Difference in Drift when using Phidget Control Panel vs. Python

[MikahB]

Have been working with a MOTO109 and now a couple MOTO110 9-dof IMUs. The documentation is pretty clear on the fact that there are two algorithm options: IMU and AHRS, and that in IMU mode the parameters (for AHRS) are ignored.

My confusion comes from the fact that when in IMU mode on Phidget Control Panel, I can watch drift (specifically heading drift) settle out once the sensors are up to temp, etc. It does drift, but it might drift 0.1 degrees in 30 minutes. It seems like tweaking AHRS parms, even with it in IMU mode, impacts the values I see in the Control Panel and lets me minimize drift.

In Python, though, I get almost 0.2 degrees heading drift in 5 minutes. It's possible that - because I can't chart the output from the heading value in the Phidget Control Panel - I'm just better able to quantify the drift with Python, but even just watching it run you can see more drift with Python than in the Control Panel.

Here's my startup routine in Python trying to minimize drift:

- Connect to Spatial and TemperatureSensor objects
- Enable heat
- Wait for sensor temp to equalize (at 45 or 50C depending on model)
- Reset gyro
- Reset algorithm
- Start reading

Should I be doing something different? Our environment is subject to occasionally fairly significant magnetic interference which makes anything using the magnetometer unpredictable, so IMU is our preference for now.

[fraser]

Regarding the IMU vs AHRS, the difference is that the Magnetometer is ignored for the calculations, however, non-magnetometer-based AHRS parameters can still be set to affect your device while in IMU.

It is possible that your set parameters are different while in python, which could affect the amount of drift you see. That's just my initial guess.

[morganC]

Hello, I have exactly the same problem with the same configuration. Can you please tell me if you solved the problem and if so how?
Thanks in advance

[AvaPickett]

Hello, I have exactly the same problem with the same configuration. Can you please tell me if you solved the problem and if so how? Geometry Dash
Thanks in advance

I had the same problem, let me know if you solved this!

[mparadis]

Should I be doing something different? Our environment is subject to occasionally fairly significant magnetic interference which makes anything using the magnetometer unpredictable, so IMU is our preference for now.

Hello, I have exactly the same problem with the same configuration. Can you please tell me if you solved the problem and if so how?
Thanks in advance

I had the same problem, let me know if you solved this!

I did some testing to try and see what could be causing this discrepancy. The first thing I checked was if there were any differences in how the Control Panel example handles things versus the python script described by MikahB. The only difference I found was that the default AHRS parameters have one small difference in the control panel versus the library default: Acceleration threshold is 0.05 in the control panel instead of the default value of 0.1. However, I did a number of test runs and found that there was no noticeable difference in heading drift when switching between those two values.

One thing I did notice, is that even after reaching a stable temperature, it still took 1-2 minutes for the heading to truly stabilize, after which point I always saw about 0.1 degree of drift over the remaining ten minutes. This makes sense- although the temperature sensor on the spatial is stable at 50C, the rest of the board and the inside of the enclosure may not be yet. It's possible the control panel appears to have better performance because it may have had more time to settle. My recommendation is to try waiting for two minutes after temperature stabilization before recording and see if that improves performance.