I keep getting Bad Compass health, have external compass but it does not want to calibrate in Mission Planner or QGC
No message in Mission Planner and in QGC mav_cmd(42424) command failed
Any idea what could be the problem? The compass is external on the i2c bus and is recognized by both Mission Planner andQGC
Put the Rover in a known orientation (pointed to a known magnetic heading) and try using
We can refine it later.
The standard calibration will require to you to rotate the vehicle through all orientations (including upside down).
Okay let me find my magnetic compass and do that
Have to do that in the morning GPS is bad this time of the day indoors.
Is there anything else you think I must look at?
I have calibrated the accelerometer and will do the compass in the morning.
The radio is setup and I now know where to change and setup things so to lets say change Acro to Auto it can be done
I drove the rover outside but only forward and it tracked straight and it turns smoothly. Havenât driven it with the one stick setup yet but will do so in the morning after calibrating the Compass.
The rover turns fairly quickly and rather sharp if you push the stick all the way to either left or right and nothing flies off so it doesnât look like she is pulling gâs so I left the parameters at default. when driving straight very minor adjustments can be made to left or right with slight movement of the stick which should be great for navigation.
Speed is low but deceptive and at full throttle she runs at a brisk pace and turns smoothly and acceleration is more than acceptable even on a slight incline. Throttle slew was left at default
This was all with the old two stick configuration, will test in the morning with the one stick configuration to see it there is any change.
One thing that I did notice was that when the throttle was pulled all the way back with the old configuration it took about 150mm for it to come to a stop, can this be changed?
I read about S-Curves and Position Controller, the targets are not that far apart, the longest unobstructed distance between 2 points are 9.6 meters, would that be sufficient to carry out the S-Curve and Position Controller calibration?
Is there anything I missed or overlooked?
Before I forget the Telemetry radios are not working and I cannot watch the acceleration and other data in real time anymore. I am busy looking for new radios. Could that data possibly be obtained from a LOG file? If so where in the file?
I see there are a number of these values
Thank you
Stopping behavior likely canât be changed, as thatâs an artifact of your motor controller. We do offer a parameter to apply reverse throttle for a braking effect, but Iâm afraid enabling that might cause some current spikes or other undesirable effects in those rather cheap controllers. I think weâd rather account for observed deceleration in the nav tuning than risk hardware issues.
While acro mode + live monitoring is by far preferable for initial tuning, we can do some post-driving log analysis with a log from an auto mission. Create one that makes a few turns (like a square-ish pattern). If the Rover will execute it (even poorly), we can learn a lot from the log. Upload the log and Iâll give a few pointers.
If your mission looks a little like a squared off figure-8 (or bow-tie) with turns in both directions, we may be able to refine the compass calibration, as well; especially if you can get it to run a couple of consecutive laps during the same log file.
Hi Yuri
I drove the rover myself, hope the data is usable.
Did not write the mission only had a small window of time to take it outside and in the haste missed it.
I will repeat the mission tomorrow if it is necessary, how can a figure of 8 / bow-tie be done in the mission planning?
I will try do a figure of 8 in acro mode tomorrow, need to practice it first.
The stopping behavior is not a problem and can be countered by reducing the throttle to lets say 75% when within a meter of the waypoint I think, there is a parameter for that. I also tested it in acro mode driving straight to a washer lying on the ground and trying to stop as close to it in relation with the GPS antenna, could well be my inexperience with RC, I will have it run past a few markers that are relatively easy to access when there are vehicles in the parking area
It also crawls rather well and very slowly if you are very gentle with the stick which could be handy in precision. With the way you set it up using only the right hand stick makes a difference with having reverse, would the autopilot, if it overshoots the target for whatever reason reverse to align itself with the target?
The autopilot control versus human should be much better controlled and precise and can be seen in all the videos with the lawn movers, scale the lawn movers down in speed and size and you should have a rover like this one slow and precise.
Well that is my reasoning behind the rover having watched lots of YouTube videos and much reading on the subject.
Before I forget that is if it is properly tuned which this one is not yet
Thank you
A couple of things that need setting before going farther:
Itâs never achieving 0.5 m/s. It looks like it can repeatably achieve and maintain just over 0.3 m/s. Below is the flattest segment I saw at full throttle, indicating 0.32 m/s, average.
Max available turn rate looks like itâs about 60-70 deg/s. Max observed was about 100, but those were very brief spikes - below is a representative sample.
GPS and compass data look very poor. It looks like there was once a compass detected, but I donât see evidence of any mag data in the log.
Thank you Yuri
The speed of .5 m/s is a calculated speed with the motor achieving the advertised 80 RPM on the shaft and the wheel diameter of 130 mm.
Drove the car myself so the control was not the best there was
Do not know what the GPS fix was besides the 3D fix led flashing, investing in new telemetry system but that will only be the week after next when the supplier has stock which will greatly assist.
The mag I also have no answer. Where is the p
Best position to mount it, presently it is mounted on the bar it is carried by. Possible that the wiring gets disturb while carrying and not impossible.
Can be mounted in the front and center, maybe a better mounting position?
Will have it drive in the morning again
Compass mounting isnât the primary issue. It wasnât detected at all during your last drive.
Mount it as far away from high current sources (battery, power wires, and motors) as practical.
