PDA

View Full Version : How to Setup Your Skylark Stabilize and RTH. (or any other plane for that matter)



Toysrme
10th May 2013, 05:34 PM
MAKE THE PLANE FLY TRUE WITHOUT THE AUTOPILOT/GYROS! That is your mantra. Zero radio trim in calm conditions.
We need to get something out of the way right off the bat. Having success using ANY form of electronic stabilization relies HEAVILY on how accurate your base-line build is. The more accurate your BUILD is (thrust/wing/stab/fin lines), the more accurate your MECHANICAL setups are (mechanically centered servos, linkages and mechanically correct trim on a centered servo.), the better your RESULTS will be and the EASIER time you will have tuning your setup. The worse electronics your build is, the harder the electronics have to work and the more sensitive to tuning changes they become. It doesn't matter if we're talking an airplane with a full blown autopilot, or the cheapest CCPM helicopter with the cheapest single axis gyro tail rotor.
Bottom line, if you are relying on electronics to make your plane fly true; your thought process is wrong. You need to re-evaluate your physical setup.
:)
So there is a learning curve to the Skylark OSD's stabilization and RTH modes because information is kind of spread about all over the place and then manuals could be more in-depth. Rest assured that they are very easy to understand, setup & work quite well. For clarity purposes this guide assumes you throw out everything the Skylark chingrish manuals have said.

Stabilization (A1) is where the OSD will attempt to keep the plane level. Like flying with Heading Hold gyros on your Pitch and Roll axis.
RTH mode will attempt to continually turn towards the launch point (aileron turns) and more or less circle the starting position once it gets there.

When these modes are active, the OSD will attempt too:
Roll & Pitch the plane so that the GPS/INS sensor is level (where accelerometers are)
Climb to the SecureAltitude setting (NOTE! This is ALWAYS in Meters. If you enter a value in Imperial mode, it will convert to meters!)
If above the SecureAltitude setting, it will attempt to maintain level. (NOTE! The OSD WILL NOT DIVE the aircraft to get to the SecureAltitude setting, but as a plane banks to turn, the nose will drop and you will loose altitude. This is one reason you conservatively limit the Maximum Pitch the OSD can use.)




Ok off we go! A lot of information correlates together.
The goal, is to let the OSD know what your joystick position is when the plane is flying true. THIS ALSO INCLUDES TRIM VALUES!!! You CAN NOT calibrate the OSD if the aircraft has not been flown and trimmed out for level flight.
-----------Servo_Calibrate-----------

Enter calibration mode with your joystick centered for a few seconds.
Give full throw for a few seconds.
Re-center the joystick and click off.
Once done enable Stabilization (A1) mode and make sure the control surfaces move in the correct direction. If so you are done. If not, re-calibrate, moving the joystick in the opposite direction.

You must understand that the OSD's understanding of "level" also has to take into consideration your stick position's definition of "level". By calibrating the OSD and NOT calibrating it with the aircraft "trimmed out"; the OSD WILL stabilize the plane to your non-trimmed direction! It's understanding of "level" is a centered joystick, and seeing your trim values as un-calibrated stick inputs, it will give the EQUAL & OPPOSITE control input as your trim.
--------------------------------------------
Understand that EVERYTHING revolves are the accelerometer inside your GPS/INS sensor. All of the OSD's stability and piloting rely solely on it attempting to orient the plane so that it thinks the GPS/INS is level.

Prop your plane up to it's straight & level flying attitude (I.E. your wing/stabilizer are level)
Mount the GPS/INS so that it is physically level on the aircraft. (NOTE! You must NOT use a surface that is wiggly under vibration. 1/2" thick squishy open-cell foams are NOT tolerable. hard mount to structural foam, or use a foam/rubber pad that is no more than 3/8" thick.)

These two simple user-failures of not calibrating TRIMMED values and not having a reasonably level GPS/INS sensor causes 90% of the stability problems on the Skylark OSD (and many others). When you enter Stabilization (A1) mode and the plane IMMEDIATELY banks and yanks, your GPS/INS is not level, or your joystick+ trim values are not calibrated correctly.



