Welcome to PhantomPilots.com

Sign up for a weekly email of the latest drone news & information

imu calibration how level does it need to be?

Discussion in 'Pro/Adv Discussion' started by cypher007, May 27, 2016.

  1. cypher007

    Joined:
    Apr 26, 2016
    Messages:
    62
    Likes Received:
    10
    ive got the horizon droop. ive tried gimbal calibration which made it better but not straight. so I did a imu calibration in my lounge on the carpet, the floor may not be 100% level. do I need to get a piece of hardwood then level that?
     
  2. With The Birds

    Joined:
    Jan 4, 2016
    Messages:
    2,577
    Likes Received:
    733
    Location:
    Australia
    You could get the hardwood or just find a level surface somewhere else.

    Needs to be as level as you can get it.

    Also needs to be vibration free.
     
  3. cypher007

    Joined:
    Apr 26, 2016
    Messages:
    62
    Likes Received:
    10
    ooh ooh just remembered ive got some granite slabs I sit my sub on with focule pods. I could level that up.
     
  4. With The Birds

    Joined:
    Jan 4, 2016
    Messages:
    2,577
    Likes Received:
    733
    Location:
    Australia
    Have you got a kitchen? The bench top should be pretty good.....

    Oh, and save the granite slab and pods for your turntable. Waste of time with the subs. Wavelengths are too long for intermod distirtion to be an issue and some LF coupled to the floor improves the experience.
     
  5. cypher007

    Joined:
    Apr 26, 2016
    Messages:
    62
    Likes Received:
    10
    my kitchen isn't that good. I use the slabs to try and stop the LF going through the floor to downstairs, missus complains.
     
  6. Pete Leare

    Joined:
    Aug 6, 2015
    Messages:
    334
    Likes Received:
    105
    Location:
    Windham NH
    Best thing I've found is a pool table if you've got one or know someone who does? Perfectly flat and level
     
  7. Mimoid

    Joined:
    Apr 18, 2016
    Messages:
    155
    Likes Received:
    82
    Even if you have a perfectly level surface, you will will find that the landing legs in the Phantom can be slightly off because of the uneven wear of the foam padding underneath the landing surface.

    The easiest method for balancing is to remove the propellers, place a glass pane (e.g. from a picture frame) directly on top of the prop shafts, and measure the top of this glass pane with a spirit level in both directions.

    You can then insert sheets of paper under the landing legs for a perfect horizontal alignment.

    Start the calibration and then try to avoid causing vibrations during the minute or two that it takes.

    Ta da!

    //Tom
     
    DroneX and cypher007 like this.
  8. alokbhargava

    Joined:
    Sep 28, 2015
    Messages:
    5,044
    Likes Received:
    1,623
    Location:
    San Francisco, CA
    Glass pane method could be the better way to level it but I never used that. I level it on my kitchen slab with 12 inch leveler. It has worked well so far.

    Your sequences should be:

    IMU calibration
    Gimbal calibration
    Gimbal roll adjustment for minor horizon tilts
     
  9. BudWalker

    Joined:
    Jun 24, 2015
    Messages:
    772
    Likes Received:
    225
    Location:
    Santa Margarita, Ca
    Can anybody explain the reasoning behind the claim that an IMU calibration has to be done while the P3 is absolutely level? I've seen this so many times that it seems that people believe it because so many other people believe it. But, what is the technical reason?

    It's often stated that the P3 won't fly straight if a slightly off level IMU calibration has been done. The reason being that in order to achieve straight flight in the X axis direction the P3 has to be perfectly horizontal in the Y axis direction. But, if the Y axis were held perfectly horizontal the P3 would wander and fly erratically. Besides not being perfectly level there other reasons the the P3 will wander, e.g. wind, gusts, or just ordinary randomness. In a sense the FC doesn't care what level is, it is trying to figure out what IMU sensor values it needs to see in order to achieve straight flight.

    Are there other reasons that an IMU calibration has to be done on a perfectly level surface?

    I think the main task of the IMU calibration is to make adjustments that minimize drift. I recently did an IMU calibration (on a close to level surface) where the before and after gyroZ drift was measured. There was 13 times reduction in the drift.
     
  10. Luap

    Joined:
    Jan 4, 2014
    Messages:
    430
    Likes Received:
    70
    Here is a thread with some further links that might help understand how the level surface reasoning came up.
    Finally fixed my Yaw Drift
     
  11. Ted4797

    Joined:
    Nov 23, 2015
    Messages:
    971
    Likes Received:
    154
    Location:
    Florida
    I use a table that looks level and have never had a problem. Maybe today I will put a level on it and see if it is really level.We nerds tend to overthink.
     
