LibrePilot Forum
Users => Vehicles - Other => Topic started by: karla on April 14, 2017, 05:49:26 am
-
Hello,
I have decided to try if I can use Librepilot FC for a bi-copter / fixed wing vehicle.
This will be a fusion between a fixed wing (with elevons) and a bi copter.
It should start/hover like a quad/tricopter and be able to make transition in to fixed wing flight and back again.
There are several examples of these flying vehicles on the internet using KK2 FCs. However, I want all the nice autonomous gps functions of Libre like waypoints etc. My main purpose is to avoid crashes during start and landings with my fixed wing. I am pretty clumsy and hope this will make it easier.
Today I have only some ideas on how it could be possible. I will put it to test in practice.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=5730;image)
These are the parts.
And this is the current UAV file i start out with.
-
Hello,
I have decided to try if I can use Librepilot FC for a bi-copter / fixed wing vehicle.
This will be a fusion between a fixed wing (with elevons) and a bi copter.
It should start/hover like a quad/tricopter and be able to make transition in to fixed wing flight and back again.
There are several examples of these flying vehicles on the internet using KK2 FCs. However, I want all the nice autonomous gps functions of Libre like waypoints etc. My main purpose is to avoid crashes during start and landings with my fixed wing. I am pretty clumsy and hope this will make it easier.
Today I have only some ideas on how it could be possible. I will put it to test in practice.
(http://PartsSetup.jpg)
These are the parts.
And this is the current UAV file i start out with.
it seems your gpsv9 is made by me, use modified openpilot's gpsv9 firmware ;D
and does it still work correctly ?
-
@Xfce Are you saying that the clone GPSv9's have different hardware and can't use stock firmware? There is a bug in the OP firmware that causes some packets to get dropped. GPSv9's obviously work with the old firmware, but if you plug an old GPSv9 into 1609 you will see that there are dropped packets from the warnings you get that have now been enabled since the dropped data is now fixed in the updated firmware.
You should consider making a version of LP firmware for the GPSv9's you sell. :)
-
@Xfce Are you saying that the clone GPSv9's have different hardware and can't use stock firmware? There is a bug in the OP firmware that causes some packets to get dropped. GPSv9's obviously work with the old firmware, but if you plug an old GPSv9 into 1609 you will see that there are dropped packets from the warnings you get that have now been enabled since the dropped data is now fixed in the updated firmware.
You should consider making a version of LP firmware for the GPSv9's you sell. :)
yes i found this, like bellow,right? i will have a try :)
(https://vgy.me/8x5mIF.png)
-
Yes. That is the issue.
So is your hardware compatible with OP/LP firmware or not?
-
And the LP firmware is compatible with OP releases, so there is really no reason to use the old OP GPS firmware.
-
Yes. That is the issue.
So is your hardware compatible with OP/LP firmware or not?
A little difference , just remove the EEPROM's code. and I will compile some other day.
@TheOtherCliff BTW, dRonin have developed a flyingpi controller(PI-HAT size) for raspberry, and i have redesign as add oplink and gpsv9 on it,do you have the plan to support it?
Sorry, seems a little off topic
(https://vgy.me/3rOPWL.jpg)
(https://vgy.me/D3YL3r.jpg)
First fly, use default parameter
https://www.youtube.com/watch?v=MNtMZkOXcp0 (https://www.youtube.com/watch?v=MNtMZkOXcp0)
-
it seems your gpsv9 is made by me, use modified openpilot's gpsv9 firmware ;D
and does it still work correctly ?
I got 4 of these, they are heavier than the original ones but work great especially after upgrading the firmware.
As mentioned here
https://forum.librepilot.org/index.php?topic=2780.0 (https://forum.librepilot.org/index.php?topic=2780.0)
and here
https://forum.librepilot.org/index.php?topic=2903.0
I feel both the magnet and gps in the units are reliable, not had any problem with them :)
K
-
Some progress as well as some problems to report here.
(please see the picture at the first post, to recall the set up)
I have now got all the servos and motors set up correctly so they move respectively spin in the correct directions as well as make corrections correctly when I tilt the flight controller around while in stabilization flight modes.
Very Good!
On the radio transmitter side its a OPlink to the Revolution built in receiver (Control channel as well as telemetry). From the OPlink module there is a bluetooth connection to the GCS. All works well, no issues.
However, a strange thing happens when I move around the revo FC too violently (not much sudden movement is needed) - then the FC reboots!
Now, If i keep everything the same but just add a USB cable between the Revo and my Computer, then this does not happen. No reboots. Even if the GCS is not even loaded on the PC.
Does this indicate there is some kind of insufficient power supply to the FC? Since it works without rebooting if hooked up to another 4.5V power supply as well.
Need to understand the cause to this problem. And how to overcome it.
Anyone?
-
Try reducing the Oplink power if the issue still.
-
Thanks. I turned the oplink power down to lowest 1.25 mW but that had no effect.
Its still rebooting when I move the FC around. But no reboot if a USB cable is connected.
b t w I tried with another Revo as well, but same result.
-
Are you sure this is a reboot ?
Connect the usb just after the reboot and look the flight time in firmware tab or PFD corner.
When using Complementary the firmware wait the board is perfectly still for Gyro init.
-
Yes the timer to the lower left corner of the PFD restarts from zero.
-
How do you power the board ?
Double check the voltage while the board is powered without USB.
Disconnect components around and check if issue still with the board powered alone.
-
Thank you again.
Its powered by a 5 in one board that can take in 2-6 LiPo cell in serial and outputs 5.0 volts to the board via the BEC that say can take 3 Amp of current. Don't know if that's enough but the same board is working well for quads with gps on and 4 motors. This build has 2 motors and 4 servos.
It does not matter if the GPS is attached or not or the motors but if I only have 1 servo (any one) then it don't reboots but as soon as two (any 2 or more) servos connected to the FC then it starts rebooting.
What to make of that?
-
Double check your power source...
You can try using a simple 1S battery fully charged as power source
-
Thanks Laurent.
Issue solved.
I am pretty sure it was the fancy Align Heli servos consuming too much energy.
When having just a small battery dc 5 volt in, I could connect only one servo in addition to the motors without the FC rebooting. Using a 3s LiPo could have 3 servos okay but not 4. Then I replaced 2 of the Align servos to some old 9 gram simple servos and now all 4 plus motors and gps works without any rebooting. I happen to have 4 of these old but new servos around being idle so they will do for this project.
Karl
-
Moving on to next challenge.
So having the motors and servos set up the way I want and confirmed they move correctly with the TX controls as well as moving correctly when the FC is trying maintain attitude, How to fasten the tilt rotors at the end of each wing tip?
Since I already have the Mako fixed wing (https://www.readymaderc.com/store/index.php?main_page=index&cPath=112_409_649 ) and like a lot the way it flies ( https://www.youtube.com/watch?v=1RHxDAjHf1c )
I thought to modify that frame but I did some research to find if other fixed wings are out there that might be even more suitable. However, I really did not find anything outstanding so decided to go ahead with Mako.
The picture attached show the standard design of the wings with a Wing Spar enforcing the 2 detachable wings from the main frame body.
Next picture show the Normal spar location on the wing and the Enforcement going along the front of the wing. Last picture show my intended location or projection of where I want a prolonged wing spar to stick out from the front of the wing (in order to fasten the rotor tilt mechanism on it).
Thats the plan, but I don't know just how to drill a whole in the wing to extend the spar plus moving through the enforcement already in the front of the wing. Well I have an idea how to drill the whole but not how to treat the enforcement. If cutting it off in two pieces it will weaken their designed capability to stiffen up the wing. Maybe better bending it up or down and let the extended wing spar pass?
-
Solution.
First tried to cut out the wing enforcement to move it up or down to let my extended wing spar pass though it. But it was glued tight all over and can't be moved. So I took my lovely Dremel out and just cut it off, feeling I will put hot glue in the cross section and make the design strong enough.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=5903)
Some pictures attached.
Next challenge is the tilt rotor design, make and implementation :)
-
So I was looking for a good Tilt design to hold the parts at wing tip.
These are the parts: motor and prop, servo and ESC.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=5934;image)
Maybe a design like this is good:
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=5931;image)
https://www.youtube.com/watch?v=xnOw3clRIoI&t=242s
or like this,
I liked this design and would have been happy to copy it completely but
the design files were not made available in spite of direct inquiry.
https://www.youtube.com/watch?v=rMGr8XQgX-0
So I had to get in to 3D design to make it myself.
Checked several CAD/CAM SW and settled for Googles Sketchup
https://www.sketchup.com not because its free but it has so many instruction videos on youtube.
So the designed 3D parts I did are 4 plus a 450 Helicopter feathering shaft with bearings.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=5936;image)
I also had to find someone who could print these designs since I do not have a 3D printer.
This site was really helpful in finding a printer :)
https://www.3dhubs.com
And I have to recommend Mr Si at this particular Print Hub in Beijing where I live, very responsive, fast and good quality.
https://www.3dhubs.com/service/printer
Then I mounted them like this:
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=5938;image)
And with a final mount on to the extended wing spar on the Fixed Wing like this:
The whole unit is 70 grams.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=5940;image)
Next challenge to complete the build.
Attached are the 4 items of 3D design I finally used.
-
I think it may be good to have the motors closer into the body? At least not all the way out at the tips. More stable, less stuntable, easier to tune.
-
yes, maybe. The first try is just to place motors as close as possible to body but just so far away the props don't hit the wings. Can be adjusted later if needed by cutting out the wings a bit to give room for the props. Will see how this works first though. Anyway i am concerned with this much mass far away from COG, 70 gram on each wing. Its interesting to see what will happen
-
Hi all,
I am running in to a wall here.
The right hand motor starts hesitantly but do start but the left hand motor sort of coughing and don't start.
https://www.youtube.com/watch?v=KmyF-0j14xU
However, when I put both motors on a short wire from ESC to the FC like 15 cm then they both behave fine.
Problem is i need them to be out on the wings like 50 cm away (too much to ask?).
Why is this and How to get around it?
Thanks anyone :)
K
btw I have checked with my voltmeter the continuity of all cables, they are fine.
Another thing to mention is that its seem its only the signal wires that seem to matter if they are short or long, not the voltage cables to the ESC and motors ...
-
Thank you again.
Its powered by a 5 in one board that can take in 2-6 LiPo cell in serial and outputs 5.0 volts to the board via the BEC that say can take 3 Amp of current. Don't know if that's enough but the same board is working well for quads with gps on and 4 motors. This build has 2 motors and 4 servos.
