Renaissance of the Yoke Drive

In September of 1980, Professor Dennis Chaddock of Quorn, England, stopped by Columbus to see my stirling engines. While demonstrating the 15cc alpha engine for him, I noticed that it seemed to take a longer time to come up to speed than the rhom­bic engines. I knew at once I’d better retest the 50cc yoke drive alpha, which had now sat on the shelf for over three years. Given a chance to properly warm up, the 50cc alpha showed very promising performance, with a free speed of 2000 rpm. I quickly built a balance shaft for the engine, and thereby confirmed that the patented balance scheme worked. My enthusiasm for stirling work was restored.

It occurred to me that a general purpose stirling engine could be designed and usefully sold as a kit in an effort to encourage more people to get involved with stirling engine development. The 50cc engine was obsolete in various ways, so a new engine was designed from scratch, incorporating everything I had learned about stirlings over the years.

The resulting engine was a 35cc alpha yoke drive engine that was without doubt the finest stirling engine I had designed. The heater and the hot pistons insulation dome were stainless steel deep drawn cups, available commercially as cases for electronic devices. The heater was of the simple annular gap type. The regenerator was wound of stainless steel foil, 0.0015 inch (0.04 mm) thick, that had been dimpled with a star wheel

Renaissance of the Yoke Drive

Renaissance of the Yoke Drive

Wr ‘

Renaissance of the Yoke Drive

Renaissance of the Yoke Drive

The V-15 engine, together (above), and apart (below).

Renaissance of the Yoke Drive

Renaissance of the Yoke Drive

The B-20 was designed to replace the V-15 kit, which had proven difficult for some home shop machinists.

Renaissance of the Yoke Drive

Renaissance of the Yoke Drive

The B-20, as incorporated into a genset by General Pneumatics Corp. (courtesy General Pneumatics Corp.).

To give a regenerator fill factor of about 10%. The cooler consisted of water-cooled slots cut into the cylinder head, connecting the compression space to the plenum beneath the regenerator. The pistons were of the clearance type, made of thin-walled cast iron, running in honed steel cylinders.

Performance was very good from the start. On the first power test the engine produced 21 watts at 2250 rpm, atmospheric pressure. After the snifters were added to the crankcase and the workspace, peak power atmospheric increased to 28 watts. Brief tests at 0.3 atm. showed 40.2 watts. Maximum free speed at this time was 3500 rpm. The cold piston subsequently seized, and the cylinders were honed out a bit more. This time clearance was sufficient to permit a smear of light oil to be used as lubrication without

Excessive drag. Free speed moved up to 4200 rpm, and power increased to 35.3 watts at 2200 rpm, atmospheric.

On tear-down, further signs of piston rubbing appeared, so additional cylinder metal was honed away. Free speed jumped to 4700 rpm, and peak power went to 44.1 watts at 2750 rpm, atmospheric. These results were extremely gratifying. This engine was substantially smaller, lighter, faster, simpler and more powerful (at a given pressure) than the 65cc rhombic. lt was indeed a turning point in my work.

I undertook my first field test with any stirling, by incorporating the engine in an outboard rig made from copper tubing, a suitable synchro drive belt, and a Sears plastic trolling-motor propeller. This device was mounted on a 17 foot canoe, and tested on the nearby Scioto river. As expected with a mere 40+ watts of power, performance was mild, but nevertheless encouraging. After 25 minutes of cruising, a portion of an epoxied-on water jacket fell off, stopping the flow of cooling water and allowing one ofthe pistons to seize. The field test was both great fun and instructive. The water jacket that had so easily come loose in the jostling of the field test had given no problem in hours of prior bench testing. l was also much more willing to push my engine when it was the means to some end in the field, rather than the focus of pampered attention on the test stand.

Putting this engine in kit form took much longer than anticipated. About 50 kits were sold, but it soon became clear that most first-time stirling engine builders needed something much simpler. For this purpose, a V version of the old 15cc alpha engine was developed. This engine, called the V-15, was popular, and it makes a very quiet and im­pressive demonstrator engine. Some builders were still having problems machining the proper piston-cylinder fits, however, so it was replaced with a 20cc yoke drive engine, the B-20, which had removable cylinders that could be more easily refinished if neces­sary.

