Advanced parameters glossary

[Otatiaro]

Originally Posted by djamgils

So you mean the proportional gain is actually a common gain and thus controller_output=Kp(e+Ki(e)+Kd(e))
or you mean that lowering Kp will lower the entire frequency response? because if controller_output=Kp*e+Ki(e)+Kd(e) then I dont see why "decrease P gain also mean decrease the other gains" Or you mean that as a result of lowering the P gain a lowering of the other gains is also needed and thus results in poor tail performance?

About the smith predictor. Arent those only used for systems with large dead times? Or is the dead time with respect to the time constant relatively large?
What was the cause for the tailbounce without the update?

Hello,

And the good answer is that if you reduce the P gain you have to lower the other gains too (for example a low P gain with high I gain will result in high amplitude, low frequency oscillations, et c'est pas beau à voir ).

I'm not using smith predictor (as least not as it is explained in the books ). We found an optimization to use when the time constant of the system is predictable, but I can't say too much of it, I hope you understand why.

Thomas.

[Diviner]

Corrado, Fabien, Thomas
I have read up on PID tuning. I have modified the tuning worksheet based on your glossary.

Could you take a look at the attached tail tuning guide and make some comments.

Specifically:

1) Tail asymmetry
2) Tail dynamic
3) Tail Expo.
4) Agility
5) Feed forward

Any thoughts would be greatly appreciated.

Keep up the good posts.

[Otatiaro]

Hello,

Just to add some more data on the stack

The wikipedia article on PID is quiet nice, but PLEASE don't start asking for ziegler nickols methods or things like that, it does NOT apply.

The main thing to understand is that we are not looking for the best response (is there any, anyway ...), but the right feeling.

The PID very basics :

- P(roportional) is the "main gain", it reacts to the error (the error being the difference between the desired speed - stick position - and the real speed sensed by the gyroscope). Too much P gain and the system will oscillate quiet fast, not enough P gain and the overall reaction is sluggish.

- I(ntegral) is the "lock gain", by integrating the error, it allows to not only fight perturbations, but also to come back where it was (without I gain in the wind, you have drift). Too much I gain and you'll get bounce back (caused by integral windup, go check on wikipedia), not enough I gain and you'l get drift or inconsistant rotation speed (piro consistancy).

- D(erivative) gain is the "smoothing gain", it basically "ho shit I'm arriving too fast on the target, time to slow down a bit". Too much D gain and the system will brake too early, not enough D gain and you'll get bounce and hard stops.

- F(eed forward) gain is the "pass through gain", the unit does pass through a ratio of the input - stick position - directly to the servos without even taking care of the gyroscope (by the way you can use the BRAIN on a flybar heli using only feed forward gain on cyclic ). Too much F gain and the heli will start very hard and the unit will have to counter the F gain to stabilize speed, not enough F gain and ... well, you'r safe, F gain is mainly about feeling.

- Agility : don't ask, it's one of the secrets of the BRAIN ... Low agility will feel like robotic, more like the beastX, high agility will feel more like a flybar helicopter, with its benefits and losses.

- Stop gain : I guess we will just throw this one away in a futre update

- Tail dynamics : think P3 on BeastX.

- Pitch and Cyclic precompensation : this is why a one axis gyroscope will never be as good as a FBL unit on the tail ... the FBL unit know the pitch and cyclic position at any time, and more pitch/cyclic means more torque to counter, hence the unit can counter it BEFORE the sensor even sensed it (before there is an actual error).


The real problem when tuning a PID (moreover a PID with some ... optimizations), is that each gain is linked to the other, you cannot tune one gain then the others, one by one, if you change one gain, it will have impact on every other gains optimum values.

