Cycling UK Forum

this page

http://www.disraeligears.co.uk/Site/Campagnolo_Sport_Extra_2180_derailleur.html

explains that the Campagnolo 'sport' series derailleurs were horrible single pulley things best confined to the dustbin of history





I have never owned, wanted to own, or fiddled about with any 'sport' type derailleur, and most of them came with a hanger bracket anyway, hence I didn't realise that the spring is anchored in the way it does. However it seems utterly baffling to me that someone would buy the most expensive, beautifully made dropouts available and then nail one of these horrors onto it....

Also it occurs to me that since Campagnolo already had the tooling to make spring-loaded top knuckles (for sport series derailleurs) this means that at any time from ~1952 they could have made a mech with sprung upper and lower knuckles if they had wanted to. As it was this 'first' was left to Simplex with their prestige model, and (AFAICT) Campag's first twin sprung mech was the first version of the Rally mech.

cheers

here (from Disraeli gears)



you can see a (~1963) guide to stripping a Campag 'record' rear mech. You can see the ball-bearing pulley design and the mounting for the cage plates. This mounting pattern was used in many other Campagnolo rear mechs, meaning that in many cases you can interchange the pulley cages between models.

I note with interest that the mech is described as being for 6s freewheels (even in 1963) but that when the (much lighter) Nuovo record mech was introduced a few years later, the offset in the lower pivot was increased, even though that was a 6s mech too. Presumably with the smaller offset the cage is swinging upwards into the low gear position, thus limiting the capacity of the mech.

cheers

Why Bother!!!

Period correct? Campagnolo introduced the Rally in 1974. By then any cycle tourist in possession of an open mind (and a remarkably small amount of money) was already using a SunTour VGT. Campag's Rally arrived on the scene too late, cost too much, and whilst it may have looked the business, it shifted worse than the Shimano Crane upon which it was modelled. (Shimano's indexed gearing ambitions had to wait for SunTour's slant parallelogram patent to expire, the rest is history!)

I remember going into Mercian's shop circa 1972, wanting something more reliable than my Shimano mech that randomly, without changing the adjustment of the stop screws, sometimes engaged but mostly either rattled uselessly against my moderately large 28T bottom sprocket, or threw the chain clean over it into the spokes! I thought I was going to have to empty my Post-Office savings book and get something by Campagnolo, but no said Jeff Bowler, bringing some other Japanese creation from under the counter: "You should try one of these. It's cheap as chips, but it works amazingly well!" Cheap as chips I could easily afford, so I took his word for it and was not disappointed. Within a couple of years, SunTour rear mechs were used by 90% of riders on CTC runs into the nearby Peak District, our favourite touring area, especially now we could all engage our bottom gears whenever we wanted, which was often!

I didn't even see a Campag Rally until the late 70s, and was not impressed. It had all the same vices as that Shimano mech I'd long ago discarded, despite weighing and costing much more! They were rarely used, and in my limited experience only by ill-advised tourists with more money than sense! So to be period correct, a 1970s bike should really have a SunTour V-GT, or Cyclone-GT etc.

For more witty observations on Campagnolo and Touring with Terror, or SunTour and the Curse of Duopar, I recommend the irreverently fascinating website: Disraeli Gears.

I was fortunate to have two of the 'original' (I'm pretty sure) version of the Rally, fitted as new equipment to a Jack Taylor 'Super Clubman' I bought in 1975, and a Bob Jackson tandem bought a year later. The original gearing with the former was 52/40 with 14-28; the latter 52/36 (soon changed to 52/34) with 14-28 (!) Both worked superbly over many years.
From what I remember everything except the long cage looked identical to the then-current Nuovo Record racing mech. I remember some years later seeing a later version still labelled 'Rally' but clearly of inferior construction. Not just the cage but the paralleleogram etc all looked cheap and nasty compared with my originals.
Although the two mechs were replaced on the bikes a long time ago (the tandem still has its Huret titanium Duopar, by the way), I'm pretty sure I still have the remains of one of them stashed away in a box. If it's another mucky day tomorrow I'll try and dig it out, and see what it has as the maximum range markings which Brucey's mentioned.

why bother? Well bikes from the 1960s and 1970s are fast becoming collector's items and having coughed up for restoration costs some folk want them to have a 'period correct' campagnolo ensemble. On a weekend toy/collector's item it doesn't have to be a 100% rational decision, any more than having the frame painted a particular colour is.

Which is all very well until you want to go up hills, at which point the gearing becomes problematic, and the limitations of 42-52 chainset and 14-26T sprockets become apparent. So if you want to stick to the brief then post 1974 a 'rally' mech is fair game but before that conversions are probably the best choice.

