TURBOGLIDE AND FLIGHT PITCH DYNAFLOW
With the introduction of the dual-stator Variable Pitch Dynaflow, GM’s “pure” torque converter automatic had reached an advanced state of development. However, Oliver Kelley’s corporate transmission group was not yet satisfied and kept working on what was supposed to be the ultimate torque converter automatic: a triple-turbine transmission.
To be clear, there were actually two such transmissions: Chevrolet’s Turboglide, introduced as an option on 1957 Chevrolets with the 283 cu. in. (4,638 cc) V8 engine, followed a year later by Buick’s Flight Pitch Dynaflow, which was standard on the 1958 Buick Roadmaster and Limited and optional on other models. Although Turboglide and Flight Pitch Dynaflow (renamed Triple Turbine for 1959) differed in layout and in many details, both transmissions were based on a common set of ideas developed by Kelly’s team and were, like the original Dynaflow and Powerglide, essentially variations of the same design.
The easiest way to conceptualize the triple-turbine transmission is as a Variable Pitch Dynaflow with an additional drive turbine rather than a second stator. The extra turbine was linked to its own set of planetary gears, the addition of which required moving both gearsets out of the converter hub and into the transmission case. Controlling those gearsets — which superseded Dynaflow’s familiar Ravigneaux gearbox — were no fewer than six clutches: two one-way clutches (not counting the stator clutch), a neutral clutch, a reverse clutch, a forward clutch, and a “hill retarder” or “grade retarder” clutch (the function of which we’ll explain shortly). Turboglide initially used cone-type neutral, reverse, and forward clutches with a multi-disc hill retarder clutch, but switched to a multi-disc neutral clutch for 1958 and adopted multi-disc reverse and forward clutches for 1959. Flight Pitch Dynaflow and Triple Turbine used only multi-disc clutches from the start.
The transmission’s two one-way clutches, which were linked to the reaction members of the two planetary gearsets — the front unit sun gear and rear unit annulus, as on Hydra-Matic — were cleverly interconnected, with the inner race of the front sun gear clutch forming the outer race of the rear annulus clutch. The forward clutch served to anchor both one-way clutches to the case, preventing either reaction member from turning backward. The rear annulus was free to rotate forward while the front sun gear remained locked, but the front sun gear could only turn forward if the rear annulus also did so. With the forward clutch released, reaction torque on the rear annulus would lock it against the front sun gear clutch, which caused both clutches to turn backward together, carrying their respective gears with them.
As in Twin-Turbine Dynaflow, the triple-turbine transmission’s first turbine was affixed to a support shell, within which were mounted the other two turbines. The support shell was splined to a central input shaft that caused the rear unit sun gear to rotate with the first turbine. The inner hub of the second turbine was attached to a hollow sleeve shaft that caused the second turbine and front unit annulus to rotate together. A third hollow shaft, located between the other two, connected the third turbine to the neutral clutch, which when engaged linked the third turbine to the planetary gearsets’ interconnected front and rear planet carriers. A flange at the trailing edge of the rear carrier allowed the carriers to drive the transmission output shaft.
The mechanics of the triple-turbine transmission were very similar to those of the twin-turbine units, but there were now three stages rather than two. At stall, most of the impeller’s torque (augmented as usual by the stator) was applied to the first turbine and thus the rear unit sun gear. This would exert reaction torque on the rear annulus, so if the forward clutch was engaged, both one-way clutches would lock, putting both gearsets in reduction. Oil exiting the first turbine would initially apply a small amount of positive torque to the second turbine and therefore to the front unit annulus. Once the turbines were moving, the oil stream exerted progressively less torque on the first turbine and progressively more on the vanes of the second. The torque exerted on each turbine was multiplied by their respective planetary gears and applied to the output shaft through the conjoined planet carriers. Turboglide’s gear ratios were 2.67:1 for the rear gearset and 1.60:1 for the front unit; the ratios for Flight Pitch Dynaflow/Triple Turbine were 2.86:1 and 1.55:1 respectively.
(We should emphasize here that while these transmissions technically had three geared ratios, they were NOT three-speed automatics. Over the years, some sources have incorrectly described them as such, which, while true in one sense, betrays a fundamental misunderstanding of how these transmissions actually function.)