The primary purpose behind all this kit activity was to interest other people in experimental stirling work, and thereby speed up the process of small stirling devel­opment. A few purchasers did try various modifications, but most were happy if their engines merely ran, and they had no interest in testing or improving performance. Eventually it became obvious that the unmachined engine kits absorbed a great deal of my time without serving their intended purpose, and so, reluctantly, they were dis­continued. The idea of introducing a pre-machined kit for a power-producing engine remained appealing to me, however.

Meanwhile, stirling enthusiast and machine shop owner John Mazur of New York suggested simplifying the 35cc design by combining the crankcase and cylinders into one unit. I redesigned the engine along these lines, and John kindly made eight unit blocks from bar stock on his numerically controlled milling machines. The modified engine was more compact and considerably simpler to make and assemble than the original 35cc.

This engine was intended to become a pre-finished version of the 35cc engine, but, after John’s untimely death, l decided to up-size the engine to 60cc to be assured of

Renaissance of the Yoke Drive

Renaissance of the Yoke DriveThe compact Unit Block engine (right), shown with the original 35cc (above).

The first of the 60cc yoke drive engines, the B-60 (left).

Getting at least 100 watts output with a simple unfinned heater.

The resulting engine, the B-60, was based on a cast aluminum unit block, which was hard coat anodized after final machining to give the cylinders a good wear surface. The water jacket on the block was formed by applying aluminum tape to cover relieved water passageways. There was provision in the block casting for the balance shaft, but none was made, as experience showed the engine could be more easily mounted on springs for satisfactory testing. The pistons were aluminum alloy, coated with a baked on teflon resin paint called "Xylan".

This engine seemed moderately successful, after a few initial bugs were sorted out. It produced 57.5 watts at 2700 rpm, atmospheric, and its free speed was just over 3500 rpm. Power tests under pressure were not completed, however, because the type 304 stainless steel heater began to scale badly after three or four hours of operation. A sec­ond heater developed the same problem. This scale, subsequently analyzed and found to be mostly iron oxide, would migrate to the cold cylinder, score the Xylan coat on the piston, and bring things to a gummy halt.

I had used 304 stainless heaters on other engines (such as the 35cc) for much longer periods of time with no such problems. Perhaps these heaters were being over­heated by the powerful burner, or perhaps the material was substandard 304. This problem by itself was not so difficult, but it triggered a major loss of morale. Why bother with these troublesome engines? I had already solved a great many problems during this program, but there seemed to be no end to unexpected new ones.

Adding to the mental chaos of this time were a number of interesting new ideas. After thinking over several cable-drive mechanisms of Jim Senft and William Beale, it occurred to me that the use of cable-driven pistons in a yoke drive engine would elimi­nate four bearings and their noise, lubrication, weight and expense. Further thought revealed that several new cable arrangements could also replace the rocking lever and its bearings.

Substantial additional reductions in engine height could come from using disc pistons with tail rods guided from below, as shown on the schematic drawing of the cable-drive system, above, but such pistons would require an excellent line (not clear­ance) seal. Mick Collins had demonstrated just such a seal in his 5cc competition stirling. These were extremely thin-edged (about 0.010 inch, or 0.25 mm) pressure-actuating cup seals made of Rulon, a brand of filled teflon. I promptly made and tested a few ex­amples, and they proved to seal beautifully with very low friction. These new ideas and tests served to restore my interest in stirling engine work.

I was well into making a guided-piston, cup-sealed version of the B-60 engine, when another idea occurred to me that was extremely appealing. If the yoke were inverted (as it had in tact been in the original 15cc yoke engine), then the cylinders and pistons could be partially cut away so they could be moved down into the drive mechanism, and thereby occupy the same space as the crankshaft and yoke (US Pat­ent 4,532,819). The height of the cylinders would thus be enclosed within the height of

Yoke mechanism, not added to it as in the previous yoke designs. Moreover, the crank­shaft would penetrate the new shorter block more or less in its center, rather than near the bottom as before, so the flywheel, too, would be contained within the height limits of the cylinders. Although the cylinders would be partially cut away, they would remain adequate to guide self-aligning pistons, without tail guides. An exceptionally compact and lightweight engine would result.