It does not matter if the GPS is attached or not or the motors but if I only have 1 servo (any one) then it don't reboots but as soon as two (any 2 or more) servos connected to the FC then it starts rebooting.
What to make of that?
That could be your problem... when you go to move the board around a lot with all four servos it might overdraw the BEC and cause it to shut off temporarily, just a thought..
-
Yes, agree kicekilla. I think that was the problem. The solution was to to change two of the servos to simpler ones that did not consume so much power. That stopped the problem.
The challenge now is to get the ESC/Motors to start and run smoothly.
Its really puzzling.
Maybe my extended wire from the FC to the ESC has a different quality?
It seems this wire is the only thing that is changed from when its running okay...
or maybe just because its 5 times longer than normal
Any ideas?
-
sorry for replying to an old (solved) problem... didn't realize there was 2 pages ::)
Interesting that it works with the short wire and not the long. Have you tried calibrating the esc's with the long wire? maybe try switching esc refresh rates.
i have heard of problems by having too long of power/3wires going to the esc but never the signal wire...
Also, good job on the project, looks fun! Ill be following.
Yes, agree kicekilla. I think that was the problem. The solution was to to change two of the servos to simpler ones that did not consume so much power. That stopped the problem.
The challenge now is to get the ESC/Motors to start and run smoothly.
Its really puzzling.
Maybe my extended wire from the FC to the ESC has a different quality?
It seems this wire is the only thing that is changed from when its running okay...
or maybe just because its 5 times longer than normal
Any ideas?
-
Haha no worries!
Wow several good ideas!
Have not tried any of that - will get to it!
Thanks
-
your video looks like a motor timing issue to me, just confusing that it works normally when the signal wires are short. Let us know how it goes.
-
you think the signal wires may have a short circuit?
-
no I don't think you have a short or it wouldn't work at all probably. I would just try the re-calibration with long wires as well as trying different refresh rates.
-
8)
-
tried re-calibration with wires that intend to use.
tried all possible update rates in the output screen 490 down to 50 as well as the oneshot etc
Right motor always work and left motor never starts.
So now I will replace for a new BLHeli and try again.
-
both motors still work normally when using shorter signal wires?
-
Regardless of the length of the wire, it sounds like one of them is broken and the other is OK.
-
yes agree. Its soldered on now.
Ready for a test later on.
I no longer have the option to check if short wires work or not.
-
ran the set up ESC calibration with a brand new ESC out of the box and the result is the same. don't spin.
I will mount a new motor on to it. If that don't work I will resign from this :(
-
Mounted the last motor I had, and it behaved just the same.
pretty low feeling now guys
-
Do these motor/esc combos work on other models? I have had motors that just don't work well with certain escs. Just confusing that they both worked fine with shorter signal wires correct? Try different output pins from the revo? Try different lipo. I know how aggravating it can be when you can't find the solution! Hang in there
-
Thanks :)
Motor/esc are sunnysky/Blheli and works well on two other quads for me. However, similar problems have occurred before when setting them up. At then I just assumed I did some bad soldering and messed up some of those small fragile looking little things on the esc. Yes they did work initially with the (signal) cables they came with. After trying different update settings now they may not work anymore. Yes I will try some of that also another brand of esc, if I can find one lying around.
-
My bet is that there is something still wrong with the wiring.
Signal wires, power wires...
Maybe swap the two ESC servo connectors around and see if the bad side starts to work. I bet the bad side stays bad.
Do you know you are getting power to the bad ESC? Does it beep? Do they both beep if you unplug the ESC servo connectors and plug the battery in?
Do you have all your output banks set to the same protocol and speed?
-
Okay I will bet you - just to make it more interesting. Will resolder the wires.
But one thing while i am at it - Is it really necessary to have both the signal and ground wire to the ESC?
(they already have ground from the power wires)
I have one quad where I never did bother with the ground and its working fine (same type of motors and ECS).
Yes, they have power and beep, both of them.
Yes, output same protocol and speed.
and I have checked the wires for continuity with voltmeter.
-
But one thing while i am at it - Is it really necessary to have both the signal and ground wire to the ESC?
(they already have ground from the power wires)
I am not an electrical engineer but signal alone might work just fine, though I would still connect "ground" which is your reference 0 voltage. If you have more than one regulator or more voltage sources these zero voltage references might differ. I use the rule "better safe than sorry" ;)
-
Not necessary to resolder if you tested them with a meter from one end to the other. Maybe you soldered in the wrong place somewhere? Maybe the ESC servo connector is plugged in upside down?
Did you try just swapping the ESC servo connectors such that left is plugged into right's place at the FC and right into left's? It would be good to know for sure whether the problem is with the ESC or the FC.
I recommend leaving all three ESC servo connector wires intact, but for most setups it should not be a problem.
I can imagine there could be a problem if your ESCs control section is isolated from the power section and you left out some wires. Opto ESCs may be isolated like that, requiring all three wires (including getting power from the receiver or FC on the red wire) to work.
-
Thank you all for attention and helping!
It seemed that after I redid the radio transmitter calibration, then both motors starts and run just fine.
I really don't see how that has anything to do with it :o
But I have made no other changes to motors, wires, LiPos, fc output ports etc
B t w they don't start all times without the ground wire connected to the ESCs signal connectors. So I will keep that wire even-though it would be much nicer with only 6 wires (2 x 3-wire servo cables) than 7 wires running from the fc out on the wings to the tilt rotor houses.
I hope they don't break down so I finally can move on with this build.
-
glad to hear it! keep us posted.
-
Last night my customized 104cm length wing spar arrived (originals are only 50cm).
I mounted it and centered it on the mako.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=5979;image)
Also drilled holes in the spar in the 'engine room'
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=5981;image)
to enter the wires from the FC and drilled holes on the wing spar tips to exit the wires that will connect to the tilt house. Then fed the wires through the holes.
This is all fine, next challenge is to do all soldering for the tilt house but also to fix the wing spar from rotating, it has to be fixed even when the tilt rotors will create a turning momentum. I don't know how to best do that yet...
-
I believe that the design of the birotor control is such that it requires the birotor to basically hover stable when the motor angles are locked (unmoving) pointing straight up. The CG needs to be well below the props.
-
I believe that the design of the birotor control is such that it requires the birotor to basically hover stable when the motor angles are locked (unmoving) pointing straight up. The CG needs to be well below the props.
some bicopters have the cg well below, but not necessary (or even preferred in my exp). On my bicopter I had the cg pretty close to directly between the motors. the servos will need to tilt the motors back and for yaw and pitch stability and motor thrust is used for roll stability.
One thing with your new spar Karla, does it pass through the CG on the longitudinal axis? if not, that is where you will have problems.
here is a video of my cc3d bicopter I made a while ago. I don't have it anymore but designing a new 250 sized one currently.
https://www.youtube.com/watch?v=Su-W-LcbEcA
-
Nice bicopter and flying site kicekilla!
That bicopter is designed to use gyroscopic forces for control right? Rotors tilting in 45 degrees angle? Thats an elegant and effective design.
This wing does not use gyroscopics for control (but will work differently, I hope :)
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=5986;image)
CG will be below Rotors but the way the Mako was originally designed the CG is 2 cm forward of the wing spar. That will mean that centre of lift will be aft of CG a bit (at least in hover mode). See the black dots on each side of the wings. I have been concerned about this but thinking I can move CG a bit aft with the LiPo placement...
btw I was clumsy enough to make a short on the power wires where it exits one side of the wing spar. I need to redo the wiring now ::)
-
Yes, mine was an OAT (oblique angle tilt) that uses the gyro forces to assist stability and control... but the normal trust vectoring is also used.. yours is a FAT (forward aft tilt) which uses the control /stability I stated before.
Slightly forward should actually work out pretty good. But keep this in mind... as your motors rotate forward the weight of the motor will shift your cg too so maybe shifting it back to the cg using lipo weight and then when you rotate the motors forward for fixedwing flight, the cg will have shifted back to the original cg or close hopefully.
Unless you designed your motor pod to pivot around its own cg (which is what I do when designing tilt mechanisms)
this is my latest tilt design:
-
just looking at your tilt pod design, it should be close enough that there wont be much cg at all. do you have a video of the tilter working? : ) I like the geared drive, I might try to do that actually. I don't know if my 3d printer can print accurate enough for the gears though.
-
I could very well be wrong, but it seems to me that if you have positive stability (CG below rotors like Osprey?) that when hovering level, the motors will basically be pointed straight up and to make it go forward, you just tilt the motors forward and the CG pendulum allows the motors to tilt forward while it is still basically level. No stabilization actually required.
With neutral stability (CG at same level as rotors) the CG is not helping. Tilting both motors and accelerating laterally does not change the bank angle, so how do you make the aircraft tilt forward or backward (to maintain level) at all? I don't think this will work at all.
If the CG is above the rotors you must constantly balance it like a top-heavy Seqway. I think the FC can balance it. I think you want Attitude mode and not Rate mode. This is the scenario that is most like a normal multicopter.
-
First to kicekilla,
Great you have experience in this area!
Yes mine are FAT and is controlled just like you said in post below.
My rotor tilt house design is near neutral to CG when its tilted around but not perfect. Its my first ever CAD-CAM design and 3D print :) I think its good enough for business though.
I will upload a video of it in action later when I have sorted the shorted slightly melted wires.
There are print shops who can print for you. In some below posts I have a link where to find one close to you and also the one who printed my gears.
Can you upload your design in a .stl file format?
Cliff,
I have no idea at this point in time but what you say seems logic :)
This build is your first case, have positive stability, CG below rotors.
-
I re-wired the shorted things this morning.
Original idea was to make the motors detachable for easy transportation, but I realized it will be just too much wiring taking too much space and create drag, so I soldered them hard in to the frame.
Here is a video of the tilt rotors and the controls in action in hover mode :)
https://www.youtube.com/watch?v=zQPpcAR2nvg&feature=youtu.be
Coming this far I am a bit reluctant to take it to the field and try it out...
But that is the next step of course.
Bit concerned about the 'coughing' of the left motor...
-
Cliff, I misspoke, my CG (along the vertical axis) was definitely lower than the motors/pivot line due to my under-slinging the battery. Not excessively but not centered. it was roughly centered longitudinally thought which gave me the close to level flight attitude.
I agree that the further the cg is lowered the more natural stability it will have, but at some point maneuverability will be overpowered probably haha.
I am curious to see what happens now with the cg centered vertically. I will probably do some tests with different cg position once I finish building my 250 bicopter.
and yes on the bicopter in the video, I always flew in attitude mode. I don't remember trying rate, but will experiment with my new one.
-
Finally took it for an intended first test hover.
