The Definitive (But Wildly Inaccurate) Guide to Brake Cable Pull Ratios
 

Background: I'm a nerd. Shimano likes to take a very narrow view of compatibility between components. Those two things combined led to this post.

It's generally well known that Shimano has been tweaking their brake cable pull over the past few generations of their road components, but the details are very hard to come by. If you search the product specifications on the Shimano website you can find out which trademarks are associated with which components. They have "SLR-EV" (DuraAce 9000/Ultegra 6800/105 5800/Tiagra 4700), "New Super SLR" (7900/6700/5700/4600) and "Super SLR" (7800/6600/5600/4500). At some point there was also plain old "SLR" but I don't know where that cutover took place. Shimano vaguely claims that each of these is incompatible with the others, though much anecdotal evidence suggestions that's not entirely true.

From various sources (mostly discussion posts on this forum and others) I've gathered that the "New Super SLR" levers pulled slightly more cable than the "Super SLR" while the "SLR EV" pulls (roughly?) the same amount of cable as the "New Super SLR" but tweaks the rate at which the pull is changed throughout the cable pull.

All of this is perhaps interesting to extreme bike nerds but mostly not very helpful without some numbers to go with it.

In another thread (http://www.bikeforums.net/bicycle-me...l#post18380942) @andr0id reported as follows:

So I took some measurements of my own....

Let me preface by saying that I am absolutely terrible at taking precise measurements. On top of that, I was using linear measurements as an estimate of lever travel that is actually an arc and picking a more or less arbitrary point on the lever at which to measure the travel (though I tried to be consistent from lever to lever). Finally, it's my understanding that the rate of cable pull actually changes throughout a full sweep and for practical reasons I measured different amounts of pull with different levers.

Nevertheless, I think the data I came up with shows a reasonable resemblance to meaningfulness.

It turns out I've got a lot of brake levers in my garage. Half of these weren't actually on bikes, so I just mounted them to a spare set of bars. For each lever I tried to measure how far the part of the lever where I'd usually be applying pressure moves, and then I measured how much cable moved during that pull (measured in various ways, depending on how the lever was mounted). Here's what I came up with:

Lever

• Lever travel (mm)
• Cable pulled (mm)
• Ratio
• MA

Gevenalle CXV0 (= Tektro RL520/Cane Creek Drop V)

• 40
• 24
• 0.60
• 1.67

Gevenalle CX2 (= Tektro 200A/Cane Creek SCR-5)

• 39
• 15
• 0.38
• 2.60

6700-series Ultegra (New Super SLR)

• 27
• 12
• 0.44
• 2.25

5700-series 105 (New Super SLR)

• 26
• 11
• 0.42
• 2.36

4600-series Tiagra (New Super SLR)

• 17
• 7
• 0.41
• 2.43

Non-group R700 (Old Super SLR)

• 27
• 10
• 0.37
• 2.70

5600-series 105 (Old Super SLR)

• 28
• 10
• 0.36
• 2.80

6500-series Ultegra (SLR?/Old Super SLR?)

• 36
• 11
• 0.31
• 3.27

SRAM S500

• 43
• 17
• 0.40
• 2.53

Avid SpeedDial (flat bar, MTB)

• 43
• 32
• 0.74
• 1.34

Avid SpeedDial (flat bar, MTB)

• 43
• 25
• 0.58
• 1.72

The pull on the Avid Speed Dial levers is adjustable, so I measured once at each extreme.

I would guess that my margin of error on these measurements might be as much as plus or minus one millimeter, but I think you can see the trend lines. It looks to me like the cable pull ratio for long pull levers is around 0.6, the pull ratio for the old "Super SLR" levers is about 0.38 and the pull ratio for "New Super SLR" levers is about 0.42.

Prior to the introduction of "New Super SLR" everyone seemed to lump all "short pull" levers together, and so I have always kind of assumed that they all had about the same cable pull ratio. Within my generous margin of error, it looks like the SRAM levers and the short pull Gevenalle levers I measured are around the same pull as the old Shimano "Super SLR" levers, so that seems to be consistent with my expectations. The SRAM levers might actually slot in somewhere between new and old Super SLR, but my measurements weren't precise enough to say for sure.

I don't know where the 6500-series Ultegra levers fit in all of this. Just looking at them I expected them to be in the "Super SLR" camp, but my measurements didn't back that up.

I'd love to see someone who is more skilled as precision measurements than I am repeat this experiment.

