I am having very high CPU usage on my flight controller.
Before I get asked yes I am using the configurator, I am at the point before a quick tune.
Param 82424.param (20.9 KB)
Set INS_GYRO_RATE to 1. 4Khz on 3 IMUs is going to kill your H7
Thank you very much!
So I am having some issues with cube orange +, @andyp1per helped get the cpu usage down from 100%. Yes I am using the configurator.
What my current issue is, I take off and the quad bascially wants to take off like there is a baro issue.
The other this is I have completed all the beg steps for configuring my quad but it has a small oscillation that becomes larger and then it become uncontrollable.
I currently have the cube on a rubber mount, I read that they do better foam tape mounted?
The props are no where near the cube. The top and bottom plate are 3.5 mm each.
The vibrations before this flight are under 6 for x and y, under 12 for z.
There is masses of oscillations on the motor outputs. I cant tell what your attitude control is doing since logging is wrong.
Vibrations are very bad - probably the worst I’ve ever seen. Have a good look around for physical vibration issues and fix them. Search for all posts about vibrations and read through them.
Dont attempt another flight until you’ve made serious attempts to reduce vibrations.
Start by changing to all these ready for the next flight:
ATC_ANG_PIT_P,6
ATC_ANG_RLL_P,6
ATC_RAT_PIT_D,0.006
ATC_RAT_PIT_I,0.12
ATC_RAT_PIT_P,0.12
ATC_RAT_RLL_D,0.006
ATC_RAT_RLL_I,0.12
ATC_RAT_RLL_P,0.12
ATC_RAT_YAW_I,0.03
ATC_RAT_YAW_P,0.30
BATT_FS_CRT_ACT,1
BATT_FS_LOW_ACT,2
FENCE_ENABLE,1
GPS_GNSS_MODE,7
GPS_GNSS_MODE2,7
INS_HNTCH_FREQ,30
INS_HNTCH_BW,8
INS_HNTCH_OPTS,2
INS_LOG_BAT_MASK,1
LOG_BITMASK,180222
PSC_ACCZ_I,0.4
PSC_ACCZ_P,0.2
SCHED_LOOP_RATE,400
Also you can try:
MOT_PWM_TYPE,5 or even 6
for DSHOT300 or 600. I’ve noticed some ESCs were not happy with DSHOT150. I guess the manufactures never tested because they assumed everyone had small racing quads and would only ever use DSHOT1200.
Shawn thanks for the reply,
I took the make shift rubber mount off and hard mounted the FC to the frame to help with vibrations (at least thats what I read is best for cube orange+) feel free to correct me.
Vibrations on previous flights are under 15, I will start digging on that though, I agree they are high in this flight, maybe it’s the folding props.
I changed to dshot 600.
I am assuming that changing the GPS_GNSS_MODE to 7 just helps the refresh rate correct? My first GPS unit is the UM982 by holybro.
PSC_ACCZ_I,0.4 PSC_ACCZ_P,0.2 These were set from hover data on the configurator.
I want to confirm that these are different (ATC_RAT_YAW_I,0.03,ATC_RAT_YAW_P,0.30)?
My SCHED_LOOP_RATE,is set to 800, should I drop that down to 400 im guessing?
Here are my PID’s and bit mask
I really appreciate your reply, any of my replies are just my wanting to better understand what effects what.
Update the values @xfacta gave you in the configurator and publish an updated .zip file with the “reason changed” fields edited. Other users also what to know what affects what. You can help them with that, and learn something in the process.
Is the standard of practice to hard mount a cube FC?
@xfacta I have updated all the setting that you suggested but I am working through all the possibilities for vibrations prior to next test flight.
Yes it is for CubePilot autopilots.
And post the updated .zip file
The PID changes are values I’ve found to be reasonably safe in most cases, since the log you provided made me think your existing PIDs might have been attained under less-than-ideal conditions.
The other changes are just sensible values. For example running a high loop rate and a bunch of extra notch filter options are unlikely to do anything for a copter this size - certainly in the state it was in.
Before layering on the performance-enhancing options lets get some good flights with nice clean data.
Dont ignore the safety-related parameters I specified - during testing and tuning is where you need them most.
@xfacta Understood about the safety settings. Before I lift off again I will need to make sure that the vib issues are something that I can pinpoint and mitigate.
The vibe issue is perplexing to me because in the prior flight they were well under the 20 number in z and under 12 in the rest.
Clearly this doesn’t work for all frames. This argument is so old it should be retired, certainly moderated. If you have a very rigid frame sure, bolt it down. Otherwise it depends on the assembly. Too many Cube applications to count have taken advantage of a compliant FC mount when their frame is a mess.
But, in addition to massive vibration you have massive output oscillation as noted. This is all tuning.
Start over.
@xfacta Working on getting vibes down, one thing that I notice is 3 IMU sensors but a large swing in vibration data. I’m guessing all three need to be under the 20ish mark.
Vibrations being under 20 is definitely an advantage to reliability and tuning.
There is no clipping, so the copter could potentially operate like this but Autotune may not work well.
Set these before continuing:
BATT_FS_CRT_ACT,1
BATT_FS_LOW_ACT,2
LOG_BITMASK,180222
Thanks for looking at the file. I’ve updated the params.
@xfacta on the subject of vibrations…
Is the “being under 20” measured after basic tuning is completed with no wind and under hover?
Reason why i am asking, when I do hover flights I notice that my vibrations are very normal (under 15), but when I do say a mag fit or decent agressive flight I see that they climb up but with no clipping at all.
Just trying to make sure.