Choosing a LoRa Module from a Power Consumption Perspective

What you will get from this article

1. You will stop judging power consumption with only one current number.

2. You will use a simple “power budget” method to estimate how many days your project can run.

3. You will be able to tell whether Sleep, RX, or Time-on-Air (ToA) is the main power drain, and then decide which specs matter most.

This article focuses on power consumption (Sleep, RX, and Time-on-Air). If you’re still deciding between LoRa vs LoRaWAN, P2P vs public networks, chip platforms, and interface choices, start with our complete LoRa module selection guide.

Assumptions and definitions

This article compares two wireless modules based on the Semtech LR1121 platform: G-NiceRF LoRa1121 and Brand A (model A-XXX). Since they use the same chip platform, the power consumption comparison is more consistent.

All data comes from the official manuals or datasheets of each module. Two points need to be clarified:

1. The Sleep current = 10 µA marked by Brand A refers to the “Software shutdown” state defined in its document. In that state, the RF part is fully off. It is not the same as the normal Sleep or Standby mode used in most real projects.

2. The G-NiceRF LoRa1121 specification provides “Sleep current ≤ 1 µA (at 3.3 V)”. RX current is given under typical conditions such as 433 MHz and 2.4 GHz. TX current is provided as a “TX power to current” reference table, which is easier to use in real designs.

If you want to verify the original sources or check the full reference list, you can request the related document links.

Start with the key power consumption metrics table

First decide what your device does most of the time. Is it sleeping, waiting for downlink, or sending packets often. Different time shares lead to different priority specs.

Key power consumption metrics at a glance (lower values are better)

Why low power cannot be judged by one spec line

Module power consumption sets the “baseline”, and usage pattern sets the “multiplier”. Battery life difference is usually decided by the time share of these three states together:

● Sleep share: Is the device sleeping most of the time

● Receive share: Do you wait for downlink, wait for ACK, do listening, or open RX windows often, and how long does RX stay on

● Transmit share (Time-on-Air): How long does each transmission occupy the air

With the same LoRa parameters, ToA should be the same in theory. With the same battery, if ToA becomes longer (for example higher SF, larger payload, more retransmissions), or if RX windows open more often, battery life can drop fast. In many cases you should first reduce consumption by “usage pattern + parameter strategy”. But power does not come only from that. When a device sleeps for a long time or receives often, differences in Sleep or RX specs can be amplified. A lower power module can then bring real battery life improvement.

First decide what drains the battery in your project

We can roughly group devices into three common types. Find which one matches your application, and selection becomes much clearer.

A. Low frequency reporting (sleep current dominates)

Report sensor data once per day or once per hour. Most of the time the device is in deep sleep.

The key spec is Sleep current.

● G-NiceRF: ≤ 1 µA

● Brand A: 10 µA (software shutdown definition)

A 9 µA difference, combined with long standby, battery self-discharge, and temperature effects, can decide whether the device can reach the target lifetime.

B. Frequent reporting or long packets (TX current and ToA dominate)

Report once every 1 to 5 minutes, or ToA is long because SF is high or bandwidth is small, or link quality is poor so retransmissions are high.

For this type, TX peak current matters, but ToA often matters more.

From the data table, Sub-GHz high power TX peaks are both around 120 to 125 mA, so they are very close. In that case, which one saves more power usually depends on:

● SF, BW, CR you choose

● Payload size

● Retransmission rate

● Reporting frequency

C. Frequent receiving or listening (RX current dominates)

You often wait for downlink control, open RX windows frequently, or need “always controllable” behavior.

The key spec is RX current.

Sub-GHz RX current < 6 mA compared with 9.5 mA is a clear gap. If RX share increases, average current difference can grow fast.

Under the same usage pattern, how much does the module difference matter

Below are two typical examples. They show the trend. You can replace ToA and window length with your real project values.

Low frequency reporting (sleep dominated)

Assumptions

● The device sleeps almost all the time.

● To see the accumulated effect of sleep current first, TX and RX are treated as very small and not included here, only the baseline from sleep current is compared.

● Battery: 2400 mAh (example)

Sleep current conversion

● G-NiceRF: 1 µA = 0.001 mA
Daily sleep consumption ≈ 0.001 × 24 = 0.024 mAh/day

● Brand A: 10 µA = 0.01 mA
Daily sleep consumption ≈ 0.01 × 24 = 0.24 mAh/day

Difference

● Daily difference ≈ 0.216 mAh/day

● Yearly difference ≈ 78.8 mAh/year

If your device truly sleeps 99.99% of the time, this difference accumulates steadily. It is the kind of gap that slowly grows over time. But if RX and TX happen often, the sleep difference can be covered quickly by RX and TX energy.

