So @58:45 you speak about the D term filtering having less detrimental effect than if you used more Gyro filtering, am I correct in saying Gyro filtering will delay everything down stream (P, I & D). Because keeping less Gyro filtering and adding more D term filtering is a viable tuning approach.
so the idea of a dterm notch filter below 100hz was briefly mentioned around 1:20:00… Seems like the consensus was that it could work for a low hz spike even if it does add latency which might not be as important for x8’s?!
I haven’t seen good evidence that it’s necessary. Maybe on very underpowered rigs, but then you gotta look at motor saturation and see if you’re gaining anything
@@Quick-Flash when I tested it on my 5inch rig, I expected it to have at least some minor impact on the shape or timing of the step response curves if it indeed added any extra authority. For this rig at least, I saw absolutely no difference in any parameters. What would be the ultimate test to see whether it has any effect?
@@TimeFadesMemoryLasts a fireman fight the fire. then i read: and pilot who knows the secrets behind filtering for propwash control. and in the video they don't know exactly ... to me these are firemen which don't know what a fire is.🙃
Well, this is a video for everyday people. Going at it from a control theroy perspective with only math doesn't make it better for anyone watching, right? They know far more than what was shown in the video, lol. I'm sure they were just worried that people got bored and left when it would have gotten too nerdy and non-practical. Firemen only need to know how a fire reacts. They don't need to know what fire really is from a physical and mathematical perspective to extinguish it.
@@7robertb That's where you are wrong. It's an instability in the PID loop, initiated by stalling prop blades. And this instability can be described by maths. I've spoken to the devs and there's already a feature in the works to minimize this instability further.
Thank you for this great livestream, lots of useful information.
So @58:45 you speak about the D term filtering having less detrimental effect than if you used more Gyro filtering, am I correct in saying Gyro filtering will delay everything down stream (P, I & D). Because keeping less Gyro filtering and adding more D term filtering is a viable tuning approach.
this is a really good point! thx
@PIDtoolbox I think Chris Rosser touched on this idea in a recent interviewer with Chris Tompson. Loved this episode thanks
so the idea of a dterm notch filter below 100hz was briefly mentioned around 1:20:00… Seems like the consensus was that it could work for a low hz spike even if it does add latency which might not be as important for x8’s?!
Gmt +7 here, sorry cant watch the live stream. But now im watching with my morning coffee 😁☕️
+1👍 💪 great to see devs. great stream as always
Great content! should have rated it!
Learning a lot from your content. Appreciate it!
Super stream! Thanks guys! A question from a beginner, maybe it's easier to filter PIDsum?
Great video
Are you also suggesting to modify the default PID_SUM_LIMIT to 1000 ?
I haven’t seen good evidence that it’s necessary. Maybe on very underpowered rigs, but then you gotta look at motor saturation and see if you’re gaining anything
@@PIDtoolbox it keeps a better motor scaling then just cutting it off early. Better scaling between axis.
@@Quick-Flash when I tested it on my 5inch rig, I expected it to have at least some minor impact on the shape or timing of the step response curves if it indeed added any extra authority. For this rig at least, I saw absolutely no difference in any parameters. What would be the ultimate test to see whether it has any effect?
@@PIDtoolbox pitch yaw and roll all at once all very quick.
@@Quick-Flash what would you be looking for and what would be the control/benchmark to see whether there was some benefit?
Super
pondering over troubled water ...
not really helpfull.
What do you mean?
@@TimeFadesMemoryLasts a fireman fight the fire. then i read: and pilot who knows the secrets behind filtering for propwash control.
and in the video they don't know exactly ...
to me these are firemen which don't know what a fire is.🙃
Well, this is a video for everyday people. Going at it from a control theroy perspective with only math doesn't make it better for anyone watching, right? They know far more than what was shown in the video, lol. I'm sure they were just worried that people got bored and left when it would have gotten too nerdy and non-practical.
Firemen only need to know how a fire reacts. They don't need to know what fire really is from a physical and mathematical perspective to extinguish it.
@@TimeFadesMemoryLasts there is no special math with regards to propwash.
@@7robertb That's where you are wrong. It's an instability in the PID loop, initiated by stalling prop blades. And this instability can be described by maths. I've spoken to the devs and there's already a feature in the works to minimize this instability further.