Introduction

1. Blanco TCS UpgradeDrive/Telescope Test ResultsNovember 2-7th, 2011M.WarnerE.MondacaR.CantaruttiG.Schumacher

2. Introduction • The approved modified Blanco TCS upgrade calls for the replacement of the old drive system, as described on the report dated Sept-29th-2011. (Ref 5) • The approach taken was first to model the behavior, of the existing telescope mount control, and then select a new amplifier capable of replacing the existing drive. • The existence of a control model provides a basic knowledge of the limitations, of the actual behavior of the real system. • To accomplish the task demanded by the upgrade, a series of tests were designed to acquire data needed to certify the fidelity of the model and the performance of the new amplifier. • This report presents the results obtained by executing those tests.

3. Goals Expected • Test the new current driver with the telescope DC motor load. • Instrument the new driver/motor/tachometer system, to measure motor torque commands, motor voltage, motor current, and tachometer outputs, at a high rate. • In addition we instrumented the capture of the new tape position encoders. This will allow to refine the system control model, before implementation, and greatly reduce integration time. • Measure mount non-linear features such as friction, backlash, and determine optimum preload current.

4. 1 - Motor and Tachometer Vendor Specifications: ftp://ftp.ctio.noao.edu/pub/warner/blanco/Blanco_motor&tach.pdf 2- Current Driver Data Sheet: ftp://ftp.ctio.noao.edu/pub/warner/blanco/az40a8.pdf 3 – Original 4m Telescope Servo Analysis: ftp://ftp.ctio.noao.edu/pub/warner/blanco/blanco_4m_analysis.pdf 4 – TCS CDR Telescope Lumped Mass Model: ftp://ftp.ctio.noao.edu/pub/warner/blanco/tcs_cdr_warner.pdf 5 - Blanco TCS Upgrade Project Report ftp://ftp.ctio.noao.edu/pub/warner/blanco/B4Upgrade.pptx References

5. General Telescope and Drive Specifications • The Dynamic Telescope requirements for both axes, needed to meet DECAM mission (2deg in 17sec): • Maximum Jerk = 0.033[deg/s^3] • Maximum Acceleration = 0.1[deg/s^2] • Maximum Velocity = 0.5[deg/s] • Motor and Gear Specifications (Ref: 1, & 3) • Torque Scale Factor = 4[lb-ft/A] • Inductance = 7.3[mH] • Resistance = 0.84[Ohms] • Gear Ratio = 144 • Current Drive and Power Supply Specifications (Ref: 2): • Nominal Current = 20[A] , 40[A]peak • Nominal Voltage = 28[V]

6. Test Plan Outline • Amplifier Tests: Current Step Response, with brakes on • Test Amplifier with actual telescope load impedance • Refine Amplifier SPICE Models • Preload Tests: Drive motors with shaped currents, in both directions until backlash is produced. • Determine optimum preload levels to avoid backlash. • Ramp Test: Ramp current until a preset velocity is reached, then remove current, and monitor deceleration due to friction. • Measure Telescope Inertia, Friction, and Imbalance Forces • Calculate Tachometer and Motor Vemf scale factors, using tape encoder as reference. • Sine Wave Tests: Drive telescope with a shaped sine wave current from 0.25 to 20Hz. • Measure telescope plant frequency response • Measure cross-coupling between axes

7. Current Driver Model/Tests • The current drive selected is an Advanced Motion Control AZ40A8, Analog Servo Drive, with 20[A] Nominal 40[A] Peak capability. • A simplified SPICE model of the current driver was assembled, to study the driver stability when loaded with the telescope motor impedance. • The driver model predicted a peaking at 70[Hz], with 25% overshoot in the step response, which has no practical implication in our application.

8. Current Driver Spice Model and Results Time Response, shows adequate performance for our application Frequency Response

9. Current Driver Measured Response withActual Motor Load (8[A] Step) The Step response matches the model, and validate the current driver, and motor impedance models. Selected driver suits the requirements needed.