Problems.
The motors did not perform much over 40-50% of rpm before disconnecting and did boot up beeps.
I tried
. recalibrating the radio transmitter/receiver
. recalibrating the ESCs
Then I disconnect all servos from the FC to only test the motors.
They did start now and then but 'coughing' like in the video I uploaded before.
However, after some 3 min of trying, then the wires on other side of the wing started to smoke and melt :(
Question: is it mandatory to keep a thicker wire diameter on the power wires from the PDB and the ESC/Motors?
I think too thin wires (servo wire type) might melt if the motors are consuming too high current, but not causing the problem.
Any thoughts?
I am thinking to try another brand of ESC (Litlebee 20A) to see if they can work better with these Sunnysky x2204S-16, 2300KV motors than the current ones (Bl Heli 20A). However, don't have much hope this will fix it...
-
Which wire got hot? (From ESC perspective) Was it signal (servo) wires or input power wire pair or output power wire triple?
Generally, you need to maintain the same size of wire as comes out of the ESC. The only ones that are allowed to be thin wires are the signal (servo) wires.
If any of the wires are too thin, it will cause hot wiires, power loss, motor stuttering and motor resets. If wire sizes are OK, the problems you mention can also be caused by using props that are too large or a battery that can't provide the desired current at full power.
I know that @Mateusz uses the LittleBee ESCs. I wouldn't think that is the problem.
-
One other thing is that it is that long ESC power wires can cause inductance problems. I have heard it recommended that if you need to extend the ESC input power wire pair a long way, that you put some ESC capacitors (same as on ESC input) (low Equivalent Series Resistance) across the wire pair in the middle of the length. I also heard that it is better to extend the ESC output wire triple (don't use capacitors there :) ) than the ESC input wire pair.
It should not be a problem to extend the ESC signal/servo cable to the lengths you need. Even to several meters.
Adding a voltage/current sensor to a system with long ESC input wires can cause ground loop problems.
-
Thanks a lot!
Seems to be a lot more to this than I first imagined :P
Hm yes the problem here might be the thin and long power cable pair going to the ESC.
. The wires getting hot (this and last time) is not the signal wires but the power wire pair going in to the ESC (V and Grd). I have these wires as thin as standard servo wires to more easily fit them inside the wing spars. The 3 wires out from the ESC to motors are fine, short and default size.
. I do have a voltage/current sensor between the LiPo and the PDB (hm, so this could cause some ground problems).
So I should use thicker, default size power wires to ESC and also consider to make the wire out from the ESC to motors longer rather than the power pair wire going in to the ESC. Maybe also remote the voltage/current sensor.
-
Your main, real problem is the thin ESC power input wires. :)
The other stuff I mention for things to keep in mind if you have future issues. I wouldn't change them until I had further issues.
-
Got it. Thanks!
This will keep me off the streets all weekend anyway :)
-
Puuh... finally done rewiring everything out to the wings.
I moved the ESCs from the tilt pods in to the body, just next to the FC.
Then rewired the tilt rotor servo triple cable and the triple fatter size wires out from the ESC to the motors.
Had to enlarge the wholes in the spar to get it through. Succeeded. Also made the tilt pods removable like the original design idea.
... also need to fix the wing spar from rotating, it has to be fixed even when the tilt rotors will create a turning momentum. I don't know how to best do that yet...
This was solved by friction of adding lots of hot glue in the wing where the spar is naked (see earlier pics). On top of the glue I pressed in the cut out foam pieces. Feels very steady and I can't force the spar to rotate even adding considerable force.
. used the same motors but a pair of brand new Blheli ESCs to start fresh
. calibrated them and motors using setup wizard for Tricopter
. all seems okay. no stuttering of motors yet.
. i kept the voltage sensor in the configuration though.
Using default PID settings from LibrePilot Vehicle setup, Tricopter.
Then set up a flight mode for Hover:
. Roll-Attitude,
. Pitch-Attitude,
. Yaw-AxisLock,
. Thrust-Manual
Took it out for a first lift off test. Trying to hover. Main idea to check CG, lift, stabilization.
Interesting initial result:
+ input controls work in correct expected directions
+ fx corrects all moves in correct directions
+ motors starts and spin smooth (up to a point, after that left motor do some stuttering in Attitude mode, not in manual)
+ tilt rotor mechanisms work fine
- seems having insufficient lift, however not really sure since if I switched to manual stabilization then seems plenty of power
- the tilt rotors Pitch was visibly oscillating, so consider to cut PIDs by half
- seems Pitch is uncontrollable, maybe CG is not far enough below the rotors, consider to rise the rotors higher or move LiPo lower.
Should I try change the left motor? It might have gotten damaged by previous experiments.
The ESC of left motor is getting much hotter than the right ESC. Both motors are hot but then again its some 35 degrees celsius here today. These ESCs have no BECs its provided by the PDB.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6035;image)
Also thinking What effect the Vehicle Set Up settings may have on motors when using 100% 50 or 25% of mixing for Roll. Could this make motors stutter more or less (when in attitude mode)? but why just one and not both?
-
. i kept the voltage sensor in the configuration though.
You shouldn't have a problem with that since your ESC input wire pair is normal length.
+ motors starts and spin smooth (up to a point, after that left motor do some stuttering in Attitude mode, not in manual)
- seems having insufficient lift, however not really sure since if I switched to manual stabilization then seems plenty of power
The ESC of left motor is getting much hotter than the right ESC.
You have plenty of power in manual mode, but not in stabilized. Motors getting hot. It also probably jumps to this reduced power just above zero throttle. These are symptoms of "invisible oscillation" but there is another possibility described towards the end. There is a very fast oscillation caused by PID too high (D term at least???). Explanation: When the throttle is very low say motors at 5%, 5% the most stabilization can do a roll is 0%, 10% or 10%, 0% but the stabilization needs to give a lot of change, it increases throttle. The biggest increase it will do is to 50%, 50% so that it can do a roll with 0%, 100%. This increases the effective throttle from 5% to 50%. You give 5% stick and it actually makes the motors go 50%.
The same happens at the high side of throttle. If you give 95%, the motors should be 95%, 95% but that only allows a roll of 90%, 100% or 100%, 90% so it reduces the effective throttle to 50% so it can do 0%, 100% or 100%, 0%. The fact that you have plenty of power in manual, but not in stabilized is saying that you have this oscillation and the side effect is that you only have 50% power.
So when these oscillations are present, the result is that no matter where you put the throttle it is actually closer to 50% than you are asking for. In very bad cases nearly all the throttle range is effectively 50%. If it takes 55% throttle to take off, it will never take off. Worse yet, if it takes 45% to take off, it will climb at just 5% throttle stick.
In your case, there is only one control that uses the motor speed to affect the control. That is (hover) roll. The other functions use servos. Reduce the hover roll PID (may be forward flight yaw?) to fix this invisible oscillation. Especially the D term.
Hover roll uses motors to affect the roll. Yaw and pitch use servos. These will probably all need different PIDs.
Also thinking What effect the Vehicle Set Up settings may have on motors when using 100% 50 or 25% of mixing for Roll. Could this make motors stutter more or less (when in attitude mode)? but why just one and not both?
These mixing sliders are used to reduce the roll, pitch, or yaw when 2 or more use the same "actuator". For you, hover yaw and hover pitch use the same servo actuators, but only hover roll uses ESCs, so I would set hover yaw hover pitch both to 50% and hover roll to 100%. Be aware that changes here affect the PIDs. Doubling the mixer is like doubling the PID so tune after changing the mixer.
I don't know why one motor stutters and not the other. Fix the PID oscillation and see if that goes away? I am guessing that the stutter does not happen in manual mode. If the stutter does happen in manual mode, then it could be the reason why you can't lift off in stabilized mode. One motor has less power, so it reduces the other motor to keep it from tipping over. Stutter can be caused by props that are too big or bad battery or bad connection (or too thin wire :) ).
In all the above when I say for instance hover roll, that is forward flight yaw. I don't know whether you have this set up as roll or yaw.
-
Wow, that was a mouthful of advice :) Will need to digest.
Thanks a lot!
-
Hello,
Done some more experiments
1. Yes, I think there was a "invisible oscillation" problem cutting down the power and thrust to 50%.
Fixed that by lowering PIDs of Hover Roll. I know its okay since the lift generated is the same with stabilization set to manual for Roll as having it set to attitude.
Good.
2. Yes, the hover pitch oscillations was more related to the yaw PID settings being too high. Could be pitch or yaw since in hover they both move the rotors direction.
Fixed now by lowering them.
Good.
3. For the mixing table, I took the advice to put hover roll to 100%, this will only affect the motor rpm and should be fine. But for the yaw and pitch its a bit different. If set them to 50/50 as recommended the yaw is to vigorous and the pitch is insufficient. So I settled for 25 respectively 75%.
Good.
For center of gravity, I think there are two issues
1. x-axis:
it was obvious it has to be centered to were the wing spar is crossing the mid section of the craft. Otherwise nose go down or up when applying thrust to motors.
Fixed by moving the LiPo. However, this is a temporary fix. As pointed out in the previous post and picture, the design of the Mako has the CG 2 cm to the stern of the wing spar (wing spar is 2 cm aft of CG). So assuming when not in hover mode the center of lift will move to the designed CG and the weight needs to be readjusted or it will be uncontrollable and flip upwards. This can be fixed by another design of the tilt pods (tilt rotor housing). Save that for the future.
2. y-axis:
I have the feeling the pitch is not really effective and suspected the rotor tilt point is not high enough above CG.
So I tried placing the LiPo lower, under the craft. And it seems better and really promising as the FC seems to be able to stabilize it.
However, my big concern now is Do I have enough lift?
At present, giving full throttle with manual or attitude mode for motors does not give enough lift to be airborn.
Anyone good at doing the math to calculate/estimate available lift in grams?
. Lightest weight of Plane excluding cameras and stuff = 800 gram (including LiPo) but I likely like to add some 200-500gram more of stuff.
. LiPo 4S 1000mAh,
. motors 2x Sunnysky x2204S-16, 2300KV
. ESC 2 x Bl Heli 20A opto
. PDB say can take 3 Amp
. Props: 2x 3 blades 30 degree attack angle 60mm long blades from center to tip.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6052;image)
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6054;image)
I am considering to fix both the CG y-axis and x-axis permanently, not by lowering the LiPo but by placing the rotor tilt higher by redesigning the rotor tilt house.
The x-axis CG can shift further stern when switching away from hover mode to forward flight, with a new design.
If I am doing that then i should also design a in a more powerful motor,
hence the need for the calculation :)
Anyone?
