Unexpected Data Rate Behaviour in LR1121 with Variable Payload Sizes

[VARDHMANPICUS]

Description:
We have observed unexpected behavior in the LR1121 radio module related to the raw data rate when using varying payload sizes. While the datasheet specifies a payload capacity of 0-255 bytes, our findings indicate that the achieved data rate is inconsistent with expectations unless the maximum payload size is used.

---

Observed Behavior:

1. When we use the maximum payload capacity (255 bytes), the raw data rate aligns with the values mentioned in the datasheet and theoretical calculations.
2. However, when we send a smaller payload (less than 255 bytes), the raw data rate decreases proportionally.

Expected Behavior:
The raw data rate should not depend on the payload size, as long as it is within the specified range (0-255 bytes). We expect consistent performance as per the datasheet regardless of the payload length.

Concern:
Generally, such behavior is not acceptable and not considered a best practice for radio modules. There is no mention of such a restriction in the official documentation, and this behavior may affect performance in real-world applications.

---

Steps to Reproduce:

1. Configure the LR1121 with the following parameters:
   • Payload Size: Variable (0-255 bytes)
   • Modulation: LoRa
2. Transmit data with different payload sizes.
3. Observe the raw data rate at the receiver end.

---

Environment Details:

• Device: LR1121
• Firmware Version: Release v2.3.0
• Payload Sizes Tested: [9, 16, 26, 32, 48, 250]
• Modulation Settings: [Insert modulation type and configuration]
• Platform: LR1121 DevKit

---

Request for Clarification:
Could you please confirm if this behavior is expected? If not, is there any known workaround or firmware update to resolve this?

---

Attachments:

• Measure Date Test Result Sheet :
  
• Expected Test Result as per the LoRa Calculator
  

[opeyrard]

Hi,
Could you please provide the way you are estimating the Data Rate (meaning are you measuring the Time on Air and comparing it to the Payload length) ?
Could you also please provide the exact configuration used (meaning : LDRO config, coding rate, Packet param especially regarding the implicit/explicit header) ?
Thank you very much for your feedback,

[VARDHMANPICUS]

Hi,

Thank you for your response.

1. How We Are Estimating Data Rate

We are using following method to estimate the raw data rate:

RX Done Interrupt Counting Method

• We count the number of RX_Done interrupts (rx_frame_cnt) within a 5-second time period.

• The data rate is then calculated using the formula:

$$ bps = \frac{rxframecnt \times {PAYLOADLENGTH} \times {8}}{5} $$

• We observe that the bps value drops when the payload size is reduced, despite keeping all other parameters constant.

---

2. LR1121 Configuration Settings (from Logs)

Common Parameters:

• Packet type = LR11XX_RADIO_PKT_TYPE_LORA

• RF frequency = 868 MHz

• Output power = 22 dBm

• Fallback mode = LR11XX_RADIO_FALLBACK_STDBY_RC

• Rx boost = Deactivated

LoRa Modulation Parameters:

• Spreading factor (SF) = SF5

• Bandwidth (BW) = 500 kHz

• Coding rate (CR) = 4/5

LoRa Packet Parameters:

• Preamble length = 8 symbols

• Header mode = Explicit

• Payload length tested = 9 bytes

• CRC mode = Enabled

• IQ setting = Standard

• LoRa Syncword = 0x12

---

3. Observed Behavior

• The raw data rate decreases significantly when using smaller payloads.

• The achieved data rate is optimal only when using the maximum payload (255 bytes).

• As the payload size is reduced, the data rate does not scale proportionally, which is unexpected.

---

4. Expected Behavior

• The data rate should not be strongly dependent on payload size within the allowed range.

• There should be minimal overhead impact when sending smaller payloads.

---

5. Request for Clarification

• Is this behavior expected based on the LoRa PHY layer in LR1121?

• If not, is there a known workaround to achieve consistent data rate behavior for varying payload sizes?

Please let me know if you need additional measurements or logs to further investigate this.

Thanks & Regards,
Vardhman

[lbm-team]

Hi,

Yes, this behavior is expected due to how LoRa PHY layer works. The raw data rate you’re calculating is influenced by time-on-air (ToA), which includes fixed protocol overheads that become more significant as payload size decreases.
Even though the bitrate (symbol rate × bits per symbol) remains constant for a given SF, BW, and CR, the effective throughput (what you’re measuring) drops with smaller payloads because:

• LoRa packets include fixed overheads: preamble, headers, CRC, etc.
• These overheads are amortized better over larger payloads.
• With smaller payloads, the overhead-to-payload ratio increases, reducing the effective data rate.

Using your configuration (SF5, BW 500 kHz, CR 4/5), the LoRa Time-on-Air calculator would show:

• 9-byte payload: ~3 ms => 23.8 kbps
• 255-byte payload: ~34 ms => 59.3 kbps

Even though the 255-byte packet takes longer, it carries much more data, so the bps (bits per second) is higher.

Your formula:
bps = (rx_frame_cnt × PAYLOAD_LENGTH × 8) / 5
is valid for estimating effective throughput, but it does not account for protocol overhead. That’s why it drops with smaller payloads.

Many thanks,
Best regards