Frequent RX windows (RX dominated)

Assumptions

● Open one RX window every minute for 1 second (for example waiting for downlink, waiting for ACK, polling).

● Total RX time per day: 1440 seconds

Using Sub-GHz RX current

● G-NiceRF: estimate with < 6 mA

○ Daily RX consumption ≈ 6 × (1440 / 3600) = 2.4 mAh/day

● Brand A: 9.5 mA

○ Daily RX consumption ≈ 9.5 × (1440 / 3600) = 3.8 mAh/day

Difference

● Daily difference ≈ 1.4 mAh/day

● With a 2400 mAh battery, 1.4 mAh/day alone is roughly like about 1700 days compared with 630 days (this only compares the RX part; real systems also include TX, MCU, regulators, and more).

As soon as RX share increases, < 6 mA compared with 9.5 mA can become a visible lifetime gap.

So the question “do I need frequent receiving or listening” directly decides whether RX current should be your first priority.

When TX peak currents are close, the key is Time-on-Air (ToA)

From the table, Sub-GHz high power TX current for both vendors is around 120 to 125 mA, so the gap is small. Then the key to “how much energy one packet costs” becomes ToA, packet frequency, and retransmission rate.

How much energy does one packet cost ?

Many people only look at the 125 mA value and ignore how long that current lasts. Below is a real calculation under these parameters (using LR1121):

Parameter settings

● Bandwidth (BW) = 125 kHz

● Spreading Factor (SF) = 12

● Coding Rate (CR) = 4/5

● Payload = 64 bytes

● Preamble = 8, Header = Explicit, CRC = On, LDRO = On

ToA result: about 2.79 s

Single transmit energy cost

Q_tx ≈ 125 mA × 2.79 s / 3600 ≈ 0.097 mAh

This is the energy for one packet. If network conditions are poor and retransmission is needed 3 times, one data report can consume close to 0.3 mAh. If you optimize parameters (for example use SF7) and reduce ToA to about 0.12 s, the energy per report is only about 0.004 mAh.

When TX current is similar, ToA difference caused by parameter settings can be more than 20 times.

How long can a 2400 mAh battery last

With the same parameters and the same battery capacity, what is the theoretical lifetime for different modules.

Scenario setup (typical industrial monitoring simulation)

● Battery: 2400 mAh (based on a 3.6 V 2400 mAh 18650 lithium battery)

● Interval: report once per hour

● Actions

1. Transmit: 2.79 s (LoRa SF12 heavy payload time on air)

2. Receive: 1.00 s (RX listening window)

3. Sleep: 3596.21 s (remaining time)

● Band: Sub-GHz 868 MHz (maximum power)

Which one uses less power

We add up the energy of all actions per hour (TX, RX, sleep), then estimate the theoretical lifetime:

Notes

● This only counts the wireless module itself. It does not include MCU wake up and sampling, sensor warm up, regulator quiescent current, power-on surge, and more. Full device lifetime is usually shorter than the table.

● 2400 mAh is a nominal capacity. Different battery chemistry, cutoff voltage, temperature, and discharge rate affect usable capacity. This article uses it for a horizontal comparison under the same supply assumption.

● The 10 µA of Brand A-XXX is based on its low power shutdown definition. In real projects, if more frequent peripheral holding or faster wake response is needed, sleep current can be higher than that value.

Where do the extra 3.5 months come from

In the heavy payload scenario where TX dominates (TX is more than 90% of total energy), G-NiceRF LoRa1121 lasts 107 days longer than Brand A-XXX (about 3.5 months).

Why can the gap be so large when TX current difference is small (123 mA compared with 125 mA). We can break down the hourly energy gap (unit mAh/hour):

78% of the gap comes from sleep current (1 µA compared with 10 µA)

● TX looks power hungry, but it lasts less than 3 seconds. The device is sleeping during the remaining 3596 seconds. The 1 µA baseline is very low, and it accumulates during those 3596 seconds, becoming the main gap (about 0.0090 mAh/hour).

13% of the gap comes from TX current (123 mA compared with 125 mA)

● TX current differs by only 2 mA, but it still brings about 0.0016 mAh/hour gap.

