Fun, Learning, Friendship and Mutual Respect START  HERE


Unregistered
Go Back   HeliFreak > R/C Helicopters > Main Forum - Helicopter Talk


Main Forum - Helicopter Talk R/C Helicopters and the people who fly them. VENDOR TOPICS DO NOT GO HERE. Full Scale Heli threads go in OT please


Reply
 
LinkBack Thread Tools Display Modes
Old 06-13-2017, 12:17 PM   #1 (permalink)
Registered Users
 
Posts: 801
 

Join Date: Feb 2006
Location: UK
Default Collective = velocity?

The 3 other stick inputs (cyclic and rudder) are managed by the FBL. You move the stick and the output is controlled, automatically adjusting for wind, RPM, etc. to give a consistent response.

Collective is the only output that is not managed by the FBL. It gives a proportional amount of pitch for the stick movement, and the result is dependant on RPM, model weight, drag, etc.

What if the collective was managed and the FBL adjusted the pitch automatically to give a velocity (in line with mainshaft) in proportion to stick deflection? Is it possible with today's FBLs?
__________________
XL Power - EGODRIFT - ManiaX - Hely-shop
facebook.com/3drchelipilot - VBar Utilities
richardv is online now        Reply With Quote Quick reply to this message
Sponsored Links
Advertisement
 
Old 06-13-2017, 12:43 PM   #2 (permalink)
Registered Users
 

Join Date: Oct 2012
Default

In order to do something like that,.. there would have to be a sensor of some sort to sense the velocity. I think that the other axis have a gyroscope type sensor,.. ie: senses rotation on an axis,.. however, since velocity of heli is not around an axis,.. the same type of sensor would not work. Not sure what type of sensor could be used, but this is my guess as to why, no.
vultage is offline        Reply With Quote Quick reply to this message
Old 06-13-2017, 12:55 PM   #3 (permalink)
Registered Users
 
Posts: 1,891
 

Join Date: Jun 2010
Default

Quote:
What if the collective was managed and the FBL adjusted the pitch automatically to give a velocity (in line with mainshaft) in proportion to stick deflection?
There are 3D maneuvers (funnels, hurricanes and many many others) where the collective accelerates in the direction flight that isn't in line with the mainshaft.
__________________

Fauropitotto is offline        Reply With Quote Quick reply to this message
Old 06-13-2017, 01:34 PM   #4 (permalink)
HF Support
 
Posts: 41,408
 
My HF Map location
Join Date: Jul 2007
Default

An accelerometer, sampled over time, would give you an average velocity for an accelerating or decelerating heli, but it couldn't tell whether you were stopped or traveling a constant velocity.

GPS and radar/lidar are about the only other options for velocity, but they couldn't tell you whether the heli was traveling parallel to the vertical axis or at some other angle, only the velocities in earth-relative axes.
__________________
Alan (just call me "Bum"; no need to be so formal)
HeliFreak Support
Protos Max Evo 700 (Brain 2), Protos Max V2 770 (iKON 2), Synergy 516 (Brain 2), Oxy 2 (Falcon 12)
Graupner MZ-32

Experience is something you gain only after you need it.
I wouldn't mind social media so much if it didn't involve people.
TheBum is offline        Reply With Quote Quick reply to this message
Old 06-13-2017, 01:34 PM   #5 (permalink)
Registered Users
 
Posts: 1,277
 

Join Date: Mar 2016
Location: Essex, UK
Default

I don't think it would be easy to fly. How would you manage transitional lift in windy conditions? Stationary flips? Tic-tocks? All of which need maximum control of the collective regardless of velocity.
__________________
Mark

So many things to fly, so little time.
Crashalotjoe is offline        Reply With Quote Quick reply to this message
Old 06-13-2017, 01:39 PM   #6 (permalink)
Registered Users
 

Join Date: Nov 2013
Location: Aberdeen, UK
Default

I've thought something along those lines myself. What modern FBL with it's 6 axis sensor can do is detect linear acceleration, in addition to angular acceleration like standard gyros do.

This 'should' in theory i think make it possible for acceleration in the main shaft (vertical if you like) direction. This in turn should allow at least some rate damping so that the any un-commanded vertical acceleration are damped out. The obvious example if if hovering and a gust of wind hits the heli and it zooms upward, this should be able to be much reduced if not totally eliminated.
__________________
Gob Kraken |Gob 700 BT 'T' | Gob 770 Comp | SoXos Strike 7 | Henseleit TDR, TDSF & TDR2 | Kasama Dune | OMP M1&M2 | UK stockist for OMPHobby M1/M2 ScotiaRC.co.uk
Smoggie is online now        Reply With Quote Quick reply to this message
Old 06-13-2017, 01:44 PM   #7 (permalink)
Registered Users
 
My HF Map location
Join Date: Jul 2012
Default

I don't really see it being of much use...