-
Two places to check for thrust flybrushless.com (http://www.flybrushless.com/tools/thrustCalc) or ecalc.ch (https://ecalc.ch/motorcalc.php?ecalc&lang=en).
Looks like around +-600g of thrust one motor.
Use ecalc as a multicopter to determine proper thrust for bicopter, idea.
Also have to figure out FF speed for stall and this link should help rcgroups.com (https://www.rcgroups.com/forums/showthread.php?1136470-Beginners-Guide-to-Motor-and-Prop-Selection)
Building a VTOL but with 78" wing span so have been trying to research lots of info before I start.
-
Just an idea but PX4.io has the E-flite Convergence coded for autonomous flight, Pixfalcon flight controller.
Its not stabilized for true hover, no tilting nacelles backward (pitch).
Might give you some ideas.
-
Thanks a lot!
More to learn :)
Your VTOL with Wingspan of 78'', thats close to 2 meters right? Pretty big!
You design it all from scratch?
-
Make sure you are getting full throttle command to your ESCs. If you log your flight, you can play it back and look at an actuator scope to make sure it is asking the ESCs for full power. You can do what you did before, and that is check the power when using Manual thrust, or when using Test Outputs on the Output page.
Make sure you don't have a throttle curve set on the Input page. Also, AltitudeHold settings and *PathFollower settings have thrust limits if you are using Altitude Hold/Vario or some GPS flight mode.
You mentioned something a while back about one motor jumping more than the other. If there is a weaker motor, remember that the FC will reduce power on the stronger motor to keep it level. That is another reason for Manual having more power than some stabilized mode.
-
Your VTOL with Wingspan of 78'', that's close to 2 meters right? Pretty big!
You design it all from scratch?
The air frame I will start with is the RQ-7 Shadow laser cut kit.
I am thinking of dropping it to 70 or 60 inch's due to the nacelles.
The design of a rotating nacelle with motor mount is the real work.
E-flite Convergence, Bananahobby V22 or Rotomast V22 style of rotating nacelle.
But yours will be flying before mine, if mine fly's at all.
-
@Mangee, yes there are several FC out there, thanks a lot! I just want to do this VTOL with LibrePilot to see how far it can go and because it offers a lot more than just stabilization. I am very optimistic at the moment but will make it run with whatever fc :)
Good luck with your build, it looks beautiful!
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6073;image)
Having read up on the links you send and exploring the site https://www.ecalc.ch I have calculated all is fine just need to change the props.
From 3 blade to 2 longer blades with higher angle of attack (6 inch long by 4.5 angle). Should provide Thrust theoretically for vertical continues flight with some 150g margin, without overloading motors, ESC or LiPo. Had the props at home :)
@Cliff, Thanks!
With the new props and new settings of the PIDs (of course), I still have some stuttering and strange behaviors of motor rpm.
But the whole project is looking more promising.
So I will really look in to your suggestions.
-
Oops ... seems I was not properly connected to the internet when I got the OK on the current set up from the eCalc site :(
This is what it says now, suggesting
. move ESC from 20 to 30 Amp
. select a stronger and slower rpm motor
. and larger prop
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6075;image)
Meanwhile, I will continue to experiment with the current setup since these warnings are really at the max throttle.
-
I wonder if a pair of 2204 motors are strong enough?
-
I can't remember, have you tried a 3s battery and still gotten the motor stuttering?
just by "eyeballing" it and my quad experience, a pair of 2204's doesn't seem like it would be enough to hover this guy, but then again that is purely just by the looks of it.
just for comparison, my 250 quad weighs about 800g ready to fly. Its running titan 2206 motors and 5030 bullnose props on 4s and hovers at about 40% throttle. Those motors are pretty powerful and I get about 8 mins on my 2200mah.
So, it might be a bit much to hover about the same weight with half the motors (smaller ones at that).
they don't have my motors in stock anymore, but these might be a decent option.
http://www.multirotormania.com/clearance-sale/1126-mrm-titan-2208-2000kv-brushless-motor.html#idTab5
look for motors that can pull about 400g of thrust at 50% throttle.
-
Yepp. Absolutely correct. Not enough motor thrust.
Thats also what the eCalc application showed clearly.
Its really great to select components by trail n error.
-
After using the eCalc App and experimenting with different motor sizes and speed (KVs),
prop size and angle, 3-4S LiPos, ESCs, and checking where available to buy, I came up with just
having to get new motors and props (10 x 6 inches).
My 20 ESCs seems okay within range.
I settled for 2216 size with 810KV
https://www.banggood.com/EMAX-Cooling-New-MT2216-II-810KV-Brushless-Motor-CW-CCW-with-1045-Propeller-for-RC-Multicopter-p-1004936.html?rmmds=search (https://www.banggood.com/EMAX-Cooling-New-MT2216-II-810KV-Brushless-Motor-CW-CCW-with-1045-Propeller-for-RC-Multicopter-p-1004936.html?rmmds=search)
Now the design looks more viable:
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6077;image)
-
I think those will give you plenty of power... if you havent ordered yet I might even look for something a little higher kv and run 8 inch props...
keep in mind, these are going to burn through your 1000mah quickly... in hover mode at least.
-
Thanks.
I am still in research mode.
I will dive in to it again.
With this smallest LiPo 1200mAh I can only hover some 3 min (according to that eCalc app table in last post), but I have larger LiPos and I just intend to lift off and land with hover, rest and most should be in forward flight.
So far i didn't notice to save energy by faster rpm. The recommendation seems on the contrary to go for the largest prop size you can fit on the frame (up to 1/4 of wing span) then figure out how fast you need to spin it to create enough lift. This is some kind of hybrid between copter and fixed wing so I am willing to try everything :)
-
Well i am going through with this selection. Lets see how it works out.
While waiting for the motors to arrive and redesigning the Rotor Houses, I am just thinking in retrospect,
so naive of me to think I could just take whatever I had at home of parts and it would really fly at first shot ::)
-
haha you wouldn't believe how many parts/equipment I have at home, yet I still find myself "having" to buy new parts because it will work better than what I have haha.
I think the combo you posted will def. lift it no problem.. I am suspect of battery size then. And I think the combo + 10 inch props might be a little bit of overkill personally, but that's just me. Are your servos going to be strong enough to rotate those motors/props well enough? just something to think about.
-
Haha thanks for sharing that. In good company then :P
Overkill... maybe, but have the option to prop up/down as needed, at least did some math/calculations this time.
... and yes the new load on the servos/tilt gear mechanism is a source of worry.
Not mainly because using a stronger motor (if used on current tilt design I am not so worried since its almost CG neutral and has no 'leverage' while it tilts around), but due to a new design that I think will permanently fix the CG placement shifting problem in hover respectively forward flight. A possibly negative effect of the new design is that the tilt mechanism must withstand more power than it can take.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6087;image)
The picture show the current design and I am just holding the new design up, its not ready.
While the current have the servo 'hanging' below the motor and the tilt point the new design have servo still below the motor but 'standing on top of' the tilt point. As an effect of this and the new 'higher' motor, the props will be spinning some 80 mm higher than currently.
Wanted, desired effects and uncertain consequences as of now:
. wanted, props spin higher above CG on y-axis to increase stability (no need move LiPo lower)
. wanted, in Hover CG should align over wing spar (center of lift, CL=Hover)
. wanted, in Hover when Pitch motor for or aft also moves CG for/aft and improves control
. wanted, in forward flight CG should move 20 mm forwards of spar to align over CL=ForwardFlight
. wanted, give room for props to spin freely without having to make 'holes' in tip of wings
. uncertain, How will this design impact where forces are applied (vectors) and what outcome?
. uncertain, Will this design increase power on servos, gears, feathering shafts (and its foundations in spar and tilt house)?
Bound to find out :)
-
looks good, excited to see more. waiting for the new parts now?
how do your gears fit to the servo splines? like just kinda mash in there or is there actually a spline printed into the part? I have tried printing some gears for my bicopter design but my printer doesn't have the resolution to print that small of a spline so I have sized the holes to just press fit, but I don't really like that as it will probably wear out.
-
i didnt even try to print those small splines.
I just glued a standard servo arm to my 3d printed tilt gear :)
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6099;image)
You usually get several different kinds shipped with the servos, I just thought the round ones had more area for the super glue to stick on to.
New motors should arrive next week.
-
I received the new motors and props yesterday :)
(and new 30 amp ESCs but have them for later, hope my 20 amp will be ok as calculation suggests).
The version 1 weight was 70grams and
ver 2 weight is 128 grams,
thats including tilt house, servo, motor and prop.
basically double the weight but also increased the lift/thrust a lot.
A considerable upgrade.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6145;image)
Mounted on version 2 tilt design.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6143;image)
Will be mounted on the wing spar soon.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6147;image)
-
uhu, when about to mount the tilt rotors on the wing spar I found out the gears are ~2.5mm too close to fit :(
trying to fix it by making the gear wheel with a bit reduced diameter, however don't know if it will be able to 3d print that small gears. If not then I will need to adjust the design of the tilt house and print them again. Think I will go fly some heli today instead.
-
The new gears that fits on the servos with smaller diameter turned out to be okay for now, but not a permanent solution.
Did mount motors and all and now there is enough lift/thrust with good margin.
A had a puzzling result when trying to hover first time.
When starting to lift off it pitches backwards and will flip over backwards unless i reduce throttle.
It is unexpected since the lift is applied vertically on the wing spar by the two motors and CG is also aligned by the wing spar, so it should not flip backwards or forwards (unless applying pitch). When it starts to pitch backwards I try to compensate with pitch forward with the stick. However, it seems to respond far too little on the stick command. I use Attitude on the pitch. I know the servos are able to pitch much further since they do when using Manual rather than Attitude.
Maybe I should try use Rate and see what happens...
Or could the board not be centered horizontally...
-
If it works OK in Manual, but has that issue in Attitude, it might be vibration. Balance the props.
-
ah will try.
-
How about air flow over the wings. Your motors are at CG but the wings are behind and under the tips of the props.
Sort of like the E-flite Convergence.
I am building a RQ-7 Shadow for vtol but its CG is down the center of a straight wing with motors at the wing tips.
All of the swept wings doing vtol on other sites are either tries or quads. Shutting down the rear motors on FF flight.
-
Thanks Mangee, Yes, can certainly be the air flow effect over the wings, that is bound to have some impact.
Yes, lots of swept wings are tri- or quads with or without a motor at rear.
Heard others say its difficult to get the frame balanced in hover like this.
However, I am trying to see if it can be done since I like the comparatively simple design if it can be set up.