Back to terminology & definitions!
I'm just going to tell you as long as you are in relatively flat ground & not drastically pushing your range, Set your SecureAltitude setting to 360ft / 110m starting out with.
-----------Control_Misc-----------
MaxRoll=15-20* (Increase in extreme conditions) How far the OSD will roll the aircraft when turning in RTH and how far the OSD lets YOU bank the aircraft in Stabilization (A1)
MaxPitch-10-15* (Increase in extreme conditions) How far the OSD will pitch the aircraft when turning in RTH and how far the OSD lets YOU bank the aircraft in Stabilization (A1). NOTE!!! When in RTH mode and the Altitude breaks the SecureAltitude by 10m, it will climb at this pitch angle.


Here are your biggies. Your main PID loop. I feel like for many people this will be their first "gyro" experience. So we need to pre-face this section with "Gyro 101-102 theory classes.
There are three main things when using any form of gyro stabilization you need to know about, and what their uses are:
Update rate - How fast the loop runs & responds while checking inputs & outputting to servos (Think action/reaction time)
Gain - How sensitive the gyro is to changes
Let's discuss their relationship. On a helicopter with digital servos or a multicopter with ESC's for control you live and die with your update rate. HUNDREDS of updates a second is preferred. Such craft typically do NOT have inherently stability and the more precise inputs you can get away sending them the BETTER they will fly. (hardware limited)
R/C airplanes do NOT typically respond in kind!!! The reason is two fold!

Our aircraft resist changes in axis and are in-motion. Setup correctly they are relatively stable left to their own device
As your flight speed INCREASES your control authority increases!!!!!!!!!!!

Back to the relationship and our example: If we have an aircraft that flies PERFECTLY at 15-30mph on a standard gyro setup with PROPER gain dailed in (no oscillating) and the update rate STANDARD (between 50-XXX/sec). What happens when this plane's flight speed increases to 60mph? 90mph? 120mph+? The answer is that it will increasingly begin to oscillate! Without being able to scale the gyro VS flight speed, this leaves us with two options:

Turn down the Gain - curing high-speed oscillation, but taking away slow-speed control
Turn down the Update Rate - lowering the ability of the plane TO oscillate at any speed

This is a LARGE mis-understanding when people are setting up their equipment for aircraft especially the terminology Skylark has chosen. Aircraft can be flown perfectly gyro stabilized at single-digit update rates. 3-6/sec is typically enough.
Back to our three main things:
"Control Bias" - The balance between how much servo throw the GYRO is allowed to use VS what the gyro allows the USER to control. I'm sorry I have to make up a term, but terms vary here. If you've ever used a standard Futaba type gyro, this would be what you are doing when you set your control limit with a pot and your turn rate via EPA.
Unfortunately, our Skylarks offer no variation in Control Bias. Measuring my own control inputs using various settings results in the Skylark OSD taking virtually all control available. It reserves roughly 9-10% of the *possible* control through for the user. Which is very little! To my understanding we have no way of changing this (Ive tried every option in the OSD's to see if they cross-reference, they do not/if they do its so little I can't measure a change with my digi calipers)
So what we've established is that a relatively high speed plane respond best with a LOWER update rate and the gain set correctly, while a relatively low speed plane will require relatively MORE gain than a high speed plane because the control-authority is diminished.


One last bit of Gyro Theory you need to know. You should NEVER EVER EVER have to reduce the control-throws to get wags cured. If you have done so, YOU HAVE DONE WRONG!

Whew! Onwards to the show!
-----------Control_PID Menu-----------
Stability- THIS IS OUR UPDATE RATE NOT OUR GAIN SETTINGS. Please let that sink in. This is not how MUCH our aircraft responds to changes but how QUICKLY and SMOOTHLY. (For whatever reason I can not explain, it also has an impact on your RTH circle size, There is a cross connection where your RTH holding pattern shrinks enlarges with this setting. At least mine does! It varies causes my Penguin's holding pattern diameter to change from Approximately 300' (+/-75') to 525' (+/-100').
If your aircraft is oscillating quickly reduce the stabilization rate.
Roll_Aile- GAIN for the Roll axis. This determines how much control throw the system is allowed to use.
Pitch_Elev- GAIN for the Pitch axis. This determines how much control throw the system is allowed to use.
Yaw_Roll- I have absolutely no idea what this does. The Dianmu does not use Yaw axis while the Standard does not document Rudder Stabilization.
Alt_Pitch- How sensitive the Stabilization is to your current Secure Altitude setting.