9% of the gap comes from RX current (< 6 mA compared with 9.5 mA)

● RX current differs by about 3.5 mA, and in a 1 second window it brings about 0.0010 mAh/hour extra gap.

If your device does not transmit continuously for 24 hours, reducing sleep baseline (Sleep current) is an efficient way to extend battery life.

Put the numbers into a power budget formula

For a quick battery life estimate, a common method is to compute average current first:

D is the time share of each state (0 to 1).

If you prefer a daily view like “how many times per day” or “how many seconds total”, you can also compute daily consumption directly:

● Daily sleep consumption (mAh/day)

● Daily receive consumption (mAh/day)

● Daily transmit consumption (mAh/day)

Where

Selection guidance

If your goal is longer battery life, you can decide which specs to check first based on the device usage pattern.

Low frequency reporting and long standby (sleep dominated)

These devices stay in low power states most of the time, so battery life is more affected by low power current.

Under the definitions from the official documents, G-NiceRF is marked as ≤ 1 µA (at 3.3 V, Sleep), and Brand A is marked as 10 µA (software shutdown). In long standby usage, a microamp level gap accumulates over time and can decide whether the target lifetime is reached.

Frequent receiving and listening (RX dominated)

If the device often receives, opens RX windows frequently, or waits for downlink control, check RX current first, especially Sub-GHz.

The documents show Sub-GHz RX current of G-NiceRF LoRa1121 is < 6 mA (at 433 MHz), and Brand A-XXX is 9.5 mA (Sub-GHz). When RX windows are more frequent and total RX time is longer, this gap more easily becomes average current gap and affects battery life.

Frequent packets or long packets (TX and ToA dominated)

In scenarios with frequent packets or long ToA, TX peak currents are on the same level (about 120 to 125 mA). Then ToA, retransmission rate, reporting frequency, and payload size can create a big lifetime gap.

A more effective method is to first optimize these “usage pattern variables” into a reasonable range, then use Sleep and RX specs to estimate how much extra battery life you can gain. In long standby or high listening share projects, the low power and RX current advantages of G-NiceRF are easier to convert into real battery life benefit.

FAQ

Can Sleep current values from suppliers be compared directly？What is the difference between Sleep, Standby, and Shutdown？

Do not draw conclusions from one number only. Different vendors may define “sleep state” differently. Commonly there are at least three types:

● Sleep: usually keeps part of the state, and can wake up through SPI or NSS events, or timers

● Standby: more awake than Sleep, wakes faster, but current is usually higher

● Shutdown: close to fully off, current is the lowest, but wake path, recovery time, and retained functions are different

Why can battery life differ by months when TX peak current is similar？Which metric should be the main focus？

Peak TX current only shows how large the current is during transmission. What decides energy consumption is current × duration.

A more practical priority order is:

● Low frequency reporting and long standby: check Sleep current (µA) first

● Need to wait for downlink or open RX windows often: check RX current (mA) + total RX time first

● Frequent packets or long packets: check Time-on-Air (ToA) and retransmission rate first, then TX current

A quick calculation is also simple:

Single event consumption (mAh) ≈ current (mA) × time (s) / 3600

How to estimate ToA quickly？Which parameters increase ToA most easily？

ToA is mainly affected by SF, BW, CR, and payload length. General rules:

● Higher SF means longer ToA (longer range and better interference resistance, but slower)

● Narrower BW means longer ToA

● Larger payload means longer ToA. If link quality is poor and retransmission happens, total ToA is amplified

Why does this article keep emphasizing RX？Does RX matter if there is no downlink control？

As long as the system needs “waiting for downlink, ACK, or polling”, RX can be a hidden major power drain.

For example, in LoRaWAN, after each uplink the end device opens RX1 and RX2 windows based on the spec to receive downlink. Even if no data is received, these windows still consume RX current. If windows open more frequently, windows are longer, or continuous listening is used (for example Class C), average current increases clearly.

DC-DC or LDO？How much does it affect battery life

Many LoRa transceivers or modules support both DC-DC and LDO power supply modes. In general, DC-DC can save more power, but it often needs an external inductor, and circuit design and EMI can be more complex. LDO solutions are simpler, but in RX and TX working states they can waste more power. Some Semtech documents also clearly recommend using DC-DC for higher efficiency (with the tradeoff of an extra inductor).