Thank you
Will do, there are a few places that it could be mounted. Will see where it is best and out of the way
Presently there is no way that the state of the GPS fix can be seen until the telemetry radios arrives.
Going to set the GPS_MIN_DGPS to 0 for now until the telemetry radios arrives. Should still be able to tune the rover and it will as it drives also improve the GPS fix.
With reference to the figure of 8 question, I had the rounded head and foot of the 8 in mind and with recently have been reading about S curves etc. can you do rounded paths or is everything in straight lines?
How do I stop the logs being written to my computer? When I tried to send you the bin file I had no space on OneDrive and discovered that between Mission Planner and QGC more than half of my hard drive consisted of log files that I had to delete to make space for the file I shared with you. Is there a way to selectively log to the hard drive similar to your LOG_DISARMED=0
No testing today, power outage since early morning. Cannot charge the batteries
Will see later this afternoon after the rain, that is if the power is restored
Hello Yuri
Took the rover out today but I somehow screwed up the log file
Sorted now, waiting for rain to finish
Hello Yuri
Why if the log_disarmed=1 when the rover is taken outside and driven for a couple of laps the log file remains at 1MB and doesnât contain data?
Have tried it a number of time outside and it remains the same
00000004.zip (468.1 KB)
Thank you
Disable this if you are Arming the vehicle and driving it.
Hi Dave
Thank you will do
Hello Yuri
I have had the rover outside a number of times since Friday, for some reason it appears not to want to arm when in Auto mode and there is a mission written to the autopilot.
It arms in Acro mode and it can be driven around
When indoors the log file is created after deleting the previous one upon arming when LOG_DISARMED = 0 and also when powered up with LOG_DISARMED = 1
Switch SWA is configured as Acro with the switch up and Auto with the switch down and SWD is configured as Disarm when up and Arm when down.
When in Auto the rover does not arm and there is no way to see why, SWA was changed to Manual when up and Acro when down
When the rover is driver outside the log file is not updated despite LOG_DISARMED=1 or 0
Without telemetry radios it seems senseless trying to test without being able to see what the state of the autopilot is.
Why the log file is not updated is a mystery, I have had it outside on seven occasions since the weekend and drove it for at least 10 minutes on each occasion in a bow-tie pattern, first in a left hand pattern and then in a right hand pattern. Also had it drive in circles both left and right hand but every time it is brought inside to pull the log file there is either no log file or it is 1Mb in size with no data.
Reformatted the SD card and also changed it twice with new 16Gb class 10 SD cards.
I might hopefully have a new set of telemetry radios towards the end of the week
Somewhere I am doing something wrong
A couple of observations driving the it with the radio.
The area where the rover is being driven is not totally flat, when driving in a straight line it tends to turn to the side that is slightly downhill.
When the stick is centered and then pushed either left or right it does not always behave the same, the one time lets say pushing the stick all the way to the left will cause the right motor to run forward and the left motor in reverse turning the rover to the left on the spot, same pushing the stick all the way to the right causing the left motor to run forward and the right motor in reverse. When the stick is centered again the motors keep running and doesnât stop. When you then push the stick left or right the motor that is supposed to run in reverse does not respond at all.
What is also weird is that yesterday and the day before the rover was left powered by battery after driving the figure of 8 and carried upstairs to see if the log file was there
As it was carried up the stairs the motors started up out of nowhere when the top of the stairs near my apartment was reached, the radio was off and the autopilot was in Auto mode and armed when connected to the computer.
Why that happened I have no explanation.
Is there anything you want me to do or check or whatever?
Thank you
Hello Anton,
I think donât start with auto mode.
First you must be sure that all is running in the manual modes like âmanualâ or âacroâ.
If you want to arm in an auto mode the rover needs to know its position. So if there is any problem it wonât arm. Also donnât leave the rover in auto mode while carrying it to some other place. As you saw it can start motors maybe as the FC detects a wrong position ant wants to return home.
I donât think that the FC only writes log files inside your home.
With LOG_DISARMED = 0 Logging only starts after the FC is armed. So if you are in âAutoâ mode and the FC decided not to arm due to some reason you donât get any log
With LOG_DISARMED = 1 Logging is also enabled before arming so you get some information why arming is not possible.
When in Auto the rover does not arm and there is no way to see why, SWA was changed to Manual when up and Acro when down
This I donât understand Do you have 3 modes âautoâ âacroâ " manual"? How you than start âautoâ mode. The FC takes always the last command.
I think it is not a good idea at this stage to start with âautoâ mode.
If you start outside with manual and drive some pattern and than you switch by your mode selector to auto mode without disarming, what happens?
Yep, without telemetry, itâs hard to troubleshoot. Just like itâs hard to tune without RC.
Manual mode isâŚwait for itâŚmanual. It SHOULD follow the slope of the terrain, and you SHOULD have to manually correct. Acro mode (when tuned) would help correct that to some degree.
@Juergen-Fahlbusch, I offered a few tips about attempting to tune in auto mode due to some constraints presented, so thatâs whatâs happening with the attempts to use it. I often tune in auto mode, but that doesnât mean itâs the best method. There are a lot of prerequisites to get to that point, and I donât recommend it as a documented practice. I was hopeful we could âskipâ to log review of an auto mission to speed things along a bit. Itâs a dice roll as to whether that will work.
Yes, I agree, but it seems that Anton has still some general problems with arming and auto mode. From my point of view this needs to be first solved.