If you followed our explanation of Twin-Turbine and Variable Pitch Dynaflow earlier in this article, you may recall that in the single-stator versions of those transmissions, oil flow from the first turbine would initially oppose the rotation of the second, a problem rectified on later versions of Variable Pitch Dynaflow by the addition of the front stator. Since the triple-turbine transmissions lacked the additional stator, oil exiting the first and second turbines at or just above stall would similarly oppose the rotation of the third turbine, reducing the net torque on the output shaft. As torque shifted from the first turbine to the second, the oil flow from the second turbine began to exert positive torque on the third turbine. (The more aggressive the initial launch, the longer this took.)
Once the speed of the second turbine reached approximately 55–60% of the speed of the first turbine (the exact transition point depending on the comparative ratios of the front and rear gearsets), the front unit annulus would attempt to rotate its planet carrier faster than the rear carrier. Since the two carriers were connected, the rear carrier was obliged to rotate faster as well. This caused the carrier to overdrive the rear unit sun gear and the first turbine, which removed the reaction torque on the rear unit annulus and its one-way clutch. The first turbine would then freewheel idly, leaving the other two turbines to drive the output shaft. The second and third turbine would repeat this process once there was enough torque on the third turbine to drive it at more than about 60% of the speed of the second (again depending on the exact ratio of the front gearset), which left both the first and second turbines spinning idly. The stator continued to provide some torque multiplication until toroidal flow dropped off enough to release the stator’s one-way clutch.
Both Turboglide and Flight Pitch Dynaflow/Triple Turbine used variable-pitch stators, but of different designs. Turboglide had a two-position stator very similar to the one used in 1957 and later Variable Pitch Dynaflow/Twin Turbine transmissions, but Buick adopted a more sophisticated infinitely variable stator. As with the two-position unit, stator blade angle was controlled by the pivoting of an annular piston controlled by hydraulic pressure. However, rather than simply flipping back and forth between two discrete positions, the infinitely variable stator’s control piston was balanced between opposing converter and throttle valve pressures that could hold the piston at any position within its range of motion. In this way, the stator blades could continuously adjust their pitch based on load. A “kickdown” valve opened by flooring the accelerator would still force the blades to their highest possible angle, just as with the two-position stator.
Reverse was an adaptation of the principle used in contemporary Hydra-Matics: allowing reaction torque on the reaction member of the rear gearset to provide reverse rotation and then compounding it with another gearset to provide reverse reduction. Since there were only two gearsets rather than three, the front unit now performed the latter chore. To accomplish all this, the neutral and reverse clutches were engaged, connecting the third turbine to the planet carriers and holding the front unit annulus in place, while the forward clutch was released so that the one-way clutches were no longer anchored to the case. The rotation of the first turbine (and thus the rear unit sun gear) therefore couldn’t apply any torque to the planet carrier, but their rotation would cause the rear unit annulus, both one-way clutches, and the front unit sun gear to all turn backward together. With the front unit annulus locked, the front planetary gearset would multiply this reverse torque and apply it to the planet carrier. Since the second turbine was connected to the front unit annulus, engaging the reverse clutch to lock the annulus also locked the turbine. This essentially transformed the second turbine into a stator, although its purpose was exactly the opposite of Variable Pitch Dynaflow’s forward stator, maximizing rather than removing the negative torque on the third turbine so that torque would be added to the reverse torque the front unit exerted on the output shaft.
The last major element of the triple-turbine transmission was the hill retarder or grade retarder clutch. As we previously mentioned, Twin-Turbine Dynaflow provided little engine braking in Drive and the triple-turbine automatics suffered the same problem. To compensate, both triple-turbine transmissions could be shifted to HR/GR, which engaged the hill clutch — locking the rear annulus — while releasing both the forward clutch and the neutral clutch to disconnect the one-way clutches from the case and the third turbine from the planet carriers. In that condition, only the first turbine could transmit any torque to the output shaft and the rear planetary unit would remain in reduction until the driver shifted to a different range.
