ITead's "Sonoff Zigbee 3.0 USB Dongle Plus" (model "ZBDongle-P") based on Texas Instruments CC2652P +20dBm radio SoC/MCU

–EDIT Begin–
Just flashed 20211207 lol I’ll let you guys know if there’s instability. So far so good. All zigbee devices came back.
–EDIT End–
I’m still on CC1352P2_CC2652P_launchpad_coordinator_20211114.hex and this has been working fine on my Sonoff Zigbee 3.0 USB Dongle Plus. Even works with the new E1 sensors along with Zigbee2MQTT.

Only issue I’ve seen is I added a transition delay to all my bulbs to be 1 second. From time to time, the bulbs would come up with a low brightness as if the transition stopped midway. I’ve since just removed the transition and no problems. Not sure where the problem lies exactly.

1 Like

FYI, ITead posted a blob that should let users themselves update the product description on already shipped dongles to supposedly match that already written in CP2102N EEPROM of the latest batch:

I tested it myself on a Windows computer and it did not work for me (it closed without any messages).

I’m sad that ITead released a blob and base a tool on this open-source cp210x-program by VCTLabs:

FYI, Koenkk has now uploaded an experimental firmware for CC2652P/CC1352P with a default set to 9 dBm (and a max of 20 dBm), see/download:

He mentions in that he does not himself have RF measurement equipment so cannot confirm RF output power is exactly as per config set in software.

Did you try that firmware already ?

how many devices are routing through it and what is the average LQI like any idea? thanks in advance :slight_smile:

I do not have time to test before next week, and regardless I do not either have any real RF measuring equipment to test whether or not this firmware produces a stronger transmission signal strength (measuring power current in mA/Milliampere) when cranking up the TX power output configuration to the higher and higher-level setting.

Just looking at reported LQI (Link Quality Indicator) and RSSI (Received Signal Strength Indication) values for Zigbee devices is just an indication of antenna gain (i.e. reception and not transmission) and not imperial proof that CC2652P is actually increasing the TX power output to extend transmission range.

Anyway, sure we can all test different transmitter power output configurations, (“transmit_power” in Zigbee2MQTT and “tx_power” for zigpy-znp in ZHA), with our devices and unique real-world setups but without proper RF measuring equipment tools that you can hook up to the SMA-port instead of the antenna and measure actual transmission power output we can only make assumptions that it “works better” if higher TX power configuration produces improved LQI and RSSI values.

Zigbee2MQTT adapter settings:

transmit_power: 5
transmit_power: 9
transmit_power: 14
transmit_power: 20

ZHA integration advanced configuration:—yaml

Also, looking at zigpy-znp configuration example sounds as if 19 dBm is highest can set for CC2652P?

      tx_power: 5

5 versus 9 in tx_power configuration

      tx_power: 9

9 versus 14 in tx_power configuration

      tx_power: 14

14 versus 19 in tx_power configuration

      tx_power: 19


My guess is that setting the TX power to somehere between 9 and 14 (and not all the way to 19 or 20) will produce the best LQI and RSSI CC2652P/CC1352P based adapters. The reason for that prediction is that the higher TX power is then the more its own transmission is also interfering with its own reception capability, as transmitting higher power will produce more signal noise.

Remmeber that raising the TX power will only make it “shout louder” and not “listen better” so it could have the negative effect in the way that it will reach devices far away but not able to hear a weak reply.

By the way, recommend hobbyist enthusiasts with time and a little more seriously interest in inexpensive RF testing hardware equipment tools check out these two blog articles:

FYI, OpenMyDog submitted a new PR for ZHA USB discovery of Sonoff Zigbee 3.0 USB Dongle Plus:

Since that post I ended up moving to Zigbee2MQTT. I have two Sonoff 3.0 Plus Dongles, one router, one coordinator. Not sure if this picture helps at all. My coordinator is in the basement and the router is on the 2nd Floor… it seems a lot of 1st floor devices are preferring the router that’s upstairs. I haven’t been too worried about LQI or anything because all of my devices are significantly faster and more responsive than they have been with my HUSBZB-1. Also less frequent random disconnects (would occasionally have to power cycle bulbs with HUSBZB-1… rarely ever happens now). I’m considering getting another router for the 1st floor but everything has been working well enough that it seems kind of unnecessary

That sounds like a bad location as coordinator should preferably be placed in the center of your house.

Thanks this is really helpful. I have been using conbee 2 with zha and 118 devices of which 62 are routers but not very effective. Im facing ocassional dropouts and non responsiveness especially with devices furthest away. My current house is a duplex and built like a tank. we are talking 5500 sqft made of concrete and rebar and all walls are concrete bricks and mortar along with a wall pass metal in between each layer of brick. each room is essentially a faradays cage. I ordered 4 sonoff dongles. Will be using one as a coordinator and 3 as routers

1 Like

That’s where my home assistant computer is and it’s on an extension… so 90% of my devices are 10 feet above-maybe 20 feet horizontally… through a wood frame house that’s only 1800sqft… I don’t have connection issues was just sharing the router connections that someone asked about

May be something you already checked but did you verify your WiFi routers aren’t using a channel that overlaps your Zigbee? Or make sure they are not on auto channel and occasional changing to one that interferes?

+1 but tip there is to only change channels on all of your Wi-Fi router and/or Wi-Fi access-points.

Recommend not change channel on your Zigbee coordinator as all Zigbee devices will pair using just any Zigbee channel. Some Zigbee devices are limited to one or a few Zigbee channels, especially when pairing. (Blakadder db does mention channel limitations on some devices when users added that information (it is only a wiki maintained by users after all → ).

More tips:


nope made sure that wasn’t the case Wi-Fi router is on channel 11 and ZigBee on 15. I think I’ve outgrown my conbee it has been very stable till now but I think it might be time for an upgrade. If my house wasn’t concrete then I bet I would get some great stability but I could be wrong

@Gumasana how can I connect sonoff dongle as a repeater only? Many thanks

New firmware 20211217 allows setting the transmit power > 5dBm (9dBm by default), ITEAD (Sonoff) confirmed this is working (using a RF analyzer).


follow the link and instructions and you should be able to flash the router firmware on your dongle

FYI, Koenkk now released the mentioned +20 dBM RF power output enabled version as “Z-Stack_3.x.0 20211217” in the develop branch and closed issues as confirmed working with improved LQI as result:

That firmware has during the weekend been confirmed working by several community members with up to +20 dBM RF power output enabled, (tested in both Zigbee2MQTT and ZHA), and according to Koenkk it has also been confirmed working by ITead’s own Sonoff Zigbee engineer using RF analyzer as well, so I think later if and when proven stable we can expect this version to ship with future batches.

Note that any releases in the develop branch should still be considered experimental until proven stable and any bugs specifically verified related to this firmware version should be reported as new issues to:

So time to crank power output up to 11 on these!


I would love to see the router firmware get an entity exposed to HA where you could be able to select the output power in dBM

+1 Great idea! Submitted a request here →

1 Like