New Vstabi tail setup

[Flybar-less]

I was able to improve the tail setup on my 599. I reduced the common gain to 50, P to 80 and increased D to 25. Also lowered cyclic and collective comp, which gave me a more locked in tail.

[rscamp]

I think Pgain and common gain are the same. Should one of those numbers be Igain?

Rob

[Flybar-less]

Common gain is a combination of PID. Just trying to figure out the right combo. I left I at 35.

[rscamp]

Ah. Okay. So what Igain setting did you use?

Rob

[Flybar-less]

35

[rscamp]

I really gotta improve my reading comprehension!

[Ticidytoc]

As of now I have my P set to 90 and common is up to 90 as well. It felt horrible down at 65.

[OICU812]

So common gain you speak of is the total of the Integrator and the Differentiatior?? How do you decide which is to go with what and what to each individually do?

[rscamp]

So common gain you speak of is the total of the Integrator and the Differentiatior?? How do you decide which is to go with what and what to each individually do?

Shawn.

I'm not sure what the difference is between Common and Proportional (P) gain. My best guess is Common gain is a coarse, overall proportional gain setting. I just don't know what else it could be but I'm not betting my life on it! It definitely has nothing to do with Igain or Dgain.

In standard PID control systems, Pgain is basicly the amount of feedback in a control loop. To me, this would be like the sensitivity setting that Common gain is described as in the Wiki. Pgain determines how hard the system tries to correct for error. The error correction is proportional to the amount of error - i.e. the difference between the actual and commanded position or velocity. In this case, the system appears to be a rotational velocity control loop so it is trying to correct the error in yaw rate. If this is correct, a larger Pgain will produce a larger correcting signal when a rotational velocity error is present. Using Pgain results in less rotational velocity error (and thus also less position error). The main caveat is it can only be so high before instability occurs. Instability occurs at lower gain if you have a tail servo that has a slower transit rate, less resolution or lower acceleration. Gear lash, "lost motion", free play in links, low rigidity/non-linearities in the control mechanism, low mechanical gain (link very close to servo pivot reduces effective servo speed) etc. all contribute to lower possible gain settings from the mechanical standpoint.

The Igain reduces error over time. It accumulates the error over time (integrates the error) and produces a correction based on this accumulated error. This makes for more accuracy in the actual commanded rotation rate. The main caveat is the slowness with which it corrects.

The Dgain increases correction with transient changes. It calculates the rate of change of rotation rate so when a change in yaw rate is commanded a transient correction signal is added to the two above to help force the motion to follow the rapid change. The main caveat is instability if it is set too high. I would imagine a high setting here can really make the tail servo act "nervous". :)

Rob

[Flybar-less]

I am running a fast nervous servo, maybe that is why I can run a high D.:) I'll see if I can run 30 or more D, and if my 3D competitor friends like it also.

The Dgain increases correction with transient changes. It calculates the rate of change of rotation rate so when a change in yaw rate is commanded a transient correction signal is added to the two above to help force the motion to follow the rapid change. The main caveat is instability if it is set too high. I would imagine a high setting here can really make the tail servo act "nervous". :)

Rob[/quote]

[Flybar-less]

Rob,
Maybe the Vstabi forum will give a rundown on how the formula works? Then you could put it in plain english for us.

[Ticidytoc]

Try Common - 90 P-90 I-35 and D-0

Let me know what you think.

[Ticidytoc]

Oh, and dont forget to set your endpoint adjustment and expo.

[Flybar-less]

I had my 600 set up that way, but D set to 2. In very FFF the tail would shake and damage the belt drive pully pin hole and other components. Also had slight tail drift during backwards loops. I have settled on I at 35. I am currently replacing some bearings on the 600 and will try a new tail setup on it, as well as a different governor.

[rscamp]

Tuning a PID loop can be a bit of an art. I've never done it on a heli - the situation is very different from the simple X-Y coordinate servo systems I've played with. In a heli there are other things to contend with like the aerodynamic interaction of the flying surfaces and the dynamics that vary with each axis. It makes my head spin! (bad pun)

Methinks I'll be doing a ton of experimentation and progress slowly when I get the VBar in the air. The nerves will be working overtime during the first spin up!

I'd be happy to help with improving the language in the technical literature but it surprisingly challenging and time-consuming. In other words, it is too much like work! ;) It can take quite a while to figure out what the translated text is really trying to say and in some cases, replace it wholesale. After this I'm not sure it is 100% correct.

