Discs or canti's on a new tourer

[MartinC]

has anyone run sums on the geometry? I confess I have not, through sloth....

Just wondering if the brakes compare well with cantis both in theory and practice.

cheers

I had a halfhearted go at this a while ago - trying to work out my options with running V's with STI's. The V MA's easy to work out 'cos the arms are a simple lever with the arm length and pad height determining it. It also doesn't vary with arm angle/yoke height. To compare it with a canti set up you need to have the precise details of the specific canti installation.

The more interesting bit is the amount of pad movement you get for the cable pull (the reason for having mini-V's) - this depends very much on the height of the pads above the bosses. The higher they are then the more pad movement you get, but the less mudguard clearance you get. I had mini-V's with 85mm arms with the pads at 30mm above the bosses, this gave a reasonable MA, a reasonable amount of pad movement (much better than full V's but not anywhere as good as V brakes with V levers) and a decent clearance for MG's over a 32mm tyre. This gave very acceptable braking performance with good pads, IMHO better than cantis.

What made me change back to full V's and Cane Creek (Tekro) Drop V brake levers was the lack of a quick release with STI's. There still isn't enough pad movement to leave enough clearance to undo the noodle unless you back the adjusters right off when you fix the cable - and this leaves you with little adjustment to take up pad wear unless you re-clamp the cable

[Brucey]

re the noodle issue;

you can file the end piece of the noodle so that you need less cable slack to disengage it. This has worked very well for some years on a friend's bike. If you put the flats on at an angle, it lessens the chances of accidentally unhooking the noodle.

I guess a typical mini-V caliper MA is about 3:1 then, although it could vary greatly with pad height.

This compares with;

~3.5 or 4:1 for a full 'V' brake caliper
~1:1 for an old-style side pull with about 50mm drop,
~1.3:1 for a typical DP with a 57mm drop

Both side pull (exactly) and DP (approximately) increase MA pro-rata with reduced drop. So I measure about 1.5:1 with short reach ultegras.

Now here is a question; even pretty crummy cantilevers struggle to go below 1.5:1 MA, yet there is a strongly held view that DP calipers are 'better' than many cantis. And few would leap to the conclusion that (say) mini Vs are about twice as good as DPs.....!

So, is this because;

a) the sums are wrong
b) it is mass hallucination
c) there is something else that systematically reduces the canti effectiveness (stiffness? flex vs pad angle?)

???

cheers

[531colin]

I just KNOW I will regret this, but.......
I just used CJ's method to work out the MA of a couple of cantis.....
This is CJ's method



herehttp://forum.ctc.org.uk/viewtopic.php?f=5&t=57410&start=30

I got the mid-profile canti I illustrated on the previous page as 3.66:1
I got a CR 720 as 2.14:1
(This seems OK, as the pads move a LOT further on the CR720 per squeeze of the lever than they do on my favourite mid profile)

And don't you need to consider the longer cable pull of a full-size V brake lever when making comparisons of MA?

Now a chap of my age ought to admit to some of his limitations, so I'll come clean, I can't do the maths....I can just about follow CJ's stuff in the thread I linked, but hell would freeze over repeatedly before I actually did anything like that myself. So I reckon the maths isn't actually wrong, its just different......you know, different conventions, working it out for one pad or the pair, or something like that.

Over to you, guys.......

EDIT.....I measured up for canti MA with the brakes "off"....I should have done it with the brakes "on" because the mid-profile canti (at least) will have a regressive action, ie the MA falls as the straddle goes up.

[Brucey]

I just KNOW I will regret this, but.......
And don't you need to consider the longer cable pull of a full-size V brake lever when making comparisons of MA?
Over to you, guys.......

re the cable pull for full 'v's, yes indeed. But all the other brakes are likely to be used these days with DP levers, so can (maybe) be compared...

also see the online cantilever brake Mechanical Advantage (MA) calculator applet

http://www.circleacycles.com/cantilevers/

difficult to get down to 1.5:1 with a canti, isn't it....?

[edit the MA convention I have used for my measurements is [Brake Cable Pull]:[Brake Block Squeeze]. I believe this to be the same convention as in the applet]

cheers

[531colin]

Brucey
Your link brings up some decorative but apparently useless lines.....I wouldn't be able to use the "applet(?)" anyway....
But I do think theres a random bit of nonsense going on....
my CR 720 pads are 3mm from the rim, brake comes on about half lever travel, MA is 2.14:1
So a DP with an MA of 1.31:1 has 5mm pad clearance? does it??

[Brucey]

5mm clearance? maybe... if you have worked out the MA of the 720s correctly... but when the brake comes on and when it is effective are two different things...

Have a go with the applet; it might surprise you. You need to scroll down and then use the sliders to adjust the geometry. It isn't perfect, because it tells you MA VS straddle height, which is similar (but not identical) to what happens when you put the brake on.

If you want me to run the 720 geometry through the applet, let me know the numbers and I'll see if it agrees.

cheers

[Brucey]

But I do think theres a random bit of nonsense going on....

