Rear mech MTB designs old Vs new

For discussions about bikes and equipment.
Post Reply
Manc33
Posts: 2234
Joined: 25 Apr 2015, 9:37pm

Rear mech MTB designs old Vs new

Post by Manc33 »

On the newer MTB mechs (M772/M972) the b-axle has no spring. The mech has only 2 springs, the p-tension and the main derailleur body spring.

On older MTB mechs (M750/M952) the b-axle has a spring. The mech has 3 springs, the p-tension, the b-axle and the main derailleur body spring.

Does this mean on the newer mechs it is relying only on the p-tension whereas on older mechs, the b-tension helped to tuck the mech in at the smaller sprockets?

After trying both designs the newer mechs "reduce chain slap" I think by having that locked b-axle but is this "locking" taking accuracy out of the shifting?

Its hard to know, but if the purpose of an old (sprung) b-axle was to tuck the mech in at the smaller sprockets and the newer b-axle design has it locked, perhaps shifting has suffered slightly with the newer design?

If Shimano can 'do away with having a b-axle spring' why has it taken all this time? :wink: I think Dura Ace now has it, the very latest ones. Are they saving weight but along with it, compromising reliability, even if perhaps only very slightly?

How have they just taken a spring away lol. :lol:

I am sure it must have served some purpose. :!:
We'll always be together, together on electric bikes.
Brucey
Posts: 44695
Joined: 4 Jan 2012, 6:25pm

Re: Rear mech MTB designs old Vs new

Post by Brucey »

going back to basics, a rear mech has to have a parallelogram or a plunger to shift the chain from side to side, and at least a single pivot to accommodate chain tension variations. Leaving the added complication of the top pulley possibly not being centred on the lower pivot (and therefore moving when different chainrings are used) aside for the moment, you can see that the top pulley is constrained to move in a fixed path if the mech has a single pivot.

Pretty much every variation of parallelogram orientation and mounting has been tried. Some examples;

-chainstay mounted rear mechs with a horizontal parallelogram (Nivex etc, plus 80s/90s suntour, plus modern Dahon mechs
-Huret allvit
- Huret Svelto (and many others similar)
- Sun Tour slant parallelogram (first seen in a pressed steel mech in the early 60's)
- Campag Gran sport

Note that many of these mechs have a swivel at the hanger (to allow wheel removal), but this does not move when the mech is in use, so it is not a second pivot per se. All these mechs are single pivot (SP) mechs.

Of these some have a path that is only suitable for close ratios, or at least doesn't help to track the contours of a typical wide range freewheel. The Huret Svelto and campag gran sport are like this.

Others have a path that does help track a wide range freewheel; the Nivex does this, as does the Allvit, and so does the Sun Tour mech.
Amazingly the Nivex was the first ever parallelogram mech, yet it has a curved path for the pulley, that can be made to suit 'real world useful' wide range gear profiles, with fixed percentage intervals between the sprockets. The Huret Allvit was not a good mech in many ways, but is almost as clever as the Nivex, being in essence a Nivex style parallelogram, but mounted on a 'normal' gear hanger instead.
The Sun Tour slant SP parallelogram has a cruder (straight) path, but it is simple and the design can easily be varied (with slant angle) to suit different range freewheels. [IIRC the Gen 1 Campag chorus mech had a unique feature; an adjustable slant angle to the mech, so that the mech would work on a wide range freewheel or a close-spaced one.]

Sun Tour's slant parallelogram patent expired in about 1983, and Shimano (amongst others) immediately incorporated a slant parallelogram element to their mechs, starting with the seminal RD-7400. Prior to that, shimano had been using an idea originally pioneered by Simplex, in which there is a sprung 'B' pivot. Shimano called this dual pivot system 'servo pantograph' or somesuch.

in theory the Dual Pivot (DP) system can allow the mech to track any profile of gear cluster. All that is required is to have the correct balance of spring forces so that there is enough upwards force to overcome the weight of the chain lower run. After that the mech should sort itself out, adjusting itself as the angle of the chain run onto the sprocket varies. However the reality is that there are several things that stop this from happening in a perfect way

1) The weight of the slack chain run varies
2) there is friction in all the pivots
3) there will be free play in all the pivots
4) the springs in both pivots are constant rate, rather than constant force.

