LibrePilot Forum
Users => Vehicles - MultiRotors => Topic started by: micaelkarlsson on February 26, 2016, 09:22:14 pm
-
I have to ask, has anyone succeeded to build a quad with Librepilot and GPS which actually are able to fly?
I have twisted the wires (esc-pdb/esc-motors), mount a 10*10 cm MU-metal plate between esc, lipo and the revo flight controller and still it fly and behave like crazy.
I think I will try to use it without PGS and wait for the new release of librepilot.
-
I have to ask, has anyone succeeded to build a quad with Librepilot and GPS which actually are able to fly?
I have twisted the wires (esc-pdb/esc-motors), mount a 10*10 cm MU-metal plate between esc, lipo and the revo flight controller and still it fly and behave like crazy.
I think I will try to use it without PGS and wait for the new release of librepilot.
Almost everyone has. A new release won't make any difference, if the vehicle is too bad to fly now, it won't be any better with a new release.
You haven't mentioned what "behave like crazy" is nor what the quad is, how it's wired, your settings, or any other details so it's a little hard to help you.
If you put a 10x10 mu metal plate on the vehicle you may have distorted the earth's magnetic field too much for the mag to work correctly. And GPS modes are heavily dependent on the mag sensor.
-
If you put a 10x10 mu metal plate on the vehicle you may have distorted the earth's magnetic field too much for the mag to work correctly. And GPS modes are heavily dependent on the mag sensor.
I don't know why people do that. It may actually cause more harm than good.
Most problems people have are related to magnetometer. Magnetometer is not only used to determine orientation in yaw, but also to show whether aircraft is up side down. You should calibrate magnetometer not putting it on the ground and far from metallic objects (except for aircraft itself), but exactly how it would fly including battery. Someone should help you telling how to rotate it, and you should use OPLink. You may also try calibrating it in another place. Also be aware that magnetometer is recording continuously and not just when you press a button (dont put it on a car temporary while in the middle of calibration). Before saving after finished calibration you should put it back horizontally. When it comes to GPS, you should avoid flying near buildings, as multipath reflections from walls may confuse GPS adding to distance. Also you should let GPS sit with battery connected and clear sky for 10-15min to update almanac before first flight. If sats are not up to date it may be getting poor quality signal. But mostly problem is a mag. If your mag is not aligned with accel or you moved it after calibration, it does not agree with accelerometer, your model might be vibrating and you might see sin-like behaviour in attitude scopes. That means EKF (INS31) receives contradicting data from mag and accel and it's not flyable.
You should use GPS with mag on a mast, to have it away from electronics that can generate magnetic fields, like high power wires or switching components such as ESCs or BEC. Also antennas such as video, telemetry etc.. may interfere with GPS. For me I get much better and quicker sats when I disconnect by video transmitter temporarily when updating almanac before first flight.
Also it's good to do full erase (to be sure no configuration artefacts are in external memory) when updating firmware, and start whole full calibration (all sensors) process from scratch to be sure everything is right and export your uavs.
Otherwise, what can I say, it works for many people, you may not even know what you did wrong, and when you start from scratch carefuly you might get better results.
-
Perhaps I will give it another try some other day. I have spent to many hours to build and tweek the flight controller. Usually I think it's fun but the last hours I think it is frustrating. When I build with APM2.6 it was only to put it together, load firmware, do some basic configuration and then it will fly. A couple of month ago I choosed between Pixhawk and Revo. Perhaps Revo has potential but I will put it away for a while.
My reason to ask this question were to actually hear if someone succeeded to build a autonmous multirotor based on librepilot 15:09. When I search for "maiden fligt revo librepilot" I only had a few hits but when I, for example wrote the same for cleanflight I had several hundreds... But I'm glad I'm beeing wrong about it and some day I give the revo another try.
-
Perhaps I will give it another try some other day. I have spent to many hours to build and tweek the flight controller. Usually I think it's fun but the last hours I think it is frustrating. When I build with APM2.6 it was only to put it together, load firmware, do some basic configuration and then it will fly. A couple of month ago I choosed between Pixhawk and Revo. Perhaps Revo has potential but I will put it away for a while.
