Repository navigation
From 4-in-1 to 1-in-4: an EFR32 board that replaces the 4-in-1 module #8017
2B-Tobe
started this conversation in
Show and tell
Replies: 6 comments 1 reply
|
Nice to see someone else working in this space, hope someone reaches out to help you with it. There is work on a successor to the MPM, as the MPMv2, which currently also uses a single RF chip, the Texas Instruments CC2652. https://github.com/raphaelcoeffic/mpmv2 |
0 replies
|
Hi Peter, hope to find you well
You sure have an interesting project there and the TI CC2652 is a good
choice.
What made me go for the EFR32MG24 was the headroom: four times the flash
and three times the RAM of a standard CC2652, and a faster core. With
multi protocols, EdgeTX support, telemetry and receiver setup in one
firmware, that extra room is quite useful.
My concept is to have one module serving as both transmitter and
receiver however it was never really intended to be pushed this far...I
have a line together with Mike Kitchen of encoders and it just happened.
In the Rx mode the I can run parallel 5 channels pwm, full ppm and full
SBus or just the regular 8 pwm.
Everything started by playing with a nRF24 that I pushed as far I could
go, with the EFR32 I have been able to improve my protocol and build the
RLINK which at the moment is the only one that can fly thru the "NO
zone" at our flying field with and w/o the dual band turned on.
Cheers and all thee best,
Tobe
…On 2026-11-10 03:40, Peter Feerick wrote:
Nice to see someone else working in this space, hope someone reaches
out to help you with it. There is work on a successor to the MPM, as
the MPMv2, which currently also uses a single RF chip, the Texas
Instruments CC2652. https://github.com/raphaelcoeffic/mpmv2
--
Reply to this email directly, view it on GitHub [1], or unsubscribe
[2].
Triage notifications, keep track of coding agent tasks and review pull
requests on the go with GitHub Mobile for iOS [3] and Android [4].
Download it today!
You are receiving this because you authored the thread.Message ID:
***@***.***>
Links:
------
[1]
#8017?email_source=notifications&email_token=CNTSIVZ7AYNRQ4DYSD7TIUL5YJ52VA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGA3TGMJXUZZGKYLTN5XKMYLVORUG64VFMV3GK3TUVRTG633UMVZF6Y3MNFRWW#discussioncomment-19507317
[2]
https://github.com/notifications/unsubscribe-auth/CNTSIV6UB6ZAT5LZEGUMBMD5YJ52VAVCNFSNUABIKJSXA33TNF2G64TZHMZTMNBYG4YTGMZQHNCGS43DOVZXG2LPNY5TCMJSGYZTMNJQUF3AE
[3]
https://github.com/notifications/mobile/ios/CNTSIV7YASGYJAUX6XSLSY35YJ52VA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGA3TGMJXUZZGKYLTN5XKMYLVORUG64VFMV3GK3TUVJTG633UMVZF62LPOM
[4]
https://github.com/notifications/mobile/android/CNTSIV7H6VEO6VLLKIB3X5L5YJ52VA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGA3TGMJXUZZGKYLTN5XKMYLVORUG64VFMV3GK3TUVZTG633UMVZF6YLOMRZG62LE
|
1 reply
|
Hi,
Thank you for taking the time to look at the project and for putting it
so plainly. I fully agree with you in principle. A binary-only RAIL is
the weakest point of the EFR32 from an open-source point of view, and I
know you and others have had this fight with vendors more than once.
The problem is that it is getting harder, not easier. More and more of
the radio layer is being locked away, at Silicon Labs and across the
industry (Nordic, Espressif and TI do much the same), much of it because
of certification. If you believe in the principles behind the GPL, that
creates a real conflict. But sometimes you have to compromise to reach
what you set out to do, and just swallow the bullet.
I also ask myself: do we always need the proprietary part? For this
project I never needed to see inside RAIL. What matters is that
everything built on top of it is open: the protocols, the timing, the
module and the controller. Anyone can read it, change it and carry it
over to other hardware. The closed part is a small, fixed piece at the
bottom that I use the way it is offered.
We also have to be realistic. The chip makers are there to sell chips,
and they won't refuse a market, including ours, as long as it doesn't
cost them their IP. Nothing comes for free anymore, even though so much
in our hobby is still done on a voluntary basis. And as chip development
costs keep rising and the competitors join forces, I expect even more to
become proprietary in the future, making it even harder to find a
suitable chip at a sensible price. I would rather use what is on offer
and keep my own part fully open than wait for a perfect chip that may
never come.