There's already systems that do controls from the ground reference frame, and altitude hold assist but that's for camera work not 3D helis.

I guess with a linear accelerometer and some fun with vectors you could subtract out the gravity component and get your absolute rate of change on the main shaft axis. You' could take the rotational rate of change data and filter out acceleration due to the FBL not being at the center of rotation.

The behavior of such a system would be really weird. I might not be thinking it through right but with the disk tilted and collective at mid stick it would freak out.
__________________
Synergy E5s 12S!
Minicopter Diabolo 550
Mikado 480 eXXtreme
GauiX3 350mm 6S
TucsonFlyer is offline        Reply With Quote Quick reply to this message
Old 06-13-2017, 01:46 PM   #8 (permalink)
Registered Users
 

Join Date: Apr 2014
Default

Yeah even in FFF you are moving in a direction not inline with the mainshaft. Basically anything besides level upright or level inverted. I think in any non-level orientation the PID loop would quickly reach min or max pitch trying to achieve the requested velocity inline with the mainshaft.

EDIT: I agree TucsonFlyer, might be able to compensate and make it work with additional complicated programming. But then we have to ask what is the point? What would we gain from this? Seems that sensor controlled collective is a solution to a problem that doesn't exist. Or more accurately a new set of problems in addition to a problem that doesn't exist! Being able to control pitch directly while in any orientation is what allows us to fly our helis the way we do.
__________________
Compass eXo - Rage 4219-1000kv - Scorpion Tribunus 120A - MKS & Xpert servos - Brain
Compass Chronos (3-Blade) - Scorpion 5020-450kv - Scorpion Tribunus 200A - BK servos - Spirit Pro

Last edited by simplechamp; 06-13-2017 at 01:56 PM..
simplechamp is offline        Reply With Quote Quick reply to this message
Old 06-13-2017, 01:53 PM   #9 (permalink)
Registered Users
 

Join Date: Nov 2013
Location: Aberdeen, UK
Default

Actually if you were simply detecting acceleration (NOT velocity) with a quite low rate mode gain i dont think it would change the handling of the heli much.
It would just dampen out un-commanded acceleration in the direction of the main shaft axis and so make the heli smoother flying. This would work equally well in any orientation and at and speed and should be especially useful on small helis.
__________________
Gob Kraken |Gob 700 BT 'T' | Gob 770 Comp | SoXos Strike 7 | Henseleit TDR, TDSF & TDR2 | Kasama Dune | OMP M1&M2 | UK stockist for OMPHobby M1/M2 ScotiaRC.co.uk
Smoggie is online now        Reply With Quote Quick reply to this message
Old 06-13-2017, 03:10 PM   #10 (permalink)
Registered Users
 
Posts: 801
Thread Starter Thread Starter
 

Join Date: Feb 2006
Location: UK
Default

The problem we're trying to solve is making the collective response consistent, regardless of external factors. Imagine heading hold on the pitch. Translational lift, bounce after a descent, thermals, etc will all countered by the system so you don't have to.
__________________
XL Power - EGODRIFT - ManiaX - Hely-shop
facebook.com/3drchelipilot - VBar Utilities
richardv is online now        Reply With Quote Quick reply to this message
Old 06-13-2017, 03:33 PM   #11 (permalink)
Registered Users
 

Join Date: Apr 2014
Default

That's the way I see it. It's one thing to still control collective directly, and then have some modifying or damped output based on another measured variable. But going to full closed loop PID control where you adjust setpoint only and the output is determined by the controller, that is going to present much more difficulty.

Doing a level climb I might take X amount of pitch. Doing a funnel may also take X amount of pitch. But the acceleration on mainshaft axis is going to be way different. So how does the PID controller know what to do in those situations? It's going to keep feeding in pitch to get the accel you are requesting, and is going to have unwanted results.
__________________
Compass eXo - Rage 4219-1000kv - Scorpion Tribunus 120A - MKS & Xpert servos - Brain
Compass Chronos (3-Blade) - Scorpion 5020-450kv - Scorpion Tribunus 200A - BK servos - Spirit Pro
simplechamp is offline        Reply With Quote Quick reply to this message
Old 06-13-2017, 04:20 PM   #12 (permalink)
Registered Users
 
Posts: 2,355
 

Join Date: Jun 2016
Default Collective = velocity?

Surely inertia/momentum would come into play? Gravity, centripetal force, gyroscopic precession - too much of a headache if you ask me lol
AndyO is offline        Reply With Quote Quick reply to this message
Old 06-13-2017, 06:25 PM   #13 (permalink)
Registered Users
 

Join Date: Aug 2010
Default

Guys

I don’t know what the goal here is. But it isn’t plausible to use this for most applications I can imagine, for a couple of reasons.