Colin, I think you might be right. I ran some numbers through the applet and -provisionally anyway- I now believe his definition of MA applies to a single arm not the whole caliper. If correct then (because the main cable tension is shared between two arms, not just one...) the applet gives MA values which need to be halved in order to be compared with the other brake values.

Re lever MAs (LMA);

A- V levers are about 2:1
B- V levers set for 'canti' pull are about 2.75:1
C- DP dropped bar levers are typically about 3:1
D- Old style dropped bar levers are about 4:1
E- Old style 'four finger' MTB levers are about 4:1

If unknown, measure the lever travel where you pull with your fingers and divide this by the cable pull acheived.
[2019 edit; 'New Super SLR' (NSSLR) levers (most new model shimano STIs from 2008) have a MA of approximately 2.5:1]

Re caliper MAs (CMA)

F- short reach side pull 1.1:1
G- DP caliper 57mm reach 1.3:1
H- DP caliper short reach 1.5:1
I- old style cantis (& many centre pulls) 0.8:1 to 1:1
J- medium profile cantis with optimal geometry 1.5:1 to 2:1
K- mini Vs 2.5:1 to 3:1
L- full v's 3.2:1 to 4:1

If unknown, measure the cable pull and divide by the 'squeeze' or total pad travel (left and right).
[2019 edit; 'New Super SLR' (NSSLR) calipers (most new model shimano DP brakes from 2008) have a MA of between 1.5:1 and 1.75:1]]

System MA (SMA) is the multiple of the two; some examples-

Full V setup; AL ~8:1
Mini V with DP levers; CK ~7.5:1 to 9:1
DP setup with 57mm drop; CG 4:1
Old school cantis and levers; DI 3:1 to 4:1
DP levers and old school cantis; CI 2.5:1 to 3:1
70s/80's race brake; DF 4.5:1
DP levers with optimal cantis; CJ 4.5:1 to 6:1

If SMA is unknown, measure the lever travel where you pull the lever with your fingers, and divide by the total pad travel.
[2019 edit; 'New Super SLR' (NSSLR) levers/calipers (most new model shimano STIs/brakes from 2008) have a SMA of approximately 2.5 x (1.5 to 1.75) = 3.75:1 to 4.4:1]

The numbers above tally reasonably well with my experience.

Obviously a higher SMA value also gives a proportionally reduced pad clearance all being equal, and will need more frequent adjustment.

Note that the same SMA value will (at a lower cable tension) give a slightly more effective brake, because cable friction losses will be lower.

Note also that a 'flexy' feeling brake may need to be run with closer pad clearances, otherwise the lever might bottom out before the brake come on with any great force. Thus stiffer brakes can be run with higher SMA (and reasonable clearances too) vs a more flexy brake.
---------------------------------------------------------------------------
2020 edit; More information about which STIs and which brake levers are NSSLR and which are not.

As of 2019 the following models are listed as NSSLR type brake pull

ST-R9100 ST-R9150 ST-R9160 ST-9070 ST-9071 ST-9001 ST-9000 ST-R8000 ST-R8050 ST-R8060 ST-6870 ST-6871 ST-6800 ST-6770 ST-6700 ST-5800 ST-R7000 ST-5700 ST-4700 ST-4600 ST-R3000 ST-3500 ST-R2000 ST-2400 ST-R460 ST-R353 ST-R350 BL-TT79
and only

ST-A070 ST-A073
are both current and use the older brake cable pull suitable for original type DP calipers.

listings for 2012 show the following NSSLR levers

ST-7970 ST-7900 ST-6700 ST-5700 ST-4600 BL-TT79 BL-4600 BL-R780

and the following with shorter cable pull for original DP calipers

BL-TT78 ST-3400 ST-2300 ST-R221

NSSLR was new for 2008, in Dura-Ace only to start with. Thereafter new model STIs in groupsets nearly all use NSSLR cable pull. AFAICT the only 'groupset' (i.e. not non-series) STI model introduced after 2008 (i.e. 2009 or later) with the old brake cable pull is ST-2300.

The quickest way of checking the introduction date of groupset parts is to look at this page

https://en.wikipedia.org/wiki/Shimano

cheers

[531colin]

WOW....did I cause all that?
So far I havn't done well with the applet, I have found a diagram with various lengths on it, but I can't find where it tells me what to measure, or what to do with the measurements....?
If you want anything measured, we have lots at Spa, here at home I only have CR720s on the back ('cos they don't squeal!) but difficult to measure with carriers in the way.
Does CJ's 2 ruler trick for canti. MA agree with the applet? (he is often right )

[Brucey]

Hi Colin,

CJ ('Cyclenut' when at home)'s post read thus;

No I can't explain Sheldon's straddle wire MA, but I can explain the MA of a cantilever brake.

In the force diagram below:

T = brake cable tension
S = straddle tension
F = force on the rim from each brake block
h = height of the block (centre) above a horizontal line through the arm pivots
a = perpendicular distance from straddle to arm pivot
b = distance along the aforementioned horizontal line, to the intersection with an extrapolation of the straddle
A = angle of the straddle to the horizontal



Canti.gif (1.67 KiB) Viewed 1246 times

(1) Balance of turning moments about the arm pivot: S×a = F×h

(2) Resolution of vertical forces on straddle yoke: T = 2×S×sinA

(3) sinA = a/b

Substitute sinA in (2): T = 2×S×a/b

Substitute F×h for S×a in (2): T = 2×F×h/b

Mechanical advantage (MA) is F/T = b/2h

Sorry Sheldon, but you were wrong!