A lot of RDs back then were rubbish; they shifted poorly or wore out too quickly. Some did both. In the late 1970s I wore out mechs from Huret, Shimano and SunTour; several broke too. It was the same story with many other parts of the bike too. However once I'd scraped up enough cash I had a bike with Nuovo Record parts on it and they were far more reliable even if the shift quality still wasn't that good.

These days I just don't do as many miles and I'm unlikely to wear stuff out (or break it) in the same way. Many of these older mechs I am prepared to give a second chance to, but only if they can have lower gears. I have a soft spot for the campag stuff because I spent so much time racing on it, so I'm interested in long arm mechs and conversions just now.

One of the things that had passed me by, until I tried some experiments not that long ago, was to what extent shift quality was variously dependant on the mech, the sprockets and the chain. BITD I knew that some RDs shifted better than others and that such things as the gap between the guide pulley and the sprockets was very important. I got fussy enough that I converted my Sun Tour sprockets to have a UG-esque profile and got better shifting that way. I experimented with chains and of course RDs too. But, after that, I just used shimano HG stuff (mostly, because it just worked) and marvelled at the shifting improvement with every passing generation, without really questioning the relative contributions of RD, chain, and sprockets.

Anyway more recent experiments have been surprising to me: Turns out that the chain and sprockets are by far the most important factors and that (provided the guide pulley has reasonably tall teeth on it, so that there is some purchase on modern chains) things like the guide pulley gap are far less important. This means that RDs which were pretty lousy BITD can shift nicely these days, provided you use the right sprockets, chain and pulleys. Years ago I'd never have guessed it would ever be that way, so some of my RDs which have been gathering dust for decades are getting another airing....

For everyday riding, touring and so forth I will happily use whatever I can get that works, more or less. But for a weekend toy/collector's item different priorities apply.

cheers

This morning I've managed to disinter both my 1975 'original' Rally mechs. The outer cage plate on both is stamped 13-36 = 36-54 (not 53). They're identical except for the cable clamp screws. One of these has a plain hexagonal head, on the other the head is slotted. The latter doesn't look quite right somehow - possibly it was imported later from a different mech.
They're both intact except for a missing lower jockey wheel, and on one the two range limiting stop screws are missing.
I don't know why I've hung onto these when I've binned many others in the meantime. Possibly because they looked so elegant - still do. It's the twin thin bars holding the cable clamp assembly away from the outer paralellogram plate, as in the Nuovo Record, which make them immediately recognisable.
Incidentally, I don't remember any of the shifting problems with these which have been mentioned, even with the greater stresses on the tandem. Possibly because I used them well inside their theoretical maximum range.

If the pulley cog gap isn't such a problem life is a lot simpler if you use a hanger adaptor to lower the mech.
Ok may not be period correct but neither are those meccanoesqe mods either.

the whole point is that these 'meccano-esque' mods are period correct; they are exactly what folk were doing BITD.

Using a hanger extension addresses half the problem in this case; the total capacity of these mechs is very small, so you soon run out of chain tensioning capacity unless a longer cage is used. BITD a common setup was to use a 14-28 freewheel with a wide range triple chainset. The main reason you couldn't run that with a standard mech was lack of total capacity.

Chris, have you any pictures of your mechs?

cheers

rjb wrote:If the pulley cog gap isn't such a problem life is a lot simpler if you use a hanger adaptor to lower the mech.
Ok may not be period correct but neither are those meccanoesqe mods either.

Brucey wrote: So if you want to stick to the brief then post 1974 a 'rally' mech is fair game but before that conversions are probably the best choice.

Brucey wrote:the whole point is that these 'meccano-esque' mods are period correct; they are exactly what folk were doing BITD.

I think this is a problem with the whole concept of period correctness, whether it be bikes, houses, film sets... If you bought a bike in say 1972 and wanted lower gears, come 1974 you might well fit a Rally mech. Or something from Suntour or whoever, or a meccano mod. So your Campag-equipped 1972 bike would be, as it were, incorrect. Just like if you were making a film set in 1972 it would be more 'realistic' to include furniture, clothes, etc from earlier eras, rather than select from a 1972 catalogue. Realism is not the same as reality... But then, that's not the point. Some would like that bike in tip-top as-new concours condition, some would be happier to replace all components with 2019 stuff as long as it works.

You can make a bike how you like of course but I think the Hetchins I linked to upthread is a good example of a mostly 'period correct' machine. This machine was built in 1970 and from that time it has had a campag mech with an extended cage on it, in two different versions. You can go overboard with all this, but fortunately no-one expects to you to ride a bike with old tyres, lever hoods or brake blocks on it, and few expect you to run with old-school sprockets and chains with 'shift-unfriendly' shapes of various kinds. The line can be drawn where is suits you really, and screw on freewheels with HG style tooth shapes work really well these days, so you would have to have a good reason for using anything else.

