E22-400M30Slora module's rssi is lower than expected

[m4kemoretime]

hello,previously, I used the RA-01S as both a receiver and transmitter. When I increased the transmission power, the current rose from a normal 100+ mA to 600 mA, and the RSSI increased significantly. However, after a while, the transmitter's power output could no longer reach its previous normal level, even when I reduced the power setting. I suspect the RA-01S's power amplifier (PA) was damaged due to excessive power, though reception remained unaffected. During retesting, I damaged the PA of another RA-01S. To address this, I switched to the E22-400M30S, which is more expensive, larger, and has better transceiver specifications. Testing showed that the E22-400M30S can sustain high power transmission for extended periods. However, I encountered a new issue: when transmitting the same LoRa signal, the RSSI of the E22-400M30S is significantly lower than that of the RA-01SP, by about 20 dB.

For example, when using the high-power E22-400M30S to transmit LoRa signals, the RA-01SP with an IPEX antenna receives an RSSI of about 0 dBm, while the E22-400M30S receives around -20 dBm. Without an antenna, the RSSI difference remains approximately 20 dB. As a web engineer, not an electronics expert, I’m unsure if I’ve made an obvious mistake. As shown in the attached image, both modules use the same code, with no antenna connected: the RA-01SP (left) and E22-400M30S (right) show a 20 dB RSSI difference. I’ve also attached my code and would appreciate any advice.

This is the complete code

#include <RadioLib.h>
#include <SPI.h>


#define SX1268_CS     2
#define SX1268_DIO1   5
#define SX1268_RESET  4
#define SX1268_BUSY   42


const float freq = 433.1;
const float bw = 250.0;
const uint8_t sf = 10;
const uint8_t cr = 5;
const uint8_t syncWord = 0x12;
const int8_t db = 20;
const uint16_t preambleLength = 4;
const float TCXO = 1.8;
const bool LDO = true;


SPIClass hspi(HSPI);

SX1268 lora = new Module(SX1268_CS, SX1268_DIO1, SX1268_RESET, SX1268_BUSY, hspi);

volatile bool receivedFlag = false;

// 接收中断标志回调函数
#if defined(ESP8266) || defined(ESP32)
ICACHE_RAM_ATTR
#endif
void setFlag() {
  receivedFlag = true;
}

void setup() {
  Serial.begin(115200);
  hspi.begin(14, 12, 13); // SCK, MISO, MOSI

  // init LoRa
  Serial.println("init LoRa ...");
  int state = lora.begin(freq, bw, sf, cr, syncWord, db, preambleLength, TCXO, LDO);
  if (state == RADIOLIB_ERR_NONE) {
    Serial.println("LoRa success!");
  } else {
    Serial.print("LoRa fail,code: ");
    Serial.println(state);
    while (true) delay(100);
  }

  // irq callback
  lora.setPacketReceivedAction(setFlag);
  lora.setRxBoostedGainMode(true); 
  lora.startReceive();
}

void loop() {
  if (receivedFlag) {
    receivedFlag = false;

    int len = lora.getPacketLength();
    uint8_t buf[len];

    int state = lora.readData(buf, len);
    if (state == RADIOLIB_ERR_NONE) {
      Serial.println("[SX1268] received:");
      Serial.print("length: ");
      Serial.println(len);

      Serial.print("data(HEX):");
      for (int i = 0; i < len; i++) {
        Serial.print(buf[i], HEX);
        Serial.print(' ');
      }
      Serial.println();

      Serial.print("RSSI: ");
      Serial.println(lora.getRSSI());

      Serial.print("SNR: ");
      Serial.println(lora.getSNR());

      Serial.print("freERR: ");
      Serial.println(lora.getFrequencyError());

    } else if (state == RADIOLIB_ERR_CRC_MISMATCH) {
      Serial.println("[SX1268] CRC ERR");
    } else {
      Serial.print("[SX1268] err, Code: ");
      Serial.println(state);
    }

    lora.startReceive(); 
  }
}

[jgromes]

the RA-01SP with an IPEX antenna receives an RSSI of about 0 dBm

I don't even want to know how much power are you blasting from the transmitter if the receiver sees 0 dBm. That's a whopping 1 mW of received power, the absolute maximum input power for the SX126x is 10 dBm. Typically RSSI would be well below 0.