In principle, this mode could be used as a low range, although in practice, doing so created too much slippage to have any performance advantage. The real purpose was to provide engine braking: The hill clutch would not unlock even on the overrun, so coasting would cause the rear planetary unit to act as an overdrive, causing the first turbine to attempt to overdrive the engine. This created a strong braking effect, but the rear unit gear ratios were so short — comparable to first gear in many contemporary manual transmissions — that using it at higher speeds was dangerous. (Causing the first turbine to abruptly turn more than twice as fast as the impeller would certainly slow the car, but could overheat the transmission.)
As with most of GM’s early automatics, the triple-turbine triple turbines had front and rear oil pumps, the latter used for push-starting and cruising. These transmissions also adopted Dynaflow’s hydraulic accumulators and Powerglide’s vacuum modulator, adjusting operating and engagement pressures based on load and selector position. The layout of the hydraulic control system, which in complexity now fell somewhere in between Dynaflow and Powerglide, required a new shift pattern: PRNDHR (or PRNDGR) rather than the GM’s previously obligatory PNDLR pattern.
Another unusual and somewhat radical move, at least for the late fifties, was the use of die cast aluminum for the transmission case and the tail housing; cast iron was used only for the hydraulic valve body. This was more expensive and posed some significant manufacturing challenges, but it saved quite a bit of weight. In fact, Chevrolet claimed that Turboglide weighed a substantial 88 lb (40 kg) less than Powerglide, which at that point still had an iron case.
The point of all this complexity is easy enough to see. Both triple-turbine automatics were what we would now call continuously variable transmissions, offering a highly respectable amount of torque multiplication over a broader range of speeds than any previous automotive torque converter. With its stator blades at their low angle, Turboglide provided a stall ratio of 3.8:1 at a nominal 1,700 rpm, better than the dual-turbine Variable Pitch Dynaflow could manage at full throttle. With the throttle floored to shift the stator blades to high angle, Turboglide’s stall ratio rose to 4.3:1 at a nominal 2,700 rpm, better than Powerglide could offer even in Low. Since Flight Pitch Dynaflow’s stator blades were infinitely variable, Buick quoted only a single ratio: 4.5:1 at a nominal 3,200 rpm in 1958, rising to 4.7:1 for the 1959 Triple Turbine, which had revised impeller and second turbine blades.
On paper, at least, it appeared that GM had finally created the ideal automatic transmission: lightweight and perfectly smooth, with ample torque multiplication. Being (marginally) less complex than some rivals, it also promised to be more reliable. Unfortunately, the reality fell short of the sales pitch.
It should be said that at least part of the problem was one of perception. The triple-turbine transmissions’ torque multiplication depended on keeping the turbine speeds (and thus the speed of output shaft) well behind the speed of the impeller for as long as possible. While that was also true of Twin-Turbine/Variable Pitch Dynaflow, the triple-turbine units’ shorter gearing made the gap between engine speed and output shaft speed more pronounced and thus more noticeable. With an aggressive launch, the speed of the third turbine and output shaft might not approach the speed of the engine until the car was moving more than 50 mph (80 km/h), which could leave the uninitiated driver fearing that the transmission was about to self-destruct. Since the lag in rotational speeds did not directly reflect the transmission’s mechanical efficiency, the slippage wasn’t as nearly dire as it seemed, but it was disconcerting, if nothing else.
As with the dual-turbine Dynaflow, the nonlinearity posed a bigger problem when it came to passing response. Unless output shaft speed fell below about 60% of engine speed, the stator was the sole source of torque multiplication for passing. That was often marginal unless the stator blades were at their highest angle, which even with Buick’s infinitely variable stator was only obtainable with the accelerator floored. Compared to the convenience of Hydra-Matic’s part-throttle kickdowns, this was frustrating, making it seem that the transmission had to be constantly thrashed to provide adequate performance. Naturally, this style of driving did nothing good for overall fuel consumption, although steady-speed economy wasn’t terrible for this era. (Buick nonetheless hedged its bets for 1959 by numerically lowering the standard axle ratio for Triple Turbine cars to 2.78, compared to 3.07 for Twin Turbine or manual shift, which improved fuel economy at further cost in performance.)
