---

On Mars: Exploration of the Red Planet. 1958-1978

---

 

 

TOP TEN PROBLEMS

 

 

 

[251] Martin began the Viking Top Ten Problems list in the spring of 1970 to give visibility to problems that could possibly affect the launch dates. Viking project directive no. 7, issued 4 October 1971, codified the concept: [252] "It is the policy of the Viking Project Office that major problems will be clearly identified and immediately receive special management attention by the establishment of Top Ten problems list." To qualify for this dubious distinction, the problem had to he one that seriously affected "the successful attainment of established scientific and/or technical requirements, and/or the meeting of critical project milestones, and/or the compliance with project fiscal constraints." Anyone associated with the Viking project could identify a potential priority problem by defining the exact nature of the difficulty and forming a plan and schedule for solving it. When Martin made an addition to his list, a person in the appropriate organization was charged with solving the problem, and someone in the Viking Project Office monitored his progress. Weekly status reports were datafaxed from the field to Langley. At Martin Marietta, William G. Purdy, vice president and general manager of the Denver Division-through Albert J. Kullas and later Walter Lowrie, his project directors-sent weekly status bulletins on the lander's top problems, since that system seemed to have the greatest number of difficult components and subsystems. In the spring of 1972, Martin told Cortright he hoped the supervisors of employees who had one of their tasks assigned to the top 10 list would not be penalized. Martin, not wanting a stigma attached to identification of a problem, was concerned that at Martin Marietta assignment of a problem might "automatically be considered as a mark of poor performance" when promotions or raises were given. Generally, the nature of the crucial problems was so complex that punishing one individual would not solve the problem. As with the gas chromatograph-mass spectrometer and the biology instrument, the novelty of the technological task was often the source of the trouble. ² Some problems seemed to stay on the manager's worry list forever. Others made repeat performances.

 

At times, Martin found it necessary to bring a particular problem to the attention of a specific subcontractor. Depending on the clout needed behind the message, Martin would sign the letter or enlist the aid of Langley Director Cortright or Martin Marietta Vice President Purdy. In extreme cases, the letter would be sent out over the signature of the NASA administrator. Early in 1973, the Viking Project Office identified six subsystems that required Administrator James C. Fletcher's personal touch: inertial reference unit, subcontractor Hamilton Standard; terminal-descent and landing radar and radar altimeter, Teledyne-Ryan; guidance, control, and sequencing computer, Honeywell; lander camera, Itek Corp.; upper-atmosphere mass spectrometer, Bendix Aerospace; gas chromatograph-mass spectrometer, Litton Systems. ³

 

Fletcher wrote the president of each company asking for his personal pledge of support for Viking and seeking his fullest cooperation in resolving the problem. The administrator usually asked them to come to Washington to discuss the issues further. By setting off an alarm in the front office, NASA managers from Fletcher and his deputy, George Low, to Jim....

 

^a Although the Top Ten Problems concept was not officially recognized until October 1971, the system was used before that date in Jim Martin to Henry Norris, "Viking Top Ten Problems." 4 May 1970, these items were listed.

---

 

 

[254]....Martin hoped to impress on subcontractors their obligations to the Mars project. ⁴ At times, the NASA administrator had to be extremely blunt. Once his letter resulted in a meeting with John W. Anderson of Honeywell, Inc., about the guidance, control, and sequencing computer. This worrisome piece of hardware, the key to the lander's performance, controlled and arranged the sequence of all lander functions from separation from the orbiter through completion of the mission. Without this brain and central nervous system, the lander would be worthless. Schedule delays and cost increases in developing the guidance, control, and sequencing computer were in large part the result of the requirements established for this piece of equipment-energy efficiency, small size, reliability, heat resistance, longevity. Each lander had a guidance, control, and sequencing computer, made up of two completely redundant computers with 18 432-word plated-wire memories. The overall computer was 0.03 cubic meter (26.7 x 27.3 x 40.6 centimeters) and weighed 114.6 kilograms. Its most advanced feature was the two-mil plated-wire memory, making small size and low power consumption possible. Either of the twin computers within the guidance, control, and sequencing computer could operate the lander, but only one would be used during descent. The computer would have to work flawlessly if the landing was to succeed. Once the lander was on the surface, either computer could control the craft.

 

As prime contractor for the lander, Martin Marietta had responsibility for the important computers. In May 1971, the firm asked for proposals from 11 firms to subcontract for the guidance, control, and sequencing computer and received 5 responses. After an unusually complicated contracting process, Honeywell, Inc., was selected as the builder, largely because of its plans to use the two-mil plated-wire memory. Work began at Honeywell's Aerospace Division in Saint Petersburg, Florida, in November 1971. Honeywell also had a contract with Martin Marietta for the development of the lander's data-storage memory, a digital-data-storage device used in conjunction with the lander's tape recorder. The data-storage memory would have the same plate-wire memory units. Combined projected costs for the guidance, control, and sequencing computer and the data-storage memory in 1971 was $6.1 million, with a ceiling cost of $6.8 million.