-----------Control_Misc-----------
MaxRoll - How far the OSD allows the plane to roll left or right in Stabilize/RTH mode
MaxPitch - How far the OSD allows the plane to pitch up or down in Stabilize/RTH mode





Random things you may not need to know, but are handy to know:
The timer starts when the aircraft has accelerated to 5m/s or moved 5 meters from the initial position. (depends on what manual/firmware you believe) Depending on firmware version, some require a 10 meter altitude changes, others do not.
If you're going to use a Skylark AAT, you need audio from the Skylark OSD to be hooked up to the VTX. if not, leave the audio disconnected.
The Filtered outputs support 12v @ 2amps and 5v @ 2amps. That's enough 12v power safely power a 2.5w vtx and 12v board camera. The 5v line can easily run 5v board cameras/vtx's, top-up battery backups (4 cell nicds) or top-off/charge devices like a GoPro. YOU MUST NOT run your R/C gear on the 5v output. Even a single micro servo can pull more than 2amps if it stalls.
The filtered outputs MUST be powered by a higher voltage battery than you intend to run as it's a simple filtered step-down. This means to run 12v you need a 4s+ lipo or 5s+ life pack.
The Skylark OSDs themself requires/will warn you of low voltage at 10.6v.
Alarm modes can be set to flash when triggered, or for menu items to remain hidden until triggered.

Toysrme
10th May 2013, 05:39 PM
My Finwing Penguin flys with:
Stability:15
Roll_Aile:33
Pitch_Elev:35
Yaw_Roll:??? Doesn't change anything when set to 0-10 or 90-100
Alt_Pitch:25
MaxRoll:10
MaxPitch:8


My .46 Nitro P-51 flys with:
8:Stability
22:Roll_Aile
20:Pitch_Elev
Yaw_Roll:??? Doesn't change anything when set to 0-10 or 90-100
70:Alt_Pitch
12:MaxRoll
5:MaxPitch



Very different aircraft. The penguin loafs along at 20-30mph and on occasion can push 75-80mph in dives. With what I call "sedate" controls. Full throws barely do rolls & hold inverted flight. The dihedral, wing area & fuse make the aircraft somewhat self-righting and susceptible to banking in crosswinds (tho nots as bad has high wing pod&booms like an easystar or skywalker). Where the P-51 is fully aerobatic and kisses 120mph level flight. It does not self-right but flies where you point.

Both work peachy keen on the OSD once setup, tho the P-51 must be setup "looser" so that it does not vibrate to pieces at high speed. Making very slow (20mph) gyro stabilized flight slightly sluggish to respond to changes in stabilized mode (RTH mode is fine as the plane will simply always bank to home)

J-DUB
18th May 2013, 06:57 PM
Toysrme,
Thanks for taking the time to put this together! Hell, I don't know if it will solve my issues but I really appreciate you putting your experiences down in print to help us out.
J.

J-DUB
18th May 2013, 07:12 PM
I have a question about "level". Do you think it's safe to assume that the bottom (flat side) of a wing is parallel with trimmed-out level flight?
Thanks,
J.

J-DUB
18th May 2013, 07:13 PM
I have a question about "level". Do you think it's safe to assume that the bottom (flat side) of a wing is parallel with trimmed-out level flight?
Thanks,
J.

chatch
18th May 2013, 07:26 PM
Lots of info here, thanks

MASHTON1138
18th May 2013, 07:57 PM
The filtered outputs MUST be powered by a higher voltage battery than you intend to run as it's a simple filtered step-down. This means to run 12v you need a 4s+ lipo or 5s+ life pack.