Good luck with your RQ-7 Shadow build!
-
Guys, I think we are approaching end of the road here.
Please feel welcome to spot if I missed options.
The problem with this design (2 motors with prop down wash partially over the front of the wings) is it cant be stabilized properly in hover mode. In vertical flight it will likely work just lovely. In hover, the lift will only work fine for the part of prop not hitting the wing thus creating a pitch up movement. This up pitch is not canceled out by a prop on the other side of the wing, on the contrary its amplified by the other motor. In the case of a straight wing with motors at CG and at tip of wings, this down wash is canceled out. So when adding throttle the plane will rise up and eventually flip over on its back. Roll and Yaw works just fine though. If fixing the pitch by moving CG a lot forward, then the forward placement will result in strong down pitch in vertical flight. Not a workable option.
1. cut off the tip of the wings
Reducing the length of wings so the props run free will create an even lift over CG, where CG needs to be. This will likely make everything work nice in hover mode. However, the reduced wing area will decrease lift, make the stall speed higher, motors work harder and cause shorter flight time in vertical flight mode. Don't desire that. Maybe I could calculate just how much lift will be lost if cutting out that area, maybe its acceptable...
2. move the motors further out until prop downwash run free of wings.
This can be done by increasing the wing spar length by 120 mm on each side. The wing spar is now 1040 adding 240 would make total length 1280, ~ 25% longer. I think that will still work fine with adjusting yaw, roll settings but will the spar hold or need to increase diameter? How will the spar respond to motors given this increased leverage? How about vibrations?
3. add a third or forth motor
Balance pitch in hover could be done adding a third motor and place it at rear using the Mako's stock mounting place for the motor. However, this would need an additional output port on the flight controller. Since all 6 are already used and I cant see how to free up one without losing vital control, this is not an option. To try make it a quad will put even more demand on output ports. It will also need to rethink how to successfully control yaw and roll in both hover and vertical modes.
4. use another airframe with straight wings
The reason I chose current fixed wing was to have the simple Elevon controls that combines Aileron and Elevator into one control and don't use a rudder. Thus in Hover mode yaw is done by tilting rotors opposite directions but in Forward flight tilting is not desired but can be done using the Elevons only, just don't touch the yaw stick.
Elevons cant be used on a straight wing. They will need separated Elevator and Ailerons and maybe a rudder control too. I am unfamiliar with that type and not sure how to set it up to work in both Hover and Vertical flight with Libre.
5. different flight controller (HW)
As far as I understand Librepilot Rhino support at least 12 output ports so just swop the Revo board to another supported board with more than 6 ports. Then do option 3 above (add one more motor).
6. develop a vehicle type switch (SW)?
Today, there is no way to switch the current vehicle type (Multirotor/FixedWing/Heli/Custom) etc, from one to another dynamically. Maybe this limitation can be done away with?
Something like in the GCS when you import a UAV-file to a board.
Maybe one additional 'uav file' can be stored in Revo memory and with a command it will be 'imported' during flight.
This would create much more options and no need for additional output ports on the boards, they can just be used in different ways.
7. different fc (SW and HW)
Use another fc more tailored to VTOLs.
Yes can do that but outside of this scope. Purpose is to see what can be done with LP.
At the moment I am not sure how to move forward.
With the current design it can't be a V-TOL. It can at best be a Short-TOL fixed wing tilt rotor. Just accepting the pitch upward and forward motion in hover and switch to half way forward when plane is starting to lift off. In half forward there is no down wash over wings anymore. The motors should have enough thrust. I might try this.
-
How about ailerons that you configure as flaps or spoilers during hover?
You might even use the ailerons instead of tilting the motors...
-
Thanks Cliff.
But I will have to put that in the category of wishful thinking :)
I would be very interested however if you could spend some minutes and share your thoughts on the option 6 below.
Karl
-
There probably are better ways to do that? It takes a long time to erase and flash a section of settings memory, and stabilization will fail for a short time with the CPU taken away for so long. That is why you are not allowed to change any settings while it is armed.
It would be feasible to store both sets in flash and switch on command, like we do with Stabilization Banks right now (code changes of course).
There have been users that wanted a transition from one mode to another, like for rotating the motors from vertical to horizontal and they wanted control throughout the transition. That seems a more reasonable way to handle it.
It seems there has not been a dev with the model, the time, and the drive to do these.
For today, the easiest way to make a switch from one uav to another would be to run two FCs and have some sort of electronic servo switch to tell your servos to switch from one FC to the other.
Also, you could start bench experiments with removing the safety checks that keep you from changing settings while it is armed and use GCS telemetry to import a uav file in flight. After importing to RAM (which is very fast), the import doesn't actually write to flash until it asks you (you could cancel that prompt or just ignore it). I don't think that many places in the flight code actually read the flash settings, they just use the RAM copy of the settings. I know that AutoTune reads the flash, but you don't need that.
-
Thanks a lot Cliff.
There are so many options and opportunities here :)
I will consider how to move forward.
This part is particularly encouraging:
Also, you could start bench experiments with removing the safety checks that keep you from changing settings while it is armed and use GCS telemetry to import a uav file in flight. After importing to RAM (which is very fast), the import doesn't actually write to flash until it asks you (you could cancel that prompt or just ignore it). I don't think that many places in the flight code actually read the flash settings, they just use the RAM copy of the settings. I know that AutoTune reads the flash, but you don't need that.
yes, I have experienced that the import to RAM is very fast. Saving to board is slower but not needed for this ends at all.
Considering LibrePilot is open source, this means it will not be a dead-end to pursue further elaborations.
-
Forgive me cause I did not read the entire thread so I am not 100% familiar with your setup, but Revo can do additional outputs on FlexiIO port too, giving few more available servo outs. That all depends on what you already have configured there.
I see however that current next does not support mixing "Output" options with anything but PPM on FlexiIO, but that shouldn't be hard to fix, at least for test.
-
Wait a second, my FlexiIO port is free and not used at the moment.
I checked with the documentation, seems you are right mr_w - I have 6 more servos and motors to play around with (the first six ports are already in use by 2 motors and 4 servos).
Very cool!
FlexiIO
It's primary use is as a receiver port compatible with PWM and PPM receivers, but can also be used to provide up to 6 additional PWM outputs, serial telemetry port, and two GPIO pins.
https://librepilot.atlassian.net/wiki/display/LPDOC/OpenPilot+Revolution (https://librepilot.atlassian.net/wiki/display/LPDOC/OpenPilot+Revolution)
-
But what is this?
I see however that current next does not support mixing "Output" options with anything but PPM on FlexiIO, but that shouldn't be hard to fix, at least for test.
So for LP Rhino its okay but in next release this will not work.
Must be for a good reason?
This is too early to ask but How to fix that?
-
This only means that you can use FlexiIO pins as either 6 outputs, or as PPM input + 5 Outputs. There is no way (but should be fixable) to use the additional UART there (for whatever reason you might need it) + 4 Outputs. Those restrictions apply to FlexiIO port only, so if you are not using anything else at FlexiIO right now, it should not make any difference either with 16.09 or in current next.
-
;D
-
Some progress here.
Had several petty problems like having the motor screws drawn too far and connect with the copper wiring and ruining the motor and having to order new ones and having to redesign the tilt rotors out of clumsiness again ::)
Anyway, I added a third motor at the rear thanks to the tips here, the additional output ports is provided by configuring the FlexiIO port. So that ESC/Motor is setup as output number 7. No issues at all :)
And it did lift off and hover today in the park :)
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6354;image)
However it was short and next time to lift off the Revo board died :(
I do not think it was caused by the 3rd ESC/Motor since i had been running it before inside many times to test it.
No, I think it might have become overheated?
The temp in the park was 33 degrees celsius (in the shadow) but it was in direct sunlight while doing these tests.
Anyway, now it just show the green power light when connected via USB cable to the GCS. The two other LEDs do not light up and it can't be recognized by the GCS as a board at all.
So I count it out as dead?
Not too alarmed about that. I had one functional Revo that I have installed and will try tomorrow again.
But temp is forecasted to be the same over 30 degrees.
Should I better give it a pass?
Could it be other reasons for the Revo to give up?
I am really eager to do more hover tests...
-
FYI: FC's should really only be powered by 5V even if the spec says more.
Outside temperature of 33C should not be a problem if FC has any ventilation. I do temperature calibration with the board going up to 70C because I see it get that hot inside it's enclosure in the summer when sitting on hot ground.
Does Windows do anything at all when you plug in the USB? In Linux (maybe Mac) the lsusb command would show a difference (new line added) when you plug it in.
-
I checked again the voltage feeding the fc and its 5.0v just like before.
So the reason for the death of the board is likely not heat stroke but something else.
Have a Mac, there was a command used in Terminal to check if there is a change/new line when the USB is inserted, but I forgot it.
-
Have ordered a new Revo unit, but meanwhile tried to get a RevoNano unit and a Oplink transceiver unit i have to work.
And interestingly, after struggling and w good help from people here, got all functions from the Revo to work with this little Nano unit:
PPM-in from Oplink aboard RX, PWM output, Volt/Current meter, External mag/compass (no need for a Revo).
Tomorrow, off to park and resume hover tests :)
-
Good to hear that you are back to the flying part of the hobby!
-
The flying joy this sunday was from my 100 sized Heli, training flips :)
The wing has got most things working now, tilt rotor mechanics, the two main motors, but do have problems with the 3rd motor, not running clean.
Dont really understand where to troubleshoot it. It calibrates fine with Vehicle setup for ESCs and also when trying it in the real configuration in Output, test outputs - runs smooth in low mid and full throttle.
But when using the TX controller only the two main motors runs fine and the 3rd (rear) don't run clean, have stuttering, often enough to be a concern and make it impossible to fly.
I could change motor and ESCs again but I doubt thats really the problem, since it works fine in Test outputs?
(I think I have gone over all the tips of causes of stuttering noted before like, too thin wires, too long wires, if connected to current sensor etc and It has been working before with the this hardware...)
-
That sometimes happens when the ESC and motor don't like each other. They can work with smooth changes, but not abrupt changes. Are you using manual thrust control or some baro control (could be jumping a little)?
If the PIDs are tuned too high it might do this. You might try manual (no stabilization at all) just for a test to see if the third motor runs smoothly, but held down and not allowed to fly. You could try Rate mode with manual thrust too and actually hover test a little bit.
-
Thanks Cliff.
I tried all tests in manual mode. Also did the ESC calibration as detailed here https://librepilot.atlassian.net/wiki/display/LPDOC/ESC+Calibration
and feel it really completed okay.
