But what happens if you just put the LoPy Antenna on one of the cheap boards?
Because the China boards + a Pycom Antenna are still (much) cheaper than the Pycom configuration that you’ve got.
Note: I’ve got both (the LoPy as part of the Kickstarter), I like the Pycom quality, their support etc.
Just pointing out that the cheapest reliable solution might be a mix of parts sourced from different venders.
My guess is that TTGO V2 will have the best LoRa performance of the Heltec and TTGO boards because its use of standard (shielded) HPDtek LoRa modules. Combine it with a good antenna (can be had for €3) and you probably have a nice combination for prototyping.
Andreas Spiess has already ordered a TTGO V2 for testing with his spectrum analyzer. Let’s hope he can manage to show the results of that in a video any soon (hopefully within weeks, if not we just will have to wait longer).
But the testing on the spectrum analyzer showed only that the antennas were wrong/useless, so it just as well can be that the V2 has the same incorrect antenna.
Aliexpress has those for that price, but without the equipment that Andreas has, or the option to test them like that, how would we now if any antenna bought there actually is good for 868Mhz?
(field testing and comparing to known nodes/antennas being the only option, but you would have to have at least 1 good reference antenna).
Early test results:
TTGOv2 using the same included rubber antenna as Heltec is ~5dBm better on LoRa 868 MHz.
The “3D” Wifi antenna of TTGOv2 works better than the Heltec’s coil, but it’s not as good as the LoPy’s onboard ceramic wifi antenna.
Unfortunately the TTGOv2 module resets after tx’ing ~200-250 packets for unknown reason
Test setup:
Power consumption, during active wifi scan, no LoRa TX’ing, OLED display off, 5V USB powered:
Result overall:
LoPy wins, 30% less power consumption with 100% same functionality.
Heltec is stable but on the RF side not as good as TTGOv2.
TTGOv2 seems to have solved the issues of TTGOv1 on the RF side, but seems it does not run stable.
You can get antennas from here:
http://www.rocketscream.com/blog/product/ism-band-3-dbi-dipole-antenna-straight/
This is the antenna that was included with both my TTGO LoRa32 V2 and my Heltec Wifi LoRa 32 boards (868MHz versions).
The Heltec antenna is 2mm shorter. This could either be by design, a production error or (non) ‘acceptable’ production margin.
See here for picture and details: The BIG and SMALL ANTENNA topic part 1
See The BIG and SMALL ANTENNA topic part 1 for a cheap 868MHz center-fed dipole whip antenna, SMA, 19,5cm (adjustable).
I use this antenna for the gateway too.
(RAK831 RPi gatewa. Note: RAK831 uses RP-SMA instead of SMA but I replaced the onboard connector. The same seller also sells the antenna in a RP-SMA version).
Hi,
This is my first post to this forum 
I really enjoy all the informations collected here : thanks for that.
I did follow a lot of this topic, together with Andrea’s review of the ESP32 ready made boards and their quite poor design…
So I decided to build my own, and as you will see it is really easy !
take an ESP32 Lolin lite V1 board
take an RMF95W lora module
take a Hallard’s shield for hopeRF
Add a nice dipole antenna
Stack all this together and you get your tiny compact ESP32 gateway 
As you can see, the Hallard’s shield and the Lolin32 lite do have the same pin spacing. And, as you can mux pretty much all the pins on the ESP32, it’s easy to adapt the software.
Only the +3.3V and the Gnd need a separate pin, but both of them are very close !
Pin Mapping is done as follow:
// if OLED Display is connected to i2c
// OLED 2 requires this updated library that is resolution independent https://github.com/DaveRichmond/esp8266-oled-ssd1306
#define OLED 1 // Make define 1 or 2 on line if you have an OLED display connected 1=128x64 2=64*48
#if OLED>0
#define OLED_SCL 25 // GPIO5 / D1
#define OLED_SDA 33 // GPIO4 / D2
#endif
and
// Use your own pin definitions, and uncomment #error line below
struct pins {
uint8_t dio0=23; // GPIO5 / D1. Dio0 used for one frequency and one SF
uint8_t dio1=23; // GPIO4 / D2. Used for CAD, may or not be shared with DIO0
uint8_t dio2=23; // GPIO0 / D3. Used for frequency hopping, don’t care
uint8_t ss=16; // GPIO15 / D8. Select pin connected to GPIO15
uint8_t rst=0; // GPIO0 / D3. Reset pin not used
// Pin definetion of WIFI LoRa 32
// HelTec AutoMation 2017 support@heltec.cn
#define SCK 17 // GPIO5 – SX127x’s SCK
#define MISO 5 // GPIO19 – SX127x’s MISO
#define MOSI 18 // GPIO27 – SX127x’s MOSI
#define SS 16 // GPIO18 – SX127x’s CS
#define RST 0 // GPIO14 – SX127x’s RESET
#define DI00 23 // GPIO26 – SX127x’s IRQ(Interrupt Request)
} pins;
//#error “Pin Definitions _PIN_OUT must be 1(HALLARD) or 2 (COMRESULT)”
#endif
I have tested it with:
Loragateway V5
Lmic as a lorawan node
lora library as sender or receiver
Range as been measured to more than 8.7km with this antenna, SF11 and BW 125
Quite happy with this result 
JP
Thanks for sharing.
for optimal result… do you mount that antenna horizontal (like on the picture) or vertical ?
I was busy to run the Heltec board in deep sleep mode. What I observed is that SX1276 receives RST signal. Either by power down or by wake up and initialization.
I have described the issue in details here - https://github.com/mirtcho/HelTec_ESP32_LoRa/issues/1
Does somebody experienced similar problem?
Is there some kind of pinmux/GPIO initialization file for ESP32?
Thanks in advance
Does anyone know how much power one can draw from a digitalpin of the TTGO V2 or heltec boards?
I need to be able to control when I power up or down some sensors, but don’t know how much I can load on digital pins.
Thanks for your help!
I have a few of the first TTGO Lora boards shown above and the LED is on pin 2. This is the first board not the 32 v1 or 32 v2.
HI,
The important thing seems to be that both sender and receiver have the same orientation.
I performed the 8.7km range with horizontal orientation.
However it seems also that it is better theoritically to have a vertical orientation with the “hot” side pointing up to avoid “grounding effect with the earth”.
JP
Hi,
Both boards are using the same ESP32 Expressif chip.
Datasheet is saying this:
Each digital output pin is associated with its configurable drive-strength. Column “Drive
Strength” in Table IO_MUX lists the default values. The drive strength of the digital output
pins can be configured into one of the following four options:
• 0: ~5 mA
• 1: ~10 mA
• 2: ~20 mA
• 3: ~40 mA
The default value is 2.
The drive strength of the internal pull-up (wpu) and pull-down (wpd) is ~75 µA
[edit] Wrong information.… It is said here that the max drain current of a pin is 12mA
Sorry for that and thanks @bluejedi
JP
OK, so for 100 mA peak, I need something else… Do you know an efficient module that would allow me to toggle energy coming directly from the 3V3 pin and can power that?
no don’t change the module.
Just add a simple mosfet on the IO pin as on this schema:
ths “switch” is done by your IO pin is commands the gate of the mosfet which itself acts exactly as a manual switch
JP
Please note that:
Well you are surely right…
I must say that this datasheet with the “drive strength” is somehow misleading 
Thank you for the link
JP
Correct, that information is not very clear.
From above link: “We’ll need to make the drive strength / DC current distinction more clear in the datasheet anyway.”
Please correct your post to prevent that others damage their board due to incorrect information.
Thanks.
