Servo Improvements

Introduction

The GBT servo system has served the GBT well throughout its early life. The currently implemented system was designed by the contractor to meet specifications NRAO provided that did not include the needs of the current GBT operating modes. The parameters of the telescope servo system were optimized for stepping and tracking sidereal sources, and for that it works well. However, our needs do not match the specifications in several ways:

OTF mapping is not very well supported
The servo includes about 20% overshoot on position changes, which makes short-duration position switching (i.e. Nodding) suboptimal. Too much time is wasted on getting and staying on-source. As such, this severely limits VLBI phase-referencing observations, particularly at high frequency where the added sensitivity of the GBT is crucial. See the attached plots below for an example from a recent VLBI run (ABG169_03 from 2007/02/01) that used a 96-second switching period, where 20% of the time is spent settling to within 4 arcsec in azimuth (vs. only 8% in elevation) in addition to 39% slewing. In this case, the switching angle was 1 degree in elev and 2 degrees in azim.
The servo gains are too low to compensate for disturbances due to wind or track effects. Increasing gain in the existing classical lead-lag system would lead to oscillations or pumping of structural modes, since increased gain also increases the bandwidth of the loops nearer to the resonant frequencies of the antenna.

Project Goals

The goals of the this project are to improve the GBT servo to support high frequency observations. Specifically, the long-term servo replacement should allow better control in windy conditions, and the short-term improvements should improve pointing in benign conditions. The goal of the PTCS is less than 1.2 arcsec of tracking error for good performance at 86 GHz. THis tracking error must be allocated among several sources of error, so the error budget for the servo is less than this, perhaps 0.5 arcsec. (N.B. I think this was in Kim's URSI presentation)

Project Charter

These issues have led us into attacking the servo system from a couple of fronts, short-term and long term. The first items are short-term improvements that work with the current system, and can be completed, with the proper staffing, by the end of the calendar year. The last item is a long-term (> 1 year from now) item.

Short-term improvements
- Servo command generation. The antenna manager sends time-tagged commands to the servo system. These commands are generated by a trajectory generation system which is supposed to minimize the jerk and acceleration imposed on the antenna. It is not clear that the trajectory generator and its interface with the servo correctly accomplish this function. During some tests last summer, it was found that no matter what trajectory was generated, the system reacted somewhat more violently than we thought it should. This is probably due to the use of 100% velocity feedforward, which is necessary for zero following error. More work needs to be done here. The effect of this defect is to add oscillation to the feed arm, and to make the system take longer to settle out. This is the Trajectory Improvement project.
- Explore adding Acceleration feedforward to the servo. The servo cannot follow a non-constant acceleration, due to the architecture of the system. Acceleration feedforward will allow the servo to follow this type command which is used with daisy scans and other non-linear trajectories.
- Tracking at low azimuth rates. This problem seems to be due to the friction and other nonlinear processes imposed on the servo when the antenna commands cross zero. A simple friction compensator has been designed and is being implemented. Initial tests look promising.
- Servo Resonance Elimination. There seem to be features in the servo errors that are not related to the structural resonances of the GBT, but rather are generated by the servo somehow. This is the Servo Resonance Suppression project.
Long-term improvements
- Servo maintainability and reliability. The servo has failure modes that are difficult to diagnose. We desire a system that is more friendly to the maintainers. This is the PLC Servo Interlock project.
- Advanced servo architecture. A number of large radio telescopes (LMT, 34m DSN, etc) have proposed advanced servo architectures. These are generally model-based systems incorporating Linear-Quadratic-Gaussian (LQG) controllers in either the position loops or the velocity loops, or both. Our existing servo system is based on an analog lag compensated rate loop and a lead-lag compensated position loop, and this architecture offers limited opportunity for improvements. We intend to replace the existing servo with a new servo system.

Project Personnel

Project Manager: JohnFord
Project Scientist: ToddHunter
Project Engineers: TimWeadon, JoeBrandt, MelindaMello
Project Team: controls consultant, new engineers

Project Charter Approval

Approval of the charter means that the program is authorized to proceed, and commit funds and labor hours to the project.

