This should be interesting

Sorry Chris but I disagree with this statement. From what I’ve studied on helicopter rotor dynamics, regardless of whether it is a flybar or a stab bar, I don’t think it serves the same function as the I term. I think it may appear to be the same because it resists rates so well. I find the use of adding pitch to the paddles very interesting. I will have to noodle through the physics on this. You essentially are giving the flybar a thrust vector. I think of the flybar as a mini rotor and it behaves similar to the main rotor. From what I’ve studied, there is no evidence that the rotor or flybar acts like an I term and opposes attitude error. I’m not saying that it won’t try to stay at a given attitude but the underlying physics is that it is resisting rate and trying to null the rates thereby resisting attitude changes. This is different from the I term where attitude error is stored and the I gain operates on that error to zero it and thus return the aircraft back trimmed attitude.

I’ll be interested to see what you find out with the DFC head.

Chris, Bill,
I have been watching your conversation, it is definatly full of information and although i have nothing to add, i do have a question.
With VFF, this as i understand it is a form of feed forward and i noticed one of you say that the guys with the downstream fbl’s were turning it all the way up. If you indeed had VFF at 100% does that mean stick inputs are passed through unadulterated? Or is the signal still going through a peice of the loop?
Im just trying to gain some insight on the parameter and how it affects the system. .
Also, with IlMI, i have it set to 0 at the moment. I saw mention of this parameter and how it can affect flight. Does this parameter affect the feedback loop when in a hover, when in flight or just on the ground.
Thanks,
Tim

Bill, look at it from a mechanical engineer’s standpoint instead of software and electronic gyros. The pilot makes an input that translates to swashplate angle. This is now the desired attitude. The main rotor angle is the actual attitude. The difference between the two is the error. The flybar takes care of the rest. It damps the input while automatically stabilizing the main rotor and keeping everything really smooth. It does the long-term correction of desired to actual. And it keeps correcting for it if it changes or an error builds between the two.

Tim, I’m reasonably sure the VFF bypasses the rate PID loop. It’s the way I fly my flybars and the way people are using the downstream FBL controllers. It just passes the input to the swashplate direct, skipping the PID’s, and lets the flybar or FBL unit do what it does. When tuning your helicopter you can put more weight on the PID loop, or more on the VFF. You need to arrive at a balance that will keep the helicopter stable in high speed flight or you’ll get what I got with my DFC heli where it went into wild pitch oscillation when I started pushing it, when I thought I had it tuned to perfection in hover and low speed. I had the P-gain too high, not enough I-gain, and not enough VFF. My final tuning with that helicopter was done in-flight trying to get rid of the high speed (almost) loss of control issues, flying it at maximum power and high speed.

It takes way more I-gain than you might expect for high speed with FBL. And a healthy dose of VFF which provides the control response needed, without the oscillation problems. That’s why I say, tune it then go fly it quite a bit in Stabilize and put it thru it’s paces before you think you really “got it”.

In stabilize mode, the VFF gain is applied to the rate command signal coming from the attitude controller. It is not your stick input.

Thanks for the replies, i think im starting to get abclearer picture of how VFF affects flight.
I think ill likely do a couple more flights and get the P gain closer to .2, or wherever it ossicilates and then add in some I. Ill have to look at the amount of VFF i have in the parameter list, i believe it is set to what was recommended here. So i should bring that down as the P gain goes up if i understand correctly.
I’m still shocked at how well it flew with only a low P gain and no I the first trip out. I truly expected much worse behavior after reading people accounts of their initial flight. Although people seem to use smaller 450 and 500 sizes birds so maybe that has a hand in it?
Is there a specific threshold I should be looking for with vibrations? What i do is go into mission planner and select the “vibrations” tab in the drop down menu. I just dont know what bad vibes would look like? Im assuming there is a numeric value that constitutes an upper limit to look for when upping the headspeed and tuning further?
Thanks,
Tim

Chris, I think I understand what you’re saying. What you have to realize is that the tip path plane angle is not controlling the attitude of the aircraft. Changes in the swashplate angle which causes changes in the rotor tip path plane angle create aircraft rates. So even though you look at it as an error in angle between the swashplate and the rotor tip path plane angle, this does not translate to aircraft attitude. For bare helicopter dynamics the swashplate controls aircraft angular rate. Therefore when you’re controlling the error between the swashplate and the tip path plane you’re actually controlling aircraft rates. So in reality the P gain of the rate controller is what does this which is exactly what the fly bar is doing.