Start with a default preset, then change (a few % at a time) ONLY one parameter, to feel the difference :

- P will make the heli feel more rigid
- I will make the heli feel more locked
- D will make the heli feel more smooth
- F will make the heli feel more aggressive
- Agility will make the heli feel more natural (more robotic is you decrease it)
- Tail dynamics will help to prevent gear stripping (lowering it)
- Asymmetry will help get same speed and reaction on both tail directions
- Pitch/Cyclic precomp will make the tail not deviate when doing hard climbouts and fast rolls/flips.

Hope this helps.

Thomas.

[GRIMX]

Nice, this is exactly what I am looking for, thanks Thomas.

[SMOKEMASTER60]

Yes, thank you.

[almigurt]

Nice!


Tyvm. May i copy your post and paste it in german forums?




But whats a p3? i dont own a beast.


Benjamin

[misterbrain]

p3 is the thrid of three potentiometers on the beast.
p1-swashgain
p2-swash direct ammount
p3-taildynamics

greets,
markus

[HighFreq]

Probably agility is an important parameter and rather easy to use and we should put it in the wizard in the setup panel where people choose flying characteristics (beginner, sport, 3d, hard 3d)

Remember:

Lower agility----->locked feeling
Higher Agility---->more natural flybarred feeling


Corrado

p.s. I know Thomas already explained it but i think this parameter deserves a special post because can change the heli behaviour quite a lot!!!!!

[alexf1852]

Corrado,

Since you have the cyclic gain adjustment (P gain) in the Basic mode screen, shouldn't changes here also make changes by certain ratios to the I, and D gains? I noticed that it doesn't. So someone could make large changes to cyclic gain and forget to also change the I and D gains.

[HighFreq]

P gain adjustment in the wizard is there for fine tuning, if the correct model is choosen (450,500,600....) than the P gain slider should only be adjusted by few % in case of oscillations or spongy feeling.


Corrado

[almigurt]

A manual with these hints 'd make lots of satisfied customers...

[gevaho]

Originally Posted by almigurt

A manual with these hints 'd make lots of satisfied customers...

Yes, or an advanced tuning guide with step by step tuning info.

But we heli-guy's love a bit of fiddling don't we

[almigurt]

Yes we do. A "little"...

[almigurt]

So agility doesn't simply reduce deg/s rate?

[alexf1852]

Originally Posted by almigurt

So agility doesn't simply reduce deg/s rate?

No. Going by the description it seems to be similar to Style setting on Vbar and to some extent pot 2 on BeastX.

[desertstalker]

Originally Posted by alexf1852

No. Going by the description it seems to be similar to Style setting on Vbar and to some extent pot 2 on BeastX.

Thats what it feels like, the lower that is the more the Brain is in control IMO, it feels solid but "distant" or "robotic" i guess you could say. I prefer a lower value, but it is personal choice.

[DannyvG]

Originally Posted by Otatiaro

Just to add some more data on the stack

this is the first time I have read a proper explanation about parameters in a FBL system.
Most others use all kind of stupid marketing terms with vague stories what they do. But this I can relate to, it gives a clear description of what the parameter does and what the change in feeling should be.
Nice work.

[maddog7]

Originally Posted by HighFreq

Probably agility is an important parameter and rather easy to use and we should put it in the wizard in the setup panel where people choose flying characteristics (beginner, sport, 3d, hard 3d)

Remember:

Lower agility----->locked feeling
Higher Agility---->more natural flybarred feeling


Corrado

p.s. I know Thomas already explained it but i think this parameter deserves a special post because can change the heli behaviour quite a lot!!!!!

I hope this is not a dumb question but if the agility is lowered is this more likely to reduce drift and an increase allow more drift?

Am I right in thinking that agility does not affect eg max pitch and roll rates, only the feeling of control over those rates?

D
:-)

[HighFreq]

It affects just the feeling in flight.

Corrado

[almigurt]

Sounds like expo to me anyway, you are doing a great Job on the brain. Thanks for all the good hints and answers.

Just please please please add a field programming unit!

Its so annoying to carry a computer to the flying field.


P.s. When will protos 700 be aviable?