Approved by PTCS Servo Team Members
- John Ford
- Todd Hunter
- Tim Weadon
- Joe Brandt

Approved by GBT Program Management
- Karen O'Neil
- Richard Prestage

Symbols:

Use %X% not approve (will display )
Use %Y% approve (will display )

Project Scope

Scope of the project refers to the activities that will be pursued.

In Scope

The scope is defined by the Work Breakdown Structure and the project activities. These project activities are documented separately, and are referenced below.

Out of Scope

Several things that will specifically not be included in the scope of work for the servo improvements are:

Feedarm motion control
RF beam position monitoring

Scope Management Plan

Changes in mission will be accepted only by the project manager in consultation with GBT Program Management (Currently K. O'Neil and R. Prestage), along with the project's team members. Increases in scope will not be accepted unless a clear source of additional resources to handle the increased scope is provided with the request.

Work Breakdown Structure (WBS)

The following will be provided by this program:

Improved tracking of daisy scans
Elimination of servo resonances
New servo algorithms for control of vibration and wind disturbances, as well as improved tracking ability
New servo platform capable of deployment of the advanced control algorithms
Reliable, maintainable servo system electronics

Project Activities

Project Schedule

The servo interlock upgrade and the friction compensator are currently being pursued. These other activities are on hold pending availability of additional effort, which is currently allocated to the pointing model and holography. We will start them approximately as follows:

Trajectory Improvement
- 9/01/2007
friction compensator
- 12/31/2007 (finish date)
Acceleration feedforward
- 11/15/2007
Servo Resonance Suppression
- 9/01/2007
New servo system
- 1/01/2008
PLC Servo Interlock project
- 06/30/2008 (finish date)

Project Status Meeting Notes

Estimated Costs

Except for the New servo system, nearly all the costs are in personnel. The new servo system will require about $50k of consulting, and $50k of hardware costs. If it is decided that the 22-bit azimuth and elevation encoders should be replaced with better performing devices, then additional costs for hardware and software interface modifications will apply.

Program Plan Approvals

Approved by PTCS Team Management
- John Ford
- Todd Hunter
Approved by GBT Program Management
- Karen O'Neil
- Richard Prestage

Current (January through March) Request for effort allocations

Management
- J. Ford 0.05 FTE
Engineering
- J. Ford 0.05 FTE
- J. Brandt 0.8 FTE
- J. Shelton 0.2 FTE
- T. Weadon 0.8 FTE
- K. Sexton 0.8 FTE
- M. Mello 0.4 FTE (back loaded)
- P. Whiteis 0.4 FTE

-- JohnFord - 19 Jan 2007

Attachment:	Action:	Size:	Date:	Who:	Comment:
azvstime.pdf	action	56958	08 Feb 2007 - 22:57	ToddHunter	VLBI fast switching: Azimuth settling time = 20%
elvstime.pdf	action	63774	08 Feb 2007 - 22:57	ToddHunter	VLBI fast switching: Elevation settling time = 8%
servo_update.ppt	action	3398656	05 Mar 2007 - 19:50	ToddHunter	powerpoint presentation for 03/06/07 PTCS meeting
PTCSMeetingnotes080826.pdf	action	19695	28 Aug 2008 - 21:01	MartyBloss	PTCS Meeting 26-Aug-08

Topic ServoImprovements . { Edit | Attach | Ref-By | Printable | Diffs | r1.21 | > | r1.20 | > | r1.19 | More }

Revision r1.21 - 28 Aug 2008 - 21:03 GMT - MartyBloss Parents: WebHome > InformationForProjectTeam > PTCSProjectPlan	Content copyright © 1999-2007 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.


NRAO Home > Green Bank \| Wiki Topic: GB > PTCS > InformationForProjectTeam > PTCSProjectPlan > ServoImprovements

Changes \| Index \| Search \| Statistics \| Go