(Can't attach files - page says "no more attachments are allowed today" - I'll try to email or upload elsewhere and link - also the 3.4.086 ikon software has no "support request")
Welp. Maybe it's opentx instead of Ethos or who knows. I'm just about done with FrSky FBus + YGE telemetry. It's still "failsafe"* city for me. But if it's "failsafed"* & I disable YGE telemetry in the iKon2 app, control is instantly restored. (More specifics on testing/behavior at bottom)
*failsafe in quotes b/c that's what's in iKon2 event log, but on recent FW testing I have not even once lost a Tx/Rx link, and all telem freezes with last values so I don't even get "telemetry lost" errors. Rx's remain "green" & transmit their on board telem.
Whatever's going on seems to be related to comms between Rx & iKon2 CH-3, and YGE telemetry.
Setup:
- FrSky x10s Express running opentx 2.3.15
- PMU was flashed with an autoupdate FW file to fix external module being powered during charging, but the FBus issues existed long before that
- ISRM is up to date (Pretty sure it's 2.1.6 but I killed the Tx batt testing)
- Archer RS receiver on 2.1.10 bound in Rx position 2
- Archer R4 receiver on 2.1.11 was bound in Rx pos 1 but had a little incident unplugging stuff while powered - DO NOT DO THAT - LOL - anyway, same issue as the RS before I vaporized a servo V+ trace
- when registering these RX's I indexed the "UID" on the register pop-up screen for each one if that matters
- both Rx's were registered & bound but were never actually operated concurrently - so no telemetry conflict there
- failsafe for ISRM set to "no pulses"
I also just added a 900MHz module for redundancy (but issues predated this):
- R9M 2019 module
- R9mx receiver
- sbus out going to sbus in on the Archer RS
- sbus out previously went to Archer R4 up until incident above
- I've also also tried the RS/R4 w/o R9mx connected
- mod & Rx FW (FCC versions) up to date as of a week ago when first got them (again, Tx batt dead so don't have exact #'s)
- FrSky GPS ADV sensor plugged into R9mx s.port
- R9mx set to s.port to utilize sensor
- R9mx had been set to FPort when trying to use GPS sensor with Archer RS/R4 using FBus (this is a major reason I was wanting to use FBus)
- updated sensor with 1.6 & 1.7 FW versions
*Just in case this can help anyone else: when updating sensor through radio s.port FW file needs to be in firmware folder - NO SUB FOLDERS like all my previous module & Rx OTA updates - and/or I shortened names to gps16.frsk & gps17.frsk (I moved/renamed at same time after countless "data refused by module" errors & massive frustration, so don't know if doing both are required)*
- "failsafe" issues persisted with & without GPS sensor
- R9M 2019 failsafe set to "no pulses"
ESC:
- YGE Saphir 125A
- FW V1.03490 as of ~Oct (I see no newer avail FW)
- Heli - ext gov
- YGE telemetry
- telemetry cable split with custom harness:
- telem wire to iKon2 4-5-6 port/pin 5
- RPM wire to iKon2 gov port
- GND wire tried both floated & to servo bus GND
- main PWM/BEC to iKon2 CH-1
- second BEC cable to iKon2 CH-2
- ferrite rings installed on all 3 cables ~3-5ish turns near plug ends
- BEC initially at 8 volts & now 7.5 volts
iKon2 std size base model:
- no remote USB or Bluetooth module added
- 3x Torq CL1208 cyclic servos at 760μs/560Hz
- 1x Torq BLS0704T tail servo at 760μs/560Hz
- also tried with and w/o a 35V 2200μF cap connected with RPM wire to GOV port
----------------
Behavior:
iKon2 FW 3.4.059
- FBus selected in advanced receiver page
- I THOUGHT this was working fine but it may have actually been a part of the terrifying fly away I had. I think this was throwing a "throttle gone in failsafe" event very intermittently, but NOT an "Rx signal gone" event
iKon2 FW 3.4.081
- FBus selected on wizard's receiver page
- not flyable at all
- constant stream of "throttle gone in failsafe," "Rx signal lost," "Rx signal ok" events
- very visible on the bench b/c it was constantly alternating between neutral pitch & whatever pitch the Tx was actually commanding as the Rx signal was lost & restored (results sucked but at the very least the iKon2 seemed to be doing exactly what it was supposed to when losing Rx signal)
- event log also peppered with ESC voltage & temperature protection events
- sensor/telemetry lost warnings on radio
iKon2 FW 3.4.086
- ESC protection events gone
- everything working fine until it isn't
- no way to reliably recreate events
- Rx failsafe
- then Rx signal gone
- then throttle gone in failsafe
- DONE
- whenever the problem hit the servos would remain frozen in their last positions
- only indication of problem on radio was that all ESC/iKon2 telemetry would freeze
- used frozen telemetry as an advantage by setting up an alarm & stopping a timer when the BEC current remained the same for 3 seconds
- problem occurred all over the place anywhere from ~seconds up to ~6-7-13-14 minutes
- thought maybe something was timing out due to inactivity so I set up a timer switch & mix to constantly vary the pitch channel between it's minimum and maximum
- seemed like this would extend the time before freezing but I think it was all confirmation bias
- only surefire way to resume normal behavior was to power cycle everything
- switching the Archer RS receiver between FBus & FPort would occasionally reset things but it wasn't really reproducible
- having USB connected seemed to cause the issue much more quickly
- turning off telemetry on the gov page would instantly restore normal operation w/o fail
- With regular FPort I was up to ~35 minutes w/o an issue before the Tx battery died on me.
