Cycling UK Forum

https://www.bikeradar.com/news/shimano-deore-m6100-12-speed/
...The RD-M4120-SGS and RD-M5120-SGS rear derailleurs are compatible with both 11- and 10-speed cassettes.
The cable pull ratio between 11- and 10-speed Shimano mountain bike groupsets is very similar, meaning that, according to Shimano either of these derailleurs will “work perfectly as an upgrade option for existing shifters on 10- and 11-speed bikes”.

...which is supported by Shimano's website:
https://productinfo.shimano.com/#/com?cid=C-432&acid=C-435

Over the years I've learnt - to my cost* - that measurements of cable-pull and unofficial compatibility charts between various brands and speeds of shifters and deraillers have been more often wrong than not.*

It appears that current wisdom has it that Shimano MTB 10spd and 11spd differ in terms of:
1. Sprocket pitch/thickness (duh),
2. Shifter cable-pull
3. Derailleur actuation ratios.**

It's not like Shimano to bless things that 'nearly' work, so my money is on the 10spd/11spd actuation ratios being identical after all.

Is there a good repeatable way to measure the end-to-end average-actuation ratio of a mech? Or has someone already done this and proven me wrong in advance

Ta,

Tom


* A 100% failure rate in my admittedly limited experience:
Shimano 9spd spacing doesn't work with a 10spd Shimano mech and SRAM 2:1. Not even close.
And nor do Campagnolo 10spd ergos quite work 7/8spd Shimano derailleur/cassette. It's out at one-end or t'other though I've been putting up with it for tens of thousands of miles...
https://www.cyclinguk.org/cyclists-library/components/transmission-gears/derailleur-gears/shimergo#Shimergo
**I appreciate the last two aren't linear...

actuation ratios are not perfectly linear through the stroke and this non-linearity varies with RD design and hence between manufacturers. When I've made 'mixed marriages' work I've usually ended up having to respace the sprockets slightly in order to match the movement of the RD more exactly. Typically even if the average is OK things get out of kilter at some point or other.

I was given to understand that the shift ratio between 10s dynasys and 11s dynasys wasn't the same, but if shimano say it is, then it is.

One way of comparing it is to simply try it and see. Another (which I have used) for comparing RDs is to take a 'power ratchet' type shifter (which pulls cable in uniform small increments) and then measure the incremental and cumulative RD movement which results.

FWIW you shouldn't lump shimano 7s and 8s in the same boat; they don't have the same spacing. Campag 10s shifting doesn't have uniform cable pulls through the range so it will never be an exact match for a cassette that does, and in addition to that the shift ratio variation in campag 10s is not an exact match for shimano. And I don't think there is any such thing as a SRAM 2:1 shifter for 9s either. Maybe I'm not clear about what exactly you have tried there (or what you find acceptable either) but 'working well' appears to mean different things to different people.

IME if you want a mixed marriage to work perfectly, hook the shifter to the mech and measure the RD movements carefully. If a conventional cassette can be respaced to match the RD movements, do that and you will have a good system. If it can't, try something else instead.

cheers

Brucey wrote:take a 'power ratchet' type shifter (which pulls cable in uniform small increments) and then measure the incremental and cumulative RD movement which results.

I will have to see if I can pick-up a cheap-y on eBay. But how do you accurately measure axial movement of the derailleur itself? I have unused Shimano 9spd,10spd and 11spd derailleurs and a Campag 10s to-hand that I could measure.

Brucey wrote:FWIW you shouldn't lump shimano 7s and 8s in the same boat; they don't have the same spacing.

Yup, understood. What I wrote was ambiguous: I have an RD-7700 derailleur on an 8spd Shimano block with Campag 10s ergos.

Brucey wrote: Campag 10s shifting doesn't have uniform cable pulls through the range so it will never be an exact match for a cassette that does,

The whole non-linearity bit is what I'm most curious about... struggling to see why it's more complex than a cos function that accounts for the slant. In the plane of the derailleurs slant it describes an arc about either of the fixed pivots, no? Although these pivots axes change their distance from the hub (as the upper knuckle rotates around the hanger bolt) I can't see why that would affect the rate of axial movement. I may be missing something though...

Brucey wrote:And I don't think there is any such thing as a SRAM 2:1 shifter for 9s either.

Yes, I'm talking cock. I couldn't find the original page(s) that declared an original 1:1-style 9-spd SRAM shifter worked perfectly with 9spd shimano block and 10spd shimano derailleur. But I see several forum posts elsewhere recommending this combination - the aim being to get a 9spd derailleur with clutch. I now see you can create a Frankenstein mech with the still available 9spd XT and punching-out the lower knuckle,clutch and cage from a donor mech which is my kind off hackery...
http://faqload.com/faqs/bicycle-components/drivetrain/derailleurs/9-speed-clutch-rear-derailleur-mash-up

Thanks,

Tom

cheers[/quote]

In the same press release they unveiled new Alivio M3100, the shifters no longer have 2 way release like M430 and M4000 and have under the bar gear indicator, it's basically a redressed Altus M2000, the rear derailleur also looks like the M2000.

the non-linearity (in campag ) arises from two sources

1) because the sprockets are unevenly spaced too (this is to do with ramp alignment in the middle of the cassette I think) (*)

2) because of parallelogram geometry.

