Thread Tools
Sep 08, 2019, 02:46 PM
Registered User
chupo_cro's Avatar
Thread OP
Question

RX lost or 5 V for FC and RX lost?


My quadcopter was working well for a year when recently motors started to stop at least once during every flight (it flies 9 to 14 minutes with one 4s 1300 mAh nano-tech). After that there is red flashing light on the FC and I have to disarm-arm to start flying again. I disassembled everything a few times, changed the wires, cleaned the PCBs etc. and still the same so I recorded the log using the built in dataflash and I can see RX Signal Received goes from 1 to 0 just before the motors stop and a few tenths of a second later RX Flight Ch. Valid goes from 1 to 0.

On the other hand, Failsafe Phase doesn't change from 0 and that is what I don't understand. I expected the Failsafe Phase 1 should start after the RX signal dissapears and Failsafe Phase 2 should start when motors stop but that isn't the case :-/

Here is the log of a flight that according to Taranis Q X7 lasted 1:31 and according to the flight log it lasted 1:44

The moment when the motors stopped is shown on the screenshot and about 4/10 of a second before is the moment when RX was lost. The last gyro bumps before the end of the log are when the quadcopter touched the ground.

Can anyone see something helpful in that flight log which would help to determine what caused the motors to stop?

I thought the FC didn't log the failsafe phases because it maybe lost the Vcc (together with RX) but then it wouldn't log the gyros after the motors stopped and besides there weren't any RX voltage dropouts in the Taranis telemetry (RxBt- is OK).
Last edited by chupo_cro; Sep 11, 2019 at 05:08 PM.
Sign up now
to remove ads between posts
Sep 08, 2019, 05:06 PM
Registered User

Check connections


I had a capacitor ground solder joint come loose but was hidden under the glob of solder. Vibration causing dis/re connections of the capacitor. Also took a look at wiring and saw lot of dirt and grass some insects. Cleaned the board re soldered everything and this year old quad flies like new. Accept for the loud beat up motors all is well. JST connections are not reliable I have found as times goes on.
Sep 08, 2019, 07:52 PM
Registered User
chupo_cro's Avatar
Thread OP
Maybe it's really because of the JST connector. I'll try to solder SBUS wire directly but the only place to solder the wire on XSR-M receiver is the JST's pin which is quite tiny. Even if I use hot air gun to remove the JST connector the solder points would be thinner than a wire. The same problem is with the UART2 connector on the flight controller. UART3 does have solder points but I am already using it for SmartPort. I could exchange SBUS and SmartPort UARTs to at least solder the SBUS wire (which is more important than SmartPort) to the FC.

I've already changed all other wires and cleaned the PCB. When I bought that FC the PB12 and PB13 pins of the STM32F3 were shorted and the dataflash didn't work. I was using this FC for a year without the dataflash and now when I was examining the PCB with the magnifier I saw the two pins were shorted (I know they were shorted at the factory because the dataflash didn't work even before I soldered anything to that PCB). After removing the short the dataflash started to work so I was finally able to record the log.

But still, I don't understand why there isn't any failsafe change in the log file after RX signal was lost :-/
Sep 09, 2019, 12:56 AM
Registered User

Hot VTX nearby Receiver...?


One more shot, VTX got older and running hotter as did my reciever sitting on top of the 30x30 stack mount VTX. I moved it away and shrink wrapped the receiver. And problem similar to yours stopped. Does flight controller fail safe when receiver looses power or locks up. Or does the receiver command a fail safe event. If so loss of power or lock up of RX wouldn't log a fail safe. GIT hub has a failsafe list and loss of power/lockup of RX is not on it. Sorry this ended in a question lolz. I'm having issues as well. cheers
Last edited by WARLocK69; Sep 09, 2019 at 02:31 AM.
Sep 09, 2019, 06:50 AM
Registered User
chupo_cro's Avatar
Thread OP
I have VTX and FPV camera but I am still practicing flying LOS and didn't mount the VTX nor the camera.

I'll try to disconnect the SBUS connector while armed to see if that triggers failsafe. I'd expect missing SBUS signal from the RX or data packet with wrong CRC should trigger it. Here it says: "Betaflight goes into failsafe when a switch turns on the FAILSAFE mode, when a receiver sends a ‘failsafe’ flag, or when any control channel signal is outside the allowed range.". The question is if missing control signal is considered to be 'out of range'.

There are similar problems that maybe can be solved by updating the RX firmware described here but my quadcopter worked well for a year before this started to happen. The only thing I changed about the time this started to happen is to configure an additional AUX channel for enabling the Airmode.

And strangely, I have a feeling that happens more often when I am using 3 of 4 batteries that I have (3x4s 1300 mAh Turnigy nano-tech and one 4s 1300 mAh Graphene).

