dirkclod
Premium Pilot
You just lost me :?N017RW said:
I had the top off already and since we were talking about this I decided to try it this way and that's what I found .
I am just not sure what you are saying .
Finally fixed my Yaw Drift
- Thread starter sgrim70
- Start date
I had the top off already and since we were talking about this I decided to try it this way and that's what I found .
I am just not sure what you are saying .
N017RW
Premium Pilot
I'm saying you want the IMU, i.e sensors or Naza, and the motors to be on the same plane as each other.
If either one is tilted compared to the other then there is an error between the input (IMU sensors) and the output (motors).
Once you achieve that then make sure the 'system' is level before doing the IMU cal.
If either one is tilted compared to the other then there is an error between the input (IMU sensors) and the output (motors).
Once you achieve that then make sure the 'system' is level before doing the IMU cal.
dirkclod
Premium Pilot
Got ya . No I was just trying to find out which way leveling worked better .I had been using the glass to do it and then tried the other way which seemed to give me better results and was just checking the glass at the end to see if it was level compared to the FC .
N017RW
Premium Pilot
If you make them the same then the sensors are aligned with the thrust.
EDIT-Add: Think of it like when your car steering wheel must be a bit off-center to go straight. You need an alignment! A bit abstract but it may help with the concept.
The last thing would be to make sure the Naza, or the IMU in the Naza, is in or at the rotational (Yaw) center point like the center of the bird but that's not possible with all the other stuff in there.
That's why the P2Assistant has that section to enter offsets but I don't think they're active for a "Phantom" Naza.
EDIT-Add: Think of it like when your car steering wheel must be a bit off-center to go straight. You need an alignment! A bit abstract but it may help with the concept.
The last thing would be to make sure the Naza, or the IMU in the Naza, is in or at the rotational (Yaw) center point like the center of the bird but that's not possible with all the other stuff in there.
That's why the P2Assistant has that section to enter offsets but I don't think they're active for a "Phantom" Naza.
After reading the full thread on the DJI forum (great thread) I think a sheet of glass across the 4 motors and a bull’s-eye spirit level will do the job the best as it’s the motors that drive the Phantom and they are the most important thing to be level.
As for the IMU being level inside I don’t think this will mater as long as the motors are level. Even if the IMU is a mile off and you get motors level and calibrate the IMU. It will set its self to the same level (ie 0 0 0 0) as the motors, all should be fine. I would think this is how DJi would do it so they don’t have to set the IMU perfectly level with the body.
let me know what you guys think.
As for the IMU being level inside I don’t think this will mater as long as the motors are level. Even if the IMU is a mile off and you get motors level and calibrate the IMU. It will set its self to the same level (ie 0 0 0 0) as the motors, all should be fine. I would think this is how DJi would do it so they don’t have to set the IMU perfectly level with the body.
let me know what you guys think.
dirkclod
Premium Pilot
Well all I know is what I posted and saw when I flew it afterwards . I have now put 3 battery's through it and it fly's better than before .
I do know I am going to just use the glass and shim it up level from now on as it's just to much trouble to take the top off to do it that way every time . On mine it did help it.
I do not know how to talk about this stuff like N017RW can as I am not that smart . I just read what other folks are saying and doing so I tried it . Why it worked for me I haven't got a clue. All I can say is it did .
I did think as long as the motors were level that's all it needed but I am seeing it differently now.
I do know I am going to just use the glass and shim it up level from now on as it's just to much trouble to take the top off to do it that way every time . On mine it did help it.
I do not know how to talk about this stuff like N017RW can as I am not that smart . I just read what other folks are saying and doing so I tried it . Why it worked for me I haven't got a clue. All I can say is it did .
I did think as long as the motors were level that's all it needed but I am seeing it differently now.
N017RW
Premium Pilot
fin032 said:
I respect your opinion.
But I disagree if you're not seeing the theory behind it.
Does it matter on a Phantom with a Naza? Who knows??? It may be academic at the 'price point' of a Phantom.
Not sure but academically speaking aligning your input platform with your output platform will result in more precision.
Hi < N017RW,
The way I am thinking is if you have the 4 motors runing perfectly level at the correct speed the Phantom should sit still in the air with no wind. To move forward 2 motors runs faster making the Phantom tip forward making it move so if you can set the IMU to 0 0 0 0 when the motors are perfectly level this could work.
I also respect everyones opinion as I am no expert on this.
The way I am thinking is if you have the 4 motors runing perfectly level at the correct speed the Phantom should sit still in the air with no wind. To move forward 2 motors runs faster making the Phantom tip forward making it move so if you can set the IMU to 0 0 0 0 when the motors are perfectly level this could work.
I also respect everyones opinion as I am no expert on this.
N017RW
Premium Pilot
Exactly what I'm saying.
The IMU and the motors must be in alignment in the 'X' & 'Y' planes. Each level to each other sort-of-speak. THEN level the system (thus both will now be level) and cal the IMU.
The slight 'Z' offset of the two is small and must be accepted in this design and form-factor.
There's no argument that coplanarity between the input and output is optimal for this type of system. How crucial?, given the quality or performance of the sub-systems is however arguable.
The IMU and the motors must be in alignment in the 'X' & 'Y' planes. Each level to each other sort-of-speak. THEN level the system (thus both will now be level) and cal the IMU.
The slight 'Z' offset of the two is small and must be accepted in this design and form-factor.