First of them is- very small accelerations are involved with relatively large errors. For example, in hover, just a bobble that would causes your machine to accelerate vertically to 0.5mph in 1 second is just 0.03 g. That happens in a space of about 5” travel. It is very difficult to read when you have a bunch of noise. Filters can only do so much, and many of these $5 sensors have problems with thermal drift and cross-axis sensitivity.

Second of them is that these machines cannot maintain higher accelerations in a pure z-axis sense. Yea, they are OK with funnels etc, but as you increase airspeed vertically, the inflow to the rotor changes its AOA. And you rapidly run out of it. Think about just a hover to a 20deg stick up collective hold. The machine bolts for a few feet, but then stabilizes at some fixed vertical rate- because AOA has dropped to the point that is just just offsetting gravity and drag.

Many of the FBLs out there are already running a full 6-axis (or 9-axis) IMU, which is to say- they are aware of their rates and accelerations in 3D space. That is needed for bailout and for some other complex things (inferring acceleration in a nose down descent vs nose up ascent, etc). So including a ‘heading hold’ mode on collective isn’t a problem in that sense- it is just that it isn’t practical at very low g (to help with bobbles etc) and it cannot work in higher g in conventional flight due to inflow speed vs blades and blade geometry (no blade twist, etc).
__________________
"The problem with quotes found on the internet is you have no way of confirming their authenticity." - Abraham Lincoln
extrapilot is offline        Reply With Quote Quick reply to this message
Old 06-13-2017, 06:43 PM   #14 (permalink)
Registered Users
 
My HF Map location
Join Date: Jul 2012
Default

It seems like two separate classes of problems HH/FBL cyclic and what's being proposed.

Your cyclic isn't in any way directly controlled by the stick. Your stick input requests a roll rate in deg/sec(or whatever unit) and then the FBL unit makes that happen. If I remember right that happens by integrating the change in roll rate to get the actual roll rate so we can use that differential between actual roll rate and requested roll rate to tell the servos how to move to keep the roll constant.

What the OP is talking about would require endpoints i.e. how fast your heli can actually climb. Now you'd have to be able to correctly track the gravity vector so you can account for it based on the orientation of the heli since we're dealing with a linear acceleration rather than a point rotation rate of change. At that point I really lose the plot.

As far as I can tell collective pitch has a non-continuous relationship to acceleration on the main shaft axis, even when accounting for the real orientation of the helicopter. I suppose it would be doable on a full scale heli with more instrumentation and lots of testing and tuning but the FBL is basically like being locked in a black box and thrown in the air. You can tell when you're accelerating but not your velocity/relative air speed and without that you can't deal with all the extra variables to collective performance that depend on it.
__________________
Synergy E5s 12S!
Minicopter Diabolo 550
Mikado 480 eXXtreme
GauiX3 350mm 6S
TucsonFlyer is offline        Reply With Quote Quick reply to this message
Old 06-13-2017, 07:42 PM   #15 (permalink)
Registered Users
 

Join Date: Jan 2012
Default

I had the same thought a while back and started a thread about it. Sounds really cool but just too messy to deal with. Also forward flight, or any direction other than along vertical axis gets really funky to deal with.
__________________
Q700
RCHN #232
quinn is offline        Reply With Quote Quick reply to this message
Old 06-13-2017, 08:47 PM   #16 (permalink)
Registered Users
 
Posts: 1,891
 

Join Date: Jun 2010
Default

Even if you set the system to counter non-vertical accelerations, can you imagine how it would handle something like a tail-slide or any other maneuver that depends on gravity to do the work independent of the collective?
__________________

Fauropitotto is offline        Reply With Quote Quick reply to this message
Old 06-13-2017, 08:49 PM   #17 (permalink)
Registered Users
 

Join Date: Apr 2014
Default

Quote:
Originally Posted by TucsonFlyer View Post
As far as I can tell collective pitch has a non-continuous relationship to acceleration on the main shaft axis, even when accounting for the real orientation of the helicopter.
That's one of the main problems. So you give a stick input for collective that now says "Accelerate at +1m/s^2". The PID controller increases collective pitch and initially is able to achieve the setpoint for +1m/s^2 acceleration. But the heli can't accelerate indefinitely. As extrapilot says the acceleration will quickly slow until the heli is at a constant velocity. At the same time the PID controller sees the actual acceleration is less than the setpoint and feeds in more collective to compensate. It may keep up for a little longer but soon you are at max collective with no more ability to accelerate. And there is nothing more the PID controller can do.
__________________
Compass eXo - Rage 4219-1000kv - Scorpion Tribunus 120A - MKS & Xpert servos - Brain
Compass Chronos (3-Blade) - Scorpion 5020-450kv - Scorpion Tribunus 200A - BK servos - Spirit Pro
simplechamp is offline        Reply With Quote Quick reply to this message
Old 06-14-2017, 04:59 AM   #18 (permalink)
Registered Users
 