In its simplest form in most rear derailleurs, the cable pulls diagonally across an articulated parallelogram. This gives an almost linear lateral movement when the cable is across the longer diagonal of a parallelogram but this accounts for only half the stroke; in the other half the cable is across the short diagonal of the parallelogram and the lateral movement becomes increasingly non-linear. In practice there are modifications of this geometry used, in which

a) the pinch bolt is connected to a cantilevered protrusion of one long side plate and/or
b) the fulcrum (cable stop) is not coincident with any of the parallelogram pivots.
c) the cable can be wrapped around a radiused piece near the pinch bolt so that its line of action varies differently through the stroke.

All affect the extent and exact nature of the non linearity, and it ain't always obvious what is going on. So the best plan is to measure the movement of the RD in response to known cable movements as previously suggested, either with a shifter that you hope is going to work or with a shifter that pulls many small uniform increments.

The simplest setup is to install an old hub or axle in a frame (with good hanger alignment), which has connected to it a perfectly orthogonal bracket. An easy way of getting this setup is to fit a single sprocket, larger than the mech will cope with (or extended with a bracket clamped exactly parallel to it), to a cassette hub, along with a stack of spacers so that it is held securely by the cassette lockring. You can then measure the lateral gap between the RD top pulley and the large sprocket, eg using digital Vernier calipers.

In practice unless this lot is offset leftwards (e.g. by respacing the hub) this won't let you measure the RD stroke into the #1 sprocket position, but you may decide you don't need this measurement exactly anyway; the RD should pull against the stop screw (the first and last pulls in most indexed shifters are slightly longer than the (trend of the) others in order that the stop screws determine the RD position in top and bottom gears).

(*) FWIW many indexed shifters don't use exactly equal/uniform trending cable pulls because of a related effect to 1) above. For example shimano have never endorsed the use of 'road' 9s shifters with MTB mechs and cassettes; the reason is that in any shimano 9s setup one of the intermediate pulls is uneven, so that it works better with the intended cassettes, which are uniformly spaced but vary in their construction (where the worst ramp mismatch is) depending on the sprocket sizes. So MTB cassettes and shifters have an uneven cable pull between sprockets 5 and 6 (or something) but in 'road' shifters and cassettes the uneven cable pull occurs between different sprockets, say 4 and 5 instead. In this case the discrepancy is very small and in practice plenty of people use non-recommended combinations and don't even notice that one of the shifts in the middle of the cassette is slightly worse than normal. if you wanted to compensate for this it would entail reducing the spacing between two of the sprockets very slightly and making a similar increase elsewhere. Or just putting up with it....

cheers

Fascinating - thank you!

I shall report back when the lockdown is over and I have some free-time again to make some accurate measurements...

I'm guessing they are going to use 2 cable grooves on the cable bolt 1 for 10 speed and 1 for 11.

pwu wrote:I'm guessing they are going to use 2 cable grooves on the cable bolt 1 for 10 speed and 1 for 11.

that is an interesting thought; I checked the dealer manual applicable to RD-M5120

https://si.shimano.com/pdfs/dm/DM-MBRD001-06-ENG.pdf

e.g. p15 and there is no sign or mention of a second cable groove...?

cheers

this looks like quite the mess of re-labelled parts

12-speed (13t jockey wheels)

XTR
M9100 1x10-45t and 1x10-51t
M9120 2x10-45t

XT (heavier)
M8100 1x10-51t
M8120 2x10-45t

SLX (worse jockey wheels)
M7100 1x10-51t
M7120 2x10-45t

Deore (same as SLX)
M6100 1x10-51t

11-speed (11t jockey wheels)
XTR
M9000 triple (SGS) with 40t, double with 42t, single with 46t, or GS same but no triple option

XT
M8000 ditto

SLX
M7000-11 ditto but SGS only



....

However M5100 is 11-speed but appears to be identical to M6100 (12-speed), with the 13t jockey wheels and a claimed spec of 11t (because 11-speed) - 51t.

Meanwhile M5120 claims 11/10-speed, but is not a new 13t (12-speed) design; it looks like a repurposed M7000-11

M5120 vs M4120 is the analogous to M6000 vs M610, i.e. M4120 lacks the clutch feature. It's rated for a double 11-42t and 11/10 speed.

So it looks like 12-speed = 11-speed = 10-speed, however some of the RDs are designed for say 50t large cogs, and others are designed for say 36t.

There is now:

9-speed 11-36t
10-speed 11-36t, 11-42t, 11-46t
11-speed 11-42t, 11-46t, 11-51t
12-speed 10-51t

So most of the supposed benefits of adding an extra gear appear to be absent, as what looks like an old SLX 11-speed derailleur is advertised as supporting 2x11-42t 10-speed or 11-speed, and 1x11-46t. So after getting everyone to 'upgrade' to expensive new systems to support 2x and 1x, the old ones could do it already.

I always appreciate your forensic analysis of Simano's groupsets thelawnet...good one