what are the negative symptoms of air and anti-gravity modes? - RC Groups
Thread Tools
Nov 10, 2017, 11:53 AM
Registered User
Discussion

what are the negative symptoms of air and anti-gravity modes?


I usually tick them on by default, but what would indicate I need to turn them off?
Sign up now
to remove ads between posts
Nov 10, 2017, 01:56 PM
Registered User
Air mode is a problem if you try to land with motors spinning because it will will try to orient itself at high power once it hits the ground.
Nov 10, 2017, 02:51 PM
Registered User
Landing with ACRO and airmode just requires that you disarm as you land. Anti Gravity mode has a level setting. Too little and it does nothing. Not sure what too much does. Purpose of Anti Gravity is to help keep the quad from pitching up or down when you punch the throttle.
Nov 10, 2017, 03:06 PM
Registered User
Antigravity boosts the I-gain on large throttle throws. Needing it is however an indicator that your tune is technically bad, which most BF tunes are, since the I-gain seems to be treated as an afterthought all the time.
Nov 10, 2017, 05:33 PM
Registered User
Quote:
Originally Posted by Glowtape
Antigravity boosts the I-gain on large throttle throws. Needing it is however an indicator that your tune is technically bad, which most BF tunes are, since the I-gain seems to be treated as an afterthought all the time.
Not necessarily so.

I-term is responsible for attitude (or angle) corrections over a "long term".

Too much I and your quad feels stiff especially when responding to small adjustments. Too little I term and your quads angle will eventually return to 0 or its last angle before your change.

You want I term right at the point where your angle is held indefinitely until your next change. Alot of people confuse I term with being responsible for fixing "drifting". This is completely opposite of what it actually does.

Anyways, normally a low I term is enough to keep your angles held at where you want it to be. But on throttle manipulation the pitch and yaw axis are affected to a point and now, because of a low I term, your quad will pitch or yaw during high throttle inputs.

To combat this you would increase I term to make it more resistant to changes but then you start getting a stiffer quad.

The compromise is anti-gravity - I term is temporarily boosted to get that stiff response during high throttle inputs and then it is low/normal during typical throttle inputs.

How fast and how much I term is adjusted is based off the gain and threshold settings.
Nov 10, 2017, 06:43 PM
Registered User
Over on dRonin, my I-gain is like 9x that of the P-gain on my 220 quad, and I can assure you, it's anything but "stiff". The high I-gain makes it converge to the new solution rather fast.
Nov 10, 2017, 08:55 PM
Registered User
Quote:
Originally Posted by Glowtape
Over on dRonin, my I-gain is like 9x that of the P-gain on my 220 quad, and I can assure you, it's anything but "stiff". The high I-gain makes it converge to the new solution rather fast.
That's not a fair comparison really. Raceflight d gains are represented in 0.xxx. Setting d gain to 1 in raceflight will send your quad into the air as soon as it is armed. Ask me how I know. Cf/bf d gains are represented in XX.

The point is how I use I term (and generally how cleanflight and betaflight apply it) is what I stated. I don't use dronin but if you are telling me that your I term is 9x that of your p gain and your quad is not "stiff" that tells me the weighting in dronin is different and again cannot be compared directly to cf/bf application.
Nov 11, 2017, 07:42 AM
Registered User
P and I term are both on the same scale. In Betaflight, in Raceflight and in dRonin. D has a different scale, because if it were also the same, it was pretty minuscule.

Means (error * kP) would the same as (error * kI) if kP = kI.
Nov 11, 2017, 01:11 PM
Registered User
As I said, I don't use dronin. And I'm not inclined to load it onto any of my boards at all. Autotune, from my research, doesn't look any more reliable than my manual pid tuning which my personal process gets me my results in a single battery pack (2 minutes or less).

Whatever screenshots and documentation I can find suggests that dronin autotune leans toward low d gains, mid p gains and high i gains with a large emphasis on noise filtering.