However, the 3rd motor still not clean running, especially when applying some input like pitch. I then moved one of the fine working large motors form the wing spar to the rear, to see if it would work - it did not. Same stuttering.
My conclusion is that the ESC driving 3rd motor should be replaced. All 3 ESCs are the same BlHeli 20A.
I also tried with then same brand of motor as the ones working fine, EMAX, result the same though.
Does it matter what motor you have connected to the ESC at time of calibration? I would assume not.
-
Do the pids have anything to do with the 3rd motor? What I'm getting at is maybe the two dominant motors need bigger pids and the 3rd stabilize motor doesn't require the amount p or I? And that is why the studdering? Just a suggestion good luck
-
There is a form of stabilization oscillation (caused by PIDs too high, but don't just reduce PIDs, it can be inconclusive) that can effectively move the throttle toward half stick. That is why I suggest a test with manual stabilization (none at all) to see if that motor is still jumpy. Just set to manual stabilization with manual thrust, tie it down or hold it, and run the throttle up to where the motor should jitter.
If it doesn't jitter in manual/manual mode then we know what the problem is. If it still jitters there, but works fine with "test outputs" then I would guess the RC control signal may be jittering for any of several reasons.... but you said the other motors are smooth while this is jittering. The third motor is entirely controlled by the normal throttle channel isn't it?
-
Thank you guys and sorry if I expressed myself unclear.
I did try it in manual stabilization for all axis, roll pitch yaw and thrust.
It gives the same jittering result as with all set to rate or attitude stabilization.
All three motors are controlled by the same RC control signal, PPM.
In the Vehicle, Custom, channel/type-table all the three motors have source curve 1, throttle. The third motor also have -64(50%) set for Pitch. If I set this to 0 it still have the same jittery result.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6376;image)
However, what does seem to have great impact is to lower the max in the Output setting all the way from 1900 down to 1700.
How to understand this? Its the same brand and type of ESCs for all three motors.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6378;image)
It seems to give enough rpm and lift even with the propeller on. If this works in the field as well, then I can live it.
-
Has motor #3 ESC been calibrated at 1000, 1900 like #1 and #2?
Also if I recall correctly, the motor (LP code) has about 255 steps running from -127 to +127. If the motor is already running +127 and you add more (like from pitch) it won't actually go higher. This should not cause jitter though.
It really sounds like the ESC/motor does not like to run at 1900, but you should be able to see jitter at 1900 with "test outputs" if that is the problem.
-
Yes, all 3 motors calibrated together 1000 to 1900 using the Link function. Followed the HowTo article to the dot.
Did "test outputs", and yes it starts jitter after 1650, but before that is smooth. The other 2 can go all the way to 1900 smooth though.
At 1650 it has really high rpm that feels like full throttle capacity of motor.
Hope this is not a sign the motor or the ESC will soon crash...
-
All my might tells me the ESC for the 3 motor is the culprit.
I have replaced it. Running tests now...
-
Yes. Seem okay now after replacement of ESC of #3 motor and redid the calibration.
All motors running smooth with similar Min, Max and Neutral settings.
Also simplified the output mixing so the pitch is only controlled by main and rear motors (no help from tilting the rotors like before).
Next is test in the field!
-
https://youtu.be/DS6Tgx63QSc
-
Just a Bicopter with a a frame that's looks like a wing...
but not flying/cannot fly like a fixed wing. :)
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6380)
-
https://youtu.be/gd4fbf9p71Y
Soon, when I turn the tilting mechanism installed.
-
Maybe...
Did you plan some gliding phases ?
Here is a good example of a working configuration:
https://www.youtube.com/watch?v=eDN7nXEO_4U
-
yes.
(https://uploads.tapatalk-cdn.com/20170722/bf1e171ec857ee546f923a66da54e550.jpg)(https://uploads.tapatalk-cdn.com/20170722/3b05780eb1be7047cdb6148ba8e5f8f4.jpg)
通过我的 Nexus 6 上的 Tapatalk发言
-
The only videos with the real one only show hovering or i didn't found a video while doing cruise flight...
Do you find the stall speed according to the wing area you have in your Rc model ?
-
The biggest characteristic of this structure is stepless Smooth transition。Because of the existence of half a rotated flight mode。
-
Great to see you back Zipray :)
Yes, get those tilt rotors in there and it will be amazing to see what will happen.
I do know the other successful amazing builds you made before, so keep it up, and show the way :)
-
karla,glad to meet you.Are you still in Beijing?Because the problem of network and language I don't often come here,But I'm willing to share with you my findings.Some of fluid mechanics and flight control parameters of the specific issues I can only express clearly in my native language,Maybe your Chinese level be better than me.First of all, you need to do a bicoper,then let it along the CG rotated and Consistently maintain the CG same as the wings.It should be pointed out that tilt rotor shaft and thrust vector is not the same thing.
https://www.youtube.com/watch?v=HTbaRLie8ws
(modified by cliff to change it to youtube.com for better viewing)
-
Yes, still in BJ. Thanks for sharing the video Zipray.
No need to explain, just good inspiration to see your builds.
That looks like a bi-copter with setup to use the gyroscopic forces to assist stabilization and control, an OAT (oblique angle tilt).
Do you intend to put that in and combine it with a fixed wing frame later on?
That would be interesting.
I have not seen anyone do that yet.
I considered it for a while, but could not figure out how to do it and put it aside.
-
... I guess if you add another motor dedicated for forward flight when using the wings, then it will work similar to video by f5soh some posts back (multi rotor + fixed wing). But making the same OAT rotors (have 450 angle) change in to some thrust vector style for forward flight, might be doable but require some complex mechanic solution?
-
Yes, use a single forward motor is a simple solution.However,The plane was inspired by the Augusta westland ProjectZero.Prototype aircraft using variable cycle/collective pitch.And I think the OAT bicopter can replace it.Doing so requires complex CFD analysis and airfoil design.In fact,you can be understood as the wings rotate relative to the OAT Bicopter.
https://www.youtube.com/watch?v=uu88P-GWrLY
https://www.youtube.com/watch?v=Cr8gPqN9YkY
-
karla,glad to meet you.Are you still in Beijing?Because the problem of network and language I don't often come here,But I'm willing to share with you my findings.Some of fluid mechanics and flight control parameters of the specific issues I can only express clearly in my native language,Maybe your Chinese level be better than me.First of all, you need to do a bicoper,then let it along the CG rotated and Consistently maintain the CG same as the wings.It should be pointed out that tilt rotor shaft and thrust vector is not the same thing.
https://www.youtube.com/watch?v=HTbaRLie8ws
(modified by cliff to change it to youtube.com for better viewing)
Is OAT just a bicopter with the motor tilt set to something besides straight forward/backward? It looks like it works well. Do you have a link to description of OAT; why to use OAT?
I have wondered why I haven't seen much about bicopter having a minimal roll control just needing some yaw mixed in.
-
Here is one description how to set it up
http://wiki.openpilot.cc/bicopter-hardware-setup.html
-
...And I think the OAT bicopter can replace it.Doing so requires complex CFD analysis and airfoil design. In fact, you can be understood as the wings rotate relative to the OAT Bicopter.
Fascinating Zipray. Good luck, will be great to hear your progress on it!
-
So glad to finally be able to fly this thing rather than just do endless soldering jobs:)
https://www.youtube.com/watch?v=zvks7SukMtI
The first lift off is without the front dome to keep the CG just above the wing spar.
I want to test the most likely configuration that will work in hover mode. And it did :D
Then, the second lift off I wanted to try how to hover when being prepared for forward flight, its with the dome on and that moves the CG to a more forward position so when the I tilt the rotors to forward flight mode, then CG will move to just above the Center of lift of the wings. About 2 cm forward.
It turned out it is a bit heavier to maneuver but it does work.
So all in all, pretty good state for taking next step and try turning the nob to make the rotors tilt forward.
That will be next test flight.
BTW I am flying in Attitude mode with complimentary stabilization, but the magnets show very green always so I have good hopes to do position hold etc in Hover mode when changing to ins13 mode.
-
Very interesting!
Now I want one. Impatiently waiting for forward transition and forward flight. :)
FYI: I have recently built a fixed wing with Revo and GPS/mag. I found that the mags on it can be affected when sensor is is 20cm from cowl hold down magnets and 30cm seems about safe. I watched mag scope while moving a the fuselage/cowl magnets around the sensor to see what is OK or bad. I also see that my lipo alarm (with little magnet speakers) needs to be over 20cm away from sensor too.
-
Haha, yes really!
The estimated stall speed for the wing is around 30 km/h, its pretty fast, so I think I need a larger field to try it out, there are too many high trees close by.
I am glad you mentioned about the magnet lock down can affect the Mag/GPS-sensor!
Been thinking of that and got strange results like the onboard mag is more accurate than the external one on the wing (should be the other way around). I will check these mag-lock things - 30cm is a lot.
-
Hope to find a better flying field tomorrow Sunday,
meanwhile relocated the GPS to avoid most disturbances, far out on the right wing side.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6436;image)
It looks like a ublox gps but its just the cover case, inside is the OP GPS, I do not have another cover case.
-
Did you know that linear antennas like your 433 antenna on the left side have a "null" off of each tip so that the way you have it mounted you will have very little connection when the nose or tail is pointed at you (and it is a long way away).
433 is a pain because of the large antennas.
Standard procedure mounts linear antennas vertically to reduce this problem to only happening when you bank steeply.
The turnstyle/turnstile antenna is a controversial antenna to try to avoid this issue.
-
Thank you Cliff, yes I do.
The good thing with 433 mHz is the penetration though, and as I am just now flying line of sight this is still such an overkill at 100mW I am sending. This is despite the fact that I not only mounted it horizontally but also I am using a 1/4 wave length antenna and not the recommended 1/2 wavelength. But its so much more convenient for me - a clumsy guy, to bring in my car without smacking the antenna around here and there :)
I feel really confident the control signal will not fail me - and believe me I have spend a lot of time on getting the OPlink to work :) Trust it with my heart now.
So, today (video is a little 'more of the same') I did find a new place to fly where I can stretch out in full forward. Its not ideal since there are some water to the left but I feel its enough of land to point to if things go bad. Good things are, there are no people, no dogs, no kids, no buildings, no cars :) Unbelievable in Beijing. However, its kinda far away so I have little time to fly there.
https://www.youtube.com/watch?v=YUI4nU3R4zU
The purpose today was to try turn the nob of tilt rotors to at least 25 or 50% forward flight.