You can't really tell if APM does something better or not, unless you find out the source of your problem. You maybe have very specific problem and switching just about everything (heart of your aircraft) may fix it, but that does not necessarily mean there was something wrong with the firmware in the first place. Could be faulty hardware, could be your setup, vibrations etc... many factors, and if you change everything you might get it working, but you will still not know why. All those factors can happen with any flight controller or setup. I am very confident with LP firmware as it just works. Also Revo was initially closed-hardware (one factory with top component quality) and was around for many years, it's very mature now and stable, very unlikely to have any firmware bugs. It just worked for many years for many people :) and if there will be bug discovered for sure not one that prevents someone from flying :)
Does it work with OpenPilot firmware ?
My reason to ask this question were to actually hear if someone succeeded to build a autonmous multirotor based on librepilot 15:09. When I search for "maiden fligt revo librepilot" I only had a few hits but when I, for example wrote the same for cleanflight I had several hundreds... But I'm glad I'm beeing wrong about it and some day I give the revo another try.
I am not sure if Cleanflight is not older than OpenPilot, it's based on MultiWii which was first firmware designed for Arduinos, then rewritten. You are very likely to get many hits just because of that.
They are at least as old as OpenPilot. If you google for OpenPilot which LP is based on, you will see much more hits, but because its so much older project.
LP just had it's first release beginning of this year, so there are not many movies around yet. Some people still use OpenPilot.
-
I had a look at your photos and a few comments
- I recognize the mast you have, if it's the one with just single screw attaching it to the frame, it is likely to rotate invalidating your calibration. There are many better masts which have two screws that attach pole firmly to the frame preventing any rotation.
- You use rubber bumpers below FC, my experience with exactly these type of bumpers is wobbly flight, I would rather try to mount the case directly with double sided tape flat to the frame. I use nylon screws on smaller builds.
- The thick foam like double sided tape attaching the shining plate to the frame also looks wobbly and may shake in flight. Try to mount flight controller as stiff and firmly as possible to the frame.
- Lots of unnecessary weight from long cables, bullet connectors may disconnect in flight. I usually cut, solder and cover the cable in shrink-tube.
Is it F450mm like frame ? I have Flame Wheel (white and red arms) 450mm diagonal flying perfectly. Does your fly ok in other sensor fusion without GPS (Complementary) ? Does it fly well with OpenPilot ? I suggest you start with Complementary and once it flies well in that mode then you can try with GPS (enable INS).
If your aircraft flies bad in Complementary, then it won't fly better with GPS, and you must get Complementary mode right first.
-
P...When I search for "maiden fligt revo librepilot" I only had a few hits but when I, for example wrote the same for cleanflight I had several hundreds...
All of LibrePilot's "maiden flights" were made under the old project name, OpenPilot. A google search for "openpilot maiden revo flight" comes up with 15,200 hits. I think the maiden flight for autonomous flight was back in 2011? My favorite was the 20km Autonomous Fixed Wing Revo Flight in August 2013. Unfortunately someone has password protected the video of that flight on the OpenPilot site. With the current regulations in the US I don't think we could legally do that long (a little over 12.4 miles round trip) a flight here.
-
With the current regulations in the US I don't think we could legally do that long (a little over 12.4 miles round trip) a flight here.
Slightly off topic... There are vids of long range autonomous flight that use a pace car with a driver in control of the car plus a passenger in control and always in LOS of the aircraft. I think that'd be legal, though IANAAL (I Am Not An Aviation Lawyer. :) )
-
I have to ask, has anyone succeeded to build a quad with Librepilot and GPS which actually are able to fly?
Here's a brand new thread with success. If the info right there doesn't help you could also write to the poster to ask for more detailed info. https://forum.librepilot.org/index.php?topic=943
-
Thanks guys!
