Thread Tools
This thread is privately moderated by mpjf01, who may elect to delete unwanted replies.
May 22, 2020, 08:38 PM
Registered User
mpjf01's Avatar
Thread OP

Testing group FRSKY ACCST 2.1.0

Looking for Failsafe with FRSKY ACCST RF Firmware Version 2.1.0.

The following is a description of a procedure to test for momentary failsafe (FS) events and possible
unwanted servo movements (USM). A basic and hopefully common procedure has been suggested to
enable easy comparison between different users and for the production and delivery of evidence that can
be used to enable further investigation and remedial action.

The procedure is intended to be general and not too prescriptive, recognising that participants will have
different models of transmitter, receiver and associated equipment and installations.

Nevertherless it has been designed to limit or eliminate some of the matters that have been blamed for
possible malfunction. We have chosen close range with proper antenna orientation, controlled power
supply and an environment where interference should not be a factor. We ask that if you need to use a
different arrangement to that specified you keep in mind the need to eliminate as many possible reasons
for malfunction as is possible.


FRSKY Transmitter.

A transmitter capable of supporting RF firmware version 2.1.0. Please make sure that you have flashed
the transmitter with the latest version of the firmware from the downloads section https://www.frsky- of the FRSKY website. Please ensure that you are using OpenTx 2.3.7 or the latest
version of FRSKYOS.

FRSKY Receiver.

Prefer model G-RX8 or RX6R but it may be useful to try other models. Must be flashed with the latest
version of firmware 2.1.0 from the downloads section of the FRSKY
website. Please use just one receiver for the test, disconnect any redundancy receivers if in a model.

Receiver Power Supply

If the receiver is not in a model, power with a separate battery such as a NiMh. Make sure that this
battery is charged. If the receiver is in a model and it is not possible to use a separate battery, use the
ESC. If using a LIPO and BEC, record the detail.


Preferably as many servos as there are PWM ports on the receiver. If you don't have 8 spare servos
then use whatever you have available. Try to retain servo wires of length supplied by manufacturer, if
longer wires are used (in a model for example) mention in comments section of spreadsheet.

Video camera/mobile phone video camera.

A recording of the testing procedure and outcomes is the best evidence possible. Others looking for
evidence from your videos will benefit from knowing where in the video to look, record the timestamps for
events if you can.

Link Quality Sensor.

This is optional. The logs of the sensor values identify FS events and may be useful in diagnosis. The
procedure for building such a sensor using an Arduino is documented here:-

Failsafe/USM detection.

Testers will need to have some mechanism to enable them to detect and record FS and USM events if
they occur. These events may be of very short duration, too short to be necessarily evident in video
footage unless arrangement are made to make the event obvious.

Failsafes may be reliably detected by the Link Quality Sensor but USM events will not be detected by
this means. FS and USM can both be reliably detected by movement of one or more servo arms. Testers
can make use the Custom Failsafe option in the transmitter firmware to force the servo arms to move
from one extreme to the other when a FS event happens. This movement can either be detected visually
either at the time or later from the video. Transmitters should be set up with a model definition that
supports control of a number of servos, eight or six depending on the receiver model. While one servo is
enough for FS detection, USM may happen on any channel (apparently) so all receiver servo ports
should be utilised if possible.

While other methods may be available they are likely to involve the use of electronic equipment that is
not available to all. This may be used in preference however, if desired by the tester.

In the past, those testing for FS or USM have used innovative mechanisms to detect (and count) FS
events such as servos moving long arms, servo arms moving other objects, servo operated hoppers
releasing one marble or ball bearing per event and such. Feel free to propose or develop something like
this if need be.

Testers may opt for something simpler, such as an existing aircraft and setup where a FS event will put
flaps down for example.

Any setup can be used as long as it allows a FS or USM event to be reliably detected and recorded.

Log Files

All of the transmitters likely to be used for testing allow for the keeping of logs of the day’s events.
Testers should ensure that logging is set up and is operational for the model to be used. The correct
operation of this function should be checked before test sessions begin. Please set logging resolution to

Log files should be retained ready to be made available for later scrutiny. Testers may find it convenient
to record and uniquely name one log file per session. This can be accomplished by copying the log file
from the transmitter after each test session and then deleting the file from the Tx. A new log file will be
automatically created when the Tx is next switched on. This may seem a bit inconvenient but will save
time and reduce confusion later.

Receiver Binding

There is some evidence that the FS events may have something to do with binding the receiver. If your
tests fail to produce any FS events it is suggested that the receiver be re-bound. It may also be beneficial
to the process to re-flash the FRSKY firmware to the receiver as well as rebinding it and to change the
receiver number. Please only change one thing at a time, obviously if one of these things changes the
outcome it will help to know which one it was.

Record whether the receiver output is HS or FS mode where this option is available.