Everything I wrote is published (module under GPLv3, controller under
MIT), and RAIL is listed separately in CREDITS.md with its own licence.
I should also say that this project was born to suit my own needs and to
work with my other projects and without ambition, not really intended
to take this direction. I share it mostly as a curiosity, because I
believe some of the solutions could interest others:
- one size fits most: the same hardware is either a transmitter or a
receiver, just by sending new settings from the controller;
- a simple dual band, with an SX127x module running a second link beside
2.4 GHz.
A second band for redundancy is becoming more and more a must in many
fields, as the radio spectrum gets more polluted. That is the idea
behind RLINK, but with my own receivers the same dual-band setup works
for the other protocols too.
Thanks again for the honest feedback.
Best regards,
Tobe
…On 2026-11-10 10:19, Raphael Coeffic wrote:
The biggest downer for me with EFR32MG24 is RAIL and binary only
distribution. This is IMHO not acceptable for a GPLv3 project. We've
fought this a couple times already and vendors don't it make it easy
on us with these conditions.
--
Reply to this email directly, view it on GitHub [1], or unsubscribe
[2].
Triage notifications, keep track of coding agent tasks and review pull
requests on the go with GitHub Mobile for iOS [3] and Android [4].
Download it today!
You are receiving this because you authored the thread.Message ID:
***@***.***>
Links:
------
[1]
#8017?email_source=notifications&email_token=CNTSIV6QG66S22TRUXB3F4D5YLMTZA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGE2DEOBTUZZGKYLTN5XKMYLVORUG64VFMV3GK3TUVRTG633UMVZF6Y3MNFRWW#discussioncomment-19514283
[2]
https://github.com/notifications/unsubscribe-auth/CNTSIV3DJME2YUPEYJBHBAD5YLMTZAVCNFSNUABIKJSXA33TNF2G64TZHMZTMNBYG4YTGMZQHNCGS43DOVZXG2LPNY5TCMJSGYZTMNJQUF3AE
[3]
https://github.com/notifications/mobile/ios/CNTSIV5UDIURRFWOM4CRJUL5YLMTZA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGE2DEOBTUZZGKYLTN5XKMYLVORUG64VFMV3GK3TUVJTG633UMVZF62LPOM
[4]
https://github.com/notifications/mobile/android/CNTSIV3KOJJODGAS723EPB35YLMTZA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGE2DEOBTUZZGKYLTN5XKMYLVORUG64VFMV3GK3TUVZTG633UMVZF6YLOMRZG62LE
|
0 replies
|
@2B-Tobe check this: zephyrproject-rtos/zephyr#17005 You might be in breach of Silicon Labs' licence by distributing the binary blog in your repository. PS: tell the LLM writing all this to stop replying with huge bloated text, it can surely do better. |
0 replies
|
Thanks for the heads up, I have updated accordingly
…On 2026-11-10 13:42, Raphael Coeffic wrote:
@2B-Tobe [1] check this: zephyrproject-rtos/zephyr#17005 [2]
You might be in breach of Silicon Labs' licence by distributing the
binary blog in your repository.
--
Reply to this email directly, view it on GitHub [3], or unsubscribe
[4].
Triage notifications, keep track of coding agent tasks and review pull
requests on the go with GitHub Mobile for iOS [5] and Android [6].
Download it today!
You are receiving this because you were mentioned.Message ID:
***@***.***>
Links:
------
[1] https://github.com/2B-Tobe
[2] zephyrproject-rtos/zephyr#17005
[3]
#8017?email_source=notifications&email_token=CNTSIV6H5CTPLM7AGRLMWST5YMELBA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGE4DOMZTUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-19518733
[4]
https://github.com/notifications/unsubscribe-auth/CNTSIV4EDFZW5CJSJGGEXPD5YMELBAVCNFSNUABIKJSXA33TNF2G64TZHMZTMNBYG4YTGMZQHNCGS43DOVZXG2LPNY5TCMJSGYZTMNJQUF3AE
[5]
https://github.com/notifications/mobile/ios/CNTSIV5B76AHEWVC3U5JDQ35YMELBA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGE4DOMZTUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSVGM33PORSXEX3JN5ZQ
[6]
https://github.com/notifications/mobile/android/CNTSIVZVAMXMWRYUVSNPLRT5YMELBA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGE4DOMZTUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSXGM33PORSXEX3BNZSHE33JMQ
|
0 replies
|
https://codex.ac6.fr/Zephyr/xref/hal_silabs-latest/simplicity_sdk/License.txt
reading this it should be ok!