As shown in the attached image, both modules use the same code, with no antenna connected:

If you are transmitting from an unconnected connector, congratulations - your PA is now on fire. Why are you even comparing the performances without antenna?

This is the complete code, you can delete the relevant content about wifi, tft, mqtt,struct...

That's something you should have done, we have no interest in digging around your code and figure out what is relevant and what is not.

[m4kemoretime]

Thanks for your reply. I have simplified my code. They are actually a simple one-way transceiver. Because of your reply, I know that an antenna must be connected for high-power transmission. I think it was my stupidity that caused me to lose two modules. The transmitter e22 uses micropython, and the transmitter uses e22, and the power is 22. Receiver A is ra-01sp, and receiver B is e22. Back to the main issue, the same transmitter, two different lora modules for reception, they are on my desk, only two meters away from the transmitter, ra-01sp can almost maintain 0dbm, but e22 is stable at -17dbm. I referred to your answer about this. The flying wires on the right side of the picture is to restore your solution. According to the design, it will only flash when sending, and this is the case. According to the official description, e22 has a higher sensitivity, so I think e22 should also be 0dbm. But this is not the case, so this confuses me.

ra-01sp
E22-400M30
As you might expect, I used the maximum transmit power in Python

from SX126X import SX1268
import time

sx = SX1268(spi_bus=1, clk=14, mosi=13, miso=12, cs=2, irq=5, rst=4, gpio=42)

sx.begin(freq=433.1, bw=250.0, sf=10, cr=5, syncWord=0xa2,
         power=22, currentLimit=140.0, preambleLength=4,
         implicit=False, implicitLen=0xFF,
         crcOn=True, txIq=False, rxIq=False,
         tcxoVoltage=2.2, useRegulatorLDO=False, blocking=True)

import network
sta_if = network.WLAN(network.STA_IF)
mac = sta_if.config('mac')

import os, struct

packFormat = '>6sI8s'
dateSN = 0

def random_bytes(length):
    return os.urandom(length)

def random_ascii_bytes(length):
    return bytes([os.urandom(1)[0] % 95 + 32 for _ in range(length)])

while True:
    data = random_ascii_bytes(8)
    package = struct.pack(packFormat, mac, dateSN, data)
    dateSN += 1
    print(f"lora send :mac: {mac}, Seq: {dateSN}, ASCII: {data}")
    sx.send(package)
    time.sleep(1)

[jgromes]

According to the official description, e22 has a higher sensitivity, so I think e22 should also be 0dbm. But this is not the case, so this confuses me.

Sensitivty of the receiver has nothing to do with the RSSI level it sees. RSSI is just the level of electrical power (watts) at the input of the antenna in a given bandwidth, that's it. You have different radios (SX127x vs SX126x), with different antenna matching networks, different receiver bandwidths, possibly different antennas etc. So I'm not sure why would you expect the RSSI values to be the same.

[serialrf433]

@m4kemoretime Stop destroying hardware and connect the antenna! On one side i spend a huge amount of time trying to repair and recover things to reduce e-waste like for example here #1421 .
And then i have to read things like this here where you destroy hardware and when its destroyed, you buy again new hardware to again destroy it. And the newer hardware you buy, you choose a model where the destruction is even faster because of higher TX output power.
Stop destroying hardware! There is a very relevant reason why people spend much money into getting a VNA device to measure antennas and their SWR to get the SWR as close as possible to 1.00 . Without connected Antenna your SWR is simplified explained at 9999999999 instead of 1.

[m4kemoretime]

If you are transmitting from an unconnected connector, congratulations - your PA is now on fire. Why are you even comparing the performances without antenna?

Luckily, the issue was brought up in that thread. I honestly didn't know that using a transmitter without an antenna could have such serious consequences.
Since then, I’ve made sure to always connect an antenna before powering on or transmitting. I haven’t damaged any devices since.
Just yesterday, I ran a test in an open area and managed to achieve stable communication over 3 km — it worked great!
If I get the chance, I’ll try to do more tests. Thanks for your reply!