Ten large smokestacks poke over the crest of a hill in the rolling country-side west of Pretoria. Nestled in the valley below -- largely hidden from prying eyes -- are the sprawling concrete buildings of South Africa's uranium enrichment plant.

This facility -- called Calindaba, which in Zulu means "that which we don't talk about" -- is capable of producing 3 percent enriched uranium that will eventually fuel South Africa's nuclear power plants.

Valindaba is also able, concedes Atomic Energy Board President A.J.A. Roux, to produce highly enriched uranium than can be used to manufacture atomic bombs.

Exactly what has been and is currently being produced inside Valindaba -- and in what quantity -- is one of South Africa's most closely kept secrets.

The installation, located about a mile from the Atomic Energy Board's separate Pelindaba research complex, is also one of the country's best guarded facilities.

A constant radar watch is kept on the airspace around Valindaba, which has been designated a security zone off limits to all aircraft. At a nearby air force base, two Mirage jet fighters stand ready to scranable to enforce that ban.

While enrichment, like plutonium reprocessing, is one of the technologies that concerns the Carter administration because of its proliferation implications, South Africa is the only developing country that has actually constructed a uranium enrichment plant.

Brazil, as part of a nuclear agreement with West Germany, plans to build a demonstration uranium enrichment plant at Resende, about 100 miles from Rio de Janeiro, which it hopes to have operating in the late 1980s.

But no other developing country is even talking seriously about this technology.

The reason is simple. While India has built a small commercial plutonium reprocessing plant for under $13 million, the cost of a small commercial uranium enrichment plant runs well over $1 billion. Uranium enrichment facilities also use massive amounts of electricity that can be spared by most developing countries.

South Africa, however, had a commercial motive in mind when it began looking into uranium enrichment in 1960.

"It was obvious that nuclear power programs of the future would be based very largely on enriched uranium reactor systems," Roux said in an interview. "Therefore, we thought that as a large uranium-producing country, instead of marketing our uranium in a crude uranium oxide form, we ought to market it in a more advanced form."

Rejecting the gaseous diffusion process developed by the nuclear-weapon states as too costly, and aware that a number of other countries were already working on the gas centrifuge method, South Africa decided to search for another process in hopes of obtaining a competitive edge.

"Fortunately, we did think of another approach," Roux said. And in 1961, South Africa began work on an aerodynamic process of isotopic separation.

By 1969, the results looked sufficiently promising that the government decided to build a pilot plant. The story of its construction illustrates the problem -- for even a relatively advanced country -- of building sensitive nuclear facilities.

"To build a pilot plant, we needed large valves -- and the technology to produce them -- to quote only one type of example," Roux said. "We had to develop compressors, and hexafluoride compressors are something different than ordinary air compressors."

Roux said that if someone had asked him in 1969 how many South African companies would have to be called on in building the enrichment plant, "I would have said 50." The actual number turned out to be 256.

But at the end of 1976, the first part of the plant was put into operation to prove in a practical way the feasibility of the enrichment process.

"It was dead on the predicted performance immediately after we put it into operation," Roux said."No problem whatsoever. So we were producing enriched uranium already by the end of 1976 -- in small quantities." By the end of 1977, the pilot plant was in full operation.

Even before the pilot plant was finished, however, South Africa was mapping plans for a large-scale commercial plant that would have a 5,000-ton annual enrichment capacity. South Africa also passed the word that it would like a foreign partner.

"We were really looking for large partners which would bring with them industrial capacity and money and manpower," Roux said. For a time, it looked like a partnership with a West German company was set. But the deal fell through, South African officials say, when the German firm insisted that South Africa take all the commercial risk.

"So then we decided to go ahead on our own," Roux said. This plan also collapsed, however, when the nuclear states got together in what became known as the London suppliers club, and agreed not to export sensitive equipment to any nonnuclear weapon state.

"That mean there was only one way open," Roux said. "Whatever we do, we must do it on our own. And to do it on our own, the obvious thing became that we expand the plant that we've got." So South Africa is currently engaged in an expansion of the Valindaba plant which is scheduled to be completed by late 1984.

South African officials refuse to provide any figures on the enrichment capacity of the Valindaba plant -- which might provide foreign intelligence analysts clues to how much weapon-grade uranium South Africa already could have stockpiled.

But Roux said that following completion of the expansion, Valindaba will be able to enrich "in excess" of the 300 tons per year required to fuel the two atomic power stations currently under construction at Koeburg, outside Cape Town. CAPTION: Picture, The uranium enrichment plant at Pelindaba, South Africa: "That which we don't talk about." Copyright (c) South African Atomic Energy Board