  12. peterepeat69

    Joined:
    Mar 18, 2016
    Messages:
    549
    Likes Received:
    110
    any table is level unless it's designed by
    Salvador Dali of course


    Sent from my iPad using PhantomPilots mobile app
     
  13. BudWalker

    Joined:
    Jun 24, 2015
    Messages:
    772
    Likes Received:
    225
    Location:
    Santa Margarita, Ca
    I looked into this. As I understand it the problem they were trying to fix is that when yawed about the Z axis the AC would move in a circle instead staying in one location. This could be fixed by doing an IMU calibration where the IMU module (not necessarily the AC itself) was perfectly level. Many people tried this and claimed that it worked; supporting the supposition that there is a real technical reason the fix works.

    I tried to recreate the problem with my P3 but wasn't able to.

    I performed a cold IMU calibration with the nose up at an angle of 1.75 degrees. The P3 was then tested by first hovering without yawing in ATTI to measure the effect of the small breeze here. This was followed by hovering in ATTI but with yawing added. I did this twice with the results

    Code:
    Test1. no Yawing, velH = 1.1.6 Meters/sec,
              Yawing, velH = 1.08 Meters/sec
    Test2. no yawing, velH = 1.09 Meters/sec,
               Yawing, velH = 1.21 Meters/sec
    
    I then did a cold IMU calibration with the right side elevated at 1.75 degrees followed by the testing I did for the first IMU calibration. The results were similar.

    I'm comfortable with the claim that, in these tests, there is no discernible effects from performing the IMU calibration at 1.75 degrees off level.

    I don't want to encourage others to IMU calibrate 1.75 degrees off level and then go fly their P3. If you do that's your choice. But, I'd be interested in results from any experiments like this or other experiments.

    Is the P2 enough different from the P3 that the outcome from this experiment would be different for a P2?

    I noticed that the P2 stuff is more than a year old. Is it possible that the newer firmware could be making the difference?
     
  14. Not A Speck Of Cereal

    Joined:
    Mar 8, 2016
    Messages:
    403
    Likes Received:
    144
    Still, it doesn't hurt to be level, even if it turns out to not be a major factor.

    But a word about some of you finding a flat place (like the kitchen table or hardwood floor) and assume that it's level: you really need to get a spirit level ... or at least a $3 line-level at the nearest hardware store. You will be surprised how many level looking surfaces are not really level.

    It may end up not mattering for IMU calibrations. It does matter for prop balancing.

    Chris
     
  15. kenjancef

    Joined:
    Jul 20, 2015
    Messages:
    1,202
    Likes Received:
    236
    Location:
    East Providence, RI
    Not sure if this is scientifically accurate... but I did an IMU calibration last night... plopped an old piece of MDF on top of my P3P, then used a bubble level app on my iPhone and sat that on top of the MDF. Was only off a hair at back right, so slid some paper under the foam pad on the landing gear and I was level... did a gimbal calibration as well because my horizon was WAY off when I took it out on Sunday afternoon. Hopefully that will do the trick.

    Not sure how accurate the iPhone level may be, but I guess I'll find out.. .lol...
     
  16. Not A Speck Of Cereal

    Joined:
    Mar 8, 2016
    Messages:
    403
    Likes Received:
    144
    NEW QUESTION:

    Yesterday, I was up in the air and noticed a skewed horizon. I mean, it was way off for some reason. So I did a quick gimbal roll adjustment and thought "I'll do a full IMU and gimbal calibration when I get home".

    Do you think it would be a good practice to zero out the gimbal roll adjustment before doing a fresh IMU calibration?

    Chris
     
  17. kenjancef

    Joined:
    Jul 20, 2015
    Messages:
    1,202
    Likes Received:
    236
    Location:
    East Providence, RI
    I probably would... not sure if it would calibrate based on the adjustments you did.

    This is how off mine was, took this on Sunday, clear calm day, no wind at all...

    DJI_0013.JPG
     
  18. Not A Speck Of Cereal

    Joined:
    Mar 8, 2016
    Messages:
    403
    Likes Received:
    144
    Wow, that's pretty severe.
     
  19. kenjancef

    Joined:
    Jul 20, 2015
    Messages:
    1,202
    Likes Received:
    236
    Location:
    East Providence, RI
    Ya think??? lol... :)

    And this was only the 3rd flight... hoping the calibration(s) worked...