Replies to This Discussion

Permalink Reply by Rob. S on June 27, 2011 at 2:31am

Just another Question to add to this
could it b my esc's at the moment its all default?

wat changes should i make to them??

• ▶ Reply

Permalink Reply by emile on June 27, 2011 at 11:06am

Please share some more details on your setup.

Like motors, escs frame,prop size etc.

E.

• ▶ Reply

Permalink Reply by Rob. S on June 27, 2011 at 2:29pm

890kv motors

35-50amp esc's

arms 28cm (default arducopter frame)

12" props

• ▶ Reply

Permalink Reply by Russell B. Sutton on June 27, 2011 at 2:41pm

try P = .3  I = .025  D = .13 for pitch & roll

• ▶ Reply

Permalink Reply by Pieter van Woerkom on June 27, 2011 at 4:42pm

Just tried this to find the PID for the sonar Altitude Hold.

By finding the maximum P where the AH starts to oscillate (copter goes up-down-up-down in rapid a/decends)

by timing, with a stopwatch for 10 cycles the average gave me Pu Unfortunately i hade a blade-off during te last test flight so could not implement the calculated PID. Still wondering about how high the gain (intmax) of this loop could be calculated.

 

 

 

[adapted from wiki]

Ziegler–Nichols method

The I and D gains are first set to zero. The P gain is increased until it reaches the ultimate gain, K_u, at which the output of the loop starts to oscillate. K_u and the oscillation period P_u are used to set the gains as shown:

Ziegler–Nichols method

Control Type	Kp	Ki	Kd
P	0.50Ku	-	-
PI	0.45Ku	1.2Kp / Pu	-
PID	0.60Ku	2Kp / Pu	KpPu / 8

_{[/end snip]}

 

• ▶ Reply

DeveloperPermalink Reply by jasonshort on June 27, 2011 at 5:52pm

Here is how I found the current gains:

This is a quick mock up in Flash of the Alt hold PID.

• ▶ Reply

DeveloperPermalink Reply by Michael Oborne on June 27, 2011 at 6:26pm

Just playing with this
p = 0.8
i = 0.02

d = 1

• ▶ Reply

DeveloperPermalink Reply by jasonshort on June 27, 2011 at 6:56pm

Whats dangerous with a high D is the noise in the Baro sensor. Worth a shot trying something higher.

• ▶ Reply

Permalink Reply by Greg Fletcher on June 27, 2011 at 7:35pm

Jason, That's one of the coolest things I've seen for a while. Thanks for that. It took me a while to figure out it was one of your interactive flash objects or what ever they're called. If you could put in real data from a throttle/thrust curve and the mass of the thing, it could be a real simulation. I have been playing with this thing for a few to many mins. It's a great training aid.

• ▶ Reply

DeveloperPermalink Reply by jasonshort on June 27, 2011 at 8:00pm

I have mass and gravity in the sim. No throttle curve but that would be trivial to add. BTW, hit the space bar to cause an engine failure... 

(you need to click on the animation to give t keyboard focus first)

• ▶ Reply

Permalink Reply by Greg Fletcher on June 27, 2011 at 8:37pm

Trivial indeed, for you:)  I have built a wooden balance rig and measured thrust. I need to build a bigger one for larger motors and I need to get a digital balance with serial output. Or buy another copy of $MotorCalc or both.

• ▶ Reply

DeveloperPermalink Reply by jasonshort on June 27, 2011 at 8:46pm

I meant trivial to replicate a curve in code. Not to measure the output of a prop! ha. I'd loose a finger trying that.

 

• ▶ Reply

• ‹ Previous
• 1
• 2
• Next ›
• Page