WTF???!?!!?!?!?!? Are you freaking kidding me? Why the hell doesn't that show up in the documentation of their product? I was about to start ranting about how bad the filtering was on the Dianmu but now I think the lines are probably from the lack of voltage. I don't own a single 4S battery all my planes and quads are setup for 3S. Now I can't run my equipment on this thing???

Wish I would have noticed statement before I bought this thing.

MASHTON1138
18th May 2013, 09:30 PM
Never mind figured out my problem. WiFi!!! UGGHGHGHGHGHGGGGHHHH!

Anyways mine seems fine with 3S powering all my 12v stuff are you sure it HAS to be a 4S or higher?

MASHTON1138
18th May 2013, 11:36 PM
nope definitely a voltage issue. When I connect a battery that is under 10.8v I get the lines back in the video. This sucks, I guess this plane will sit for a while until I can get 2 4s batteries for it. That shit is so gay. I wish Skylark would get someone who knows how to write instructions to write their manuals and descriptions of the products so important stuff like this is put out and you don't have to dig through 10 different forums and thousands of posts to find the info.

chatch
18th May 2013, 11:40 PM
i'm the opposite... the only 3S batteries I own are in the ground station :p

MASHTON1138
19th May 2013, 12:04 AM
Yeah screw this ****ing thing. I am so pissed off right now. I hate the fact that Skylark can't put shit in their manuals. I can't use the ****ing 12v out of the damn thing and I am not buying 4S batteries just so I can. So now I have to rewire my cam and vtx with a filter inline so I can use the batteries I planned for. The damn thing also doesn't correct properly in Attitude mode during my testing. I manually roll the plane left and the imu tells it to keep rolling that way (left elevon up right down), but everything works fine under manual control. When I have it in A1 the right elevon is perfectly adjusted where it is in manual but the left is about 13mm lower. Pitch is fine if not a little off but at least it works in the right direction. Calibration steps followed exactly and I have these issues, look for solutions to the issues on here and RCG execute those solutions and they mess up the parts that did work right.

The best part is I bet the Trace IV I got for another plane has the same step down voltage out as this thing does so I can't use the output on it either. WTF was Skylark thinking on this one.

chatch
19th May 2013, 12:21 AM
switch your elevon output servos. one plugs into "aileron" and one plugs into "elevator" which are mixed. swap them and the gyro will now put out the correct movements. now go through the calibration process again

MASHTON1138
19th May 2013, 12:38 AM
You mean on the receiver?

chatch
19th May 2013, 12:48 AM
1 2 input -----> osd ---->1 2 output

switch output

1 2 input ----> od -----> 2 1 output

chatch
19th May 2013, 12:50 AM
or pull the opposite way during calibration. instead of full right do full left

doonunder
19th May 2013, 05:08 AM
Yeah screw this ****ing thing. I am so pissed off right now. I hate the fact that Skylark can't put shit in their manuals. I can't use the ****ing 12v out of the damn thing and I am not buying 4S batteries just so I can. So now I have to rewire my cam and vtx with a filter inline so I can use the batteries I planned for.



My PZ0420 and dianmu happily worked on the 3s through the stepdown on the current sensor.

doonunder
19th May 2013, 05:12 AM
switch your elevon output servos. one plugs into "aileron" and one plugs into "elevator" which are mixed. swap them and the gyro will now put out the correct movements. now go through the calibration process again

Ans although a tricky process, depending on which way you start with the stick up, down, left or right will change the direction of control. So if you do ail and start with left then move right then after calibration the gyro reacts the wrong way, then redo calibration again starting with stick right then left.

Now I'm confused whether the calibration calibrated gyro direction or control direction....

MASHTON1138
19th May 2013, 09:20 AM
My PZ0420 and dianmu happily worked on the 3s through the stepdown on the current sensor.

I can too but if I put a battery with low voltage on it the picture goes to crap and I get pulsing in the feed. I don't plan on running it at less than 10.9v but you never know what may happen.