This would make sense as to why your i gains are higher than your p gains especially when noise filters are constantly employed.

My quads are pid tuned with dynamic filter on and hard filters off. The net results are high p gains, mid d gains and low i gains coupled with the anti-gravity feature settings (gain and threshold). I have yet to burn a motor for being too hot.

In the end I still believe dronin's weighting/scaling of their pid gains is different than cf/bf especially when I have come across information that suggests that autotune will never yield the same result twice for the same build. There is enough variances in the autotune process for me to stick with my manual pid tuning. This is also why I don't use other peoples pids.

As a side note, boris tried implementing an autotune (not dronins autotune) feature on a previous bf build and also decided it just was not reliable enough to keep working on.

If you like dronin and it flies for you more kudos to you but I still think dronin relies heavily on noise filtering which can be resolved by building the quad to a better state.
Nov 11, 2017, 08:19 PM
Registered User
dRonin runs a PT1 filter by default, just like BF does (again), nothing else. I'm not sure where that idea of heavy noise filtering comes from. However if you're fancy, you can raise it up to the 8th order, for whatever reason, which turns it into a Butterworth.

As far as autotune goes, as opposed to Boris, the author of the initial implementation (dR has an improved one since recently) did a whole lot of mathematical work, of the kind you don't find in BF (the attitude EKF of dR/TL is equally complex). It's essentially a paper.

https://github.com/d-ronin/dRonin/bl...Derivation.pdf

In the end, I don't really care who flies what, anyway. I just brought it all up to point out that the BF tunes are in general bad, if it requires antigravity. The I-term isn't a bad thing, it gets however applied badly. It's just easier to crank up P, then unf*ck with with D, and then just get I up enough the quad maintains attitude in rate mode, than tune it correctly with whatever methods there are (including Ziegler-Nichols, which would likely also result in high I-gains).
Last edited by Glowtape; Nov 11, 2017 at 08:31 PM.
Nov 11, 2017, 10:10 PM
Registered User
Bf filter default is still on biquad iirc. PT1 still has to be applied.

Your statement about cranking p, fixing with d and then adjusting i to hold attitude is precisely how I tune. But, there are people who don't know when to stop adjusting p and everything else is correlated in the end.

I basically started with this resource on pid tuning https://github.com/betaflight/betafl...D-Tuning-Guide and have since refined my personal process to get my results. The difference in my tuning is that I start with i and d at 0 when tuning the p gain. The bf wiki mentions high i gain resulting in "robotic" (or stiff) characteristics. I have seen it first hand.

As for the bf tunes being generally bad, I would agree with you to a point. The default tune is meant to be noob friendly and get the quad in the air. Tuning should still be done eventually.

Ant-gravity doesn't exist to fix bad tunes. It exists to "have your cake and eat it too". At least that's my opinion. I don't care to fly with high i gain all the time (I personally don't like how the quad feels/responds) but I also prefer high i gain in the moments where I want full throttle bursts.

Like you said, I also don't care who flies what as long as at the end of the day I'm in the air with a quad I like flying.
Nov 13, 2017, 09:37 AM
Registered User
the filter default is biquad on BF, as far as I know

I only recognize a few of the other words used, but I'm learning


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Question What Are The Negatives To Shopping Online? traceyl Life, The Universe, and Politics 92 Aug 23, 2016 10:03 AM
Discussion So, what are the symptoms of balsa dust allergy? merlinmurph Hand Launch 18 Apr 10, 2008 04:38 AM
Help! What are the symptoms of a Dying HBK2 motor? OGBrain Mini Helis 19 Jan 10, 2008 05:54 PM
Question What are the symptoms that occur when the BEC is overloaded mpcrash Castle Creations 7 Oct 04, 2006 12:06 AM
BEC Failures-What are the symptoms? Ah Clem Micro Helis 1 Jul 02, 2005 02:27 PM