I made a new calibration and when I was done and put the quadcopter one the floor, the horizont gyro told me it was about 15 degrees left (or right) and pointing a bit up. I have done exactly as the video descibe for calibrating the FC. Perhaps I have some hardware problem? It seems I are the only one with problem, all other people hace success with revo/libre combo?
I hade a APM on the same frame before I upgraded it to Revo and with APM it flew almost perfect.
I have thought about change from Libre to cleanflight or betaflight and see if it is any diffrence. I have not yet test with OpenPilot.
jbarchuk: My GPS works perfect it is a UBLOX M8N, just plug and play.
Mateusz: Thank you for your comments. I will give it a try to remove the rubber bumbers and put the FC directly on the plate. The GPS antenna doesn't move a bit :-). It has four screws on the bottom plate. I think it is a 450' frame with 12'' props, 3S lipo 5000mAh, 25-30amp Esc (simonK). It oscillate when I don't use GPS but when I made the new calibration I give up again. :-)
Thank you all for the comments, it encourage me not to completly give up revon but I will order a PixHawk soon and try it instead of the Revo.
-
I made a new calibration and when I was done and put the quadcopter one the floor, the horizont gyro told me it was about 15 degrees left (or right) and pointing a bit up. I have done exactly as the video descibe for calibrating the FC. Perhaps I have some hardware problem? It seems I are the only one with problem, all other people hace success with revo/libre combo?
....
Mateusz: Thank you for your comments. I will give it a try to remove the rubber bumbers and put the FC directly on the plate. The GPS antenna doesn't move a bit :-). It has four screws on the bottom plate. I think it is a 450' frame with 12'' props, 3S lipo 5000mAh, 25-30amp Esc (simonK). It oscillate when I don't use GPS but when I made the new calibration I give up again. :-)
Hi again,
What do you mean by horizon gyro ? Gyro always drifts, that's why accel and gyro must be combined. Usually gyro is more precise but noisy and dirfts. Accel is less accurate but does not drift. You should be able to fly first with "Complementary" attitude estimation algorithm. (See attached screenshot 1). In that mode only Accel+Gyro is used.
After calibration, saving powering off/on you should see something similar. (See screenshot 2 how it should look like).
Where gyro scopes are noisy, accel gives negative z and flat curve and top right corner is what is estimated so your attitude scopes should also show more or less flat lines. If that is not the case, something is definitely wrong.
If your flight controller is dancing or vibrating while you fly, you will also confuse sensors, that's why it's better to mount flight controller directly on the plate, flat, rigid and stiff.
If you still have problems, please revert to OpenPilot, if it does not work there, try contacting your seller where you bought Revo. Firmware is definitely fine, so it's either user error or faulty hardware.
Please make a movie how it flies and post it.
-
I don't get any useful GPS / MAG values either.
Tried almost everything. At least I think, I did.
Am pretty familiar with librepilot by now. Got my 450 Heli flying perfect with a CC3d and a revo.
But...
...I'm trying to set up a new revo using GPS and onboard MAG on the desk, no esc or wiring disturbing the setup. Have all the above described problems too! :(
I'm using Beitian-880 GPS because I discovered these GPS are phenomenal fast when starting and usually have a fix indoors within a minute. On my desk I usually have 12 to 15 satellites in view. These are values i collected when working with apm and pixhawk or the u-blox software.
Attached and configured to Revo with LP the GPS needs beteen 0,75 and 2,5 hours to get a 3d-fix. Even when it shows for example 12 satellites in GC-GPS-Display it states sats used: 3.
When I finally have a 3d-lock and the gps status goes to green usually the levelling of the board goes crazy, tilts or flips by 20 or more degrees. If I move the revo, the gps in GCS immediately goes red and shows no fix, the control led on the GPS itself shows stable 3D-fix.
I don't know what do, but wait for a new release. Hoping ;)
btw. It's not the hardware, as I have all components two times and changed everything from cable to revo.
@micaelkarlsson: you are not alone ???
-
I don't get any useful GPS / MAG values either.
Tried almost everything. At least I think, I did.
Am pretty familiar with librepilot by now. Got my 450 Heli flying perfect with a CC3d and a revo.