Several problems were also immediately apparent. Any balance shaft on an invert­ed yoke engine would now be below the cylinder block, and thereby add back some of the unwanted height. The cut away areas would make the cylinders difficult or impossi­ble to hone. Counterbalance masses mounted on the crankshaft would have to be hung outside of the cylinders in order to clear them.

The potential problem that worried me the most, however, had to do with the geometry of the yoke drive. Like the conventional crank and slider mechanism, the yoke drive has a dwell at one end of its stroke and a snap at the other, caused by the yoke’s varying angularity. By inverting the yoke, relative to the pistons, the dwell is moved from bottom dead center to top dead center. In fact, the piston phasing is slightly differ­ent throughout the cycle. My initial investigation into this difference lead me to believe that engine power with this new arrangement might be 10% lower at any given pres­sure level. Only later did I realize that the inverted yoke phasing looked at least as good as the 90° V-2 phasing, and this realization largely dispelled my concern. Subsequent test results reveal no decrease in performance from inverting the yoke.

Soon enough, other aspects of the inverted yoke design began falling into place. The stroke-multiplication effect of the yoke mechanism means the crankthrow diam­eter needs to be only 71% ofthe actual piston stroke, making possible a strong, one- piece stepped crankshaft, supported by bearings at both ends, which is nevertheless small enough to fit within the main needle bearing in the yoke. Counterbalance mass may readily be hung outboard of the main bearings of the rigid crankshaft. Assembly and disassembly could be extremely simple and rapid. The spool-like cutaway pistons are very lightweight, so the balance shaft can be permanently omitted. The triangular braces for the yoke arms are now stressed in tension, rather than in compression, per­mitting a lighter yoke design. The cylinders can be left intact, honed, hard anodized, finish honed, and only then cut away as necessary. One by one, the problems began to disappear, and additional advantages emerged. The new engine was named the model C-6O, and its overall size would be similar to that of the original 35cc engine.

At this point, my enthusiasm for this new approach was so high that the C-60 was designed and machined in the matter of a few months. The cylinder block was cut from 6061 aluminum alloy bar stock, and hard anodized for wear resistance. The cooler, cylinder head, and regenerator (as well as the first burner) are from the B-60 engine. The heater is identical to that of the B-60, except for being made of type 310 stainless, for higher scaling resistance.

The pistons are made from aluminum alloy bar stock, with thin strips of etched

Rulon LD epoxied onto the wear surfaces. The Rulon is then machined to final size, and grooved axially so its high rate of thermal expansion will not cause it to buckle circum­ferentially as the piston heats up. Diametrical clearances on the pistons of about 0.003 inch (0.08 mm) have proven adequate. Piston sealing is provided by thin-edged cup seals also machined from Rulon LD.

Originally both pistons were similar, and employed separate connecting rods. Sub­sequently a rocking type piston, which is a piston and connecting rod combined into one piece, was installed in the cold cylinder. This modification resulted in a piston that is easier to make, lighter in weight (40 grams vs 80 grams), and quieter in operation. It has even contributed a few watts to power output, probably due to reduced friction.

The yoke and rocking lever are also machined from aluminum alloy bar, with drawn cup needle bearings pressed in. The crankpin bearing in the yoke was initially a full complement needle bearing, and it was axially located on the crankshaft with a shoulder and a retaining ring; however, this arrangement resulted in fretting of the bearing against the retaining ring after several hours of operation. Apparently the needles were skewing, due to their short length relative to the crankpin diameter (ratio = 1 to 2), and this skewing repeatedly drove the bearing against the retaining ring. The substitution of a longer bearing (bearing length to crankpin diameter ratio = 1 to 1.25) with caged needles solved the problem. The yoke now floats on the crankpin, with its axial location determined solely by the rocking lever, and the shoulder and retaining ring have been eliminated.

With the new bearing installation, cupped grease catchers were also added to the ends of the main yoke bearing to catch the small bits of grease that were being thrown form the bearing onto the cylinders, thus solving another early problem.