EDIT: See post 18 in this thread for what I consider to be more reliable data.

Well done!

While accuracy is generally a good thing, I don't think you need to worry too much. Between hand strength, brake track surfaces, weather, pads, compression of casings etc etc there are so many other variables that I think you've done enough. This would be a be a good chart to indicate which combos that are good, possible and hopeless matches.

I've long suspected that brake part matching mostly isn't an either/or scenario anyways.
Back when I took a more active interest, I determined the (added) pull of a Travel Agent and compared it to the difference in pull between the Canti/Caliper levers and V-brake levers that I had available. And despite TAs being generally considered quite OK, it wasn't a close match.

Thanks for starting.

I've been meaning to start tracking shifter cable pull ratios. Most of it is online, but I wanted to verify the data myself. I'll try to build a rig to do it shortly.

Here is Sheldon Brown's Cable Pull Device.

http://www.sheldonbrown.com/images/cable-pull-tool.jpg
Sheldon Brown's Internal Gear Hub Cable Pull Crib Sheet

One of the issues with road levers is setting up a standard mount. Less of an issue with MTB levers.

It would also be good to convert cable pull to brake caliper movement.

I'm really surprised at the speed dial lever. I had assumed the dial would move it from V-brake long pull spec to cantilever short pull spec. Instead it seems to go from V-brake to some wacko extreme.

Yeah, that does seem wrong, and it's entirely possible that I mis-measured that one. It was one that I mounted on an empty bar, and having a different lever shape It didn't lend itself to consistent reference location. It may be entirely wrong. I have a short pull flat bar lever that I could add in as a sanity check for the SpeedDiall -- maybe I'll do that this weekend.

Thanks for this. Not so much for the actual data but more for the cause of not asking others and doing the measurements yourself. This is a lesson for others to follow. BITD John Schubert did much the same with the then available drop bar levers and calipers. Turned out the Universal levers had the least leverage. The next step is to decide what kind of performance one wants in their brakes. Max leverage/clamping pressure, the least leverage but the most solid lever feel, or something in between. Andy.

Your data shows the true "long pull" levers (the Gevenalle CXV0/Tektro RL520...) and the Avid Speeddials set to their V-brake setting do indeed have a lot more cable pull and lower mechanical advantage than any of the caliper brake levers. The caliper levers do vary a bit in pull and mechanical advantage but, except for a couple of outliers, are in a fairly narrow range. I can't see any of them being unusable with any of the current or former brake caliper models.

The Travel Agent gives about a 2:1 Pull ratio. The straight ones may be discontinued, but the 90 degree travel agents are designed to replace the noodle with the Linear Pull brake levers, and should bring one close to what is needed.

My pull measurement is pretty accurate. I put a piece of masking tape on the TT under the cable and another piece of tape on the cable and marked the starting and finishing points.

The lever pull was less scientific. I just squeezed until it touched the bars mounted in a normal position.

The different curves of the 3 levers I have are pretty inconsistent.



I also use a travel agent on my tandem to change my 10 speed Ultegra STI ratio to work with my V brakes.

I would definitely agree with this. I think there's more variation in leverage based on where different people grab the same lever than there is is between different levers in the short pull group.

Thank you for this. It's a good reference, even if the measurements of the lever are subjective.

Yeah, measuring lever travel was my biggest uncertainty too. For the levers that were mounted on bikes I used a digital caliper to measure the change in spacing between two reference points at the brake (for instance, between the anchor bolt and the opposing brake arm on caliper brakes). I feel reasonably confident of those measurements. With the brakes that weren't on bikes I squeezed the lever, pinched the cable were it exiting the brake then released the lever and measured how far I could pull the cable back. Those are a little sloppier, but still not as uncertain as my measurements of lever travel.

Perhaps a better way to measure is at the caliper. You should be able to measure the pad motion and dial in 5mm of cable adjustment using a barrel adjuster. For a given caliper you're measuring the pad displacement as function of cable pull.

Ah, I think you're on to something

Stick a 5mm shim between the pad and rim on each side. Then remove it and see how far the tip of the brake lever moves until the pads touch.

That along with the masking tape behind the cable gets you all three measurements.

You're measuring the ratio of the caliper but you're not learning anything about the lever. The other end of the system...

If you used the same caliper and changed the lever then the levers would be compared. Andy.

You'd still need an accurate way to measure the lever travel. It just gives you a less arbitrary end point. It may still be an improvement.

