycle. 1.0× = full workload, 0.5× = half utilization, 1.3× = over-subscribed concurrent tasks.

Should I include external-powered cameras?

No. External-powered cameras (e.g., PoE cameras) have their own power supply and should not be included in the node system budget. Size the PoE switch separately.

What PSU wattage should I buy?

Buy a PSU rated at least to the 'PSU Recommendation' value. We include 20% safety margin + 25% headroom by default. Ensure sufficient amperage at your chosen voltage.

How do I size a battery for extended runtime?

Specify desired runtime hours. The calculator accounts for battery efficiency, DC cabling loss, and depth-of-discharge to recommend safe capacity. LiFePO₄ offers better cycle life than Li-ion.

What is confidence score?

Confidence reflects estimate reliability. Typical inference (1.0× util) = 85%. Other workloads = 65%. Custom workload/util = 45%. Validate critical designs with real measurements.

Can I use this for continuous 24/7 deployments?

Yes. This calculator supports system budget sizing for 24/7 edge deployments. Account for environmental cooling needs (passive vs active), power redundancy (dual PSU), and thermal management in your design.

How accurate are these estimates?

±15% under vendor-specified conditions. Custom workloads or high utilization may vary by ±25%. Always validate with real measurements before finalizing designs.