So. Yeah. I'm willing to do testing so I'm taking suggestions. But really don't want to flash Ethos ATM. That's about my only limit b/c it seems like FPort is workable & I want to fly.
Regards,
Steve
Here are a bunch of event logs & configs. Apologies if there is a bunch overlap. I'm not sure what data gets saved in the background for support purposes so I was trying to be very thorough.
<a href=" removed link " target="_blank" rel="noopener">Files for support shared on Google drive
Regards,
Steve
Cut & paste link: removed link
Edit: link to get around auto removal b/c I don't know if you can see original:
drive(dot)google(dot)com/drive/folders/1vZWWCd4qRRm2P-3RlmcL0IpoinLNmxtO
Thank you for the report and the thorough testing
But especially thanks for the files sent, which are essential to try to reproduce the phenomenon and be able to do the necessary analysis.
You are very welcome!
Let me know if you'd like me to flash anything & get more event logs and/or specific logging parameters if they'll help out. Especially on the 081 fw since I didn't save an event log. (You can also reach me through Facebook messenger for some more real time interaction/testing if that'd be beneficial - I copied my first post to forum in the MSH group under "Steve Berger" to reach the folks that don't follow the forum if you need to find me - then we can work out a time since I'm in California)
Ever since moving to FrSky I've been using integration & pretty much forget to save logs when I finally do connect USB. (I'm about to post a semi related feature request in the appropriate forum)
1 note about the 059 logs: the crash was hard enough that the inertia of an Ec-5 connector disconnected the battery for me. But there was definitely that first failsafe before the "output limit reached" train usually accompanies crashes. I try to not have rotation rates high enough to cause these warnings during regular flight, but a violent spool up associated with bailout after the first failsafe may have caused the "rudder out limit reached."
Again, please don't hesitate to ask for specifics. I've got the time & want to help make brain/iKon the best it can be.
Regards,
Steve Berger
Does the recent Brain firmware 3.4.094 release resolve the FrSky Fbus Failsafe issue?
After a little testing:
Having USB & app running causes a loss of ESC telemetry relatively quickly (~2-8 minutes) & only a power cycle fixes it.
Very first test run with app running I also saw one Rx failsafe event, but no throttle gone in failsafe & control didn't lock up like in previous versions.
Without USB & app it appeared to run fine for over an hour.
Config & event log attached. Most recent Rx failsafe was first test, then hour long w/o USB, then a USB powerup & initialization, followed by 2 runs where telem was lost with USB connected. I flipped the setups soon after the telem loss if that's of any use.
Regards,
Steve Berger
Testing myself on my Protos 380. YGE 65LVT, iKON2, Archer RS, X20S and ISRM all latest firmware.
Everything seems to work fine except as srberger mentioned, USB (including remote USB module) seems to overload something. Perhaps the F4 CPU cannot handle so many UARTs at once, especially with F.Bus being a faster baud rate?
Anyways. All sensors seem to work mostly reliably with latest Brain FW, except one. VFAS (I renamed to VBAT but same thing) will often drop out for seconds at a time before returning.
Occasionally, it will drop out long enough to exceed the "sensor lost" delay (default 5s in Ethos, not sure about OpenTX). In that case, it will say "sensor lost", before after a few seconds (up to 10s sometimes) the sensor returns. Happens a few times per minute in my testing.
Also when running the integration script, I hear sensor lost a lot more often. Not a deal breaker since it's understandable that sending a bunch of data over telemetry might make the actual telemetry sensors drop out, but this happened on the OpenTX script and eventually was fixed, so maybe it's fixable here too?