Introducing the UDB6mini

169 views
Skip to first unread message

ben levitt

unread,
Apr 24, 2020, 7:17:50 PM4/24/20
to uavdevboard
Hi uavdevboard folks,

First of all, I hope you're all staying healthy and sane during this crazy time.

A few months, I was preparing to produce another batch of UDB5mini boards, and realized that the MPU-6000 had been discontinued.  So after researching the available replacements, I prototyped a few boards with the ICM-20689.  I bench tested and flight tested them, and sent one along to Bill to play with as well.  It works great, and has the advantage of having really low gyro drift!  I can imagine this could open up some fun new possibilities, like more accurate automatic takeoff without a magnetometer.

So I produced a batch of this new board, which I'm calling the UDB6mini.  The code to use the new sensor is already in git's master branch.
And you can leave the BOARD_TYPE set to UDB5, since the new chip is auto-detected at run-time.

If you're interested in trying out the new boards yourself, you can order online at my little web store, here.  This shop, Octopilot Electronics, just sells my UDB5mini and UDB6mini boards.  It's just me, and isn't meant to turn a profit.  For legal reasons, I set it up as an LLC, and the prices I charge are meant to just barely cover production, my prototyping costs, and my legal costs to set up and run the store.  I do this for myself and for this community, so that we can all continue to have fun with MatrixPilot, and everything else that we build based on it.  And if I keep selling enough boards, I can keep running the store.  :)

I wish you all health, sanity, and great flying!

Best wishes from Mountain View, California,
Ben

Roby

unread,
Apr 25, 2020, 11:24:28 AM4/25/20
to uavdevboard
Hi Ben,

Thank you for your supporting the uavdevboard community keeping it alive. Your new product is amazing. The bias drifts are very low compared to the MPU6000.
Hoping to put my hands on one in the near future!!

Best Wishes,
Roby

Jack Robinson

unread,
Sep 26, 2020, 9:31:00 AM9/26/20
to uavdevboard
Hello Ben 

The ICM-20689  datasheet telling that the maximum output rate is 32KHz, however, 8kHz  for MPU6000 can you use this advantage(32kHz) in raw data reading from IMU? 
and what about noise sensitivity for the newest IMU they are used if the outputs rate is increased? knowing that there are IMUs other than  ICM-20689  that has better noise parameters like  ICM-42688   
and what about programming UDB5 mini from the master MatrixPilot version so what option can I edit to use MPU6000 in the master version?  


thank you,
jack 



 
 


ben levitt

unread,
Sep 26, 2020, 4:49:48 PM9/26/20
to uavdevboard
Hi Jack,

We could definitely make use of the faster refresh rate, and it would be interesting to see whether that helps with accuracy during fast rotations / maneuvers.  I haven't tried that yet, but would welcome your experiments if you want to try before I get to it!

As for the setup in options.h: You should leave the BOARD_TYPE set to UDB5 for the UDB5, UDB5mini, and UDB6mini.  The sensor type is auto-detected and set up accordingly.

Best wishes,
Ben

--
--
---
You received this message because you are subscribed to the Google Groups "uavdevboard" group.
To unsubscribe from this group and stop receiving emails from it, send an email to uavdevboard...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/uavdevboard/a3e922ae-ff06-4627-8758-7eace79c6d9bn%40googlegroups.com.

Jack Robinson

unread,
Sep 27, 2020, 4:40:31 AM9/27/20
to uavdevboard
Hello Ben 

It's a pleasure to makes my experiments under your advice and your interest (: 

refer to your experiences what would be the procedures to start with these types of experiments?  
I'm hoping to test my own RC plane in fast and sharp maneuvering and rotations and I wish good results. 

Thank you, 
Jack 

ben levitt

unread,
Oct 13, 2020, 12:53:06 PM10/13/20
to uavdevboard
Hi Jack,

To look into increasing the sensor read rate, you could start by searching for "SAMPLE RATE" in libUDB/mpu6000.c.  (This file is used to control the ICM20689 as well.)

After changing that, you could log the raw output and check if noise looks any different at higher sample rates.  If the raw data looks ok, we could raise the DCM update rate as well.

Ben


Reply all
Reply to author
Forward
0 new messages