An often-repeated myth is that racing bikes before 1950 used half-step gearing. This is because the front chainrings often were relatively close in size, as they are on a half-step set-up.
However, both Koblet's bike in the 1951 Tour de France (see VBQ vol. 2, No. 2), as well as the 1960 Olmo in the upcoming issue, show that the gearing was far from half-step:
Koblet used gears like 48x52 with a 14, 16, 17, 18, 19 freewheel or a 48x50 with 14, 16, 18, 22, 24 or similar stuff.
Charting the first gear ratios show a lot of duplication:
48 52 14 93 100 16 81 88 17 76 83 18 72 78 19 68 74
The Olmo, according to the original sales sheet, used a 48x51 with a 13-26 freewheel. Assuming this was a 5-speed freewheel, the Olmo charts to:
49 51 13 100 106 16 81 86 19 68 72 22 59 63 26 50 53
Toward the middle of the range, it is almost half-step, but the chainrings really are too close in size. You get five pairs of gears with big jumps in between. (Half-step means that all gears are evenly spaced).
Charting Koblet's gears shows it is even further from half-step, yet they obviously spent lots of time thinking about their gears: Most racers then used different sizes of chainrings almost for every stage, much to the chagrin of the mechanics (which were provided by the organizers back then).
Also, the recently posted report from Jobst Brandt showed no evidence of half-step gearing, as far as I could tell. (The fact that he changed the freewheel ratios without changing chainrings indicates that it cannot have been half-step, as the two are interrelated with half-step: Chainring size difference must be half (in %) of freewheel cog size differences.)
So if it wasn't half-step, what was the reasoning behind those gears? Anybody know? Maybe Jobst would know. Does anybody have a contact for him?
Jan Heine, Seattle
Vintage Bicycle Quarterly