I excluded the E32 as it uses the older SX1278 module while the E22 mounts the newer SX1268. Looking at the datasheet, the range tests were performed with different data rates, which justifies the extra kilometers (8 vs 12km) of the SPI version.
Same specs, different marketing strategy?
Going to the E62, the datasheet reports
Module E62-433T30D is a full-duplex wireless transceiver module with TTL level output, operates at 425-450MHz (default: 433MHz) […]
Features TDD (Time Division Duplex). User can transmit data while receiving, not have to wait for the end of receiving
I am not aware of any similar full duplex LoRa implementations, although concurrent TX/RX can be faked by the UART internal buffering.
If you have already tried one of these modules let me know.
What is the difference between the E22 UART vs the SPI version?
E22-400M30S: this is a SPI module which has no MCU, no firmware. User need to do developing works before using this module. It’s max tested range could be 12km if user can make good design of their own MCU, PCB and firmware.
E22-400T30D: this is a UART module which has Ebyte MCU, Ebyte PCB and Ebyte firmware. After making balance between the cost and performance, we make it’s max tested range to be 8km.
Is the E62 really full duplex?
E62-433T30D is full duplex module. If ESP32 can be combined with E62-433T30D successfully, this system can surely send and receive data at same time.
I’ve been using the EByte E22 modules for QMesh. IMHO, the big benefits of the E22 over other 1W modules is that it has a TCXO (important for lower speeds/longer transmission times) and an external LNA that, according to the datasheet, improves Rx sensitivity by ~2dB.
I had a buddy of mine test the output of and E22 module. The 1W power output is accurate, and its higher-order harmonics are within FCC regulations (as a ham, I felt I had some duty to make sure that the harmonics were within regs).
The -150dBm Rx sensitivity is for the lowest bit rate (SF12, BW=7.8KHz). The TXCO should make this possible (I don’t think the HopeRF module has a TCXO), but do note that you’re moving data at a blistering 18bits/s. Normally, the SX127X and SX126X have a best sensitivity of -148dBm; I’m assuming that the 2dB improvement in sensitivity Ebyte claims is due to the external LNA.
re: turnoff TXEN before turning on RXEN
I think we are doing this properly in the normal case (because we always call setStandby() before starting receive or xmit. But you reminded me of something that made me nervous, which is: In the constructor we are setting TXEN/RXEN as outputs before setting the driven value to be zero. I’ll fix. thanks.
(btw qmesh looks pretty slick and one of the workitems for the 1.5ish timeframe for this project is to possibly switch to it as our transport - but other than ‘this looks cool’ I haven’t thought about this yet)
Hello. Thanks a lot, I saw your post right away. One of these days I will receive ESP-32 boards and will be screwing E22 to them. Already made a large collinear antenna. I think it should make a good repeater.
The only caveat, I live by the river and around the hills is my house, the worst place for a repeater)