Layout of equipment

The transmitter and receiver should each be placed on a non-metallic surface around 1m high about 4m
The transmitter antenna should be left in a sideways orientation if a movable antenna such as in a X9D+.
Use internal antenna if one is fitted (X12S, X10 and so on).
Orient the receiver antennae in a V shape pointing upwards with the flat side of the V facing towards the
If the receiver is in a model, place the model in any specific orientation relative to the transmitter and
keep that constant for all tests. Put model on a stand or non-metallic surface 1m high.
The working environment can be anywhere convenient. Record rough details – eg urban back yard,
Urban inside building, rural flying field, etc.



Do a short test run beforehand to check that everything works as desired. Force a FS event and check
that the detection mechanism works as intended. Look at the video record and make sure that you can
use it to detect FS events if required to later, especially if not planning to observe the test live.

For every Test
Check that Tx and Rx batteries and video recording device have sufficient charge to complete the test
Place the transmitter in normal full range mode and power up the transmitter then the receiver. Make
sure that the servos/control surfaces move under transmitter control.
Make sure that the FS detection device (if used) is switched on or operational
Start the video recording device.

The test has started.

Allow the test to run for at least 20 min.
Observe and count any incidences of FS or USM. USM may look the same as FS, record as FS.
Record which of the events were accompanied by a “telemetry lost/telemetry recovered transmitter
Live recording may be replaced by later observation of the video record if required.

Additional Testing

It may be worth repeating the test with the Tx in Range Test mode if full range tests indicate no issues.


For each test record on the spreadsheet

Tester identification. Use RCG or other forum handle, email address or anything that will enable
someone else to ask questions of you in the future if required
Your Test number
Transmitter model
Receiver model
Receiver number (the number allocated in OpenTx)
Whether newly bound for the test or not (Y or N)
Binding HS/FS
Whether new firmware flashed before the test or not (Y or N)
Receiver power source (NiMH, LIPO, ESC)
Whether BEC used with LIPO (Y or N)
Servo model(s)
Test duration
Number of FS Events detected
Number of telemetry lost warnings
Number of USM events detected
Timestamps in video for any events of interest
Log File name
Any comments
Sign up now
to remove ads between posts
May 22, 2020, 08:46 PM
Registered User
mpjf01's Avatar
Thread OP
Test rig showing failsafes. From G-RX8 test 2.

VID test 2 edited (0 min 12 sec)
May 22, 2020, 08:53 PM
Registered User
mpjf01's Avatar
Thread OP
It's impossible to test for every possible scenario. The simple test procedure that we outlined as a first step was a bench test intended to identify major issues with the 2.1.0 firmware to the point that we could be confident to commence flight testing.

This identified and proved the existence of unwanted FS with the G-RX8 and RX6R. In doing this, as an example of the time involved, I tested 3 or the G-RX8 receivers with a X9D+ 2019, a X9D+, a X9D, a X-Lite, a X12S a X10S and a XJT module. I did not test non-EU firmware, Marco is doing that as well as testing the receiver operating in HS/FS modes. There is also a bit more work to do looking at the receiver when operating in a redundant setup.

I have not been able to provoke any problem with RXSR, RX8, RX6, RX4, RX8R, but I have more to do with these as I've only tested with one transmitter.

While we are each operating pretty much independently at the moment, others in the group had also been testing with other transmitters in the FRSKY range - although we don't own the full set between us.

I have yet to convert my X10S to the new ISRM RF module but I read that it's firmware is now sound - so I may do that today.

The bench testing involves putting a receiver on a stand and a transmitter at a set distance and leaving both running for 20 minutes. The receiver should be connected to something that will visually show unusual behaviour. We were particularly looking for FS and USM but there are other potential outcomes. Videos are taken and log files kept.

After testing the G-RX8 is complete, the next focus as far as I am concerned should be the RX6R, RX4R and RXSR. I only have one RX6R, the sole RX4R I have was faulty when delivered so I can't test that, and I have 3 RXSR.
May 22, 2020, 09:49 PM
Registered User
Great rig!

I agree with testing RX series - my favorites would be RXSR (for FC and redundancy use) - and RX8R PRO (for one gaser I am preparing)
But seems that RX6R and RX4R will go first ..

Then - S8R (sadly we can't test S6R with LQ - no SBUS) - not for stab needs - but for setting Auto Level in case of FS (need to check this again in specific thread..)

huh, morning again - will check - well, today, but later - need to rest - enough stupid mistakes ..
May 22, 2020, 10:08 PM
Registered User
mpjf01's Avatar
Thread OP
BTW the silver can shaped device in the centre is a mechanical counter. I have had this for so long I don't recall where it came from.

It counts up by one when it's activating lever is hit by the servo arm when FS is detected. I don't need to be present for the whole of each test.
May 23, 2020, 01:22 AM
Registered User
Ha, ha - incredible - we have similar "toys" !!

My counter comes from period between two jobs, when I worked in office equipment repair shop - although electronic cash registers and typewriters began to prevail at that time - electro-mechanical beasts were my favorites - this one comes from either old Hugin (more likely, because it is still working like a clock) or our local "Digitron Buje" - can't remember anymore ..