However, as I started out it was a lot more wind that last time and the pitch was going back and forth and it did not feel comfortable. Before doing smt out of comfort zone then at least need to have a safe retreat place, like a comfortable hover. So did fly some 4 rounds and two LiPos to get PIDs better in this wind. Not really happy with result though, will take more time. I do have a good feeling about it though.
BTW, I am using Attitude with INS13 now for stabilization. The external mag is beautifully aligned.
-
You might try AutoTune, but look at the new PIDs before flying them.
-
Interesting idea.
However, when I set Autotune as a Flightmode I get the red Config error warning.
I assume this is because this is a Custom craft?
-
That sounds like a good guess.
I haven't checked the code but I wonder if disabling sanity check would allow you to do it...
-
;D yes it worked.
But will autotune work?
I don't know how it can figure out correct settings - its really magic :)
But if you think I can give a try, then can I also give it a try on my Heli's?
Too good to be true?
-
The main problem with AutoTune seems to be bad results afterward, not crashing while flying the AutoTune flight.
It does induce 75% of full control step outputs. I would guess a typical heli would handle that OK, and I would try it on a smaller cheaper heli, but I wouldn't want to be the first one to try it on a Trex 700.
I would definitely try it on the tiltrotor but check the PIDs before flying them. Especially check for PIDs that are zero.
If it doesn't break because of the large control inputs, then it should be safe to try it up fairly high with a finger on the FMS, ready to switch to Attitude mode.
-
Thanks again Cliff.
This is a bit odd, the next day when I tested and wanted to use Autotune I could get around the Red Config error on the flight mode page, but the Health gadget had a red Config Alarm. Will not arm. The only way to get it green again was to remove AutoTune as flight mode. After doing that, there was still an Orange warning on the ?-box. That did not go away until I put the sanity Checks back in operation. Maybe there is another workaround I don't know. Would be interesting.
Anyway I tuned the Pitch manually and got better hover stability.
Unfortunately I leave tomorrow on a long trip and can't bring the wing along.
I don't expect I can update this project until end of September
- Turning the nob for forward transition and forward flight seems a cliff-hanger ;) until then - sorry.
-
Hey Karla (and many others)
It has been a little while since our last comms. You are making some serious progress! The Mako looks great.
I saw this machine advertised today. Check out the GCS view in the video. Looks familiar to me.
https://www.youtube.com/watch?v=cCm-IvRo2rc&t=31s (https://www.youtube.com/watch?v=cCm-IvRo2rc&t=31s)
Video of it hovering.
https://www.youtube.com/watch?v=Dl2_uPTDrug (https://www.youtube.com/watch?v=Dl2_uPTDrug)
But have not seen any transitioning yet.
MJ
-
Hey Emjay!
Thanks and good to hear from you mj.
Yes, thats really interesting, surely based on LP, encouraging :)
Hope they publish a video when they fly hover mode and also the transition to and from it.
K
-
Oh these guys seem serious about this, aiming for supersonic, manned flights!?
Anyway, it seems to me from the videos that for hover they are not using OAT (oblique angle tilt), nor FAT (forward aft tilt), but using the small rudder below the motors for stability and control.
Interesting.
Seems they got the hover okay and also the forward flight okay, but how about the transition?
... and using CC3D according to the videos...
-
They are using a revolution board and LibrePilot
CC3D do not support current sensor.
https://www.youtube.com/watch?v=cM7VG31ctwE
-
Hope everyone had a good summer!
At least I had :) Now back and have resumed this project.
First thing I took it to the field and finally turned the nob for forward transition and forward flight. The tilt rotors worked fine and the wing started to accelerate forward. Then there was a mishap, the gear on the starboard tilt house derailed in a yaw correction of a wing gust. Likely due to the servo was not totally fixed and had moved. The motor tilted forward and hit the wing, resulting in a crash. Not so bad but the motor had broken off from its fix and broken the tilt house.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6730)
To fix damages and resolve this from happen again I:
. 3d printed a new tilt house
. enhanced the motor fix to the tilt house with longer screws and stronger fundament
. glued the servos with hot glue before securing with the normal 2 screws
. redesigned and 3d printed 'fatter' gears that will sit on the wing spar (prevent derailing)
It turned out well, and its working better and more reliable and feels durable.
Next thing was to fly it around in hover mode to make sure it worked. Then the motor on the other side started to run unclean and stutter. Had to fix that. Recalibrated all ESCs and eventually replaced the ESC. Still some stutter. Then by accident I moved around the wire (signal wire pair) running from the FC to the ESC and the stutter stopped. It seems the wire was touching the Power Distribution Board and got some disturbances there? I don't know if that can happen but rerouting the wire made the stutter go away. So I assume it sorted for now.
Will resume things after the Chinese Holidays.
Still have a good feeling about this design.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6732)
Anyway, this summer I had a chance to fly some of my other standard fixed wing, using hand launch and landing in the grass. Getting better after training at not crashing at launch or at landings :)
That was one of the drivers for me to explore a VTOL craft to begin with.
Edit:
I forgot to mention the third issue. I was so happy and impressed with the configuration of the Revo Nano board getting three functions out of the FlexIO port: PPM, A 7th port out to control rear motor/ESC and the volt/current meter input. However, I was using a OPLink receiver for the PPM and it occasionally ended up in failsafe mode, then coming back online again. I could not figure out why or repeat the problem. When recalibrating the transmitter manually it first became better for a while but then failsafed next time. In the end i could not even complete the manual transmitter calibration process. Channels were jumping and jittering around after I had already done the min and max movement with the sticks.
(actually I tried with another OPlink rx as well, but it did not solve the problem, so its maybe some issue after the rx)
I just did not have the patience to spend more time on it, but replaced the Revo Nano + OPLM Rx for a Revo board I had idle.
Then all is stable and no failsafe.
Took some days to complete though.
-
Heyjo :) some progress here,
Today I did turned the nob for forward flight and it did fly away!
I made the first left turn since the field is very small but I could not make it climb fast enough over a tree so it was crash, however mild and not fatal. Sorry i do not have a video.
Will rebuild and retry next weekend.
-
:'(
My Beijing park announced they have now banned all flying aircrafts.
Have to find a new field.
Spend all Sunday browsing around some 20+ km radius.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=6766;image)
Think I found a place, tucked away from everything enough I hope, however need to buy a pair of sturdy rubber boots to be able to walk out and pick up my VTOL in the muddy field, if have to land not VTOL.
-
Last weekend I was at the site at two different times, but both times it was just too much traffic with people walking, on bikes, scooters and cars and trucks around the field. It does not feel safe or comfortable for test flying. Its really difficult to find a place for fixed wings here. Its easier with helis and quads. I will continue but now considering to bring the plane to my fields in Europe in December or Hainan in Feb. :(
-
Will not be flying and testing it this time in Europe, time too short this trip :'(
Next shot in mid feb in Hainan :P
-
...
I have now got all the servos and motors set up correctly so they move respectively spin in the correct directions as well as make corrections correctly when I tilt the flight controller around while in stabilization flight modes.
Very Good!
...
Hi. Sounds like you have done a great job. Did you have to modify the code to get the results you were after or were you able to do it all using the "Custom" tab in the configuration user interface?
I checked the Wiki but couldn't really tell what is possible with the "custom" tab. Can we turn on the accel during hovering and turn it off when we are at 45 degrees tilt or in forward flight modes?
-
The simplest way to turn the accels on and off would be to use Attitude mode (accels on) and Rate mode (accels off)
Attitude mode for hovering, and Rate mode for forward flight (aircraft probably rotated 90 degrees).
-
@ TheOtherCliff- Thanks for the accel switching advice. Much appreciated.
@karla- Would like to see a video or screen shots of the configurator program of how you set you bird up if you get a chance.
-
Hello RcV,
Not sure I did a good job at all, still waiting for an opportunity to flip the switch to forward flight, and then we will see.
But, I have done no change of any code, its all inside the configuration of LP
-
@karla- Would like to see a video or screen shots of the configurator program of how you set you bird up if you get a chance.
Here is the configuration file attached. Its on LP ver 16.09.
I only use two flight modes, only difference is that yaw is stabilized in hover and its not in forward flight.
Later I see if I can explain it better, maybe after a successful test flight :)
-
Thanks karla for the info.
Took a closer look at the configurator program from some old youtube videos. Looks like there might not be enough mixing capabilities for what I was thinking of. I would like to adjust the transmitter influence and sensor influence separately for each axis. Also would want a minimum of 7 modes.
Hover, 45, and Forward Flight. Hover to 45 and 45 to Hover, along with 45 to Forward Flight and Forward Flight to 45 mode, so I can trim all aspects.
Would also need to the ability to trim each axis for each mode.
Will keep an eye on this project though.
-
Not sure what you want to do but there are 6 flight modes and each of them can have separate trim/stabilisation for each axis. How about simplifying your set up and just do one trim for hover and another for forward?
After several tests I put the transition from Hover to Forward on a transmitter nob that can be turned from 90 to 180 degrees seamlessly to have less abrupt motions on my tilt servos.
-
...How about simplifying your set up and just do one trim for hover and another for forward?...
Thanks again for the help karla. Much appreciated.
I have done several tilt rotor VTOL airplanes using modified MultiWii but I never got good results when using the accel for auto-level so I'm looking for a new FC to play with. My new little T-28 from E-flite with SAFE-Mode, in other words it has an accelerometer for auto-level, is so easy to fly in very gusty winds that it's sold me on how helpful an accel can be if implemented well. I've tried to fly it will safe mode off and it just becomes uncontrollable, which really shows me how much work the accel is doing for me.
I have seven modes that get triggered automatically based on the tilt servos position. I found that when switching from one mode to the other, the aircraft wants to pitch quickly either up or down. The 7 modes appear to cover most situations so far.
Hmmm. Perhaps if the accel is good enough, I don't need so many modes to trim out a level transition?
Control swapping is a must for me though. My brain can't do the mixing of hover roll turning into forward flight yaw and such. I think I recall you posting that this FC doesn't do that yet?
-
Oh, good to know you and that you already did several VTOLs!
Is the fixed wing frame you use this one:
https://www.youtube.com/watch?v=N2JZSOEjwtI
How do you fly it (like hovering?) and what do you wish the FC to do for you?
Really not familiar with this :)
Best
-
Sorry for the confusion. I only use the T-28 as a regular airplane, but it uses an accelerometer that makes it super stable is all I meant. Not sure if the one you posted is the micro version with "Safe" technology or not. Mine is all yellow in color.
This is one I built-
https://www.youtube.com/watch?v=5LCW_I5eu7A
It only uses gyros though and only the P parameter of the PID is active. I disabled I and D since they didn't appear to do anything useful.
