RCA Computer
| Revision as of 07:26, 18 March 2007 (edit) Paul (Talk | contribs) ← Previous diff |
Current revision (07:28, 18 March 2007) (edit) Paul (Talk | contribs) |
||
| Line 38: | Line 38: | ||
| [[Image:RCA-4101C.jpg|right|thumbnail|230px|RCA-4101C processor organisation: Image - RCA manual]] | [[Image:RCA-4101C.jpg|right|thumbnail|230px|RCA-4101C processor organisation: Image - RCA manual]] | ||
| - | Early in 1970 the FPQ-6 computer was further upgraded by replacing it with the RCA FC-4101C with integrated-circuitry logic and 16K of core memory expandable to 32K. It had 32 data bits per word with 2-bit parity; a memory recycle time of 2 microseconds; and a ‘load’, ‘store’, and ‘add’ instruction time of 5 microseconds – excellent at the time but still only with a little more memory than the first office and domestic personal computers appearing seven or so years later. [7] The ‘old’ computer was reinstalled at Tananarive. | + | Early in 1970 the FPQ-6 computer was further upgraded by replacing it with the RCA FC-4101C having integrated-circuitry logic and 16K of core memory expandable to 32K. It had 32 data bits per word with 2-bit parity; a memory recycle time of 2 microseconds; and a ‘load’, ‘store’, and ‘add’ instruction time of 5 microseconds – excellent at the time but still only with a little more memory than the first office and domestic personal computers appearing seven or so years later. [7] The ‘old’ computer was reinstalled at Tananarive. |
| The program had been written by GSFC programmers and was named, by the critical local team, Carnarvon Real-time Acquisition Program (C- - P). The station was asked to give it an alternative name but no one seems to remember what that was; it is sufficient to say that the CRO FPQ-6 crew was not impressed. A lengthy review of the new RCA 4101C program was accomplished by the crew and a substantial number of program errors were identified and corrected. A team of four GSFC programmers arrived to review the station’s 26-page report; they endorsed the CRO findings -a final acknowledgement of local programming skills. [8] | The program had been written by GSFC programmers and was named, by the critical local team, Carnarvon Real-time Acquisition Program (C- - P). The station was asked to give it an alternative name but no one seems to remember what that was; it is sufficient to say that the CRO FPQ-6 crew was not impressed. A lengthy review of the new RCA 4101C program was accomplished by the crew and a substantial number of program errors were identified and corrected. A team of four GSFC programmers arrived to review the station’s 26-page report; they endorsed the CRO findings -a final acknowledgement of local programming skills. [8] | ||
Current revision
- Antenna structure
- Tracking and Ranging
- Brief system details
- RCA Computer
- Key Q6 Mission Activity
- Research activity
- BDA, CRO & RCA: Q6 partners
- Other Q-6 tales
Back to Station Equipment
The RCA FC-4101A computer was the ‘heart and brain’ of the FPQ-6 radar. In 1964, and for some time after, it was the only operational radar with a general purpose computer integral to its tracking function.
All known mechanical and dynamic tracking inaccuracies were programmed into the computer to provide immediate correction of range and angle outputs, compensating for such errors as out-of-level azimuth-plane and elevation-gear irregularities, dynamic lag, and ‘reflector droop’ effects on null shift and beam squint.
To make it easier for human operators, the 4101 displayed all tracking data on decimal displays. It also predicted the target’s continuing path and range so that if track was lost in low signal-to-noise conditions the antenna would automatically look-ahead under computer control and then perhaps reacquire the target. In 1967, when the computers memory was extended to 8K and an Inter-range Vector (IRV) capability was added, the look-ahead function was even more accurate as it was based on the predicted orbit.
| Peter Main expresses caution about assuming 4K was a small memory: “… we must remember that in those days memory sizes often were not expressed in bytes (IBM's System 360 computers were just starting to make bytes popular), but most likely in words. 4K words for the 4101 would nowadays be called 32K bytes - not quite so tiny.” [1] |
It was surprising how much was accomplished using the 4101 computer’s 4K memory; upgraded in early 1967 to 8K. A parallel bus-organised computer, it used a magnetic core memory to minimise the use of transistor flip-flops registers and had only four module types utilizing NOR logic for all its logic functions; it was quite a sophisticated machine. Its 32-bit words were divided into two 16-bit groups each containing 15 data bits and one parity bit. The modules were built to BMEWS (Ballistic Missile Early Warning System) standards using stainless steel metalwork – built to last.
