Preliminary Tests

The torque transducer incorporated in the drive shaft connecting the motor with the target wheel records both the torque applied to the target wheel from the motor, and the speed of rotation of the target wheel. Altogether, there are four channels returning data from the torque transducer: Torque 1 (returning torque values at a rate of 2400 Hz); Torque 2 (torque values at a rate of 600 Hz); Angle; and Speed. Data can be recorded over a maximum time period of 180 s. During the initial test runs, the magnet was kept turned off.

With the motor turned off and the wheel stationary (and with zero current in the magnet coils), the torque transducer still returns non-zero values for the torque. These values indicate the noise level, which is of order 1 Nm.

Fig.5.10shows the results of a typical test run, with the torque and speed recorded over a period of 60 s. Initially, the motor controller was set to drive the wheel at a nominal speed of 33 rpm. After about 28 s, the motor was switched off, so the wheel slowed to a stop; after a further few seconds, the motor was switched on again, and the wheel was driven at a nominal rate of 15 rpm. We see that while the speed of the wheel was set at 33 rpm, the torque fluctuated between +4 Nm and -4 Nm. Also, the speed of 33 rpm was not actually achieved: instead, the speed fluctuated between about 29 rpm and 30 rpm. The reasons for this are not clear. At the moment the motor was switched off, there was a peak in the torque of -8 Nm, from the inertia of the wheel driving the motor. While the wheel was stationary, there was a small level of noise on the torque data. There was a peak (of about 4 Nm) in the torque as the motor was re-started; then there was again a fluctutation in the torque data, but between lower limits than before, corresponding to the lower speed of rotation of the wheel.

From the speed and torque data returned by the torque transducer, it is easy to calculate the power required to keep the wheel in motion. For example, in the test run shown in Fig. 5.10, with the speed of the wheel at a (nearly) steady value of about 30 rpm, the maximum torque value was about 4 Nm: therefore, the maximum power needed to keep the wheel in motion was about 12.6 W.

Note that when changing the speed of the motor, the rate of acceleration or decel- eration can be specified. Although an “instantaneous” stop command can be issued, in

N m ) To rq u e (N Time(s) 0 10 20 30 40 50 60 0 5 10 15 20 25 30 TimeHsL Speed H rpm L

Figure 5.10: Data from torque transducer channel Torque 1 (top) showing the target

wheel rotating with nominal speed set at 33 rpm; then stopping for a few seconds; and then finally restarting and accelerating to a speed of 15 rpm. The bottom plot shows simultaneous data from the torque transducer Speed channel.

e (N m ) To rq u e Time(s) p m ) ed ( rp S p e Times (s)

Figure 5.11: Data from torque transducer showing the target wheel accelerating from

rest to a speed of 198 rpm; maintaining this speed for about 20 seconds; and then finally decelerating to rest. The top plot shows the torque; the bottom plot shows the speed.

practice this would be dangerous if the wheel were rotating at high speed, with possible damage being caused to the torque transducer or the motor.

The low-speed tests showed that the system behaved qualitatively as expected; however, we found that the rotation speed of the wheel did not necessarily match the speed set for the motor controller. Some regular oscillations in the speed were visible in the measurements from the torque transducer. Also, the torque values returned by the torque transducer showed large, rapid fluctuations, which needed to be averaged out to obtain meaningful values to compare with theoretical predictions.

Further tests were carried out with the wheel rotating at a higher speed. Fig.5.11

shows results from a test in which the wheel was accelerated to 198 rpm, maintained at that speed for about 20 seconds, then decelerated to a stop. During the early stage of the acceleration, the torque reached a maximum value of about 8 Nm. However, as the

wheel reached a speed of 174 rpm, there was a peak in the torque (which reached more than 20 Nm), before the torque returned to a similar level as during the early part of the acceleration. Similar behaviour was observed during deceleration, with a similar peak in the torque at the same speed of 174 rpm.

Note that although there are still clearly large fluctuations in the torque readings, the average value is clearly positive during acceleration and steady rotation, and neg- ative during deceleration.

In order to confirm the large value reached by the torque at 174 rpm, the motor controller was programmed to accelerate the wheel to this speed, maintain the same speed for a period of about 18 seconds, then decelerate the wheel to a stop. The torque data collected during this run are shown in Fig.5.12. We see that the torque reaches an even higher value than before, about 30 Nm. This is about four times larger than the usual value reached during acceleration up to 198 rpm. Although the reasons for the large increase in the torque at the particular speed of 174 Nm are not fully understood, the increase is reproducible, and likely to be the result of a mechanical resonance. A more detailed model of the system, including many aspects of the motor and drive shaft as well as the wheel itself, would be needed in order to understand the resonance properly. However, this would be very difficult to do because of the complexity (for example) of the motor, and predictions of such resonances from a mechanical model would likely not be very reliable. This issue is of some potential concern for the ILC positron production target, since if a strong resonance occurred close to the nominal operating speed of 2000 rpm, it is possible that some damage could occur to the system. In the case of the experiments described here, it was important to monitor the torque data closely, and avoid operating for any length of time at speeds where resonances were observed.