Maybe we could split up the job. ;)

BTW. I tried to improve the language a bit in the VBar Wiki. If anyone would like to review it and/or improve on it further I'll email the file. The logical final step would be to replace sections in the actual wiki.

Rob

[rscamp]

It definitely has nothing to do with Igain or Dgain.

A correction. Apparently Common gain does have something to do with Igain and Dgain in the sense that it increases ALL types of gain (P, I & D) by an equal amount. Reference - V-Stabi forum.

Rob

[Flybar-less]

I just wanted give you guys an update on tail settings. I got to fly a number of flights back to back with the Logo 599 and the TRex 600. The tail on the Logo with the Vstabi settings I have been using on this heli are very predictable. Basically I have been trying to set up the Vstabi gyro with as little collective and cyclic comp as is possible, and let the Vstabi gyro do more of the work. It seems to help the tail hold in hard 3D transitions.

Collective comp 20
Elevator comp 5
Aileron comp 10
Common gain 50-55
P 80
I 35
D 25-30
Stick Lim 100
Stick gain 115
Deadband 5
Smoother off

[Flybar-less]

Hey Rob, can you explain this english in layman's english? And maybe explain how it may apply to Vstabi?

Derivative

With derivative action, the controller output is proportional to the rate of change of the measurement or error. The controller output is calculated by the rate of change of the measurement with time.
dm CONTROLLER OUTPUT = DERIVATIVE ---- dtWhere m is the measurement at time t.
Some manufacturers use the term rate or pre-act instead of derivative. Derivative, rate, and pre-act are the same thing.
DERIVATIVE = RATE = PRE ACT
Derivative action can compensate for a changing measurement. Thus derivative takes action to inhibit more rapid changes of the measurement than proportional action. When a load or set-point change occurs, the derivative action causes the controller gain to move the "wrong" way when the measurement gets near the set-point. Derivative is often used to avoid overshoot.
Derivative action can stabilize loops since it adds phase lead. Generally, if you use derivative action, more controller gain and reset can be used.

Tuning a PID loop can be a bit of an art. I've never done it on a heli - the situation is very different from the simple X-Y coordinate servo systems I've played with. In a heli there are other things to contend with like the aerodynamic interaction of the flying surfaces and the dynamics that vary with each axis. It makes my head spin! (bad pun)

Methinks I'll be doing a ton of experimentation and progress slowly when I get the VBar in the air. The nerves will be working overtime during the first spin up!

I'd be happy to help with improving the language in the technical literature but it surprisingly challenging and time-consuming. In other words, it is too much like work! ;) It can take quite a while to figure out what the translated text is really trying to say and in some cases, replace it wholesale. After this I'm not sure it is 100% correct.

Maybe we could split up the job. ;)

BTW. I tried to improve the language a bit in the VBar Wiki. If anyone would like to review it and/or improve on it further I'll email the file. The logical final step would be to replace sections in the actual wiki.

Rob

[Ticidytoc]

Basically I have been trying to set up the Vstabi gyro with as little collective and cyclic comp as is possible, and let the Vstabi gyro do more of the work. It seems to help the tail hold in hard 3D transitions.I'm lost here.

Are you saying the collective and cyclic comp doesnt work? Mine is set lower than stock, but not as low as yours, I am also not running the same power combo.

Is this with the stock tail or with your moded gear ratio and tail grips?
I would like to compare notes, so if you dont mind, could you go into more detail as to what "problems" you are having with what particular transition.

What traits does it exhibit?

Also, what are your flying conditions? What altitude do you fly at?


Thanks.

[Flybar-less]

One situation would be a tail slide from 50 feet with a gentle cyclic input from 30 to 15 feet to level flight backwards, the tail moves without being initiated by the stick. Just feels a little disconected compared to what I am used to.

The setup I am currently running works for me now on the 599, which is stock Logo 500/600 except the tail boom and belt length.

The 600 is now built up with a speed up tail gear (9 VS 10) and a Quick UK tail. Recieved a Kontronik 55 to replace the PJ, so maybe this weekend I'll have some results.

[rscamp]

Hey Rob, can you explain this english in layman's english? And maybe explain how it may apply to Vstabi?