…On 2026-11-10 13:42, Raphael Coeffic wrote:
@2B-Tobe [1] check this: zephyrproject-rtos/zephyr#17005 [2]
You might be in breach of Silicon Labs' licence by distributing the
binary blog in your repository.
--
Reply to this email directly, view it on GitHub [3], or unsubscribe
[4].
Triage notifications, keep track of coding agent tasks and review pull
requests on the go with GitHub Mobile for iOS [5] and Android [6].
Download it today!
You are receiving this because you were mentioned.Message ID:
***@***.***>
Links:
------
[1] https://github.com/2B-Tobe
[2] zephyrproject-rtos/zephyr#17005
[3]
#8017?email_source=notifications&email_token=CNTSIV6H5CTPLM7AGRLMWST5YMELBA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGE4DOMZTUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-19518733
[4]
https://github.com/notifications/unsubscribe-auth/CNTSIV4EDFZW5CJSJGGEXPD5YMELBAVCNFSNUABIKJSXA33TNF2G64TZHMZTMNBYG4YTGMZQHNCGS43DOVZXG2LPNY5TCMJSGYZTMNJQUF3AE
[5]
https://github.com/notifications/mobile/ios/CNTSIV5B76AHEWVC3U5JDQ35YMELBA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGE4DOMZTUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSVGM33PORSXEX3JN5ZQ
[6]
https://github.com/notifications/mobile/android/CNTSIVZVAMXMWRYUVSNPLRT5YMELBA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGE4DOMZTUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSXGM33PORSXEX3BNZSHE33JMQ
|
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
From 4-in-1 to 1-in-4: an EFR32 board that replaces the 4-in-1 module, and is the receiver too
A 4-in-1 module needs four radio chips. This one does it with one, and the same board is the receiver too: from 4-in-1 to 1-in-4.
Hello all,
I'd like to share a project that may interest anyone running a 4-in-1 multiprotocol module, or anyone who likes to build their own radio gear. It is a small RF board built around the Silicon Labs EFR32MG24. It plugs into the JR bay of an EdgeTX radio and works there as a MULTI module. The same board, with the same firmware, also works as the receiver in the model.
What it is
The EFR32MG24 is a microcontroller with a 2.4 GHz radio built in, here on a small ready-made RF-star module. Unlike most RC radio chips, its radio can be reconfigured in software for different modulations, data rates and packet formats. That lets one board speak several RC protocols:
• FrSky D8
• FrSky D16, ACCST v1 and v2, EU-LBT and FCC
• DSMX
• RLINK, the project's own resilient two-way link (more below)
• nRF24, a simple fast link for nRF24L01-based receivers
In an EdgeTX radio
You select MULTI as the external module, exactly as for a 4-in-1. Nothing changes in EdgeTX itself.
• The protocol list: the radio reads it from the module, so it offers only what the module can fly. It uses the standard Multiprotocol numbers (FrSkyD, DSM, FrSkyX, FrSkyX2), so existing models carry over. The project's own links appear as "MiTo", with RLINK and nRF24 as sub-protocols.
• The status line: shows the module and its channel order.
• Telemetry: found with "Discover new". You get RSSI, TRSS, TQly and the receiver battery (RxBt), and EdgeTX's usual RSSI alarms work as normal.
Wiring is four wires from the bay:
• pin 1 (MULTI frames) to the module's serial input
• pin 2 (battery) to its supply
• pin 3 to GND
• pin 5 (S.Port) to the line that carries telemetry back
I have it running in a Taranis Q X7 on current EdgeTX, flying RLINK with telemetry on the radio's screen.
RLINK: a link built to stay up
RLINK was written for this board and uses the freedom of its radio fully. The aim is simple: a frame should get through even when part of the band is busy.
• Every frame is sent twice. A frame goes out every 11 ms, and a second copy follows 2.5 ms later on a channel about 40 MHz away. The two copies hop independently over 39 channels each, so one blocked channel or a short burst of interference does not lose the frame.