Exacerbating this exasperation was the fact the triple-turbine transmissions had no Low range. If the 1.82:1 ratio of Dynaflow’s Low gear was less than ideal, it nonetheless provided immediate relief for any shortage of midrange punch and, with typical late fifties axle ratios, it could be used up to about 60 mph (97 km/h). Turboglide and Flight Pitch/Triple Turbine had only the hill retarder/grade retarder, which was similar to Dynaflow and Powerglide’s Low range only in its position on the selector and was intended for slowing down, not for accelerating. Anyone who shifted from Drive to GR thinking it would improve passing or hill-climbing power was quickly disabused of that notion. (The owner’s manual cautioned against engaging the hill clutch at more than 40 mph (64 km/h), lest you overheat the torque converter.)
As for reliability, it was initially quite poor for both Turboglide and Flight Pitch Dynaflow. One problem was the aluminum case; although aluminum transmission cases would become very common just a few years later, aluminum die castings of this size and complexity were still at the bleeding edge of GM’s metallurgical capabilities (a problem that also dogged the early Buick/Oldsmobile aluminum V8s). On early units, it was not uncommon for the case to crack or split, particularly if the transmission was overheated. It also appears that Chevrolet, at least, underestimated the demands on the clutches — particularly in the area of heat dissipation, which was the primary rationale for the subsequent switch from cone to multi-disc clutches. Even then, the clutches had to be beefed up several times and their engagement pressures increased (among various other changes). Many of the early issues had been addressed by 1959, but neither transmission ever lived down its checkered reputation.
Even if the triple-turbine automatics had been 100% reliable, we suspect that many buyers would have had difficulty seeing the point. That a great many American new car buyers of the time preferred automatic transmission is beyond question, but the need for multiple automatic transmission options was a good deal less obvious. Both Powerglide and Variable Pitch Dynaflow/Twin Turbine certainly had their flaws, but by the late fifties they were well-proven and worked well enough for many customers. The operating principles of Turboglide and Flight Pitch/Triple Turbine are complex enough to mystify even many automotive writers, so it’s easy to imagine the befuddlement of contemporary buyers trying to decide whether the triple-turbine transmissions were worth the attendant price premium. Turboglide’s continuously variable smoothness was a relative novelty for Chevrolet, but for Buick buyers, the dual-turbine Dynaflow, which was also functionally a CVT, was just as smooth. Therefore, the pricier transmission’s notional advantages were probably lost on all but the most technically savvy shoppers.
The upshot of all this was that most buyers shied away, which made both triple-turbine automatics costly failures. Since they shared very little with other Chevrolet and Buick transmissions (although Chevrolet later borrowed some Turboglide pieces for Powerglide), the tooling bill was immense — Buick alone spent a reported $86 million (around $730 million in 2016 dollars) — and warranty costs were high. The extensive changes necessary to address the various reliability problems can’t have helped; we don’t suppose that repeatedly redesigning Turboglide’s clutches was cheap.
Chevrolet, at least, was better able to absorb that expense. For Buick, the failure of Flight Pitch Dynaflow/Triple Turbine was yet another in a long list of calamities to befall the division during this period, doing serious damage to both sales and market share. The new transmission was certainly not the primary culprit — bigger issues included a newly recessionary economy, unpopular styling, and an assortment of assembly woes — but it added yet more red ink to the ledger at a time when Buick could least afford it.
The triple-turbine transmissions also marked an inauspicious period in the career of O.K. Kelley, who had left the Engineering Staff to become Buick’s chief engineer in August 1957. Less than two years later, Buick’s financial woes led to a major shakeup of the division’s upper management, beginning with the replacement of general manager Ed Ragsdale with Edward D. Rollert that April. Kelley departed about seven months later to a new post as chief technical adviser for GM’s Defense Systems Division. Even before he left, Buick terminated production of the Triple Turbine transmission, which vanished at the end of the 1959 model year.
Chevrolet continued to offer Turboglide through the 1961 model year, perhaps in the vain hope of getting their money’s worth. Experience with Turboglide did help Chevrolet engineers develop the Corvair Powerglide and the successful aluminum-case Powerglide (introduced in 1962–63), so it wasn’t a total loss, but all in all, it was not a particularly successful experiment. Looking back on it now, it seems like an intriguing idea that was under-developed and over-sold.
GM’s experience with the triple-turbine automatics was unhappy enough that these transmissions had no direct successors as such. (Kelley also designed a quadruple-turbine transmission, but nothing came of it.) However, some of their design elements did find their way into subsequent GM automatic transmission designs, as we’ll see in the next section.