I discovered earlier that in the recent update of the Hydra-Matic article, I had made an embarrassing error in the description of the 1937–1939 Automatic Safety Transmission. The error is now corrected, but if you want a more detailed explanation, see below.
The error, which I discovered in trying to fix an awkwardly worded (and technically inaccurate) sentence in the Automatic Safety Transmission sidebar (which appears midway through page 2), was based on a fundamental misunderstanding of the mechanical layout of that transmission. While this was very embarrassing, it was a perhaps understandable side effect of trying to cross-reference several closely related but not identical patents (not all of which reflected the final production unit) with a couple of small and unlabeled cutaway diagrams of the production transmission. This was compounded by the fact that the diagrams of the production transmission were reversed 180 degrees relative to the patent illustrations. Since none of the schematics actually depict the clutch assembly, the discrepancy is not immediately obvious unless you have a good idea what you’re looking at.
The Automatic Safety Transmission’s two planetary gear units look very similar externally, since each is surrounded by a big brake drum. From that, it’s easy to assume that both units are laid out the same way, although obviously they have different ratios. However, the elements of each unit are actually quite different. To whit:
- The first unit comprises a single set of epicyclic gears (one annulus, one sun gear, three planet gears, one planet carrier); the brake drum is connected to the sun gear. The annulus drives, the planet carrier is driven, and the sun gear is alternatively braked (for reduction) or locked to the carrier (for direct drive).
- The second unit comprises two sets of interconnected epicyclic gears (two sun gears, two sets of planet gears, two planet carriers, two ring gears). The brake drum is connected to the first annulus. Both sun gears drive. The first annulus is alternatively braked or locked to the first sun gear. An extension of the first planet carrier (which is driven) becomes the second annulus, which then drives. The second carrier is driven and drives the output shaft.
The first production Hydra-Matic, the version used by prewar Oldsmobiles (the Model 180) had a simpler second unit with only one set of epicyclic gears. However, the early Cadillac version (the Model 250) again adopted the compounded rear gearset, although the ratios of the second unit were different from the Automatic Safety Transmission and different from the Oldsmobile Hydra-Matic. That variation was the point of using compounded planetary gears for the second unit — any time you compound two sets of gears, the net result is equal to the product of their respective ratios, so using multiple gears gives you a wider choice of ratios. (It also makes it easier to vary the ratios for different applications without making the individual gears dramatically larger or smaller, which might entail more elaborate tooling changes.)
Based on the various patent filings, the engineers involved in the development of the Automatic Safety Transmission and Hydra-Matic seem to have been on the fence about the use of the compounded gears; some design iterations used them and some did not. After the war, Cadillac switched to a simpler non-compounded rear gearset. By the 1949 model year, Detroit Transmission had standardized that layout with a common set of internal ratios, presumably in the interests of simplifying production. [Author’s note: This is an addendum to my original erratum; like I said, this point has been confounding me.]
As a side note, the fact that the prewar and postwar Hydra-Matic had different internal layouts and different ratios was mentioned only in passing in the sources to which I had access (and then only in contemporary shop manuals). Furthermore, schematics or cutaway diagrams of the early single-coupling Hydra-Matic might show either layout depending on when the illustration was created and which version of the transmission it was intended to depict. I’m not sure later writers necessarily grasped the difference, which is why I’m harping on it here.
I’m embarrassed by the confusion, which was exactly the type of mistake I was hoping to clear up in the new draft, but if you’ve ever tried to puzzle your way through a patent application and its accompanying illustrations (much less a bunch of them), you may find the error understandable, if not excusable.
Author’s note, Feb. 3, 2016: I have further revised the above (and the original article) because I realized part of it was still wrong. The clutch packs serve to lock the driving member of each gearset to the respective reaction member, rather than locking the reaction member to the driven member as originally stated.