• The link learns which channels are bad. The receiver counts losses per channel. A channel that loses 40 % or more is dropped from the hop sequence on both ends, up to 12 channels per pool. After a minute it is tried again, so a Wi-Fi network that goes quiet gives its channels back.
• Listen before talk. Before every packet the radio checks that the channel is clear. A busy channel means that copy is skipped, and the other copy carries the frame.
• Power follows the need. The transmitter adjusts its power from the receiver's reports. It goes up quickly when the signal weakens and comes down slowly, aiming for a comfortable margin rather than full power all the time.
• Two-way, all the time. Every fourth frame the receiver answers with its link quality, the signal at both ends, lost frames and its battery voltage. The transmitter warns with beeps for a weak link or a low receiver battery, and EdgeTX shows it all as sensors.
• Per-model binding. A receiver bound to one model only answers when that model is selected, so the wrong model on the field gives no response instead of flying on the wrong settings.
Add the optional second band (below), and the receiver has two completely separate radio paths to its servos.
One board, two jobs
The board in the radio bay and the board in the model are the same hardware with the same firmware. As a receiver it drives eight servos. For flight controllers, gyros or servo buses it can instead give five servos plus SBUS and/or PPM. SBUS comes out inverted as other receivers send it, with no extra inverter.
Setting it up with the controller box
A board needs to know whether it is a transmitter or a receiver, and which protocol to run. That is stored in its own flash memory and survives firmware updates. The easiest way to set it is a small controller box: an STM32F103 with a 128×64 display and five keys. It plugs into the board's serial pins and sets:
• the role: transmitter or receiver
• the protocol
• for a receiver, the output layout
You don't need a PC, a terminal program or jumpers. It's handy for a receiver already installed in a model, or for a module about to go into a radio. A terminal on a PC or a set of jumpers works too, if you prefer.
Serial is the key, but PPM works too
With serial MULTI frames, from an EdgeTX radio or a serial encoder, the radio chooses the protocol per model, and bind and range check come from the radio. If your encoder only has a PPM output, as many older transmitters do, the module takes PPM instead. You then set the protocol once with the controller box, and the module flies it.
An optional second band
The board has room for a LoRa-class radio module beside the EFR32. With it fitted:
• the transmitter sends the same channels on a sub-GHz band alongside 2.4 GHz;
• the receiver listens on both;
• if 2.4 GHz goes quiet, the servos follow the second band within about a tenth of a second.
The 459 MHz version (UK model-control band) works now with the RA-01 module. 868 MHz (Europe) and 915 MHz (US) are options with the SX1276 version of the module, within your local rules. It is switched on per model: on the MiTo protocol with the Option value, and on FrSky or DSM through the option field.
This follows the same idea as ExpressLRS dual-band setups, but it is the project's own link and does not talk to ELRS receivers.
A few honest notes
• Failsafe: set on the receiver the module's own way. EdgeTX's failsafe setting is not used, because EdgeTX resends it every few seconds and the receivers here would store the stick positions of that moment.
• DSMX: carries no telemetry back.
• Black-and-white radios: they reach the MiTo protocol through the module's status, so the telemetry wire on pin 5 must be connected.
Open source and credits
The module firmware, documentation and hardware files are released under the GPL v3, and the controller under the MIT licence. [link]
The project stands on a lot of open work, and I want to credit it properly:
• DIY Multiprotocol TX Module (Pascal Langer and contributors): the serial frame and telemetry formats the module speaks, and the DSMX channel computation.
• Deviation: the DSM2/DSMX framing and code tables.
• The Multiprotocol and Deviation projects and the wider open RC community: the FrSky D8/D16 protocol descriptions the implementation follows.
• EdgeTX and OpenTX: their MULTI driver and telemetry handling were studied so the module meets the radio exactly as a 4-in-1 does; no EdgeTX code is copied.
• Silicon Labs: the SDK and RAIL radio library.
• olikraus: u8glib, which drives the controller's display.
• Mike Kitchen's VintageTx: the starting point of the companion MiTo encoder, which brings vintage transmitters onto the same system.
RLINK itself, the second-band link and the controller are the project's own work.
Where it goes from here
What began as a way to fly my own transmitters has grown into a complete system:
• a module that takes the place of a 4-in-1;
• receivers built from the same board;
• a link of its own, with a second band beside it;
• a controller to set it all up.
It has reached the point where it is wider than one person can carry. I'd like to take it further with others, and there is room for many kinds of help.
Tobe
All reactions