10. Preload Determination Tests • The preload value is the torque needed to compensate for the motor/gear friction. • The optimum preload value was found by driving the telescope, with both motors in both directions, until gear backlash occurred, and then determining the current preload values. • These tests indicated that a 2[A] preload will prevent backlash, under all conditions for both axes. • The gear play was also determined from the tests, using a nominal bull-gear diameter of 3.622 [m]

11. Motor Preload The telescope is driven by two motors on different direction to avoid backlash Motor 2 Motor 1 or Motor 1 Motor 2

12. Gear Backlash To test the backlash effect we drive the motors in the same direction forward and backwards. Preload is applied Motor 1 Motor 2 Motor1 changes polarity releasing preload, causing backlash Motor 1 Motor 2 Both Motors pushing on the same direction Motor 1 Motor 2

13. RA Backlash Tests (3[A] Peak with 1[A] preload) Backlash causes jumps in position, this indicates a gear play ~0.94[mm] Backlash appears as spikes in the measured velocity

14. DEC Backlash Tests (1.5[A] Peak with 0.5[A] preload) Backlash causes jumps in position, this indicates a gear play ~0.63[mm] Backlash appears as spikes in the measured velocity

15. Current Ramp Tests • The current ramp tests was used to determine the telescope inertia, peak friction, and imbalance as shown, by using the peak acceleration, and deceleration values. • This kind of test will be utilized in the future to support balancing of the telescope. • The Current Ramp Tests were performed by driving one motor, with a current ramp to a maximum of 8[A], with a constant opposing 2[A] preload, on the other motor. When the tachometer indicated that a maximum velocity of ~0.5[deg/sec] was reached, the drive current was lowered to 2[A] preload, causing the mount to coast to a complete stop.

16. Ramp Tests Fundamental Equations of Motion

17. Inertial Moment Determination Applying motor torque in the same direction we obtained or After an adequate velocity was reached, we released the motor reducing it’s current to zero, so we obtain: or Because, immediately after releasing the drive, the position and speed of the telescope are the same, assuming that friction and position are also the same. Eliminating the friction and imbalance Torque between the two measurements, we obtain finally:

18. Friction and Imbalance Torque Determination Knowing the inertial moment J, and considering that the position and velocity of the movement is almost the same, we could consider friction torque and unbalance torque almost equal when the motors are released on both movements directions. and So, we will obtain and Then and

19. Motor Constant and Gear Ratio The motor torque is obtained from the current measurement using the motor constant provided by the manufacturer and the gear ratio of the telescope. (Ref 1)

20. RA Ramp Test Motor18[A] in 2[s] Ramp, with 2[A] Preload Motor ON Motor OFF Peak Acceleration (α₁) Peak Deceleration (α₂)

21. RA Ramp Test Motor2 8[A] in 2[s] Ramp, with 2[A] Preload Motor ON Motor OFF

22. DEC Ramp Test Motor1 8[A] in 10[s] Ramp, with 2[A] Preload Motor ON Motor OFF DEC1 Tachometer oscillation (described later)

23. DEC Ramp Test Motor2 8[A] in 10[s] Ramp, with 2[A] Preload Motor ON Motor OFF

24. Ramp Tests Experimental Results

25. Ramp Tests Summary • RA Telescope Inertia = 3.34e6[lb-ft-s^2] • Baseline =3.5e6[lb-ft-s^2] , from telescope design (Ref 3) • Torque for 0.1[deg/s^2] = 40.48[lb-ft], 10.12[A] • Total current assuming worst case Imbalance = 0.5*Friction, and 2[A] preload: 10.12+1.5*3.68+2 = 17.64 [A] • DEC Telescope Inertia = 1.05e6[lb-ft-s^2] • Baseline=1.23e6[lb-ft-s^2], from telescope design (Ref 3) • Torque for 0.1[deg/s^2] = 12.72 [ft-lb], 3.18[A] • Total current assuming worst case Imbalance = 0.5*Friction, and 2[A] preload: 3.18+1.5*3.51+2 = 10.44[A] • These tests indicate that the new current drives, with 20[A] nominal, have ample margin.

