Talk:Queerbox

ZF?

The page is in the category “ZF Friedrichshafen transmissions” but it’s not stated what their involvement was. Mr Larrington (talk) 09:58, 25 September 2021 (UTC)

In the Lotus 16 article, it's stated that a ZF limited-slip differential was used. BMJ-pdx (talk) 15:58, 31 December 2023 (UTC)

pradowaits

This should be added somewhere. It was on the Lotus 12 article.

"Queer Box" should refer to the gearbox, not the entire car

It's implied that the car "earned the nickname 'Queer Box'". But according to Innes Ireland, it was the gearbox that was the "queerbox". He describes it as being "built on motorcycle principles" -- the gear selection being done by pins between the shaft and the inside diameter of each gear. That makes them highly stressed, and Ireland says it worked fine when brand new, but when worn, it was, as he describes it, a "box full of neutrals". (This is from the "Lotus Story" video set, Part 2.)

Unfortunately I can't edit this into the article, because it says "the car had a (long-undiagnosed) oil starvation problem", and it's unclear whether that means the engine or the gearbox. If it was the engine, then it's not the cause of the moniker "Queer Box" as stated; if it was the gearbox, that would certainly contribute to wear. So I don't know how to make the correction.

Can someone provide more info on the oil starvation?

BMJ-pdx (talk) 16:35, 1 August 2017 (UTC)

The 'Queerbox' was the transaxle of course. Nor was it a gearbox, it was a transaxle. I've not read anything about Lotus or Chapman that ever really explained what the problems were - all the good writing is in relation to Keith Duckworth, who started out working for Lotus and was told to fix it, before leaving to start Cosworth. According to Duckworth, because Chapman wouldn't pay up to make the Queerbox fix that Duckworth wanted.

There were two problems: one with the gears, one with the final drive. The gear shifting problem was about the length of the splines, the crownwheel and pinion failure problem about poor lubrication.

Duckworth fixed the crownwheel problem by arranging thin tinplate deflectors and moving the oil jets, so that the oil went where it needed to. It was then difficult to assemble, because there wasn't yet an appreciation of how crucial this was, but it was a fix. Duckworth had to go to ZF in Germany (who were building them) to explain this - BTDT, I pity anyone who has to explain any sort of engineering fix to a German engineer.

The shifting problem was because the Queerbox was another of Chapman's brilliant brainfarts: utterly innovative, impossible to build in a production car, and sometimes workable enough to be a winner on a racetrack - or at least, until the wheels fall off. The Queerbox was (more or less) a motorbike sequential shift box - except that the shift mechanism was moved from outside the gear clusters to inside the input shaft gears (yes, inside the gears - there's a reason it was called the Queerbox). This avoided the whole motorbike business of overhung forks and that drum with the cam tracks on it. Because that never worked on motorbikes, did it? Here's a Queerbox internal photo - all I can find. Well, it's the gears at least: [1]

The pinion shaft gears are selected by the mechanic, popped onto the shaft and fastened in place. Easy. The input shaft gears are placed correspondingly, on the input shaft. Where they can rotate freely. The gears are in constant mesh, but are free on the input shaft. On a motorbike gearbox (or something sensible like a Hewland), there are dog clutches to engage one gear at a time to the shaft. These clutches are poked into place by forks from the side, all controlled by the cam drum.

For the Queerbox though, the gear selection is squeezed into a cylindrical sleeve between the input shaft and the gears(!). This 'Selector Sleeve" is splined internally to lock it to the input shaft (but sliding back and forth) and it has a ring of dogs on it, which can engage one gear at a time - the sleeve slides along the whole length of the gear stack. There's also a "false neutral" cheesemetal (zinc-aluminium, later bronze) spacer between each gear position, because if two gears ever engaged at once, the selector dogs would just get chewed to pieces. The proper neutral is at one end,

OK, so I lied. First gear is different. That's a sliding mesh gear, because it's used for reverse too (and can't be engaged except when stationary). But I'm not explaining how 1st gear works without gin and pictures.

It's brilliant. Compact, lots of ratios, easy to change ratios and the absolute minimal movement to make a gear change. So why doesn't it work? When it's perfect, it mostly does. But the gears have no real bearing on the selector sleeve and there are no clearances, so the selector sleeve has to position itself perfectly in the middle of each gear. There is no location stop for this. There's barely any location control relative to the casing, and nothing between the sleeve and the gear. So if the gear's axial location becomes at all vague (and it does) then the sleeve is flapping around trying to find it, where it might not be. The gears don't even run in a single position on a sensible shaft - they do their best to find some sort of bearing surface on that selector sleeve, and it's not even one position on the sleeve, because the sleeve is being continually slid back and forth. So any wear on the sleeve and the gears can move around - radially, or nutating to tilt sideways. And the clearance from the gears to the cheesemetal spacers (and there are a dozen interfaces here, in a distance of only a few inches) is prone to wearing too. So the Chapman ur-gearbox and its precise fit just isn't a realistic comparison to a practical working gearbox, with the gears flapping about and missing the dogs. There are no baulk rings on these dogs, so the dog has to act as one itself. The big problem is that if they don't engage fully with the gear, the gear tends to kick the dog back out of engagement again, and into the adjacent false neutral.

Duckworth's fix for this was to make the gears wider - or at least their base thicker - giving a more stable bearing on the selector sleeve. I think it was him who swapped the zincoid cheese for bronze, making the whole box a few grams heavier and making Chapman impose a regimen of pre-race bulimia on drivers to compensate. Maybe Duckworth's fix would have worked too - but the trouble is that it made the gear stack longer, thus needing a longer gearbox casting, invalidating all the existing gearboxes and parts - something that Chapman just wouldn't pay for. Andy Dingley (talk) 17:52, 1 August 2017 (UTC)

Dubious -- lower polar moment of inertia

The "Lotus Twelve origins" section states that a transaxle configuration was chosen because it would result in a lower polar moment of inertia. That seems dubious, because in a front-engine car, a conventional gearbox location would be immediately to the rear of the engine, hence close to the fore-aft center, whereas with a transaxle, it would be close to the rear axle. (On the other hand, with a transaxle, the driver could perhaps be more forward, i.e., closer to the center. Maybe it boils down to which is heavier, a gearbox or a driver.) BMJ-pdx (talk) 15:37, 31 December 2023 (UTC)