 

A preliminary design review for the computer was held on schedule in April 1972. At this review, plans called for development testing to be completed by December 1972, following which Martin Marietta and NASA personnel would hold the critical design review. Because of difficulties in fabricating the sense-digit transformers, the plated wire, the memory tunnels, and memory planes, the critical design review was rescheduled for March 1973. As problems continued with component deliveries and memory fabrication, the date for the review was slipped several times. Finally held in August 1973, the critical design review indicated that the design was acceptable in theory, but more development tests were required. Because Martin Marietta needed early delivery of the computers to keep lander fabrication and testing on schedule, Honeywell had to proceed with [255] a parallel program in which the development units and the flight-style computers were built at the same time.

 

Throughout 1973, Honeywell had difficulties with the plated-wire memories. Engineers could not produce sufficient quantities of the plated wire with the proper magnetic characteristics, and they had problems in fabricating the matlike tunnel structures in which the wire was manually inserted. The magnetic keepers applied to the exterior of the sandwiched memory planes also became troublesome for Honeywell. The subcontractor faced another setback when faulty plated-through holes were found in many of the printed circuit boards, which had been purchased from a commercial supplier. These 18- x 23-centimeter Honeywell-designed circuit boards had up to 16 layers of circuits and 3000 plated-through holes for making electrical connections. A great many of the original circuit boards had to be scrapped and reordered from another supplier. ⁵

 

The various problems with the guidance, control, and sequencing computer led NASA Administrator Fletcher to ask Honeywell Vice President Anderson in for a serious talk. Anderson had previously met with Fletcher on 15 February 1973, and subsequently Deputy Administrator Low told Fletcher that some of the computer problems had apparently been solved as a result. But, still unhappy, the administrator wrote again: "During our meeting I was....disturbed by the inference in one of your remarks, that Honeywell is unable to put forth its best efforts on this job because of the type of contract....I hope I was mistaken in my impression that this is so, and I trust that Honeywell will fully live up to all of its obligations." Only Fletcher could talk this firmly to corporation executives. Jim Martin, for all his crustiness, was not in such an authoritative position. ⁶

 

Anderson responded in the positive manner Fletcher was seeking: "In spite of my comments of philosophical concern, I had hoped to have left you with the conviction that Honeywell was applying the best of its resources in a prudent and expeditious fashion. I believe it would be agreed by both the people from NASA and the Martin Company that we are going to find solutions to problems." ⁷ However-despite all the efforts of the agency, Martin Marietta, and Honeywell-the guidance, control, and sequencing computer, which first made the top 10 list on 21 July 1972, was not removed from the chart until 15 January 1975.

 

By late 1973, Honeywell had exceeded the $6.8-million ceiling by nearly $3.5 million. Working under a fixed-price contract, the contractor had no profit incentive to improve the situation. Martin Marietta took several steps, at NASA's urging, to improve the Honeywell operation. The contract was changed from a fixed-price to a cost-plus-incentive-fee contract, with Honeywell accepting a $3.5-million loss. The project was rescheduled, and its cost reestimated to $24 million. Honeywell doubled the number of employees assigned to the computer, as special teams worked to solve specific problems and expedite production. Alternatives to the two-mil plated-wire memory were also examined. While the engineers in Florida [256] attacked these problems, Martin Marietta began a backup program to develop an alternative memory system. ⁸

 

Also in late 1973, Martin Marietta established a resident management office with a staff of 20 at Honeywell's Saint Petersburg factory. This team, and NASA personnel from Langley, assisted Honeywell's managers with scheduling, managerial, and engineering tasks. Everyone was clearly concerned about the fate of the computer. Langley Director Ed Cortright told Deputy Administrator Lose on 30 October 1973 that a meeting with top computer industry experts indicated that Honeywell's problems were not unique. "It appears that the major difficulty is one of schedule time available to reliably produce and test computers needed to support the building of flight" landers. Money might buy more people, but neither people nor dollars could purchase more time. ⁹

 

In January 1974, the Viking Project Office team decided to change all the flight-model computer memories to a new two-mil plated wire known as "coupled-film wire" because it had a second magnetic layer. It was easier to produce, had higher electrical output, was less subject to mechanical damage, and was affected less by temperature changes. Honeywell became more optimistic about meeting schedules. The first flight-model computer was delivered to Martin Marietta in April 1974, nine months late according to the original schedule. Although this proof-test-capsule model had a great many deficiencies, it did permit Martin Marietta to go ahead with its lander tests. Unhappily, delivery schedules continued to slip during 1974 as Honeywell faced more technical difficulties. Faulty components were uncovered. One lot of transistors was rejected. More unsatisfactory printed circuit boards came to light. ¹⁰

 

Continuous monitoring of the subcontractor's troubles was rewarded, however, in late 1974 when the computers were finally ready for delivery. On 15 January, Jim Martin received the following message from Walt Lowrie at Martin Marietta:

 

Oh ye of little faith-We gave birth to the last computer today, I don't know how you feel on the subject but it would appear to me that this top ten has now died of old age.

 

Seriously-although the path has been extremely tortuous I really feel we now have an excellent computer on Viking." ¹¹

 

Martin removed the lander computer from his list of major problems. Thus it went, step by step-problems identified and then solved. At this stage in Viking's existence, there was very little glamor, just long hours, hard work, and an occasional antacid.

 

---