But...
...I'm trying to set up a new revo using GPS and onboard MAG on the desk, no esc or wiring disturbing the setup. Have all the above described problems too! :(
I'm using Beitian-880 GPS because I discovered these GPS are phenomenal fast when starting and usually have a fix indoors within a minute. On my desk I usually have 12 to 15 satellites in view. These are values i collected when working with apm and pixhawk or the u-blox software.
Attached and configured to Revo with LP the GPS needs beteen 0,75 and 2,5 hours to get a 3d-fix. Even when it shows for example 12 satellites in GC-GPS-Display it states sats used: 3.
When I finally have a 3d-lock and the gps status goes to green usually the levelling of the board goes crazy, tilts or flips by 20 or more degrees. If I move the revo, the gps in GCS immediately goes red and shows no fix, the control led on the GPS itself shows stable 3D-fix.
I don't know what do, but wait for a new release. Hoping ;)
btw. It's not the hardware, as I have all components two times and changed everything from cable to revo.
@micaelkarlsson: you are not alone ???
suzali, as far as I understood he wrote "It oscillate when I don't use GPS".
It would be good to confirm that you got your aircraft flying in Complementary mode and you only have problems with GPS+Mag ? Complementary does not use Mag or GPS. INS31 requires both Mag and GPS correctly functioning. So to start with most critical things
- Mag should be calibrated outdoors, not on desk. Far from metalic elements, using oplink. Mag is calibrated correctly if it remains green all the time, when you are outdoors. Switching orange/green is not acceptable.
- OP GPS v9 unit is rotated at 0 deg and works fine by default, other GPS units may need rotating mag by 180 deg to match with accelerometer, depends how mag is located in your GPS unit. Misaligned mag with accel will lead to oscilations.
- Built-in mag is not recommended.
- GPS must be used outdoors, and only outdoors preferably far from buildings. Multipath reflections from walls can result in measurements that once show position of 10m in one direction or the other. Also GPS must be given ~15min of completely clear sky to update its almanac. This is because most units use super capacitor which lasts 5-6 hours when charged. It's not a GPS in your phone which uses WiFi and cellular network to increase precision or fetch sats. Many people don't know this.
You must be aware that it takes only 1s for a lock when both Mag and GPS become green, but this might be by accident and you wont be able to fly indoors with INS31.
When you do calibration outdoors, and get stable (not switching) Mag and GPS you should see in Scopes steady straight lines in attitude.
If you don't get stable signal there, then something is wrong, but the least thing in the very long chain is the firmware.
-
...Attached and configured to Revo with LP the GPS needs beteen 0,75 and 2,5 hours to get a 3d-fix. Even when it shows for example 12 satellites in GC-GPS-Display it states sats used: 3.
If your gps is reporting to the FC that it's only using 3 sats then the FC won't use the info, the default LP settings are to wait for the gps to be using 7 sats before it relies on the data. That's set on the system tab in settings - gpssettings. If I'm working indoors I'll often temporarily set the number of sats lower and minimum DOP higher while I'm working. I wouldn't fly with flaky settings like that but indoors even a full size gps has problems getting and maintaining a lock.
The number of sats your gps is using has nothing to do with LP. Nor does the time it takes to get a 3D fix, that's all in the gps itself.
When I finally have a 3d-lock and the gps status goes to green usually the levelling of the board goes crazy, tilts or flips by 20 or more degrees. If I move the revo, the gps in GCS immediately goes red and shows no fix, the control led on the GPS itself shows stable 3D-fix.
There is no "3D-fix" led on a gps, the led is actually "timepulse" and by default it outputs 1 pulse per second as soon at the gps has enough of a marginal fix to read time from the satellites. That's way before it has a good enough lock to navigate by.
-
I think I probably could have it flying rather good in Complementary mode if I tweak the PIDs.
The flight controller and the thin mu-metal plate is tight togheter with the frame but I didn't removed the rubber dampers. I really don't Think thats the issue.
