This document serves as an estimate for the cost of upgrading the control system for the McGraw-Hill 1.3 meter telescope at Kitt Peak. The current control system resides on an old sun 2 workstation and is written in C. The system controls the telescope through an auto guider function only, and displays information about the location of the dome, and telescope position (RA and DEC). Given that the Hiltner 2.4 meter telescope uses the same encoder system, it is straight forward to adapt the 2.4's telescope control system (TCS) for use with the 1.3 meter telescope.
Several issues have to be dealt with in this adaptation. First, the C code on the sun must be ported from unix to OS/2 warp operating system on a DX466. This C code is needed in the new system because the telescope uses a slightly different mounting than the 2.4 meter telescope and the transformations from encoder counts to telescope position are different. The second issue that must be addressed is that the WWV clock is currently controlled from the sun-2 and transmitted around the site via ethernet. There are several possible solutions for this problem. For the time being, the old sun-2 will used for clock time until a more permanent solution is found. A third issue is that a back plane must be constructed which takes the encoder connections into the counter/timer cards of the new system, and guider and dome signals to the telescope. Lastly, the dome must be fitted with a control system that allows the TCS to communicate direction and speed.
As noted before all code created for the 2.4 meter telescope will be used for the 1.3 meter telescope. This means that code which controls the telescope's RA, DEC, etc. will remain in place but commented out. If in the future the telescope is retrofitted with hardware to allow it to be computer controlled, the software should be able to be adapted to use the new hardware by uncommenting the original code. Any modifications to the code to allow it to control the RA, DEC and any other new features can be handled in a new contract.
I anticipate work occurring in four stages. Stage one consists of ordering and collecting equipment needed in the form of a computer and counter boards for monitoring the various signals going to and coming from the telescope's encoders and motors. During this stage a backplane will be constructed which transfers the encoder signals to the computer's counter timer and D/A boards. Stage two consists of transferring the old source code's functionality to a C++ style program and expanding the capabilities of the old software. Stage two programming will involve adapting the 2.4 meter TCS to the new backplane configuration as well as design of the new display. Stage three consists of installing the full system and debugging any glitches that might arise during the installation. Stage four consists of a software support period. The following is a rough time table for when these stages would occur.
| Stage 1 Hardware Acquisition/Design | Feb-Mar '96 |
| Stage 2 Port of existing software | Mar-May '96 |
| Stage 3 Installation/Expanded functionality | Jun-Aug '96 |
| Stage 4 Software Support | Aug '96-Aug '97 |
The computer requires two types of hardware interface cards in order to perform the required functions: three counter/timer boards and one digital to analog converter board. These boards will control the telescope through: 7 or more Digital I/O lines, one 8 bit Digital to Analog line, and twelve 16 bit counters (two 16 bit, four counters, and two 32 bit, eight counters).
| Required: | Needs: |
4 Encoders |
|
| RA & DEC | 2 32 bit up/down counters |
| Dome, Rotator, Focus | 2 16 bit up/down counters 4 possibly 5 autoguider signals 5 TTL |
Dome Control |
|
| Fwd and reverse signals | 2 TTL |
| Analog Motor | 1 D/A (10V output) |
| 7 | TTL |
| 2 | 24 bit counters |
| 2 | 16 bit counters |
| 1 | D/A line |
Much of the software has already been developed during the construction of the 2.4 meter telescope's control system. The main development of software including porting of old software can be accomplished in Tempe. However, I have found that at least half of the development time will be spent at the telescope itself. Work that must be accomplished consists of porting the existing transformation and display code on the sun 2 to the DX66 machine.
| Graphical interaction/ display of telescope parameters. |
| DEC,RA,Dome, and Focus (focus when hardware permits). |
| Dome Control. |
| Auto Guider |
| Communication link through serial port. |
Software Development Cost 96hrs@50/hr |
4,800.00 |
It is estimated that it will take about a month or so to install, educate, and answer questions related to the operation of the system. Installation will occur during the day and should not interfere with normal operation of the telescope. Problems that were not anticipated during previous testing and design will be addressed during this time and corrected. If serious problems should arise in the abilities of the program, installation may be delayed until additional development is completed in Tempe.
After the installation is complete, time will be spent training various persons on the operation of the program, including calibration, diagnostics, and normal operation. Also, it may be necessary to educate someone about the design of the software.
Installation/Education 72hrs@50/hr |
3,600.00
|
If it is necessary, I will be available for further development and improvement of the system on an as needed basis. As with any program of this nature, it may require an ongoing commitment of?support through the first year of operation.
1 Year telephone support |
no cost |
||
On site help |
400.00/day + expenses |
This system can be implemented within a period of three to five months and can incorporate the functions and capabilities of the current system. All of the astronomy can be gleaned from the code that already exists on the old system and on the 2.4 TCS and transferred to the new. Some support will be required from a knowledgeable astronomer in the form of phone conversations, but should be very minimal. The existing hardware muscle box, encoders, and limit switches will be used allowing for the old system to be used as a backup in case of computer failures. All software will be written in well documented object oriented C++ style on a 486 type computer.
Total cost for implementing this system is as follows:
Software Estimate 72hrs@50/hr |
8,400.00 |
All of the estimates represent guesses at the amount of time it will take to complete the system. However, should it take longer to complete, no additional cost will be charged. The cost will remain the same given early completion. One half of the software costs, ($4200.00) must be paid up front with the remainder due upon delivery. Final delivery will be completed when all parties have agreed that the work has been completed.
Any questions concerning this document can be directed toward Robb Lovell by phone at (602) 965-9251, (602) 967-2616 or through e-mail at lovell@mythos.fa.asu.edu.
Back to Robb's Home Page