Posts: 2,597
 

Join Date: Apr 2013
Default

Putting it at mid-stick would immediately counter any velocity and put it into a hover.
__________________
700E Pro DFC, 550SX v2, 450 Pro DFC, mCPX BL, DX8
Crowbar is offline        Reply With Quote Quick reply to this message
Old 06-14-2017, 05:17 AM   #19 (permalink)
Registered Users
 
Posts: 801
Thread Starter Thread Starter
 

Join Date: Feb 2006
Location: UK
Default

Quote:
Originally Posted by Crowbar View Post
Putting it at mid-stick would immediately counter any velocity and put it into a hover.
Combine this with a sprung collective and it might be the solution.
__________________
XL Power - EGODRIFT - ManiaX - Hely-shop
facebook.com/3drchelipilot - VBar Utilities
richardv is online now        Reply With Quote Quick reply to this message
Old 06-14-2017, 07:41 AM   #20 (permalink)
HF Support
 
Posts: 6,850
 
My HF Map location
Join Date: Jan 2013
Default

Its been posted before on this thread, but collective does not equal velocity.

Aileron modifies the roll of the model irrespective of it's position in space.
Elevator modifies the pitch of the model irrespective of it's position in space.
Rudder modifies the yaw of the model irrespective of it's position in space.

All the above can be handled by 3 axis gyroscopes on the model. Forming this into a picture of the models current rotation relative to a horizon is very difficult, but not impossible (this is what rescue FBL units do).

Gravity exerts a downward force on the model irrespective of the models pitch, roll and yaw.
Collective exerts an up or downward force roughly orthogonal (at right angles) to the models current roll and pitch. I say roughly as there can be a rotational component to this force based on the pitch and roll being commanded at the time.

Accelerometers on a model only track acceleration in relative axes to the model. Forming this into a picture of the models velocity in all directions is remarkably difficult.

There are also things like translational lift and torque, but I'll ignore those for the moment.

If you want a FBL unit to automatically hover a model it must:
  1. Orient the main shaft orthogonal the to the rotation direction that requires velocity alteration.
  2. Apply appropriate collective to provide thrust counter to accelerate the model against the velocity vector it is trying to negate.
  3. Repeat steps 1 and 2 until the accumulated accelerometer data registers the model as stable.

The issues with all this is:
  • The time it takes to orient the main shaft (through aileron and elevator) to make it orthogonal to the required direction to which acceleration needs to be applied.
  • The feedback on the collective required to counter the velocity.

And this is if you want it to stop anywhere (neutral velocity). You would then extrapolate this to have your collective stick be the forward/backward velocity vector.

If you want full 6 independent axis command (rotational - (pitch/roll/yaw) : translational - (forward/back)/(left/right)/(up/down) ) you would need a 6 axis controller, not our existing 4.

Imagine if each stick had a rotational component as well (right controlling rotational (twist for yaw), left controlling translational (twist for up/down)). This would take quite a while to master. Also the onboard gymnastics to facilitate this would be incredible.

There is a FPV game that uses 6 axis control for a space ship. It's called Descent. It takes a while to get familiar with the controls (but awesome once you do). You can do maneuvers in Descent that helis can only dream of. I cannot imagine what it would be like to try control a craft THAT maneuverable without being inside it.
__________________
TRex 700E Pro DFC HV (BD3SX) - Gaui X5 (6S/BD3SX), Blade 130X, RealFlight, Heli-X, Taranis+DSMX
ArchmageAU is offline        Reply With Quote Quick reply to this message
Reply




Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the HeliFreak forums, you must first register.
Please enter your desired user name, your REAL and WORKING email address and other required details in the form below.
User Name:
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself. Use a real email address or you will not be granted access to the site. Thank you.
Email Address:
Location
Where do you live? ie: Country, State, City or General Geographic Location please.
Name and Lastname
Enter name and last name here. (This information is not shown to the general public. Optional)
Helicopter #1
Enter Helicopter #1 type and equipment.
Helicopter #2
Enter Helicopter #2 type and equipment.
Helicopter #3
Enter Helicopter #3 type and equipment.
Helicopter #4
Enter Helicopter #4 type and equipment.

Log-in


Thread Tools
Display Modes

Posting Rules
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On




Copyright © Website Acquisitions Inc. All rights reserved.
vBulletin Security provided by vBSecurity v2.2.2 (Pro) - vBulletin Mods & Addons Copyright © 2021 DragonByte Technologies Ltd.

SEO by vBSEO 3.6.1