FWIW I have not used the first two types of 'rally' mech back to back as it were, but all things being equal I'd expect the shift to be better with the twin-sprung design.

cheers

rjb wrote:If the pulley cog gap isn't such a problem life is a lot simpler if you use a hanger adaptor to lower the mech.
Ok may not be period correct but neither are those meccanoesqe mods either.

Pulley-cog gap still matters, in spite of the best modern designs of teeth and chain, believe me.

I lowered the gears on my 'Audax' bike by fitting an 11-36 cassette (10-speed) and mounting the Ultegra mech on a Wolf-Tooth hanger extender. It did the job and shifting was fine at the big bottom end of the cassette, but the greater distance between the mech and the smaller sprockets made the shifting a bit mushy at the top end, more mushy as the chain wore a bit (but still within 0.5%) so that it was sometimes necessary to overshift and back one.

I've since replaced that mech and extender with a Deore XT Shadow mech. And although the mech is made for 9-speed (so may be less precise and has a wider cage than ideally guides a 10-speed chain), and although I used a more flexible helical wound piece of cable casing for the shorter end-piece required for this mech: the shifting is uniformly more precise now, across the whole cassette. Given all those 'althoughs' the only reason this can be so is that this mech's steeper parallellogram slant (and unsprung, effectively fixed top pivot) ensures that the guide pulley more closely tracks the profile of the cassette.

It should be noted that those revolutionary SunTour mechs of the 1970s did NOT have a sprung top pivot. They didn't need one because the guide pulley tracked the cassette profile without it. And better without it. The sprung top pivot of contemporaneous Shimano and Campagnolo mechs let the guide pulley back away from the cassette under pressure, instead of stubbornly sticking close and shifting that chain! This sprung top pivot, a necessary feature in the absence of a naturally cassette-tracking parallelogram, was IMHO the cause of these mechs' inability to deliver bottom gear in the very circumstance when it was most needed! For reasons best known to themselves, but I guess probably because it makes for a slightly quieter transmission, Shimano retained the sprung top pivot when they copied SunTour's parallelogram slant, though they adjusted the spring tensions to largely disable its effect. Sram copied the SunTour concept fully, with an effectively fixed top pivot. (Not actually fixed because it nevertheless allow the mech to be pulled back manually when getting the wheel out, but effectively fixed at all other times by the chain-tensioning action of the mech's strongly sprung lower pivot.) And Shimano have eventually followed suit, starting with their 'Shadow' mechs for MTBs. The sprung top pivot is not a good thing in my opinion, merely the relic of an inferior and long ago superseded design of mechanism.

I had to opportunity to ride a vintage Raleigh Professional the other week, with full Campagnolo Record groupset. From circa 1980.

I have to say it was an absolutely rubbish ride, and the gear shifts were awful, slow, unreliable and very rough. Give me Shimano STI anyday. Even the first 7400 Dura Ace shift was incredible compared to vintage campagnolo. We've come a long way, in a short time for sure.

Polisman, what chain and sprockets were fitted on your test ride? [One thing you have to watch out for is that when test-riding an older bike/groupset, that the parts are really matched properly and in good condition; there is a lot of worn-out/mismatched old rubbish out there.]

Although they built a few NR-styled levers with retro-friction innards this wasn't a standard feature until C-record came along, and the usual campag NR friction gear levers (which could have metal on metal or metal on plastic friction parts BTW) were as draggy one way as the other, which goes a long way towards making the shift feel less good than it should, even if you are 'used to' DT friction levers, which few are these days. What often puts the nail into the coffin of the shift feel is that, if you are a strong rider on a lightweight (flexy) steel frame, frame flex in the BB area (in combination with the cable routing) can 'pull the lever back' for you, unless the friction mechanism in the lever is set abnormally tight. BITD most of my contemporaries used Simplex retrofriction levers which were a significant improvement. However I even had a set of these where the spring slip clutch lost it's preload, and these levers would be 'pulled back' in hard efforts regardless of how tightly the friction mechanism was done up.

SunTour 'powershift' levers were available from about 1970. The Hetchins upthread is fitted with them. In theory a small ratchet, pawl and spring can introduce more failure modes, but I can report that I have only ever seen one or two such levers where the mechanism has failed. The shift feel is pretty good with these levers, and some of my chums used to 'count the clicks' with these levers and therefore avoid much mech trimming.

