When Coleman Raphael was working for Republic Aviation Corp., the company brought in an outside consultant to draft a new five-year plan -- an elaborately detailed scenario bound in leather with gold lettering on the cover.

Before the plan's prophecies could be carried out, Republic's business fell into a slump, sales plunged from $300 million a year to $100 million and the company was gobbled up by another firm.

Coleman Raphael learned a lesson.

As chairman of Atlantic Research Corp., he writes his own corporate five-year plan and keeps it in a loose-leaf notebook.

"In this business things change too fast to put them in a leather binding," he said recently as he talked about the management philosophy that has made ARC a $75-million-a-year Washington business.

"The best way to learn how to do things right is to work in a system where everything is done wrong," added Raphael.

Raphael recently reorganized top management of the firm, relinquishing the job of president to William H. Borten, who had been executive vice president and treasurer. W. Gerald Hamm, formerly vice president and general manager of the propulsion division, became executive vice president.

"As chairman of the board," Raphael explained, "I believe there isn't much that I should have to do. I should be free to do anything I want.

"I can take a day away from the desk and know that nothing is going wrong. Therefore I can devote my time to the things that I want to."

That means that even though the majority of Atlantic Research's business comes from production rocket motors, Raphael can pursue more down-to-earth projects like creating a mixture of coal and water that can be burned like oil, or developing a method for disposing of cancer-causing PCB's.

Raphael's sometimes serendipitous approach has turned Atlantic Research into a high-technology miniconglomerate. The company, whose rockets have helped launch every U.S. earth satellite, now has its own satellite businesses.

Around the rocket business spins a data communications divisions devoted to printing, manufacturing special vehicles, producing bottle-cap liners, making air purification systems for submarines and performing research on contract.

But rockets are the main reason Atlantic Research's profits have increased from 6 cents a share in 1976 to $1.30 this year.

In much the same way that Atlantic Research rockets spin the nose cones off giant space probes, Atlantic Research was spun off from its parent, Susquehanna Corp., effective Dec. 31, 1976. In its first four years as an independent operation, revenues more than doubled -- from $27 million to $54 million -- and net income increased from $241,000 to $1.6 million ($1.25 a share).

ARC's sales for the first nine months of 1980 were $55.5 million -- greater than for all of last year -- and earnings totaled $1,166,000, or $1.01 a share. Analysts have projected revenues of $75 million for 1980; ARC officials expect that when final figures are released next month, that target will be surpassed.

Rocketry has consistently produced 55 percent to 60 percent of Atlantic's revenues, profits and growth, providing a base for diversification into fields not dependent on government contracts and defense budgets.

Federal rocket contracts will continue to produce that share of the company's revenues, Raphael predicts, even with an aggressive diversification program aimed at an annual corporate/growth goal of 20 percent.

The company's rockets are essential to more than half-a-dozen weapons systems that are likely to benefit from the Reagan administration's pledge to increase defense spending.

As a subcontractor, Atlantic Research produces part of the MX missile, Trident missile, Cruise missile, Viper antitank rocket, Stinger antiaircraft rocket, the Multiple Launch Rocket System (MLRS) and some other weapons still in the development stage.

Atlantic Research doesn't make any complete rocket systems, but it provides either the propulsion, guidance or launching rockets for several main-systems contractors. The company has the technical capability to become a prime rocket contractor, but then would have to compete with the firms that are now its partners, explains rocket division chief James R. Sides.

For the MX, Atlantic makes explosive charges to blast the lids off the missiles' underground shelters. The Trident missile is steered to its target by ARC maneuvering rockets.

For the submarine-launched Cruise missile, Atlantic Research makes an underwater rocket that ignites after the missile is launched from a torpedo tube and flies it out of the water, where an air-breathing jet engine takes over. The land-based Cruise missile is blasted up to the speed at which its jet engine ignites by another ARC rocket.

Stinger, Viper, MLRS and a proposed new artillery rocket known as "assault breaker," are propelled by Atlantic Research solid-fuel rocket motors.

Most of the rocket motors are assembled at ARC's 500-acre complex in Fauquier County, 45 minutes from the corporate headquarters in Alexandria.

