Is it normal that a 12 V DC motor changes its RPM depending on its orientation?

Question

I got a 40mm brushless DC fan from SHENGSHIDA which is rated at 12V and 0.15A. I supply it with 5V from a Raspberry Pi.

I noticed that the rotational speed of the rotor changes depending on the position to the ground. When the airflow is directed toward the ground, the RPM is noticeably lower than when the airflow is directed toward the sky. (In both cases, the airflow is not obstructed.)

Is this a normal behavior for this motor?

Answer 1

With - forgive me - crappy fans that use sleeve bearings and as little axial support as they can get away with, yes.

Answer 2

A 12 volt motor running on 5 volts does not have enough torque to overcome the changing friction of the sleeve bushings and thrust washers.

Answer 3

Running a 12V motor from Arduino in a terrible idea, as those pins should be used to power a small relay to turn on full 12V to the motor. With 5V, fan may not even start to spin and will never reach the RPM for which it's designed.

Answer 4

• When the fan is oriented downwards: the weight of the mobile mass is directed down, and as a result there is very little radial force on the bearings. There is little friction, so the motor is sufficiently capable to drive a high rotational speed.
• When the fan is oriented 90° from the vertical: the reaction forces in the bearings are directed radially and with the maximum magnitude of all configurations to hold the shaft whose weight attracts it both down AND tilted down. The friction torque is maximum, so the load torque on the motor is high and the motor, uncontrolled, suffers a loss in speed just like anybody who's dragging a heavy load up a slope.
• When the fan is oriented upwards, you should end up in pretty much the same configuration as the downwards case. The only difference lies in what, in the motor or the external bearings, is used to hold the shaft axially. In your case, the fan may be stopped axially by a shouldered bushing or the chassis itself in the downwards (intended) configuration, and only by the back of the motor shaft in the upwards (not intended) configuration. This would result in a bigger surface and radius of friction in the downwards configuration than in the upwards configuration - without assuming anything else on the materials involved -, resulting in a lower speed when oriented downwards. Consider tearing it down to find out.

Note: This ends up being more of a physics/mechanical design question than electronics though.

More details:

1. The shaft is held by the chassis using bearings. The friction torque between the shaft and the chassis at the location of the bearings increases with radial force - proportionally, when using plain bearings ("bushings"), as is done in cheap equipment.
2. The speed of uncontrolled brushed DC motor (as is used, again, in cheap equipment) varies linearly with the load torque - from the no load speed to 0 RPM at the max (temporary) torque capability. Both the no load speed and the torque capability vary linearly with voltage.