The other thing that is a frequent stumbling block with old campag mechs is the match between the guide pulley teeth and a more modern chain. This isn't widely appreciated, so let me explain; 1960s or earlier mechs had round pulleys and the NR mech (and others through the 1970s and '80s) used pulleys with teeth that were a good deal smaller than those found on some other mechs. In both cases the pulley only has any good lateral traction on the chain if the inner side plates protrude well beyond the roller diameter in the chain. In most full-bushing chains, this is the case. So with this setup you have a chain that doesn't want to shift very well, but that is pushed by the RD because it has good traction on it. The usual result is fairly clunky but basically reliable shifting. However if you use a typical bushingless chain (which was commonly the case from the late 1970s onwards, from the introduction of the Sedisport chain) then you need more pronounced teeth on the pulleys if you want to be able to push the chain around; the inner sideplates do not stand proud of the rollers, so there is virtually nothing for the pulley to gain traction with. So with this combination you have a chain that ought to shift slickly, but the mech can't get hold of it. BITD I used sedisport chains and I found I only got acceptable shifting with a NR mech if I used a different top pulley and/or I used the best then available (e.g. UG type) shaped sprocket teeth.

Shimano's SIS indexed system, when it was launched, permitted the use of Sedisport chains; it was even printed in shimano's instructions; since this was the de-facto chain that most keen racers (and pros) used, this was one less thing to inhibit uptake of the system. The upper pulley on RD-7400 came with taller teeth than in campag mechs, an almost indestructible ceramic bushing, and the pulley itself was made of something incredibly hard-wearing, which keeps its tooth shape (and its lateral traction) much longer than almost anything else. Unfortunately it is also rather noisy in use, so I found this RD OK for road racing but a bit annoying when I used it in time trials; the noise was more like that you would get on another system that wasn't set up right, which was rather off-putting on tailwind legs. As CJ says this mech design was able to use both the (now patent-expired) SunTour slant parallelogram and a twin sprung 'servo-pantograph' layout.

In the very latest indexed systems shimano have started to use guide pulleys with even taller teeth which -even when the teeth are a bit worn- retain high levels of lateral traction on the chain. These pulleys often improve the performance of other RDs, provided the pulley gap is large enough to allow them to be used; if the gap is too small the sprocket teeth are more likely to hit the pulley teeth. I've seen mechs of this sort (with new chains and 11s sprockets) shift just as well (even on the small sprockets with a large pulley gap) with a hanger extender as without; amazing. I don't expect that performance to last once the chain gets worn but it is pretty unbelievable that it works that well to start with.

When it is in good condition I expect a twin-sprung (servo pantograph) mech to downshift better than a single pivot mech because when the chain is initially derailed, this is more easily accomplished; all that is required is that the mech body can start to deflect backwards (which takes little force) and the shift should be initiated. You can see that during every shift both upper and lower pivots articulate when the mech is in good condition. However the mech also moves whenever the bike goes over a bump, which means that between the two things, both pivots potentially see quite a lot of wear. A very common failure in twin sprung mechs is that the upper pivot (which is built compact for packaging reasons, even if this means the bushings are not very good) wears and makes the whole mech rather baggy. In the 1970s the upper pivots were less well engineered and (I found) wore out rather quickly. Single pivot mechs, in use, have one less pivot that moves and wears as you go down the road; simple as that, so often last longer. However even if the spring rates in the pivots are not perfect (and they are not; the same compact upper pivot housing contains a spring with a steeper rate than is ideal, whereas it ought to be closer to the spring rate of the lower tension spring) the effect of the twin sprung design is to extend the range of sprocket sizes over which a slant parallelogram design might be effective. You can see this in some shimano mech specifications; modern single pivot mechs have a minimum low gear sprocket size as well as a maximum, and they are not far separated. Twin sprung mechs have a wider range of low gear sprocket sizes over which they work well (for a while at least, until they wear).

So whether it is SRAM, SunTour or more recent shimano, if you use a single-pivot slant parallelogram mech, it works best and works well for longest if the cassette/freewheel profile matches the slant angle. In practice if you are touring and want low gears, it probably will do; you are most likely to choose the widest range cassette that the mech will comfortably work with.

But, such is the powerful effect on shift quality that

- the correct upper pulley
- modern chains and
- modern sprockets

can exert, even fairly archaic mech designs can now shift better than they ever did BITD; I was surprised just how much better they can (with the right setup) work these days.

cheers

Here are some pics of my two.

Note the wear on the top edge of the inner cage plate - frequent close encounters with the freewheel. An illustration of CJ's point about the top spring, perhaps?

ChrisButch wrote:Note the wear on the top edge of the inner cage plate - frequent close encounters with the freewheel. An illustration of CJ's point about the top spring, perhaps?

I reckon.

I guess when you have a side plate which stands up like that, it can maybe touch the neighbouring sprocket if the intervals between the (large) sprockets are smaller than a certain amount. I would suppose that the 'rally' mech would originally have been intended for use with a 'half step plus granny' type freewheel, in which the intervals are large.

FWIW those mechs do look like the very first version, with the unreinforced neck...?

cheers