The last point is subtle but important. The springs in the pivots are fighting one another and must be made to fit inside small housings. This means they cannot be made constant force, and instead of the mech self adjusting to different sprocket sizes, it doesn't do this perfectly and indeed it cannot do this. This means that the DP method (in this form) only works over a small range. So adding a slant parallelogram to a DP mech simply means that the DP part has less work to do and can therefore work better. Campag, Shimano, Huret, Simplex, Sun Tour, (and others) all produced 'slant DP' mechs in the 1980s onwards. They worked OK too, but very commonly wear out (or seize up...) on the top pivot.

[BTW the flaws in basic mechs may instead be addressed by adding a second parallelogram; this gave rise to the Huret Duopar, the Sun Tour mountech, amongst others. These worked brilliantly when new, but were complex (expensive) and ultimately unreliable; all those extra pivots just wore out prematurely.]

In theory a DP mech will 'kick' on the upper pivot during some shifts, which may allow a smoother shift, even when the top pulley is running very close to the sprocket. In practice this isn't essential with modern chains/sprockets, plus the need to accommodate very different gear clusters with a single mech has been reduced; bikes are more specialised these days and fewer riders change the gearing appreciably; (they are more likely to buy another bike in fact.... :roll: ). This means that the slant SP mech has come back into play; despite being suited to a small range of different gear clusters, and having an 'imperfect straight path' they are simpler (cheaper) to make and can wear better than slant DP mechs.

It is as well to note that much of the improvement in shift quality in recent years has come from subtle changes in chain and sprocket shaping. This can allow slick shifts even when the top pulley isn't that close to the sprocket.

A slant DP mech can allow fast shifts with a wide range of chain and sprocket combinations, by running a close pulley. However these shifts may not be slick and some may be baulky if you try to 'force' a bad combination. If you try to run a slant SP mech with the same (bad) chain and sprocket combination, you will also get poor results. You don't need much wear in any of this kit to reduce shift quality again, either.

My personal take on all this is that modern gearing can work beautifully, but it is getting less and less versatile and it mostly wears (to the point that the shift quality is lost) too quickly. The 'dream' of a truly versatile rear mech (that copes with a lot of different gear systems yet is simple, cheap and robust) has been abandoned by the manufacturers. I think this is a mistake, and that it is still attainable. I even have some ideas of how this could be done.

However, in the meantime, I'm also becoming less and less interested in changing gear at all. I quite often ride for hours and only use two or three gears. In hilly terrain, I am happier than I used to be with just a few widely spaced gears. When I ride with other folk, they seem to be endlessly prodding at their gear levers for no good reason, which is just annoying, even if you are only watching it!