I gave up and ordered a complete kit based on Pixhawk 2.4.6. I think I will give it a try. I put the Revo away for a while :-)
I already have two quadcopters flying really good with APM 2.6 and now it's time to check Pixhawk out :-)
-
Another successful movies, author says "VtolPathfollower all defaults"
Position Hold
https://www.youtube.com/watch?v=03t_RqmJHsQ
Autonomous flight
https://www.youtube.com/watch?v=7Z_fjSfCmbM
Sounds pretty windy, but looks good.
I think I probably could have it flying rather good in Complementary mode if I tweak the PIDs.
Regarding your case though, I don't see how GPS would fly any better, if it does not fly well in Complementary in the first place. If your control PIDs are off, no matter what you do with sensors they won't be any use if your control loops are not tuned.
-
Another successful movies, author says "VtolPathfollower all defaults"
...
You don't recognize the backyard from other videos? That's by Laurent, f5soh in this forum.
-
You don't recognize the backyard from other videos? That's by Laurent, f5soh in this forum.
The point is that all those modes work in LibrePilot 15.09 and there is nothing wrong with the software. What might be incorrect though is expectation that GPS is able to magically fix everything, including tuned PIDs for control loops. It really must fly stable in Complementary mode in the first place, before adding more sensors. More sensors wont help if your control is lousy. The LP software is pretty advanced and sometimes some tuning is necessary. It is the case for every build, that sometimes default parameters are not optimal for this build, as you can't just cover all, and require some extra care. If it works out of the box well lucky you, if it does not then it needs some patience and learning. Maybe some other system will have default parameters fitting your build, but you don't learn to solve problem this way, instead you changed conditions without knowing why it didnt work in the first place. If the problem happens again with other flight controller using similar algorithm (very likely uses PIDs and EKF), you still can't solve it. If I wanted an aircraft that is plug&fly I would buy DJI, guys there are really smart and make awesome products. It's just not my cup of tea. If you want help on forum, please ask you are welcome and everyone here is glad to help, but if you come to say it does not work, I give up or buy another product... it feels hopeless and strange ;)
-
You don't recognize the backyard from other videos? That's by Laurent, f5soh in this forum.
8)
I need to choose another place to fly :D
-
I'm using Beitian-880 GPS ... If I move the revo, the gps in GCS immediately goes red and shows no fix,
GPS problem (indoors there are multipath issues and you may loose your fix) or wiring problem (connection comes loose).
Red simply means that the GPS is telling the FC that it doesn't have a usable fix.
-
I'm old and stuborn. I put togehter a hexa copter during weekend (kongcopter) and attached the flight controller to it. I have adjust the pids (read another thread here in forum) but I have the feeling it behave exact the same as with my small 450 quad. This weekend I was out and flew my APM2.6 based quad and brought my hexa but after a friend looked at it and we did some testning. I hold it over my head and he tried to controll it but we both had the feeling it it would crasch. Perhaps it is the antivibration plate who make the motorn sudently speed up and behave nervously?
Could i post a logfile here and someone could take a look at it?
-
Not sure this is the best place for APM log analysis.
The sure thing, regardless of board/system used: Never hold a multicopter over your head, that not safe !
-
Not sure this is the best place for APM log analysis.
The sure thing, regardless of board/system used: Never hold a multicopter over your head, that not safe !
No, I didnt mean the APM log. I meant Revo log. If I run the motors and collect some info in the logfile. If someone could help me analyze them.
-
This weekend I was out and flew my APM2.6 based quad and brought my hexa but after a friend looked at it and we did some testning. I hold it over my head and he tried to controll it but we both had the feeling it it would crasch.
Sorry i understand you flew a APM with your friend :)
Whats your issue ?
-
ok just to give my experience
i managed to get it to work
https://www.youtube.com/watch?v=9aCUt8MUrMo
Openpilot revo running 15.02.01
V9 gps , zuul frame, 1900kva motos, 6" props, 4s, Kiss esc.