And here is my "rig" - inspired by Andrew's - still need to add cross member for multiple servos ..
Among "too many" other interests - I am also a railway fan (sorry - Trainz !!) , so ..
May 24, 2020, 04:13 PM
Registered User
Dear all,

Here's a link with the results of the test we received so far. Just 4 contributors... it would be nice to receive more!

It seems to me that we have 4 diferent showcases.. I think it is difficult to aggregate data until we get more.

It would be interesting to compare G-RX8 release firmware with the latest one published on Git-Hub.

Today I made several flights with the latest firmware, still rock solid in my case (ACCST. LBT) as the release one.

The link to the file is:

Ciao Ciao

May 24, 2020, 06:21 PM
Registered User
Ciao Marco!

Note taken on oficial FW for G-RX8 ..

I think I am close to iron out initial hurdles - and hopefully will send you first bach of data soon ..
May 24, 2020, 07:19 PM
Registered User
mpjf01's Avatar
Thread OP
I have added later test results to the spreadsheet. I'll be testing RX6R with different Tx today.
May 25, 2020, 01:22 AM
Registered User
mpjf01's Avatar
Thread OP
Checked the FRSKY Github for updates, downloaded the latest RX6R firmware, then 3x20 minute tests. 2 FS events detected. Got part way through editing the video and took a break.

Went to Github again to get G-X8R ACCESS firmware to help me answer a question from Andrew and discovered new firmware for the RX6R fixing the FS issue had been added (as well as the G-X8R).

Waste of a morning.
May 25, 2020, 06:48 AM
Registered User
I was just preparing to do more tests (with RX6R and G-RX8) - when few last posts here and on "alert ACCST 2.1" thread made me think ..

I was playing "smart" when answering one inpatient RC (turbine) friend that we need to be organised ..

Are we ?

I was very pleased when reading last post from @mpresi (Marco) here - that would be interesting to compare last official and last beta FW - in this example for G-RX8.. - I said - now I know what to do tomorrow ..

It could be template for future test batches ..

Then again - there would be hiccups - like @mpjf01 found new beta release just after test ..
(and I found new beta for ACCST RX6R .. - but in my case it would nicely fit in batch - if I manage to do all..)

Any one of us could just do what is most interesting for him and send data - but this is not very organized ..

Weather is calling out - in that case we will obviously fly only the appropriate version - for me LBT - others FCC ..

And I am no really keen to be beta FW tester with model in air - at least not before bench test give good enough results ..
But in bench test I could do both FCC and LBT .. (and collect more data for Marco - when needed - not just depending on my "hunch")

ACCST or ACCESS - depends on hardware that we have - I sent the list of mine to @mpjf01 in my initial PM (could do it again)..

It would be really helpful if we are in sync with FrSky - it is easy to criticise them - but they can or not read RCG - it is not their official channel..

They opened github page recently - and I have opened account on their site long time ago - but will be first to admit - there is so much (too much) threads in RCG alone - that never used it after few first (not so successful) attempts to get info ..

Is anyone in contact with them ? If we agree to test beta FW - it would be beneficial to do it in sync with them ..

In short - I would like if Marco continue with - "this is "hot" now - anyone with appropriate HW and available - do this .."
(here? or in Google? - please your guide too)

I am really enjoying this project - if this work out - finally something that will bring back confidence in my gear - thanks to all for your help and support so far ..

These are just some of my thoughts over "morning" coffee - pardon my (french) English writing skills - but hopefully you understand my thoughts - and I am looking forward to hear yours ..

back to work .. (the coffee break is over)
May 25, 2020, 10:34 AM
Let's fly !
DaySpeed's Avatar
I appreciate all the work you are all doing, this is great!

I started doing testing according to your post however, I added a RB-20 with single S8R since I will use that configuration.
I will also test without the RB-20.

I can also test RX6R, RXSR, X8R and X4RSB but I have no G-RX8 or LQ sensor.

If my results would be of interests, I could publish them.
May 25, 2020, 05:55 PM
Registered User
mpjf01's Avatar
Thread OP
I'm sure that we would all appreciate as much help as possible with what is a significant task.

As far as I know none if us yet have done much with SxR receivers but there are a few posters on FRSKY's Github with an interest, so I would check there if I was you and see what progress has been made if you have not done that already.

Testing with RB20 also has to be done, I have mine connected to 2xRXSR receivers at present but some of my servos don't seem up to the task and have failed and I have to try to work out what I have set up wrong with that. A lot more testing would be helpful.

You can send your results to Marco or post on Github, or both.

Quick Reply
Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Question Futaba 10J or FrSky 2.4GHz ACCST TARANIS X9D for FPV racing rolsen66 FPV Racing 10 Jan 06, 2016 06:50 AM
Discussion Speedix ARF S250Q w/CC3D or KK2 compatible with FrSky Taranis X9D Plus 2.4GHz ACCST R retnevel Radios 2 Nov 04, 2015 02:24 PM
Help! What receiver to get with FrSky 2.4GHz ACCST TARANIS X9D Grifftech Multirotor Drone Talk 7 Jan 23, 2015 12:23 AM