Not sure if the links are working correctly to the flight videos? Ill have to log on to my computer and see there, android does some weird things occasionally.
Tim

Well, the links work after opening them in a new window. I wonder if uploading to youtube then linking is better than using google drive?
Tim

https://drive.google.com/file/d/0BxUn7rkPzp9sRG1CeGpjX2tFUkk/view?usp=sharing

The photo is one of my earlier creations for a video platform. That was a 700X by Blade that the overtly large gimbal is attached to. Lets just say it was a good proof of concept. Later on I modified the mount and it hangs under a Trex 800 which handles the load far better. I wanted the ability of full 360 degree gimbal rotation vs the narrow field you get when using a front mount gimbal. I have a Tarot T1000 Octacopter that also has a gimbal like that one but as I prefer helicopters I was determined to make it work.
Now i’m hoping the Pixhawk will handle all that weight underslung from the center without toilet-bowling slowly in a hover.
Tim

Tim I can’t see the links in the first two posts. I can get to the google drive link ok.

While you lost me with that explanation, I do agree that the flybar does rate P.

Tim, just be careful with that. You might be able to hover it. But I doubt it’s actually flyable, at least not very well. Take it out and slam it around a little bit and you’re likely to lose control of it with the rates set close to the defaults.

The basic tuning procedure is to turn off the rate I and D and turn up the P until it gets unstable, then back it off to where it seems stable. Then adjust the VFF to get decent control response so you can actually fly it. Then turn up the rate I up to about .15 or so to start. Most people don’t run any rate D with helicopter with FBL.

Then I would fly it awhile and slam it around a little bit. Slam it into the ground hard and lift off again and see if it upsets it. Fly it around trees in gusty wind. Hover it sideways and scrape a skid and see if it goes into roll shakes. Take it out and make a couple high speed passes with it to see if stays stable. That gives you time to tweak it and get used to it, look at the logs and see if there’s any problems, etc…

Then try Alt Hold once and see if it holds altitude ok. If it does, try Loiter, and turn up the I gain until it seems to hold position pretty well in the wind. Just be aware that in Loiter it is sluggish. Loiter is not designed to be flown at very high speeds or any sharp moves. It is more of a stationary hovering mode. There’s some adjustments you can make to Loiter to make it “sharper” on handling though.

I don’t know if Position Hold works in helicopter or not in the latest code. I crashed one 600 with Position Hold before and I consider it a “bad” mode for helicopters. It tries to use the rates in Alt Hold when you move the sticks, and change back to Loiter when you let go of the sticks. I think it was invented to try to compete with DJI with the way their “GPS-Safe To Fly” stuff works. It might be the best thing since sliced bread for multi’s, but it doesn’t actually let you fly a helicopter - it’s a wonked up mode if you ask me. I don’t recommend using it, but you can if you want, at your own risk.

As far as vibration, it’s just like any FBL unit. The less you have, the better it will work. I guess the “guideline” is about +/-3 on X and Y and below 15 for Z for really good performance in Loiter. Mine are all higher than that and yet they seem to fly pretty well. I think the guideline on Z vibes is that it has to be below about 25-30 or it will have problems holding altitude.

Uggh, they are all google drive links but for some reason its being a pain with the videos.