There's no argument that coplanarity between the input and output is optimal for this type of system. How crucial?, given the quality or performance of the sub-systems is however arguable.
fin032 said:
Here's what I think. fin032 you've got me rethinking all this.
It is possible DJI levels the NAZA, calibrates the IMU, puts the top on and ships it to us.
That would suck because the first time we do the Advanced IMU Calibration, without leveling the NAZA, we induce the yaw drift.
I've read a number of posts which state very clearly "this is the solution".
So, fin032, I have to ask what makes you question the conclusions which so many believe?
Was it in the DJI thread or this thread?
Just when I thought it was all settled... :roll:
Do we have a volunteer to test the theory by tipping their NAZA Module a bit?
Think of it as 3 levels
1) Motor level. Very important as this is what controls where the Phantom moves.
2) Naza mod level. I don't see how this maters as its set as level with a calibration. I think when people are doing well levelling the Naza module it’s because DJI have got the level of it very close to the motor level.
3) 0 0 0 0 point after calibration, think of this as an invisible line. This is the point the Phantom thinks its level.
If you get the motors 100% level meaning the point where the props cut in to the air is level and you calibrate at that point bringing the 0 0 0 0 invisible line will be 100% in line with the motors and 100% level its self. If you think of it that ways there no need to have the Naza module 100% level.
Now if you think of it the other way.
If you make sure that the Naza module is 100% level and calibrate at that so the 0 0 0 0 invisible line point is 100% the same level as the Naza module but the motors for talking sake are 10 degrees off from the level of the Naza Module and the 0 0 0 0 point the Phantom will think its level but the motors and props will be cutting in to the air 10 degree off. I would think the Phantom would be all over the place with that.
I know the Naza module would never be 10 degree out from the motor level but just trying to show what should happen if it was. I will try sticking an image up tonight to try show what I mean. Just remember the 0 0 0 0 invisible line point is where the Phantom thinks it level.
I may have got this wrong but it’s just my opinion.
It would be good for someone to try push there module way off and level the motors to see what happens but not sure what the limits of this would be meaning how far it could be off and still set a 0 0 0 0 point. Not for me
1) Motor level. Very important as this is what controls where the Phantom moves.
2) Naza mod level. I don't see how this maters as its set as level with a calibration. I think when people are doing well levelling the Naza module it’s because DJI have got the level of it very close to the motor level.
3) 0 0 0 0 point after calibration, think of this as an invisible line. This is the point the Phantom thinks its level.
If you get the motors 100% level meaning the point where the props cut in to the air is level and you calibrate at that point bringing the 0 0 0 0 invisible line will be 100% in line with the motors and 100% level its self. If you think of it that ways there no need to have the Naza module 100% level.
Now if you think of it the other way.
If you make sure that the Naza module is 100% level and calibrate at that so the 0 0 0 0 invisible line point is 100% the same level as the Naza module but the motors for talking sake are 10 degrees off from the level of the Naza Module and the 0 0 0 0 point the Phantom will think its level but the motors and props will be cutting in to the air 10 degree off. I would think the Phantom would be all over the place with that.
I know the Naza module would never be 10 degree out from the motor level but just trying to show what should happen if it was. I will try sticking an image up tonight to try show what I mean. Just remember the 0 0 0 0 invisible line point is where the Phantom thinks it level.
I may have got this wrong but it’s just my opinion.
It would be good for someone to try push there module way off and level the motors to see what happens but not sure what the limits of this would be meaning how far it could be off and still set a 0 0 0 0 point. Not for me
dirkclod
Premium Pilot
fin032 said:Think of it as 3 levels
I understand what ya are saying But I also know this .
I have done all 3 ways and using the level inside of Mine makes mine perfect !!!
I thought I had it figured out when I first used the glass on top to level it and then flew it as it was better .
But when I level from the inside it was perfect !!! Now this might be with My bird and a friend told me last nite that
at a certain point we can go overboard on all of this and he is right ! This isn't the Space Shuttle
Maybe they when at the factory and set it up the first time have it upside down on the top of the prop shafts when they
calibrate it :idea:
Maybe I try that next :lol:
N017RW
Premium Pilot
dirkclod
Premium Pilot
N017RW said:
Yea but my problem is I don't know what to do with it or how to explain it when I find it ! :? :lol:dirkclod said:fin032 said:Think of it as 3 levels
I understand what ya are saying But I also know this .
I have done all 3 ways and using the level inside of Mine makes mine perfect !!!
I thought I had it figured out when I first used the glass on top to level it and then flew it as it was better .
But when I level from the inside it was perfect !!! Now this might be with My bird and a friend told me last nite that
at a certain point we can go overboard on all of this and he is right ! This isn't the Space Shuttle
Maybe they when at the factory and set it up the first time have it upside down on the top of the prop shafts when they
calibrate it :idea:
Maybe I try that next :lol:
The main thing here is someone has worked out a problem that has been bugging people for a wile and the good thing is there are now 3 options to try. Some folk might like one way and others will be happy with the other. Glad yours is sorted.
dirkclod
Premium Pilot
Me not just right no way
Has anyone tried Fin032 's idea of calibrating with the help of a glass sheet sitting on top of motors ?
dirkclod
Premium Pilot
Yes and ya can see on the 1st page what I found by doing it .
That is the way I will be doing mine from now on as taking the top off just to much trouble every time .
Thanks !
I am going to try it as soon as I will get a glass sheet.
Although my Phanton is stable enough the extra precision is always welcome
Similar threads