I was frustrated with the thing last night because of all the issues with the IMU correcting in the opposite direction for roll. I am gonna just run it like it is on 3s and work on getting the IMU correcting in the right direction.

MASHTON1138
19th May 2013, 09:23 AM
Ans although a tricky process, depending on which way you start with the stick up, down, left or right will change the direction of control. So if you do ail and start with left then move right then after calibration the gyro reacts the wrong way, then redo calibration again starting with stick right then left.

Now I'm confused whether the calibration calibrated gyro direction or control direction....

You you are saying you have to move in the exact same direction that you went in the calibration for your first movement after exiting calibration needs to be in the same direction? If that works I am going to scream. These are the little details you need in an installation manual!!!

J-DUB
19th May 2013, 03:19 PM
Toysrme,
Problem solved...A1 is rock solid now!
Thanks,
J.

MASHTON1138
19th May 2013, 08:56 PM
Well the only thing I can do to get this thing to roll properly under autopilot is to turn the GPS/INS so the plug points to the rear of the plane, the pitch works fine either way. I know I saw somewhere that others were having to do the same but now I can't find the posts and if there was a fix to this.

doonunder
19th May 2013, 09:08 PM
Mashton did you try to calibrate aileron 2 times both directions?, ie first time left then right and second time right then left, after each time check balance mode and if gyro corrects the right way. If you get correct gyro reaction but the aileron control is reversed then you can reverse the channel on your tx.

Also for a week I could not work out why my calibrations were not sticking, I was using autopilot channel to complete calibration instead on switch channel.....

doonunder
19th May 2013, 09:10 PM
Also make sure you either not handle the main board white connectors too much or hot glue them as the plastic socket and plugs come apart after a few weeks.

MASHTON1138
19th May 2013, 09:14 PM
I did try that. When I calibrate one way it will actually mix the the aileron/roll channel as elevator/pitch then go the other it does the opposite, then if I do the elevator/pitch one way it may or may not screw the mix up completely by turning the channels backwards again. I dunno what to do next. I think I am gonna shoot a vid and post it so you guys can maybe point out what I am doing wrong.

doonunder
19th May 2013, 09:16 PM
You have to keep trying different combinations as it's not intuitive and I can't explain it but one combination will eventually be the right one.

MASHTON1138
20th May 2013, 10:31 PM
Okay that problem is solved it required swapping the servos outputs and about 23 different combinations of stick movements. I can't even remember now which one worked but I got it after about 2 hours of playing with it tonight.

New issue now that I have the INS working in the proper directions. The servos are calibrated to read 0 in the calibration menu. When I switch over to "Fly" the OSD shows the plane in a roll about 18* from level. I know there are posts about this so now it is time to study those LOL. I think this thing may be better suited for a traditional airplane given how everyone who uses it on a Skywalker or the like seems to have it working straight out of the box and all the issue having peeps are using it on wings.

doonunder
20th May 2013, 10:41 PM
Ok the obvious thing first, is the GPSin adjusted (the physical level of unit) till it reads 0 on pitch and roll on osd whilst plane is in it's level cruising attitude?

MASHTON1138
20th May 2013, 11:19 PM
It was level to the wing, if I moved it until it read 0 I had about 45* of angle on the roll axis. I did this (http://fpvlab.com/forums/showthread.php?11370-Dianmu-OSD-pitch-and-roll-values-incrementing-by-itself.&p=208496&viewfull=1#post208496) and it fixed it. I have about 4 of the USB cables so now one is permanently modded for the GPSINS.

Toysrme
24th May 2013, 01:31 AM
J glad you sorted it out. most reasonable is to go by the wing, but it can vary. Glad u got it worked out tho!


Havent flown the osd on a wing yet, sorry. One thing that can happen to a setup for planes mixing A/E on two servos is requiring one servo to be installed inverted from the other instead of mirroring them. (Instead of both servo arm horns in or horns out, both arms on the left or right).
Ive run into that more than a few times on wings/deltas, canardavons, hell even getting ccmp helis programmed LoL! Aileron is fine but elevator is fudged. Flip/spin the correct servo and you fix it.