The crankshaft is machined from low carbon steel, case hardened to Rc 60, and finish ground. lts axial location is maintained by a ball bearing slipped onto the rear of the crank, and located between a shoulder and a clamped-on balance bob weight. The front main bearing is another needle bearing. The crankcase seal is a simple Rulon lip seal, held by a retaining ring in the front bearing case.

The C-60 ran well from the start. It produces a steady and reliable 100 watts at two atmospheres pressurization on air. Maximum recorded free speed is 4002 rpm. I have put 21 hours of operating time on the engine, and Sunpower, Inc., has put another 50 hours on it.

On one occasion the engine failed to run properly after having been apart for inspection. Free speed was down, and power was off at least 25%. The engine was again taken apart, and after considerable scrutiny the only anomaly I could find was that the foil regenerator was wrapped the opposite way around the inner sleeve.

This regenerator is made of stainless foil 1 inch wide and 0.001 inch (.025 mm) thick. It is dimpled with a seamstress’s tracing wheel to create about 0.009 inch (.229 mm) spacing between wraps. I had previously wrapped it with the dimples facing the

Renaissance of the Yoke Drive

Four illustrations of the C-60, showing, (upper left) the piston assembly with the new rocking compression piston; (upper right) the block with the original pistons and crankshaft; (lower left) the engine with a clutch and reduction gear set, ready for mounting in a bicycle; and (lower right) the heat exchangers, which are the same as used on the B-6O.

Inside, but on this occasion I had wrapped it with the dimples facing outward. On this particular foil the dimpling process seems to have caused more distortion than usual, and, when wrapped with the dimples outward, the foil showed a slight waviness, which no doubt caused nonuniform flow. When the engine was rebuilt with the foil wrapped the other way, full performance was restored. I had exclusively used similar foil regen­erators for years before this instance, and yet had not noticed such a problem.

With a nice little 100 watt engine, it was time to conduct another field test, this time on an old ten speed bicycle. A clutch and reduction gear drive was made to con­nect the engine to the crankset via a suitable freewheel sprocket. In this way, five of the bicycle’s extant speeds could be used. Engine speed control was a primitive spring-load­ed flywheel brake controlled by a motorcycle twistgrip on the handlebar. Engine cool­ing was provided by an engine driven water pump and an onboard aluminum radiator. Unfortunately, there was no provision for engine pressurization, so power was limited to about 50 watts.

On October 19, 1986, what may have been the worId’s first stirling powered bicycle was ridden for about 5 miles. With the limited power, the performance was modest.

As in the previous field test, I noticed how much more willing I was to push the engine hard when it was outside doing something useful, than when it was on the test stand. Also apparent was the importance of small details, such as the quality ofthe water pump, the tubing connections, the burner control, etc.. It was one thing to develop a good stirling for test purposes, and quite another to develop one for everyday hard work.

As always, I had many new areas I wanted to explore. Unusual speed control de­vices and wick-fed kerosene catalytic burners were two among many others. But what I actually did was more sensible, and that was to further develop the heat exchangers of the C-60.

It was apparent, and no surprise, that the C-60 was running out of power by 2.3 atmospheres pressure when charged with air. It was also likely that the limitation was in the heater, since it had somewhat less surface area than the cooler. But the cooler was also limited, by its very design; there is only so much surface area one can fit into a flat cylinder head connecting two closely spaced parallel cylinders. It would be point­less to make a superiorfinned heater for the engine only to have performance now be cooler-limited. The plenum located between the cooler and the regenerator of the C-60 (and the B-60, B-20, and 35 cc engines before it) was also undesirable. This volume was always filled with gas at the wrong temperature for what was happening in the engine at any given time. During expansion, for example, it was filled with cooler gas than dur­ing compression. Obviously, such a plenum, if it could not be eliminated entirely, should at least be moved to the other end of the cooler.

These considerations lead to the next version of the 60cc engine, the model D-60, and this engine was indeed sketched out before the C-60 was even finished.

The major design differences in the D-60 engine are an internally finned heater

Renaissance of the Yoke Drive

The original shouldered crankshaft is shown above the modified crank that solved the fretting problem (above); the C-60 mounted in a bicycle for testing (below).

Renaissance of the Yoke Drive

And an annular cooler.