So I'm interested in trying similar builds with a better FC. Is additional VTOL code functionality being actively worked on or has it remained the same for a while? Glad to help with configuration and feature suggestions if helpful. Can't help on the code side though I'm afraid.
-
If a car cruise control (often a PID controller) only had P term then when you set it to 60 it would do maybe 55 on flat ground, 50 climbing a mild hill and 65 coming down the hill.
For quads, if you don't have enough "I" term the most obvious symptom is that the quad won't hold correct pitch angle when flying forward fast.
D term is mainly used so you can run slightly higher P. That one you can actually live without.
-
Is additional VTOL code functionality being actively worked on or has it remained the same for a while? Glad to help with configuration and feature suggestions if helpful. Can't help on the code side though I'm afraid.
To my knowledge there is no code developed specifically to be used in any VTOL situation.
However, the LibrePilot code as is, has plenty possibilities to be configured for a multitude of VTOL uses though.
Once you get the hang of the GSC, Configuration, Vehicle, Custom screen, then you can really experiment with different setups (and not only VTOLs).
-
One of the main things that VTOL needs is a smooth transition from nose up mode to nose forward mode and all that entails such as mixing in some changes to Rotate Virtual (say add 90 degrees to pitch) and mixing two "mixing tables" (say basically aileron/elevator/rudder controlling the rudder/elevator/aileron when in hover mode and controlling aileron/elevator/rudder when in forward flight).
It's been discussed, but needs a developer who wants to work on it.
-
...For quads, if you don't have enough "I" term the most obvious symptom is that the quad won't hold correct pitch angle when flying forward fast...
Thanks for the info. I thought it was strange that it worked with no I and D also. As soon as I put in some I it would just drift all over the place. Don't really care to mess with that old code after experiencing what can be achieved these days.
-
...For quads, if you don't have enough "I" term the most obvious symptom is that the quad won't hold correct pitch angle when flying forward fast...
Thanks for the info. I thought it was strange that it worked with no I and D also. As soon as I put in some I it would just drift all over the place. Don't really care to mess with that old code after experiencing what can be achieved these days.
...One of the main things that VTOL needs is a smooth transition from nose up mode to nose forward mode...
I did one tail sitter VTOL flying wing model. Don't think I got any video but it did the entire transition. Was using the same FC though, so just gyro with the P PID. Really want a good accel to add into the mix. I'm working on another Osprey type bicopter VTOL so I don't need to rotate the assumed "level" angle on this build.
...the LibrePilot code as is, has plenty possibilities to be configured for a multitude of VTOL uses though...
Sounds good. I'll take a look at the custom tab some more.
Thanks for letting me interrupt your thread karla. You have both been very helpful in answering these questions for me and providing great suggestions. I'll keep an eye on this project.
-
Thanks for letting me interrupt your thread karla. You have both been very helpful in answering these questions for me and providing great suggestions. I'll keep an eye on this project.
Haha, no worries, not interrupting, visitors welcomed!
Provides inspiration to this very slow project. I lack suitable flying fields to test it. Its easier to find a spots for quads and helis.
Good luck :)
-
Hello all and Merry Christmas :)
Just arrived to a place where flying is legal.
It looks like this VTOL has gone to the dark side after so long with no opportunities to stretch out from hover to straight forward flight :) just look at his eye brows.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=7957)
Here is a very short video of transition.
I changed the transition from hover to forward from a on-off switch to an accessory turnable nob, in order to have some better control of how much the tilt rotors should tilt.
https://www.youtube.com/watch?v=RgPc2U7CKP4
I turned it some 50% from vertical towards horizontal and it took off just fine.
Will need to do a lot more testing about this.
At least I have some time here to fly.
-
How do you handle the difference between tilted forward thrust and tilted vertical thrust? Do you switch to manual (disabled) stabilization for forward flight?
-
At the moment, I use Basic Attitude Estimation Algorithm with stabilization mode for Thrust, kept at Manual.
That is both for hover and forward flights.
Actually, the only difference in stabilization from hover to forward, is that the yaw axis is at AxisLock in hover and at Manual in forward flight.
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=7962)
The reason for this Yaw stab settings is that in Hover mode it really helps not have to think about the yaw direction. And in Forward mode, I don't use the yaw stick at all (rolling with the elevons makes the wing turn), plus the fact that, if the FC would try to stabilize the yaw, then it would mess up the flight completely since it would now affect the roll axis.
This sounds like I know what I am doing, but I am really just experimenting :)
But it make some sense to me at least.
The flying behavior will be the judge.
Too windy today to fly.
Hoping for tomorrow.
-
So it sounds like you have elevon servos set up for roll (/pitch), but you also have left-right motor differential set up for roll (and differential motor tilt for yaw) and all of this (except motor tilt in switch position #2) is always active? Are the elevon servos stabilized or just manually connected to the RC aileron channel somehow?
So in forward flight, you probably get yaw effect when you use the roll stick and don't even need elevator for turns? So even if you used Rate instead of ATtitude, it would be hard to do a roll?
I like flying fixed wing in Rate or Manual mode (to allow loops and rolls) more than ATtitude mode. And because I have flown unstabilized fixed wing for many years. :)
An interesting coding project might be to automate all this:
- in hover flight mode, differential motor thrust gives roll, so it is connected to roll stick and roll stabilization, also differential motor tilt gives yaw so it is connected to yaw stick and yaw stabilization
- in forward flight mode, differential motor thrust gives yaw, so it is connected to yaw stick and yaw stabilization, also differential motor tilt could give roll, so it could be connected to roll stick and roll stabilization; it could also be locked pointing straight forward.
- these act like two flight mode switch positions but don't use the flight mode switch, they use a transmitter knob to control the transition from one mode to the other: tilts motors proportionally and also effectively proportionally mixes everything between hover mode and forward flight mode
- different PIDs for the different flight modes, just like when using the flight mode switch, PIDs mixed by the knob too
- options for elevon or elevator/aileron setups, maybe even V tail
Also: I see that you have what looks like a 433mhz antenna mounted horizontally on the wing. I guess you understand that your range will not be good to the front or back (antenna nulls) and that the ground antenna should also be horizontal (for polarization issues) (also bad). It is OK at high power and close range, but both ground and air antennas should really be mounted vertically. :)
-
Cool idea Cliff,
So it sounds like you have elevon servos set up for roll (/pitch), but you also have left-right motor differential set up for roll (and differential motor tilt for yaw) and all of this (except motor tilt in switch position #2) is always all active? Are the elevon servos stabilized or just manually connected to the RC aileron channel somehow?
All above correct!
Always active, and yes, elevon servos mixed in with stabilization.
So in forward flight, you probably get yaw effect when you use the roll stick and don't even need elevator for turns? So even if you used Rate instead of ATtitude, it would be hard to do a roll?
I like flying fixed wing in Rate or Manual mode (to allow loops and rolls) more than ATtitude mode. And because I have flown unstabilized fixed wing for many years. :)
Yes, should get yaw effect since turning left (roll/elevons left) will increase right motor and decrease left motor speed, resulting in left yaw motion. A roll could be twisted, but I really do not know. For loopings, they might be quicker and tighter due to the tail rotor pitch (could also configure it as a reversible motor to facilitate that).
An interesting coding project might be to automate all this:
- in hover flight mode, differential motor thrust gives roll, so it is connected to roll stick and roll stabilization, also differential motor tilt gives yaw so it is connected to yaw stick and yaw stabilization
- in forward flight mode, differential motor thrust gives yaw, so it is connected to yaw stick and yaw stabilization, also differential motor tilt could give roll, so it could be connected to roll stick and roll stabilization; it could also be locked pointing straight forward.
- these act like two flight mode switch positions but don't use the flight mode switch, they use a transmitter knob to control the transition from one mode to the other: tilts motors proportionally and also effectively proportionally mixes everything between hover mode and forward flight mode
- different PIDs for the different flight modes, just like when using the flight mode switch, PIDs mixed by the knob too
- options for elevon or elevator/aileron setups, maybe even V tail
I think I know what you mean...
The current Mixer table (actuator mixer matrix?) looks like this and delivers all that you pointed out.
Ch 1 and 2 - left and right front motors
Ch 3 and 4 - left and right motor tilt servos
Ch 5 and 6 - left and right elevon servos
Ch 7 - tail motor
(https://forum.librepilot.org/index.php?action=dlattach;topic=3334.0;attach=7964;image)
Now, you suggest to simplify this table to take care of only Hover flight and then create an additional table simplified to take care of only Forward flight. Call them mix matrix A and B. Then to create a function that will transition Ch 1 to 7 from 100% A to 100% B proportionally by turning of a transmitter nob.
You mean something like that?
Sure! the interest for VTOLs is there and the available FCs to handle it are not many.
The KK2 board that most VTOL vehicle projects on the internet seems to use, are getting more and more difficult to get hold of? Someone did his own FC to fix this, but anyway, seems easier to add on to a mature fully gps enabled sw like Librepilot, than build from scratch.
Also: I see that you have what looks like a 433mhz antenna mounted horizontally on the wing. I guess you understand that your range will not be good to the front or back (antenna nulls) and that the ground antenna should also be horizontal (for polarization issues) (also bad). It is OK at high power and close range, but both ground and air antennas should really be mounted vertically. :)
I know. Thank you. Its just a very convenient and robust mounting during testing.
-
I haven't tried to figure out everything that would need to be changed.
I believe it would need two read only mixing tables (the two endpoints) plus the current mixing table which would be some linear combination of the two fixed tables. It would also need to mix everything from two switch positions, including the PIDs.
The mixing table takes care of outputs. On your aircraft, the FC is at the same attitude when hovering as when in forward flight, so that may be most of everything you need (plus PIDs if you want them too).
It would be nice to allow for a hovering airplane that does not have tilt rotors. It would hover with nose pointing up but fly with nose pointing forward. I imagine it would be nice for this to allow the roll stick to do a hovering airplane rudder yaw in hover mode and a flying airplane aileron roll in flight mode. That would require remapping the inputs and stabilization as well.
Of course, the easy way to do this (non-tilting motor) doesn't require any code changes. Attitude or rate mode in hover and rate mode in forward flight. The pilot would have to use the correct sticks. Yaw in hover becomes roll in forward flight, etc.
Finally, the transition knob could be replaced with a knobless timed transition that is activated when you use the normal flight mode switch to switch from hover to forward flight mode.
This would allow other such mixings as well:
- X to + quadcopter
- flying a quad backwards or sideways like it was flying forward (we already have CourseLock)
- (auto leveling) fixed wing knife edge or inverted flight
- attitude mode in either hover or forward flight