PIDtoolbox
PIDtoolbox copied to clipboard
bw1129
Function Request
Would it be possible to add the capability to process multiple files at once, then dump the resulting step-response to a GIF file?
Example: https://youtu.be/IaulQLG__eA
Spent the last two days making that and I'm really close to the ideal Servo PID's.
Thanks in advance!
-RS
Do you mean you would like Ptb to compute a single step response from data from several logfiles? or do you want an animated Gif? Im guessing the latter
Well that would be very cool too but what I was trying for was being able to run multiple flights at differing PID values. Then be able to see all of the logged data as a Step-Response GIF or video. This would streamline the tuning process by allowing someone to make small changes overtime and then see the overall results.
On Fri, Dec 6, 2019 at 4:12 PM Brian White [email protected] wrote:
Do you mean you would like Ptb to compute a single step response from data from several logfiles?
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/bw1129/PIDtoolbox/issues/31?email_source=notifications&email_token=ACPI7PPR5RSHYKCDFJOMG7DQXJTXDA5CNFSM4JWZ2A22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEGEMCTI#issuecomment-562610509, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACPI7PODCWSKYB3U6RFLNQ3QXJTXDANCNFSM4JWZ2A2Q .
Either that or have the ideal PID's calculated using data from multiple logs.
On Fri, Dec 6, 2019 at 4:12 PM Brian White [email protected] wrote:
Do you mean you would like Ptb to compute a single step response from data from several logfiles?
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/bw1129/PIDtoolbox/issues/31?email_source=notifications&email_token=ACPI7PPR5RSHYKCDFJOMG7DQXJTXDA5CNFSM4JWZ2A22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEGEMCTI#issuecomment-562610509, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACPI7PODCWSKYB3U6RFLNQ3QXJTXDANCNFSM4JWZ2A2Q .
Yes, that would be cool. It would be a bit cumbersome to program at the moment, just because of the way Ptb is structured, but certainly a great idea. If one could for example just go out and collect 5-10 logfiles, with various PIDs, it would be nice for Ptb to produce a complete function describing the changes in the step resp as a function of PIDs.
Yeah, thats exactly what I've been doing with my Tricopter tuning. https://www.youtube.com/watch?time_continue=61&v=KszUlSj1dPc&feature=emb_logo While this flies pretty good, the tail-wag is ever present...
On Fri, Dec 6, 2019 at 4:40 PM Brian White [email protected] wrote:
Yes, that would be cool. It would be a bit cumbersome to program at the moment, just because of the way Ptb is structured, but certainly a great idea. If one could for example just go out and collect 5-10 logfiles, with various PIDs, it would be nice for Ptb to produce a complete function describing the changes in the step resp as a function of PIDs.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/bw1129/PIDtoolbox/issues/31?email_source=notifications&email_token=ACPI7PLQXMTL54IK5YZNMYLQXJW6DA5CNFSM4JWZ2A22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEGEO7RI#issuecomment-562622405, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACPI7PLLR5TUD5C5US6H2O3QXJW6DANCNFSM4JWZ2A2Q .
Sorry, wanted to link this as well... https://rotorbuilds.com/build/18949
On Fri, Dec 6, 2019 at 4:40 PM Brian White [email protected] wrote:
Yes, that would be cool. It would be a bit cumbersome to program at the moment, just because of the way Ptb is structured, but certainly a great idea. If one could for example just go out and collect 5-10 logfiles, with various PIDs, it would be nice for Ptb to produce a complete function describing the changes in the step resp as a function of PIDs.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/bw1129/PIDtoolbox/issues/31?email_source=notifications&email_token=ACPI7PLQXMTL54IK5YZNMYLQXJW6DA5CNFSM4JWZ2A22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEGEO7RI#issuecomment-562622405, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACPI7PLLR5TUD5C5US6H2O3QXJW6DANCNFSM4JWZ2A2Q .
Cool! Makes sense that it’s a small copter. You’d certainly want a light tail so the servo could move very quickly. It really has to be limited in its resp time to change directions though. This is why I would presume that it’s be better controlled from I gain the way BF currently recommends yaw tuning. While yaw authority would be high for a tri copter, the change in yaw direction would takes some time, so stopping a fast yaw maneuver would be too slow for p gain to be very effective.
Well Yaw-I doesn't work the same on a Tricopter, simply because Quads use differential torque, while Tri's much thrust vector. Thus a Servo is completely driven by Yaw-P, while Yaw-I acts more to help raise it's response to "1" then to stabilize the copters strait-line flight. Yaw-D can be used to a limited degree but often can cause response problems due to the added loop / filtering delay compounding the Servos slow response rate. Keep in mind that our worst ESC's are 10-20 times faster then any Servo on the market.
On Fri, Dec 6, 2019 at 4:51 PM Brian White [email protected] wrote:
Cool! Makes sense that it’s a small copter. You’d certainly want a light tail so the servo could move very quickly. It really has to be limited in its resp time to change directions though. This is why I would presume that it’s be better controlled from I gain the way BF currently recommends yaw tuning. While yaw authority would be high for a tri copter, the change in yaw direction would takes some time, so stopping a fast yaw maneuver would be too slow for p gain to be very effective.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/bw1129/PIDtoolbox/issues/31?email_source=notifications&email_token=ACPI7PIZAH5HX763KZZRZUTQXJYI5A5CNFSM4JWZ2A22YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEGEQGOA#issuecomment-562627384, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACPI7PNPEHSQ35N5A6WECL3QXJYI5ANCNFSM4JWZ2A2Q .