The angle of the straddle does come into it, because flatter makes b longer. But when you start to apply the brake, a flat straddle rapibly becomes less flat and b shrinks dramatically, which is why brakes that rely upon a flat straddle have a regressive action - loads of MA when you want movement, followed by movement when you want MA!

You can also see from this formula, that when the brake arms are horizontal, b is simply the length of the brake arm and the straddle angle becomes irrelevant. It follows that straddle angle becomes less important and the braking becomes more linear (which is generally a good thing) when the arms become closer to horizontal.

This formula and a consideration of how b varies as the brake is applied also demolishes the notion that there is something special about having the straddle at right-angles to the arm. But that may nevertheless be the optimum arrangement for some designs of brake and lever.

I cannot disagree with any of this; I only looked at the derivation today for the first time, and it is both elegantly done and correct AFAICT. The only quibble is, I suppose, that this method does not help you clearly and easily see when the brake action is likely to be regressive, progressive or neutral. But then neither does the applet TBH. I don't know why it doesn't work for everybody BTW; that needs a computer expert I'd say (not me!). Next step is to do as you suggest, and run some numbers in the applet vs CJ's neat method.

[531colin]

At the very end of the thread http://forum.ctc.org.uk/viewtopic.php?f=5&t=57410&start=30 CJ posts a correction about 2 brake blocks halving the MA .....or something like that!!

[Brucey]

Mmmm, well I think that last bit is wrong TBH. CJ's derivation allows for T/2 early on and is correct as it stands AFAICT. I think the applet gives values double what they should be because it doesn't allow for this correctly.

The CMA is such that cable travel x cable force = pad travel x pad force. (otherwise you have not conserved energy). And CMA = (pad force/cable force ) = ( cable travel/total pad travel). Thus my overall definition of CMA would (say) give an overall CMA of 1:1 with horizontal arms if b = 2h. Any movement at arm ends (equal to the cable pull) is half as much at the pads, but there are two pads....

Do you see what I mean?

[edit- I've just checked on the applet (which is incredibly annoying to use BTW, at least with my browser settings) and it says the MA is '2' for the flat arm b= 2h geometry. So it is (for the whole brake anyway) quite wrong then.]

[edit 2- I've been in touch with Benno (of applet fame) and he tells me that the applet will only display correctly if you have at least 800 pixels resolution vertically. I can just work the alpha slider (it is half-visible) if I set my monitor to 960 pixels vertically and remove the tab list display by using F11. We are having a discussion re the MA of the brake; the latest update is that no-one is convinced as yet! ]

cheers

[gregoryoftours]

You could use HS33s, or 700c version if you need them - less mainenance than cantis, and completely sealed, decent power without the necessity of uncomfortably stiff forks (not great in the wet though as all rim brakes).
EDIT - Ah, you'll be using them with drop levers - there is this: http://www.bikerumor.com/2012/04/19/mag ... ted-forks/
Expensive, I guess.

[CJ]

At the very end of the thread http://forum.ctc.org.uk/viewtopic.php?f=5&t=57410&start=30 CJ posts a correction about 2 brake blocks halving the MA .....or something like that!!

I've now retracted that correction, with thanks to Brucey for pointing out that I was right all along!

Unfortunately Snibgo and a discrepancy with the results produced by that impressively complex (but actually erroneous) Circle Cycles website thingy, caused a uncharacteristic crisis of confidence. Silly me. Of course I was right.

[bealer]

Copied Spa and got my Tektro RX5 mini-v's all hooked up. They fit in fine with plenty of space to put in mud guards. They're very powerful, forgot how good v's can be. I think I could have got by with Calipers to be honest. The v's are as powerful, if not a little more than my Croix De Fer's disc brakes (they're a bit spongey).

My new bike!



About 1kg heavier than I wanted, at 11.3kg. Not bad for touring trim though. For race trim I switch the touring wheels (with heavy dynamo) and put on a lighter set with 25C tyres. The Brooks sometimes come off too, bring it down to 10.2kg.

[CJ]

I like to run 37mm tyres on my tourer, so as to be able to ride potholed country lanes and gravel rail-trails etc. whilst looking at the view. And the Spa frame & fork has room for them if the mudguards are snugged up to the bridge/crown. But then you definitely cannot use mini-Vs, so I'll be sticking with cantilevers. And so long as I have strong enough hands to lock either wheel should I wish to, I'm perfectly happy. And I do like to be able to get the wheel out without having to fiddle with adjusters.

Dual-pivot sidepulls meanwhile, the 57mm reach versions of, are surely the brake of choice for tarmac-only tourers running tyres up to 28mm with narrow mudguards.

This seems to leave rather a small niche for mini-Vs: just those who want four more millimetres of tyre cushion than a 28, but never any more than 32mm.