IDG-500 vs ADXRS613

[Tom]

One of the main reasons the gluonpilot carries an ADXRS613 gyro for yaw, is because it's most widely used and claimed as one of the best in the MEMs gyro world. Of course it's also one of the most expensive ones

Anyhow, I'm currently disappointed by it's noise performance (see picture - y-axis = °/sec)

When adding the gyro's raw output on my scope, I clearly see it oscillating around it's resonance frequency (14.5kHz).

Does anybody have similar experience?

[Tom]

It appears to behave as white noise when integrating these rotation rates, so no problem:

For your reference, I included the results when integrating:
Note the pink "pitch acc" spikes that happen when roll is near 90 degrees: numerical instability! The integrated angles that do not come from the accelerometer don't suffer from this (note that I used quaternions here). Also note that no (kalman, complementairy) filtering has yet been applied! I'd say those gyros and quaternions handled my rolling of the unit pretty well

gluonangles.png (20.35 KiB) Viewed 15794 times

roll = integrated gyro values (and that's it!)
roll acc = uses only the accelerometer

[asifjahmed]

Tom -

Impressive results so far with your gluonpilot IMU!

Especially since that is currently without filtering!

Any word on the batch of gluons?

Thanks,

-Jamie

[Tom]

Hi Jamie,

The production prototype works perfectly, so I told my assembler to go on and assemble the batch. On Wednesday his company will open again from Christmas vacation, so I expect to have the modules a week later. Then I will test them and ship them to every one of you who bought one Which means I will probably ship them around mid January. It will depend on the mailman when you get them. I suppose a about a week to the US.

Tom

[estevanoveja]

Hey Tom!

This is my first message on this forum so I will start saying "Congratulation"... things in here are pretty amazing.

I'm currently developping an autopilot for a hexacopter (yes, I know everyone is doing quadricopters, but I'm french so I have to do it... different ).

I was wondering about the MEMS sensors and especially the gyros : which one to choose?

I've been crawling the web for a while, and I've seen a lot of different opinions.

What where the key features that motivated your choice for the IDG-500 and the ADXRS613 ?

Thanks in advance for your answer.

Stephane.

[Tom]

Hi Stephane,

The choice for the IDG-500 is obvious: roll and pitch in 1 chip! The sensitivity and drift are not too bad either. The max roll rate should be at least 300°/s for my use in small UAV's. The 500°/s in the IDG500 is still OK.

Now for the ADXRS613: The yaw rate in planes is a lot slower than the pitch and roll rates. Also, we don't use (for the moment) any magnetometers to compensate the drift. This means I needed a better gyroscope with more sensitivity for yaw. As you might know, the ADXRS613 is more expensive than the IDG-500 but should be more stable as well, and less affected by noise.

[estevanoveja]

Now for the ADXRS613: The yaw rate in planes is a lot slower than the pitch and roll rates. Also, we don't use (for the moment) any magnetometers to compensate the drift. This means I needed a better gyroscope with more sensitivity for yaw. As you might know, the ADXRS613 is more expensive than the IDG-500 but should be more stable as well, and less affected by noise.

I see your point, but compared to a GPS aided solution to compensate the yaw drift, does it really worth the extra $$$ ?

[Tom]

Oh but you're missing the point here:
The yaw gyro is needed for stabilization! GPS heading is way too dependend on the wind... so doesn't work good for stabilization owhen there is wind.
Image you have a roll angle of 30 degrees, and you change the yaw around the z-axis for 30 degrees. What happened? Yes, you pitched up as well! So you see the yaw-rate from the gyro is very important for stabilization, especially for the pitch angle.

[estevanoveja]

I guess there is a misunderstanding.

Of course, the yaw gyro is needed to stabilize the plane !

I was just wondering what are the benefits of using an expensive ADXRS613 rather than a more casual yaw gyro (let's say with equivalent specs than the IDG-500 but on the yaw axis) coupled with a GPS to compensate the drift. I know that an ADXRS is drifting more or less 10 times slower than a IDG-500 equivalent, but does it worth the price ?

Anyway, a magnetometer based solution to fix the yaw drift would just be great (actually, it's the only missing part to turn the gluonpilot into the perfect board for quadri/hexa/whatever-copters or anything that flies, floats, rides...) and it wouldn't increase the price that much. The cool thing is that parts like the HMC5843 are working on I2C, so depending on the application, you can choose to solder it or not without needing several versions of the board... but i guess i'm quite off-topic right know.

[Tom]

When I bought the components the ADXRS613 was only slightly more expensive than the IDG-500. Now prices dropped again. If I'd remove the ADXRS613 it would be more because it's 5V, and this annoys me. Also, the cheaper gyros behave really bad when there are vibrations, which is bad for quadrocopters.

The HMC5843 is not on the board because
- I'm afraid there would be too much magnetical noise on the board.
- There is an I2C port available especially for the HMC5843/2, so you'll need to buy a breakout board and place the sensor on a good place.
- A plane doesn't need it, and those sensors are not cheap.

I don't have a lot of experience yet with magnetometers... but maybe you have?