T3

Invensense gyros in the UAV DevBoard

UAV Devboard users,

The UAV Devboard, version 3, is now available from SparkFun. It is the same as version 2, except it replaces the LISY gyros with Invensense 500 degree gyros, which are practically vibration proof. Also, now you can use higher rotation rates in your flights without saturating the gyros. The previous gyros had a maximum range of 300 degrees per second.

The board has been flight tested by the team. All firmware for the UDB2 will work on the UDB3 without changing any parameters except the board type.

In MatrixPilot, you should select RED_GREEN_BOARD (even though the board is entirely red!) as the board type. If you are using any of the heli firmware, the board type depends on what version you are running. In the latest heli firmware, set the board type to INVENSENSE_BOARD. In either case, the configuration file that should be pulled is "IXZ500.h"

The "bigger and better" version of the UDB is still in the design pipeline, and is expected to be available in the spring of 2011.

Best regards,
Bill
E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones

Comments

  • Thanks Morli
  • I've ordered extra batteries. Have you managed to find higher power batteries that fit? or would that alter the cog of make it too heavy? Anyway thanks for the help I look forward to some UDB fun
  • Thanks Morli, I will order one of those
  • T3
    Carlito,

    It is not possible to buy only the gyros board to update the UDB2.

    In any case, it is very difficult to do, there is signficant risk of damaging the board while you try to remove the gyros that are there.

    Sorry.

    Bill
  • Morli,

    Is it this one

    Thanks
  • "We only briefly considered the ITG3200. In terms of drift and working range, the IXZ500 seemed a better fit, with a range of 500 degrees/second, which is adequate for most aircraft"

    Worth to take a closer look. ;-)

    I tried it with a BMA180 in an ARM based IMU application. First impressions are very positive. Low noise, low drift. Nearly impossible to turn the board by hand fast enough to loose position - while it's easy using the 500 degree/sec izx etc. line.
    (The BMA180 datasheet is simply terrible, detailed description of bits but no sign of the register address the bits belong too, nor the position of the bits inside the byte, billion details but not a clear description on typical use etc.)

    Jozsef
  • Bill,

    Thanks for the tips. Clever stuff, aside from waypoint accuracy, it sounds like the turns are also going to look great!

    Fab
  • T3
    Hi Fab,

    Regarding the stock Easy Star working out of the box with a stock UDB....

    There is one modification that I made to my Easy Star, I attached a playing card to the rudder.

    That said, the Easy Star is a very stable plane to begin with, the UDB makes it even more stable, so a stock Easy Star should work ok with a stock UDB.

    The only thing you might have to do is set the switches for the servo reversers on the UDB, because they will be in random positions when you receive the board.

    Also, as we speak, I am testing a MatrixPilot revision that will improve the smoothness of waypoint navigation. I have added a feedback term into the rudder deflection proportional to the amount of roll. It reduces the effective time constant of the steering, so you can increase the steering gain without inducing overshoot in the turns. So far, there is large improvement in the smoothness of the turns. I want to do more testing under various conditions, and plan to revise the firmware this weekend.

    Best regards,
    Bill
  • William,

    I've retracted my response, sorry was in a hurry and I shouldn't have responded with so much tosh!

    I should have just said that I will look into the things you mention and that the new UDB board seems great.

    Vibration seemed to be the only concern amongst UDB users and it is good to know that is resolved.

    When I recieved my first UDB I found the installation instructions were easy to follow, the Pic3 worked like a dream. I was "up & running" in no time at all.

    Can you please confirm that a stock Easy Star worked out of the box with a stock UDB?

    Well done to the UDB dev crew
  • T3
    Hi Fab,

    Regarding the sensors that we selected for the UDB, we deliberately chose only accelerometers, gyros, and GPS. We adopted a "minimalist" approach with the hardware to keep down the size, weight, and cost. We then took on the challenge of achieving as much performance as possible through clever algorithms.

    For example, MatrixPilot uses a combination of a wind estimation algorithm and dead-reckoning computations to produce a high-bandwidth estimate of airspeed, without the need for an airflow sensor. The wind estimation algorithm computes the wind by comparing changes in the ground velocity vector with changes in aircraft orientation. Dead-reckoning computes 3D ground velocity from accelerometer and gyro signals, with GPS providing drift compensation. Then, the air velocity vector is computed by subtracting wind velocity from ground velocity.

    We are quite satisfied with the performance of MatrixPilot firmware running on the UDB. Recently, both Rick Kuebler and myself achieved totally autonomous flights. Rick was using a FunJet with a top speed of 150 miles per hour. I flew both an EasyStar and a GentleLady. In the case of the EasyStar, I was using a stock UDB.

    Also, Rick has made flights with an AcroMaster in which he used the UDB to stabilize his plane in rather unusual orientations, including inverted flight. I suggest you take a look at some of the Blog posts listed on Rick's page. His accomplishments are impressive.

    Best regards,
    Bill
This reply was deleted.