Dispersed around the wooded grounds on the outskirts of Gainsville are 150 small buildings. They are flimsy plywood structures, most of them built inside individual horseshoe-shaped earthen bunkers. Should a rocket go off accidentally, the buildings will explode easily and the bunker will divert the blast away from neighboring workers.

For the high volume of Stinger and Viper production, Atlantic Research has opened a second factory, in Camden, Ark. There the company leased a World War II weapons facility built for handling and storing explosives.

U.S. and Allied military orders for the two shoulder-fired rockets are expected to be so large that the Arkansas plant's production will soon exceed the present output of the entire propulsion division, Sides said.

Growth of the rocket-motor business should enable Atlantic Research to continue its rapid growth, the Baltimore brokerage house of Baker, Watts & Co. said in a recent research report on the firm.

"We believe that sophisticated weapons technology expenditures will continue to expand throughout the 1980s, and that ARC will continue to participate in these programs," Baker, Watts said. "Renewed emphasis on defense spending should provide a continuing base of earnings."

The firm warned, however, that comparisons between earnings per share in 1979 and 1980 "will not be impressive" because the company issued additional stock last year, diluting the effect of growth in net profits.

Baker, Watts predicted Atlantic Research will report profits of $1.30 a share for 1980 and $1.55 a share this year and recommended purchase of the shares "as a long-term holding in speculative accounts."

Analyst William Socha said ARC's data communication and electromagnetic engineering divisions also have strong growth potential, but he was less enthusiastic about some of the company's minor business lines.

Commercial printing brings in almost $2 million a year, but the Baker, Watts study said it "will not have significant bearing on ARC's future profitability."

The custom-vehicle business offers only "moderate growth possibilities in 1980," the analyst's report added. Raphael himself admitted the operation produces little profit and loses money on many of the specialized buses and trucks it builds.

Like the printing business, ARC's special-vehicle operation was started because the company was doing the job for itself and Raphael figured he could make a little extra money -- or at least pay some of his own costs -- by taking on some outside work.

An even less likely adjunct to a rocket and electronics business in ARC's subsidiary Tri-Seal International. It makes plastic liners for bottles and jars, selling billions of the simple products to companies like Revlon, Clairol and Procter & Gamble.

Atlantic Research acquired the firm in 1978, picking up the Long Island, N.Y. manufacturing plant from an owner who was retiring for health reasons. ARC paid most of the price in notes and earned its cash investment back in the first year of business. Tri-Seal is now a $2 million-a-year operation and with a newly beefed-up marketing program could grow very quickly, Raphael said.

Regardless of their origins, these small subsidiaries don't make a lot of sense for a high-technology company like ARC, some analysts who follow the firm believe.

They also have reservations about the company's potential for improving profit margins. Atlantic Research earns about 3 cents on every dollar of business it does, and Raphael acknowledged margins will never get much above 4 percent. That's because the government limits the profit on the production, research and service contracts that make up the bulk of the company's business.

Asked recently by an investment adviser if he would consider selling off some of the subsidiaries, Raphael said he might "if I had something I wanted to do with the cash."

One possible use for additional capital, he added, would be to invest in development of a process called ARC-coal, a slurry of finely ground coal and water that is meant to be a substitute for heavy fuel oil.

Atlantic researchers have shown that the fuel works in a small oil burner. The Department of Energy recently gave the company a $447,000 grant to do further work on the process.

The next step, Raphael explained recently, is to produce several thousand gallons of ARC-coal, so the product can be test-burned in a large utility company boiler. That will require a bigger investment in coal grinding and slurry mixing equipment than Atlantic Research can afford.

The need for additional capital and its aggressive growth pattern raise the possibility that ARC could become a takeover candidate. Acquisition of the company would be difficult without the cooperation of Raphael and the other top management members.

Raphael owns 235,000 shares of the company's stock, about 16.3 percent. Borten controls another 106,000 shares, or 7.4 percent, and the total holdings of officers and directors add up to 421,000 shares, or 29.3 percent. Another 5.9 percent is held by a Netherlands Antilles investment firm called Quantum Overseas, N.V.