PDA

View Full Version : COTS gyros for 2011


Jon236
01-16-2011, 08:23 PM
The KOP supplied gyro chip cannot be purchased if you need replacements; something which makes me nervous. I would prefer to ensure that once we have developed code involving a particular piece of hardware, that we could replace that piece if needed.

I see a number of gyros available at SparkFun. I would appreciate comments and recommendations from the community.

Mark McLeod
01-16-2011, 08:32 PM
I'd use one of the 300 or 500 degree models.
You want to use one that's as close to, but higher than, your maximum expected robot turn rate.
The lower rotational ones have finer precision, but overload quickly when used on an FRC robot.
The extreme higher ones, e.g., 1500 degrees per second, don't give you the resolution you'll need to control driving accurately.

The Sparkfun models are mostly analog, and that's easy to integrate into our way of doing things.
Be sure to purchase an evaluation or breakout board, NOT the chip by itself.

You probably only need a 1-axis model. For driving you'll be measuring yaw, but that's determined by the mounting orientation.

Michael Simmond
01-22-2011, 10:54 PM
I agree that designing around an obsolete chip on an obsolete PCB is not very good engineering.

We purchased two models of gyro from SparkFun: ADXRS613 and LISY530AL.

The ADXR output is ratiometric to the supply voltage, so you could carefully read the supply voltage on a second analog input to improve accuracy. It also has a temperature output that can be useful as a linear correction for thermal drift of the signal zero. These two features add more complication, but they reduce drift sources. It makes for an elegant solution.

The LISY uses 3.3V power, so you will need a regulator. It gives an absolute voltage output (not ratiometric to the supply).

As far as I can tell, the cRIO analog inputs are fixed at 0-10V, so for best voltage resolution you should consider using a rail-to-rail output OpAmp running from 12V to add some additional gain to either gyro.

There was some concern voiced about having at least 300 deg/sec range. But in testing the ADXR (with 150 deg/sec range) we found that nothing bad happend when we saturated the output with rapid rotation. We just straighened the robot out with the joystick and kept on truckin.

We need the gyro feedback signal to reduce unwanted rotation when trying to drive straight. Otherwise the holonomic wheel tends to make the frame orientation wander substantially.

We have not even tried the LISY yet: The ADXR is doing everything we need.

BugByte
01-24-2011, 07:29 AM
I hadn't thought of that problem but I'll bring it up to my team today. Especially after we burned out our gyro at a competition last year... Thanks for the reminder.

Mark McLeod
01-24-2011, 09:24 AM
in testing the ADXR (with 150 deg/sec range) we found that nothing bad happend when we saturated the output with rapid rotation. We just straighened the robot out with the joystick and kept on truckin.
It depends on what you've chosen to use the gyro for.
As a simple assist in driving straight you can use the momentary values coming from the gyro and ignore extreme values generated by it on turns or during collisions.

Using it to command turns in autonomous or with field centric driver controls won't work if you saturate the gyro. It won't know where it is.
Robot-to-robot hits can over saturated the gyro, as well as, quick turns.

Michael Simmond
02-01-2011, 09:51 AM
Good point Mark.
We were only planning to do vehicle-centric drive at first, but field-centric looks like it will be easier for the drivers to deal with. Too bad SparkFun does not offer the 250 deg/sec version of the Analog Devices chip.

Ether
02-01-2011, 10:07 AM
As far as I can tell, the cRIO analog inputs are fixed at 0-10V

I am told (http://www.chiefdelphi.com/forums/showpost.php?p=1011552&postcount=2) they are -10V to +10V.