Ttgo T-beam


(Rem73) #83

Hello,
After many investigations i have the solution to improve the poor level of transmit power on Lora TTGO version 1 at 868 Mhz ( I suppose it is version 1.) The lora TTGo like this :

There are 2 issues :
Software :LMIC library is not clear on this side : Check radio.c file, and go to SX1276 power definition ( about line 476 ) : add this at the end of configuration file: writeReg(RegPaConfig,0b01111111) : It suppose you have a hardware witout Paboost connected.

Hardware:Is low cost, interresting but poor RF design. I remove all the parts at he RF filter. I apply the Semtech ref design to changes 8 capacitors and inductors, and use Pa output without Paboost option
https://www.semtech.com/uploads/documents/SX1276MB1MAS_e315v01a_prod_folder.zip

Measurement are performed with radio laboratory equipement in conducted condition : about ±0,1 dBm absolute accuracy. That’s my job.

Before : 2dBm : about 1,6 mW ( with LMIC modification describe above)
After : 14dBm : 25mW. ( with LMIC same modification describe above)

Conclusion : some issue around lora TTGO could be the result of wrong parts…Simply apply Semtech refdesign could solve the issue, without hard optimisation, without network analyser and PCB effect compensation.

I don’t check the receiver parts. May be some errors on design again…

You cannot have this kind of issues with RFM95 module.


#84

buying a new one is probably cheaper… time is money :wink:


(Verkehrsrot) #85

but sure that the newer versions does not have these or other flaws in RF design?


(DarkMan) #86

nice. tnx.


(Rem73) #87

“buying a new one is probably cheaper…” : No!! You don"t undestand.
I’m pratically sure that the mistake is not on MY board but on ALL TTGO V1 at 868Mhz:
There is a mystake in the layout: use of PAboost with wrong parts. Or it is only a BOM mistake
Understanding PA SX1276 is not obvious: many frequencies and BOM changes, many power configuration, 2 différents layout available, use or not EMI filter…not so easy as connecting an Oled Display.

Any way , if you are not able to do the measurement with specific hardware, you can not see the difference. 10dB Could be over-pass by good or bad cable and antenna. You will see it only for test at the limit range, with same antenna…


#88

ok, therefore you can’t say that changing a few components will help if there is a fundamental design error.


(Rem73) #89

To be clear : If you change theses components ( not easy ) and modify LMIC lib, you can improve de transmitted power, to be close to the semtech specification and RFM95


(Ud Lo Ra) #90

There is a video by Andreas Spiess about the ESP32 issues due to bad RF design . However, I agree that is likely better to buy an ESP32 alone + RFM95w instead.


#91

tnx… Andreas was spotted at The Conference yesterday ! :sunglasses:


(Ori) #92

I wonder what is this UFL connector for, does anyone know?UFL


(Eric VdB) #93

Wifi ?


#94

It is meant for connecting an external Wifi/BT antenna but that would require moving (de/re-soldering) a 0 ohm resistor (on the left of the U.FL connector) from one location to another to switch the RF connection.


(Ori) #95

Not good for the Lora antenna then…


#96

That is unrelated to the LoRa antenna (or at least as much related as the on-board Wifi/BT antenna).


(Ori) #97

Any luck/update with regards to the GPS and sleep mode?


(Bercman) #100

Hello!

I’m new to the LoRaWAN and TTN community.

I’ve been trying to get my TTGO T-Beam to work (connect to TTN) for several days now, but have had no success whatsoever.

What I’ve tried so far:

  • Downloaded LMIC library and tried out the following pinout settings according to the best TTGO T-Beam pinout information I’ve found so far (incl. soldering LoRa1 to pin 33 and LoRa2 to pin 32 on the board itself):

const lmic_pinmap lmic_pins = {
.nss = 18, // LoRa chip select on ttgo t-beam
.rxtx = LMIC_UNUSED_PIN,
.rst = LMIC_UNUSED_PIN, // Not required
.dio = {26, 33, 32}, // To LoRa DIO0 (interrupt), To LoRa DIO1 (interrupt) & To LoRa DIO2 (interrupt)
};

  • Updated SPI settings in hal.cpp according to the correct SPI pinout:

static void hal_spi_init () {
// SPI.begin();
SPI.begin(5,19,27); // These specific pins are for TTGO T-Beam only!
}

  • Tested LMIC AGP example, paying very close attention to little- and big-endian formats. This is what I see (no joining feedback, the last lines just repeat after every 60 seconds as defined):
    Baud rate 115200

ets Jun 8 2016 00:22:57

rst:0x1 (POWERON_RESET),boot:0x33 (SPI_FAST_FLASH_BOOT)
configsip: 0, SPIWP:0xee
clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
mode:DIO, clock div:1
load:0x3fff0018,len:4
load:0x3fff001c,len:928
ho 0 tail 12 room 4
load:0x40078000,len:9280
load:0x40080400,len:5848
entry 0x40080698
Starting
11124: EV_TXSTART
Packet queued
206222: EV_TXCOMPLETE (includes waiting for RX windows)

  • Tested LMIC OTAA example, paying very close attention to little- and big-endian formats. This is what I see:
    Baud rate 9600 (the last lines just repeat)

⸮as⸮⸮J&&⸮SY
Packet queued
4918: EV_JOINING
4937: EV_TXSTART
414734: EV_TXSTART
843576: EV_TXSTART
1302692: EV_TXSTART
2665242: EV_TXSTART

  • OTAA example. Baud rate 115200

ets Jun 8 2016 00:22:57
rst:0x1 (POWERON_RESET),boot:0x33 (SPI_FAST_FLASH_BOOT)
configsip: 0, SPIWP:0xee
clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
mode:DIO, clock div:1
load:0x3fff0018,len:4
load:0x3fff001c,len:928
ho 0 tail 12 room 4
load:0x40078000,len:9280
load:0x40080400,len:5848
entry 0x40080698

  • Peeled off the LoRa module sticker to make sure I do not have the 433 MHz LoRa module as I did have with other LoRa modules I failed with as I live in Europe. Although it is not the 433 MHz module, I found HPD13A-915 written on top of the module. Knowing that the TTGO T-Beam advertises itself to be compatible with 868 MHz and 915 MHz at the same time (?!), might this prove to be a problem? I do have to admit I’m too novice in this to make sure, so any feedback on this would be fantastic.

Going through numerous topics on TTGO T-Beam and different examples, the board itself works and I’ve managed to upload code onto it, but I cannot seem to connect with the TTN. Any help or extra questions to give input to would be greatly appreciated!

PS. I am using The Things Gateway for testing.

Thanks!
-Tarvo.


#101

see


(Verkehrsrot) #102

If you have the older TTGO T-Beam Version you need to wire Lora Pin DIO1 to CPU.

You can also check paxcounter software which comes with tailored settings for TTGO T-Beam, perhaps it works on your’s. If yes, you know that you did something wrong.


(Bercman) #103

Thanks for the tip. Could You please elaborate to where I would have to wire the LoRa DIO1 in case I need to?
I’ve had a look at the Paxcounter before, but haven’t really tried uploading it to the T-Beam that as it seemed a bit too complex code to begin my exploration with, but I will try it out. Hopefully it reveal something. :slight_smile:

EDIT: Paxcounter works fine, I was able to connect to TTN and forward data. So what mightbe up with the previous codes? :confused:


(Deux Vis) #104

Well first make sure you have a version which needs the wiring.

Can you find any version marking on the board ? Please copy it here.

Later : I just saw your edit. You definitely do not need more wiring if paxcounter works.