The internally finned heater is actually an intermediate design, since a heater finned on both the inside and the outside will probably ultimately be necessary. Never­theless, it was of interest to explore whether the C-60’s apparent heater limitations were on the outside (flame to heater head) or the inside (heater head to working gas).

My belief was that the inside surface area was the limiting factor, so corrugated fins were formed from nickel 200 and furnace brazed to the inside wall of the type 310 stain­less steel heater can.

Before the actual heater assembly was brazed, several flat test assemblies were prepared and brazed with both silver-based and nickel-based fillers. The silver filler made nice large fillets, but it also partially blocked a few fin passages. The nickel filler was very clean and neat, but formed little or no fillet, leaving some doubt about the adequacy of its thermal contact. Nevertheless, the nickel filler was used for the actual heater, with satisfactory results, although one portion (under 10%) ofthe fins is unat­tached to the heater can.

The resulting heater is slightly shorter in length than that of the C-60, so its outside surface area is only about 80% that of the C-60. The internal finning, however, gives an inside surface area that is 4.8 times larger, and a dead volume that is over 3 times larger, than in the C-60.

The annular cooler permits a great deal more surface area than the flat cylinder head coolers ofthe previous prototypes, but it does tend to make the engine taller, since now the cooler is positioned vertically instead of horizontally. Mechanically this change offers certain advantages; for example, the cylinders may be located closer to each oth­er, and much less cutaway clearance is needed for the yoke. ln constructing the D-60, unlike the C-60, no portion of the block needed further relieving after final honing.

The D-60’s cooler has 160 internal fins to provide 1.38 times more surface area than the C-60 cooler, but this is still only half the area of the D-60’s heater, and is probably less than is desirable, since the uncooled compression cylinder runs somewhat hot at 175°F (80°C).

These fins are machined internally into the cooler wall in order to assure the best possible thermal bond. A tool-post mounted internal fin cutting machine was made expressly for this job, by mounting a small slitting saw on an arbor in an arm, and driv­ing it through a miniature synchro belt. This setup was sufficiently rigid to machine the aluminum alloy cooler, but it would not have been suitable for machining heater fins out of stainless steel or nickel

In other details the D-60 engine is the same or very similar to the C-60, with the exception that considerable effort was made to keep its total weight low. The C-60 weighs 3.4 Kg (7.6 pounds) with burner and flywheel, while the D-60 weighs 2.2 Kg (4.75 pounds), also with burner and flywheel, Further significant weight reduction is possible.

The D-60 is a wonderful engine. Whereas the C-60 is essentially a 100 watt engine,

Renaissance of the Yoke Drive

The original D-60 (above left) and the subsequent version with alternator and bulbs (above right). The D-60 cooler, crankshaft, rocking lever, and yoke (below). Opposite: The ma­chine devised to cut the internal fins of the cooler (top), the D-60’s piston assembly (bottom left), and the internal fins for the heater (lower right).

Renaissance of the Yoke Drive

Renaissance of the Yoke Drive

Renaissance of the Yoke Drive

The Philips MP1002CA air engine of about 63cc swept volume (left) is shown on a common scale with the D-60 (right). (drawing of Philips engine courtesy of Professor Allan J. Organ).

At 2 bar pressurization, the D-60 is a 200 watt engine at 3 bar pressurization. Even before its heater is at a visible red heat, the D-60 clearly outperforms the C-60 at its best. At similar temperatures and pressures the D60 is considerably superior, producing, for example, 1.5 times the power, at 1.3 times the speed, of the C-60, at 2 bar pressurization. The limitations that are beginning to show at 3 bar are probably related more to the limited surface on the outside of the heater and cooler, than to any internal limitations.

Although no outdoor field tests have been conducted with the D-60, it was mount­ed on the same outboard rig used to test the 35cc, and run in the shop’s basin, where it could be pressurized. Nothing very scientific was learned from this amusing exercise, but 200 watts of power does move a lot of water around a small basin in a hurry.

The completion and testing ofthe D-60 represented a definite high point in the program. FoIIow-up work concentrated on making a version of the engine suitable for limited production and sale to other interested parties. Before getting into the history of that enterprise, it would be a good time to describe some totally unrelated efforts to make small engines.

Добавить комментарий

Ваш e-mail не будет опубликован. Обязательные поля помечены *