Derivative

With derivative action, the controller output is proportional to the rate of change of the measurement or error. The controller output is calculated by the rate of change of the measurement with time.
dm CONTROLLER OUTPUT = DERIVATIVE ---- dtWhere m is the measurement at time t.
Some manufacturers use the term rate or pre-act instead of derivative. Derivative, rate, and pre-act are the same thing.
DERIVATIVE = RATE = PRE ACT
Derivative action can compensate for a changing measurement. Thus derivative takes action to inhibit more rapid changes of the measurement than proportional action. When a load or set-point change occurs, the derivative action causes the controller gain to move the "wrong" way when the measurement gets near the set-point. Derivative is often used to avoid overshoot.
Derivative action can stabilize loops since it adds phase lead. Generally, if you use derivative action, more controller gain and reset can be used.

In an earlier post I described it like this:

The Dgain increases correction with transient changes. It calculates the rate of change of rotation rate so when a change in yaw rate is commanded a transient correction signal is added to the two above to help force the motion to follow the rapid change.

The description above is over-simplified but more applicable to the heli rudder application. In the V-Stabi application, I believe the control system is basically a rotation rate feedback system so I try to put everything in these terms.

The key thing with derivative correction is it is larger when the rate of change is greater. So if you make a quick yaw rate change (move the rudder stick quickly) it results in more signal going to the rudder servo to make it move. This makes the rudder quick to respond. It does a similar thing when you stop the motion of the stick, only this time it results in more signal to make it stop changing the yaw rate. A very slow acceleration of the rudder stick would result in only a very small signal going to the rudder servo from the derivative output which would be swamped relatively speaking by the regular proportional gain correction signal.

It is sometimes called "pre-act" or whatever because it is a leading or advanced signal. If the input signal was a sine wave (i.e. you were wiggling the rudder stick smoothly back and forth) the derivative correction signal would be advanced 90 degrees. In a sense, this derivative signal "anticipates" the correction that is required.

In the most basic terms I can think of, proportional correction is like a linear spring forcing the tail servo and thus the heli to follow the commanded heli yaw rotation rate. There will always be some lag or error in a system with only proportional gain that can be reduced by other means. Integral correction gradually forces the heli yaw rotation rate to agree more accurately with the commanded heli yaw rotation rate. This is basicly finding the area under the curve describing the error (hence the term integral) and applying it to correct error. Derivative correction forces the heli to respond more quickly to changes in the commanded heli yaw rotation rate for the reason described above.

I think I am describing this correctly, but if anyone is more knowledgeable on PID control theory please chime in!

Rob

[OICU812]

We need to translate these one step further to dummy it up for foolz like me. :o

[rscamp]

One situation would be a tail slide from 50 feet with a gentle cyclic input from 30 to 15 feet to level flight backwards, the tail moves without being initiated by the stick. Just feels a little disconected compared to what I am used to.

The setup I am currently running works for me now on the 599, which is stock Logo 500/600 except the tail boom and belt length.

The 600 is now built up with a speed up tail gear (9 VS 10) and a Quick UK tail. Recieved a Kontronik 55 to replace the PJ, so maybe this weekend I'll have some results.

Here's a couple of thoughts...

I think in very general terms, the amount of mixing needed from collective and cyclic to rudder is less when the tail is more tightly controlled by the PID controller. So if you have a very good tail control system you need less mixing. This could be part of what you are experiencing.

However...

If this problem does not go away when you turn the mixing off then I think it is more likely this behaviour is caused by some crosstalk between two axes of your gyro. The most obvious way this would occur is if the gyro was not perfectly orthogonal to the heli mechanics when it is sitting in the frame.

Rob

[rscamp]

We need to translate these one step further to dummy it up for foolz like me. :o

Well, my thumbs are dumb as all get out. Is there anything you can do to smarten up my thumbs? :)

From a practical standpoint, it is more important to know the effect of changing a value than to know what it means. Sometimes I think I understand the theory but it still doesn't do what I expect in practice. :(

Rob

[OICU812]

I guess what it is, is simply that I would be looking for example:

If tail rebounds on left piro then adjust this value down and vice versa for right piro, if tail wags in FFF start by adjusting this value. All the other words in their sentences are straight out jibberish, all I need is the meat and the patatoes man, keep it simple I say! :D Thanks for all your efforts however on the Vbar for dummies like me translations! LOL