cheers
-------------------------------------------------------------------------------
~~~~~~~~~~~~~~~~~~~~~~Brucey~~~~~~~~~~~~~~~~~~~~~~~~
User avatar
barrym
Posts: 634
Joined: 22 Jun 2012, 10:05am
Location: Corsham - North Wilts

Re: Rear mech MTB designs old Vs new

Post by barrym »

Brucey wrote:However, in the meantime, I'm also becoming less and less interested in changing gear at all. I quite often ride for hours and only use two or three gears. In hilly terrain, I am happier than I used to be with just a few widely spaced gears.
cheers


+1

I've been in 4th on my 1 x 7 for months now :) , just summoning up the enthusiasm to remove the mech, cable and shifter. I've calculated the chain length and reckon it'll be fine without a tensioner :idea:
--
Cheers
Barry
Threevok
Posts: 195
Joined: 30 Sep 2016, 3:11pm

Re: Rear mech MTB designs old Vs new

Post by Threevok »

barrym wrote:I've been in 4th on my 1 x 7 for months now :) , just summoning up the enthusiasm to remove the mech, cable and shifter. I've calculated the chain length and reckon it'll be fine without a tensioner :idea:


Go on, do it. You know you want too. :twisted:

I love my single-speed
User avatar
barrym
Posts: 634
Joined: 22 Jun 2012, 10:05am
Location: Corsham - North Wilts

Re: Rear mech MTB designs old Vs new

Post by barrym »

You're right. Even when I'm browsing eBay, I only look at single speeds, or if something with gears catches my eye, I first look to see what dropouts it's got:-)
--
Cheers
Barry
Threevok
Posts: 195
Joined: 30 Sep 2016, 3:11pm

Re: Rear mech MTB designs old Vs new

Post by Threevok »

barrym wrote:You're right. Even when I'm browsing eBay, I only look at single speeds, or if something with gears catches my eye, I first look to see what dropouts it's got:-)



I'm actually considering building a 3x1 for my next project - just for fun 8)
User avatar
Gattonero
Posts: 3730
Joined: 31 Jan 2016, 1:35pm
Location: London

Re: Rear mech MTB designs old Vs new

Post by Gattonero »

Manc33 wrote:.... Are they saving weight but along with it, compromising reliability, even if perhaps only very slightly?

How have they just taken a spring away lol. :lol:

I am sure it must have served some purpose. :!:


I don't know, I never have this attitude of thinking that a manufacturer is always trying to cheat or loose its customers (as "users").
Shimano has removed one spring when has done the "shadow" derailleur, changing a design that was over 20 years old. You know, at some point you have to draw a line in the sand? And this is surely not a tragedy, especially when considering that "traditional" design is still widely available in one or two tiers down of groupsets; plus the new design made possible to use the "clutch" mechanism that is a pretty good thing when going off-road: no more chain slap, no chain loss, and 1x front ring setup is very reliable.
It is by riding a bicycle that you learn the contours of a country best,
since you have to sweat up the hills and coast down them.
Thus you remember them as they actually are...
Brucey
Posts: 44695
Joined: 4 Jan 2012, 6:25pm

Re: Rear mech MTB designs old Vs new

Post by Brucey »

good point about the clutch but also it arguably shows that Shimano are looking over their shoulders at SRAM, does it not?

cheers
-------------------------------------------------------------------------------
~~~~~~~~~~~~~~~~~~~~~~Brucey~~~~~~~~~~~~~~~~~~~~~~~~
User avatar
Gattonero
Posts: 3730
Joined: 31 Jan 2016, 1:35pm
Location: London

Re: Rear mech MTB designs old Vs new

Post by Gattonero »

Shimano has long been far behind in the Mtb world for their top-tier. Although the XTR Di2 is phenomenal.
It is by riding a bicycle that you learn the contours of a country best,
since you have to sweat up the hills and coast down them.
Thus you remember them as they actually are...
Manc33
Posts: 2234
Joined: 25 Apr 2015, 9:37pm

Re: Rear mech MTB designs old Vs new

Post by Manc33 »

On the old Shimano derailleurs the mech body itself clamps to the bike's derailleur hanger, making it all completely solid. With that setup the derailleur has no play in it.

On the newer MTB ones (and now from the look of it the latest Dura-Ace) it has a metal bracket that can move, that clamps to the bike, with the mech attached via that attached bracket - making it a looser connection. That's not my opinion, its physics. :P

I prefer the old way because the mech attaches more solidly. Time will tell I suppose, but I have seen complaints about those types of fixings on the M810/M772/M972 and so on and that connecting piece having a little play in it, but it physically has to because it is able to move (or you'd have a mech that could not move anywhere vertically).

One other thing about it is the newer design "locks" the mech there, meaning it has no spring in the b-axle, meaning on the smaller sprockets, you're relying on the design of the derailleur alone to make it shift well there, it can't tuck in... I think this is what they advertise as "avoid chain slap" sigh. :roll:

On a sprung b-axle the old way, the mech does tuck in at the smaller sprockets, that is the purpose of it being sprung in the first place. Maybe Shimano forgot why there was a spring at the b-axle :P
We'll always be together, together on electric bikes.
Post Reply