"There is supposed to be some content here, but the publisher doesn’t allow it to be displayed in a frame. This is to help protect the security of any information you might enter into this site.
Try this
Open this in a new window"
That’s the message I see in the post window on my end, if I click on the blue words “Open this in a new window” It will open another browser window to the video.
I have a feeling YouTube is going to be the way to go with videos in the future. :confused:
Tim

Chris,
As it stands now I am interested in “Loiter” and “Stabilize” mode pretty much exclusively. Loiter for holding position while doing photography work and stabilize for the videography shots. Maybe in the future after I have a more complete understanding of the system I will dabble with “Auto” missions. Doesn’t loiter also hold position, but allows you to adjust that position actively with the sticks if needed? If it does than “position hold” sounds like it wouldn’t be of use/redundant anyways all problems with the mode aside.

Also, when I say “I’m still shocked at how well it flew with only a low P gain and no I the first trip out. I truly expected much worse behavior after reading people accounts of their initial flight”, I am using the term “Fly” very loosely. I guess I expected a wobbling nightmare that was all but impossible to hover? I was pleasantly surprised and felt a little silly about going through all the measures I did for the maiden, although I probably would do the same again for safety’s and my wallets sake.
I am for sure going to put it through its paces and knock the skids and check higher speed flight etc coming up, if the darn rain will ever stop. I’m sure I will have to re-tune some things a lot once it gets a gimbal attached as well, but for now I am focusing on getting the bare airframe flightworthy while I get a feel for how Pixhawk does things.
Tim

Yes, it does. But it’s quite sluggish usually. It’s definitely not a flight mode you’d want to try to fly the heli with to do any real flying. Alt Hold works awesome with a heli. Can fly around with it and not have to worry about collective management, and you get normal pitch, roll and yaw response. If you want normal collective response in addition, then use Stabilize.

Chris,
Did you know that you can adjust the maximum speed and several other characteristics for the Loiter mode. I was looking through the parameters and found these. I put the description from the code after the parameter.
WPNAV_LOIT_JERK,1000 - maximum jerk which in cm/s/s/s
WPNAV_LOIT_MAXA,250 - maximum acceleration in cm/s/s. Higher values cause the copter to accelerate and stop more quickly.
WPNAV_LOIT_MINA,25 - minimum accleration in cm/s/s. Higher values stop the copter more quickly when the stick is centered, but cause a larger jerk when the copter stops.
WPNAV_LOIT_SPEED,500 - maximum speed in cm/s

I think the one that is most helpful would be WPNAV_LOIT_SPEED which allows you to set it for higher loiter speeds. Again this is a position hold mode and your stick commands velocity. So with this parameter set to larger values, your stick can command higher speeds. Just thought I would share since it seems that you do not like the mode. Not sure if you played with these paraemters.

I’ve played with the Loiter params on many different aircraft, including multi’s, and it’s just not a flying mode. It’s exactly what it’s named - Loiter. It’s for stationary and very limited maneuvering type flight. It’s designed to be very slow and smooth for stationary or hovering camera work.

It’s not that I don’t like the mode - I do. I use it for various things, but inside the limitations it was designed for. I set the max speed no higher than 3 m/s and have used it for doing aerial lifts with a 20 foot tether. It takes all the work out of flying the helicopter so you can concentrate on your load on the tether and gently maneuver the helicopter to place your load and drop the tether.
I lifted fairly heavy parts to the top of a 190 foot grain elevator leg with it before. Placed the toolbox and parts on the work platform on the leg in 20mph wind, accurate to within a foot, guided by hand signals from the crew working on the leg on the platform 190 feet up.

While it’s not a good flying mode, its holding power in the wind is amazing for stationary work. If you turn up those params you suggested, Bill, it becomes way too jerky, rendering the mode useless for what it was designed for.

Well, after hearing an in-depth description of loiter, it certainly sounds like it suits one of my needs exactly. Ergo aerial photography work and possibly cable style shots with video.
Eventually, with a different camera setup, i would like to get into site mapping using the auto mode, but i feel that is a ways off.
Tim