What I'd really like is to be able to produce a graph of lever travel vs. cable pull for each different lever.

I was poking around Google, and I stumbled across this article from Park Tool:

Travel Agent? Installation and Adjustment - Park Tool

The article points to an obvious simplification of the problem: measure the distance from the pivot point to the cable anchor (more or less) and you can calculate the cable pull from that. The 5700/6700 levers complicated things a bit in the the cable rests on part of the lever behind where it is anchored, so I had to measure to the point where the cable is free to move independent of the lever body.

Anyway, after measuring the radius of the virtual circle about which the fixed cable point pivots, you can calculate the amount of cable pulled in a 20 degree swing of the lever (for instance) by the formula:

C = 2 * ∏ * r * 20 / 360

or, if you're a math geek and see that the cable pulled will actually be a chord and not an arc:

C = 2 * r * sin(20/2)

which works out to be pretty close to the same thing.

From there, the kind of cable pull versus lever travel I was trying to measure before is mostly a function of lever shape and where you squeeze it.

So, grabbing my trusty digital caliper and returning to the garage, I found the following:


(Model

• Pivot radius
• 20 degree pull)

Gevenalle CXV0

• 32.5
• 11.3

Gevenalle CX2

• 20
• 6.9

6700-series Ultegra

• 23
• 8.0

5700-series 105

• 23
• 8.0

4600-series Tiagra

• 23
• 8.0

Non-group R700

• 22
• 7.6

Shimano 5600

• 19
• 6.6

Shimano 6501

• 18
• 6.3

SRAM S500

• 22
• 7.6

Avid SpeedDial

• 27
• 9.4

Avid SpeedDial

• 33
• 11.5

Shimano R550

• 20.5
• 7.1

Alivio MC11

• 22
• 7.6

I have a lot more confidence in these measurements than I did in my earlier ones. Most of this seems consistent with my table above, with a few notable exceptions.

1. My earlier Speed Dial measurements appear to have been way off (as was suggested previously).

2. The SRAM levers seem to be closer to the "New Super SLR" pull that to the old "Super SLR". My earlier SRAM measurements were probably bad too.

3. Despite my absolute certainty to the contrary, the R700 levers do not have the same cable pull as the 5600. This shocked me because they look identical holding them side by side. I measured both of them half a dozen times trying to find an error, but it seems to be what I've got here. Still not convinced, I dug up an old Shimano specification chart, and sure enough the R700 are listed as not "Super SLR". It's like the Michelson-Morley experiment all over again.

I also added a Shimano R550 lever (short pull, flat bar) and an old (pre-V-brake) Alivio MC11.

I used one going from a Campy road lever to full-sized V brake and it worked extremely well. Don't have any actual measurements but I can take some. I'm now running Mini-Vs and don't need the cable pull change, but you can set it for 1:1 pull and it acts as a roller noodle with an adjuster. This gives a noticeably better feel than a regular noodle. I was surprised, it feels almost like a rod linkage or something, very direct.

Can you get your hands on a 5800/6800/9000 levers for comparison with newest shifters? I'm wondering if you can also do the reverse and measure the distance from the pivot to the cable pull on the caliper, which should give you the total mechanical advantage.

It's interesting that SRAM is slightly different but still close enough for compatibility.

Sorry, no, I don't have any of the new series of levers.

Measuring the arm of a V-brake or mini-V is pretty trivial. I think dual pivot caliper brakes would be a bit more complicated, but you could probably still figure out a number from a couple of measurements.

Why not just measure the triangle between the pivot, cable anchor, and the entrance to the housing run (I'm not sure what to call this point). You can then extrapolate the cable pull for any angle of brake lever travel.

I think for any practical angle, the formula above will be a very good approximation, probably well within the margin of error for the measurement. In any event, as others have commented, I"m not sure how much precision is really necessary here. My main goal in this was simply to quantify the relative differences between levers, and this certainly seems to do that.

The one big mystery in leaves in my mind is why the 2:1 ratio for the Travel Agents works as well as it does. It seems like it would be too much pull. If my numbers are even in the ballpark I would expect 3:2 to be a better ratio for Travel Agents. Yet I've used them and they worked very well.

Just measured my Campag veloce 10 spd (and these are identical to all 10sp campag levers I think) and the pivot radius is 24mm.

My measurement system is I mount the brakes and levers I want to use and then try them.

By riding.

If they stop to my satisfaction then the measurements are good, if not, they are bad.

Haven't made any mistakes yet...