GPS is enabled as assisted flight option,
this way its only active when sticks are neutral.
allows me to fly in my usal mode, and have a safety net if i lose control :-)
it was a long process to get it working ,
it came down to getting the calibration process done 100% and keep the orientation of the quad consistent through out the process.
go to an open area to get a faster GPS fix.
doing all this multiple time using a laptop in an open filed is not ideal
so an android GCS will really really help.
-
Unfortunately GPS setup is a picky technical thing to get it right. A correctly set up quad with an easy GPS flight mode is much easier to fly than a quad without GPS, but you need to get it flying without GPS before adding GPS, and it is very important to be able to fly it without GPS in case there is a problem.
Instead of using Attitude mode with GPS Assist, I recommend that you try Velocity Roam mode. Truly anyone can fly that the first time they try.
-
...The flight controller and the thin mu-metal plate is tight togheter with the frame but I didn't removed the rubber dampers. I really don't Think thats the issue.
Have you tried mu metal with an external mag ? Mu-metal is supposed to do an excellent job in magnetic shielding
-
Mu metal is a bad good idea but in our application we still need to measure the world magnetic vector.
Put external Mag away from strong currents and mag is not disturbed anymore.
-
Hi Lazaros !
No, I have not tried external mag. I have a UBLOX m8n and if I understand it right I should wait for next release of libre to be able to use that external mag wich is included in the GPS.
However I bought a Pixhawk and have mounted it on the hexacopter and what I can see so far everything is ok. Everything works perfect out of the box, same as with my two other APM flight controllers. Pretty easy to install and configure and they work with GPS. I admit I'm not enough skilled to build and configure librepilot based multirotors.
Perhaps I save the Revo to another day or perhaps I sell it. It's soon summer here I want to spend most of the time outside and fly.
-
Regarding your Revo problem.
The LibrePilot software uses 3D magnetometer (all three axis of space), in contrast APM uses only 2D magnetometer.
It makes no difference with Complementary, but EKF benefits greatly from getting more data as it makes estimates more accurate. On another hand, less data as in 2D case, makes algorithm more robust to noise.
If you don't include noisy measurements, your estimates won't be affected. The EKF is designed to deal with noise but to certain level, it just mixes best of all sensors. If one of your sensors gives total crap it just wont make it better. As people hinted on forum, your nu-metal may distort subtle magnetic earth field, your internal magnetometer might be affected by switching components or power wires.
The solution for this is either not to use EKF, or use external magnetometer, twisted wires if necessary, proper outdoor calibration and no nu-metal thingy, that may distort subtle earth field.
-
....no nu-metal thingy, that may distort subtle earth field.
Even if you use only external mag and use mu-metal to shield distribution boards and ESCs ?
If the mu-metal is located well below the external mag (15-20cm) will it still affect earth's
magnetic field at the point of measurement of external mag?
-
Even if you use only external mag and use mu-metal to shield distribution boards and ESCs ?
If the mu-metal is located well below the external mag (15-20cm) will it still affect earth's
magnetic field at the point of measurement of external mag?
I have no definite proof whether it helps or not. But also I haven't seen any methodological experiment showing that it is beneficial for magnetometers detecting earth magnetic field. The only source that mentions using those in that combination I found are anecdotes on forums.
Since there is no logical explanation, and all reasonable sources of information suggest that it may only distort magnetic field, I would not use it.
You can't isolate (shield) from magnetic field easily.
Only options are redirecting it in a smart way, staying away from it or putting 3-4 meters of concrete block wall like at large hadron colider. Earth magnetic field is static and way more subtle than what wires or switching components produce.
I am not sure how big is your sheet, but try downloading digital compas application to your phone, and placing it next to that sheet or pice of metal. Does it point north correctly and in a stable way or freaking out ? If you have your setup working, don't bother, but it wouldn't be my first choice to start with, if I had magnetic field interference problems.
-
I am currently in the process of building my own 3D printed quad..still collecting
components...and trying to organise space/room to fit them :)..therefore I am
by no means an expert, thats' why I am asking after all 8)
Thanks for the valuable feedback...and sorry for hijacking the thread :) :)