BTW, what exactly is the purpose of activating the failsafe using the switch when you could just use the switch for disarm? In which situation would one want to use the switch for activating the failsafe instead of disarming the quadcopter?

What issues are you having?
Last edited by chupo_cro; Sep 09, 2019 at 06:56 AM.
Sep 09, 2019, 07:05 PM
Registered User
chupo_cro's Avatar
Thread OP
Here's what happened when I removed the JST connectors from the receiver and FC to solder the wires directly. I removed the connector from XSR-M and soldered the wires, then I removed the JST from SP F3 and soldered the SBUS to UART2 RX but... :-) Before I glued the wire the contact (which is even smaller than on the receiver) broke. I couldn't scratch the surface to solder the wire because there was VIA right next to the where there was a pad. And it was a blind VIA so I couldn't trace the signal so I had to chech all possible UART2 RX pins on STM32F3 to find the right one but every wire was too thick to solder it directly to the pin of the processor. Then I soldered 0.1 mm wire to the UART2 RX pin and used STM32's top as a 'soldering surface' to solder the normal wire to the one I soldered to the pin, glued everything together and it works :-) I really don't understand why they didn't put soldering pads instead of those JST connectors :-/ Today I'll test if that solved the problem.

Here are the pictures:
Sep 10, 2019, 08:12 AM
Registered User

TX RX thru holes


Aren't those TX RX thru holes by the removed JST a duplicate of the of the JST connection. I've soldered to a damaged JST just like that. Some flux and iron at 300 you can touch the board and not damage it. It can handle more heat than the solder. Flux should make solder wick and make a connection. Great job though. Oh wait I see now that's your only connection to RX 1, nice job scrape some silicon away and expose something to solder. I'm gonna give that a whirl if the issue comes up again.
Last edited by WARLocK69; Sep 10, 2019 at 08:28 AM.
Sep 10, 2019, 09:31 AM
Registered User
chupo_cro's Avatar
Thread OP
Unfortunately UART2 RX is not exposed anywhere except through the JST connector. Those through hole UART pins are UART1 and UART3 and UART1 is the same as on the USB connector so it can't be used while connected to the Betaflight Configurator and there aren't any UART2 TH contacts on the PCB.

There was nothing to scratch because the UART2 RX pad for soldering the JST connector goes straight to the blind via and a few mils around the via is already the ground plane, that can be seen in the last two pictures. The only thing I could (maybe) solder to that via would be the same 0.1 mm wire but it was easier to solder it to the pin on the STM32.

That could be a really good PCB if they put the solder pads for soldering wires insetad of SMD pads for soldering the JST connector. I bought it to build a cheap racing quadcopter for learning LOS acro and I thought the quadcopter will last maybe just a few flights. My plan was to use a cheap quadcopter for learning and then to build something more expensive but now I really don't see any reasons for using F4 or F7 FC. Faster ESC protocols, faster looptime and lower latency might be important for racing competitors but for 'normal' flying F3 FC is more than enough. These FCs are now only $15 with barometer and magnetometer and only $10 without them so I'll buy a few more.

I still didn't test if soldering the wires directly solved the problem, I'll try to do the tests today or tommorow.
Sep 17, 2019, 02:54 PM
Registered User
chupo_cro's Avatar
Thread OP
Quote:
Originally Posted by WARLocK69
I had a capacitor ground solder joint come loose but was hidden under the glob of solder. Vibration causing dis/re connections of the capacitor. Also took a look at wiring and saw lot of dirt and grass some insects. Cleaned the board re soldered everything and this year old quad flies like new. Accept for the loud beat up motors all is well. JST connections are not reliable I have found as times goes on.
You were right!! :-) The problem was one of the JST connectors. I don't know which one because I removed both (from the receiver and from the FC) but since I soldered the wires directly the quadcopter was in the air for about 1.5 hour and the motors didn't stop even once.

Thank you for your help!

Regards
Sep 17, 2019, 09:49 PM
Registered User
You went above and beyond buddy... glad to see your flying... I just meant the plugs themselves not the plug ins... although I have got a few of those broken off from crashes. Some were the only way to connect and some had a copy solder pad. But finding the pin on the CPU and soldering to that... NICE. Some paths to the CPU might have a diode, invert-er, or resistor. Saw ya had some skills though and thought you would be flying. Congrats...!
Sep 18, 2019, 01:24 PM
Registered User
chupo_cro's Avatar
Thread OP
Thank you :-) I work with electronics for years (mostly designing embedded devices from scratch - circuits, PCB-s and firmwares) and I was pretty sure there was nothing in series or parallel to RX pin of the processor. For example, if it was I2C pin then there might be an externall pull-up resistor connected to Vcc (but not necessarily because thare are internal pull-up resistors too which can be activated in software) and in that case there could be even a capacitor (depending on bitrate). And there are only 4 choices of pins for UART2 RX on STM32F3. After I located which pin was used it looked as the trace from that pin is probably going somewhere below the chip where there must be a blind via connected to the trace in the layer 2 or 3 (because the via next to the RX pad was blind too and since the plane in top layer is ground plane, there is probably a ground or a power plane in the bottom layer). I wish I was that good in flying LOS acro as with electronics :-)