Read Bob Hocking’s detailed memories of the RCA-4101
and Peter Main’s skilful modification to a modification
Programming the RCA 4101
Trevor Housley explains: “The program was written in machine code although we did have access to a compiler but, to use it, we had to go to [DoS], Salisbury, S.A., because it needed an IBM 7090 series computer to run … In between, we did ‘Manual Assemblies’; … we wrote the programs in assembly language and then manually converted the assembly code to machine code. Working so close to the hardware gave us an incredibly detailed knowledge of the whole thing so that we were able to do things that one would not dream of doing today. For example, on one occasion, the spacecraft passed directly overhead so that the antenna went into ‘Plunge’ which is when the elevation exceeds 90 degrees [with] … the azimuth reading … out by 180 degrees from what it should be … [W]e had a problem because we needed to report elevations to the M&O at certain times and I quickly wrote a few lines of code and fed them into the paper tape reader and the displays suddenly showed the correct readings. We could do that due to our intimate knowledge of what was going on. If there is a computer controlling a modern radar system, it is unlikely that one could do things such as this.” [2]
Machine language programming was a skill developed by a number of FPQ-6 programmers; the 4101 proving to be an excellent training ground for computer personnel. Several of them moved on to the station’s MSFN computer section and later were keenly sought-after by a burgeoning Australian computer industry.
Upgrade of memory to 8K
In early 1967 an RCA programmer, Dick Wattis, came to CRO for a major software change which introduced the ability to handle both NORAD Orbital Elements and Inter Range Vectors – two different formats displaying the seven parameters necessary to define an Earth orbit. The program, written in advance, was debugged at CRO. Dick and Trevor Housley went to Salisbury to compile the program tape for the RCA-4101. [3]
After nine months of joint development work between the CRO and BDA FPQ-6 teams using the new orbit definition capability, BDA’s real-time tracking data was used to develop orbital parameters for a CRO track half an orbit later. [4]
Good technical work by CRO FPQ-6 led to a lasting requirement by US Air Force OV Network Control to provide pointing data to other stations to track other satellites in the OV series. This resulted from CRO's support of the polar-orbiting Cannonball-2 and Musketball ‘Orbiting Vehicle’ sub-satellites in August 1971. [5].
The capability was again an advantage in November 1972 to rescue a Canadian Communications Satellite, Telesat-A, which went astray shortly after launch. FPQ-6 was able to acquire it on the first pass and to develop new orbital parameters in real-time to enable successful acquisition by Hawaii and Bermuda later in that first orbit; a synchronous orbit was ultimately achieved. [6]
Upgrade of memory to 16K
Early in 1970 the FPQ-6 computer was further upgraded by replacing it with the RCA FC-4101C having integrated-circuitry logic and 16K of core memory expandable to 32K. It had 32 data bits per word with 2-bit parity; a memory recycle time of 2 microseconds; and a ‘load’, ‘store’, and ‘add’ instruction time of 5 microseconds – excellent at the time but still only with a little more memory than the first office and domestic personal computers appearing seven or so years later. [7] The ‘old’ computer was reinstalled at Tananarive.
The program had been written by GSFC programmers and was named, by the critical local team, Carnarvon Real-time Acquisition Program (C- - P). The station was asked to give it an alternative name but no one seems to remember what that was; it is sufficient to say that the CRO FPQ-6 crew was not impressed. A lengthy review of the new RCA 4101C program was accomplished by the crew and a substantial number of program errors were identified and corrected. A team of four GSFC programmers arrived to review the station’s 26-page report; they endorsed the CRO findings -a final acknowledgement of local programming skills. [8]
References
[1] Main, P., email to PD, 08 March 2007
[2] Housley, T., email to PD, 23 August 2005
[3] Housley, T., email to PD, 08 February 2007
[4] NAA: PP538/1, C287, CPR, January 1968
[5] NAA: PP538/1, C287, CPR, January 1968
[6] NAA: PP538/1, C622, CPR. October 1972
[7] Apple 2 (4K) 1977, Atari 400 (8K) 1978, Vic20 (5K) 1979, <www.oldcomputers.net>, 20 March 2006
[8] NAA: PP538/1, C622, CPR, July 1972
We thank Ken Anderson, Bob Hocking, Trevor Housley and Peter Main for their immense and patient contributions to this section.