26. Tachometer Performance • Tachometer performance measurements are needed in order to establish the suitability of the devices, to allow a smooth tracking and offset motions behavior. • Scale Factor from Vendor Specifications (Ref 1) • Tachometer = 32.67[V/deg/s] +/- 10% • Motor Back EMF Voltage (Vbemf)= 12.56[V/deg/s] +/- 10% • Measured Tachometer and Vemf Scale Factors: • RA Motor1: Tac = 27.88 [V/deg/s] ; Vbemf = 14.27 [V/deg/s] • RA Motor2 : Tac = 27.86 [V/deg/s] ; Vbemf = 13.84 [V/deg/s] • DEC Motor1: Tac = 22.9 [V/deg/s] ; Vbemf = 14.72 [V/deg/s] • DEC Motor2: Tac = 15.4 [V/deg/s] ; Vbemf = 14.65 [V/deg/s] • Back EMF Voltage is a parameter needed to select the power supply voltage.

27. Tachometer Problems • DEC Motor1 Tachometer has large periodic error at a frequency of 36[deg] on the motor, corresponding to the brush spacing. This causes a 2[Hz] oscillation at 0.5[deg/sec] on the telescope, rendering it useless for motion control. • RA Motor2 Tachometer has large noise due to poor contact because it is coated with oil. This causes loss of servo control. • In addition the measured gains are below the vendor specifications.

28. DEC Motor1 Tachometer Periodic Error Problem Old Drive, goes into a 2[Hz] limit cycle during slews, as measured. DEC1 Tachometer Periodic Error increases at higher velocities, to 0.05[deg/s] (10%) at 0.5 [deg/s], this causes servo loop to oscillate.

29. RA Motor2 Tachometer Noise Problem Oil bead on brush RA2 Tachometer noise increases at higher velocities. This causes a loss of feedback during slews, triggering the brakes, in old drives. Noise could be reduced by periodic cleaning.

30. Telescope Plant Sine Wave Response • These tests are needed to establish the final control model to be implemented. • The results shown compare favorably with the lumped mass model (Ref 4). And also help in understanding the response of the new tape encoders. • A shaped sine wave sweep current command was used to excite the telescope, and the response of both tachometers and tape encoder was recorded, for RA a 0.2-10[Hz], and for DEC a 0.4-20[Hz] sweep was used. • The cross-coupling between tape encoders at each axis was also measured, for completeness.

31. RA Torque to Velocity 0.2-10[Hz] Sine Wave Response, 8[A] peak, 2[A] preload • Tachometer Response • from Lumped Mass • Model Ref(4) • Response is scaled in • [deg/s/A]

32. DEC Torque to Velocity 0.4-20[Hz] Sine Wave Response, 4[A] peak, 2[A] preload • Tachometer Response • from Lumped Mass • Model Ref(4) • Response is scaled in • [deg/s/A]

33. Cross Coupling between RA and DEC axis measured by tape encoders. When RA is driven, DEC shows a cross-coupling response peak at ~2[Hz]. (DEC brakes on) When DEC is driven, RA shows almost no cross-coupling response. (RA brakes on)

34. Test indicated that the new current driver performance, and design margins are good, for operating the telescope with increased inertia expected for DECAM. • The optimum preload levels were determined, and backlash measurements indicated normal gear wear. • The Telescope inertia, was close to the baseline values, and measured friction is acceptable. The next step is to be fully characterize friction at all telescope angles. • Tachometer performance shows several deficiencies, resulting in the need to upgrade to new units. • The frequency sine wave tests indicated a good agreement between the actual measurements, and the lumped mass models, this will lead to establishing the final control architecture to be implemented. Conclusions

35. Further Steps • Selection of new tachometers, will be done once we understand the mechanical interfaces needed for their installation. • The velocity loop will be programmed based on the results exposed in this report. • With the velocity control closed, will refine the characterization of the telescope dynamic behavior. • Finalizing the previous tasks will allow the implementation of the position loop, leading to a complete control of the telescope, with the new system.