You said you have some issues with your quadcopter too. I asked you what issues are you having in one of the posts (I once posted two messages in a short time here in this thread so you might have missed that message), maybe we can solve it?
Sep 22, 2019, 09:57 AM
Registered User
My problem was sudden reboots and the quad dropping 50 meters onto cement. Capacitor disconnection hidden under glob of solder. I determined the vibrations caused 100s of dis/re connections of the Capacitor. Also alot of unplugging replugging of tiny JSTs caused insulation failures near the plugs themselves. I'm a long sighted INJUN just turned 50. Gotta have glasses to see anything that small. Once I got some reading glasses to work with, I'm back flying. Father was an IBM engineer. In the 70s he'd wake me up in middle of the night to go fix a computer and install state of the art 15mb hard drives the size of a small refrigerator.
Sep 22, 2019, 02:48 PM
Registered User
chupo_cro's Avatar
Thread OP
:-) I didn't see such a huge hard disks myself back then when they were used but I do remember them from books and a TV. I started with ZX Spectrum in early 80s programming in Z80 assembler when I was 10 (and later learnt 6502 on Commodore 64) which was an excellent starting point when I 10 years later started using microcontrollers. But similarly as you, I started to learn electronics already in 70s too, using some learning kits which I still have.

My first hard disk was 252 MB. It was at the time hard disks were mostly 40 MB. And it still works, I did raw backup of that disk a few years ago. The sound was as if something will break apart but the disk worked and I can recognize file names and contents of some files with the hex editor. The computer was Cyrix 386 DX 40 MHz with 1 MB RAM :-)

You said the capacitor was the cause of the FC reboot because of the vibrations. That might be the reason but maybe there was a cold solder on the battery connector contacts which you maybe resoldered when repairing the capacitor? It is similar to my problem which was solved after removing both JST connectors but now I will never know what was the cause (RX JST connector, FC JST connector, the plugs, the wires, ...).

However, I now have another problem because of very stupid mistake. I started to practice rolls but forgot to activate airmode and the drone stopped rotating after about 45░ roll when I cut the throttle and it fell on the road. I thought the 'motors stop problem' returned and did the same one more time and then the drone fell on the battery strap. Then I realized I forgot to activate the airmode and after that the drone was flying well. But when I (after the flight) wanted to remove the battery I saw all PCB-s are hanging on the motor wires (it is 4 in 1 ESC and the motor wires are holding the ESCs preety well, everything works well even without the screws) because all 4 standoff screws were broken.

I fixed the standoffs and during the first test flight I noticed a very short drop of the throttle (maybe just a few 10ths of a second). Then I one more time activated the dataflash logging to try to see if it is 'RX lost' as was before and this time I noticed a short throttle up. When I checked the flight log I saw there is a Setpoint [throttle] = 98% event at the time when I did not move the throttle stick.

So now I have to solve another problem - occasionally valid but wrong SBUS data on throttle channel. My first guess would be the potentiometer in the transmitter used for the throttle axis but the transmitter is pretty new (Taranis Q X7) which still wasn't working for more than 20 hours. On the other hand, SBUS has a simple packet validation (just parity and 2 stop bits) and a bad connection could generate the valid packet with the wrong data so I will (once more :-) )disassemble and reasemble everything from scratch. I might even swap the UART channels for SBUS and telemetry (Smartport) to move the SBUS channel to more reliable pads and the telemetry to the port that was damaged.

BTW, I soldered the 0.1 mm wire without a magnifying glass :-) I sometimes use a small 2 cm diameter mini loupe to check the connection after I solder it.


Quick Reply
Message:

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Question 5 v power for vtx from rx? William Bosco FPV Equipment 5 Dec 03, 2016 11:41 AM
Question FC or RX to control a servo? Desync Beginner Multirotor Drones 8 Jan 24, 2016 01:40 PM
Discussion use 5.8ghz a/v tx/rx designed for home use for FPV? rudyauction509 FPV Equipment 10 Nov 01, 2013 10:05 AM
Mini-HowTo 105.5˚ or 110˚ V-Tail Clamp & "Universal" Deck Mount for Top Mount V-Tails 8thelephant Sailplane Talk 6 Jul 28, 2013 08:10 PM
Discussion Which FC to get? (AIOP v Ardu 2.5, RCTimer v WiteSpy) mageus Multirotor Drone Electronics 15 Jan 24, 2013 03:33 PM