Chain Length

[hairylegs]

Silly question. I changed my standard crank to a compact. Should I have taken a few chain links out? Also switched out my 11-25 to a 12-23.....

All seems ok, except large-small chainring sometimes spins out on the small ring.

[Ken Robb]

[QUOTE=hairylegs

All seems ok, except large-small chainring sometimes spins out on the small ring.[/QUOTE]

I don't know what this means.

[hairylegs]

Sorry! When derailling to the lower chainring from the big one (and in a smaller cog) the chain goes from the big chainring to the small, but it doesn't "grab" or lock in. If I switch to a larger cog it'll eventually snap in. Make sense? Hard for me to describe!

[fierte_poser]

so you are using an old chain with a new crankset and a new cassette? how many miles on the old chain?

[Mike748]

If you went from a 53 to a 50 on the big chainring, and from a 25 to a 23 on the rear, you could take out one link. I wouldn't bother unless the rear derailleur can't take up the slack when in the small/small gear.

What you are describing can be fixed by adjusting the front derailleur limit screw to allow it to move inward a little more.

[Volant]

Have you wrapped the chain (around largest chain ring and largest sprocket) and added the obligatory 2 links to make sure the chain is the correct length?
http://www.sram.com/_media/techdocs/roadchains__10spd__2_07.pdf

[John M]

Sorry! When derailling to the lower chainring from the big one (and in a smaller cog) the chain goes from the big chainring to the small, but it doesn't "grab" or lock in. If I switch to a larger cog it'll eventually snap in. Make sense? Hard for me to describe!

Agree with Mike748, you need to adjust your front derailleur. This is likely not a chain length issue.

[AndrewS]

Have you wrapped the chain (around largest chain ring and largest sprocket) and added the obligatory 2 links to make sure the chain is the correct length? This is the correct technique for setting up SRAM derailleurs, not Shimano or Campy. Shimano you put the front and rear in the highest gears (chain outboard, front and rear), and make the jockey pulleys vertical. Campy I'm not sure. But the rear derailleur determines how long the chain is for the gearing you have.

[Dave]

The brand makes no difference. If you want the absolute maximum wrap capacity from any RD, shift to the little ring and smallest cog. Make the chain as long as possible without hanging loose or rubbing on the RD cage.
This one length should handle any cassette within the RD's wrap capacity. In some cases, shifting to the second largest cog should be tried, since the little/little should really never be used and even if it is, a little chain slack won't hurt a thing.

With Campy 11 for example, they offer 11-23 to 12-29 cassettes. Only one chain length is needed with any of them.

The other methods can be vague. With the big/little method the RD cage can be tilted forward with one length and back if it's an inch longer. Depends on the chainstay length.

The big/big plus 1-inch method can also be ambiguous. If the ends of the chain that come together will not match up, you have to add 3 links or 1-1/2 inches, or the chain will be too short. This method produces the shortest possible chain, so later changes to a cassette with a larger cog may require a longer chain.

Often overlooked is chainstay length. The perfect length will add up to 2T of wrap, compared to the worst-case length.

[AndrewS]

Sure it makes a difference. What you stated will certainly work, but that doesn't mean it is the optimal amount. With Shimano and close ration set ups (12-21 with 52x42) it is going to make the chain rather short and put a lot of wear on the RD pulley spring as well as move the top pulley (the one that does the shifting) further away from the cogs than it should be.

There's a difference between doing it right and doing it "good 'nuff". SRAM derailleur pulley cages pivot from the top, Shimano from mid cage. They each deal with chain slack and top pulley position for different gears in different ways. I see no reason not to respect the design specifications of the engineers who built them.

[Dave]

AndrewS...

Sorry Andrew put you are wrong. You have a lot to learn. I got my mechanical engineering degree back in 1981 (when I was 28). When did you get yours?

The little/little method always produces the longest possible chain length and the minimum tension in the RD cage spring. How much tension that spring sees depends on how much time you spend in the largest cogs. The big/big method will suggest the SAME chain length if the cassette's largest cog creates the maximum chain wrap.

The big ring/vertical cage method's suggested length varies little since the smallest cog usually ranges from 11-13 and those 2T only make for a 1/2 inch chain length difference. The minimum length change is 1 inch. The chainstay length length also causes up to a 1/2 inch length variation. It's still ambiguous in many cases. If in doubt, you make the chain longer and then use the little/little method as a double-check. That method should have been used in the first place.

[AndrewS]

I was a pro bike mechanic for 5 years and have been building bikes for 20. And my penis is quite large. How about you?

Now that we've dealt with credentials, let's get to the problem:

The problem with setting it up for small/small is that it will not produce enough chain in some wide gear set ups to prevent the chain from over tensioning in accidental Big/Big shift. And in close ration set ups is going to produce a lot of slack throughout the shift range, resulting in more chain slap and noise from the low tension on the return side.

And the Big/Big will can locate the derailleur pulley cage more forward THROUGHOUT the gear range, not just in some combinations. When you're only using a portion of the pulley range you are putting more strain on the springs than necessary.

Any of these methods CAN work and certainly do most of the time. But since engineers (you know who I mean) actually went to the trouble to figure out the best way to set up their long or short cage derailleurs, why are you (as a non-derailleur engineer) suggesting that we use a general rule of thumb instead of the manual????


Speaking from my experience as a bike mechanic (which you aren't), I would not have been able to set up my mountain bike with its compact triple and Deore short cage derailleur had I used your methods - the chain would have been too short or too long.

PS: My brother is a mechanical engineer and patent attorney. Somehow, he still manages to read manuals and ask me for bicycle mechanic advice. Weird, eh?

[Dave]

AndrewS...

WOW! A bike PRO bike mechanic for 5 years. That's right up there with oil change technician. High school kids get paid to work on bikes so they are also PRO bike mechanics.

I love your statement "Campy I'm not sure. But the rear derailleur determines how long the chain is for the gearing you have." If that were true then the big/big method that bypasses the RD would not be valid for all RDs would it? In fact, it always suggests the shortest possible chain (that will wrap the big/big) for any brand of RD or any drivetrain for that matter.

I'm not a bike mechanic? I've been wrenching on them for 25 years, but only as a hobby. Thank goodness I never held a job that paid that little! I was rebuilding car engines when I was 16 and machined aircraft parts from age 18-23. Both are far more complex than any bicycle mechanic job. After I got my BSME (working part time as a machinst while in college), I was responsible for creating the manufacturing processes for nuclear weapons parts for most of my career.

You're making up new rules about the RD subject. First you mention what the manufacturers recommend, then you switch the subject to special cases. I'm speaking strictly in terms of RDs with enough wrap capacity to handle the cassette being used. That is also what the manufactuers speak of - not some setup that would never be manufacturer recommended. None of them would ever suggest using a RD with insufficient wrap capacity and a cog that was larger than recommended. If you deliberately use a RD with insufficient wrap capacity, then using the big/big method will prevent an over-extended RD in the case of a shift to the big/big. It will also produce chain slack in several of the smallest cogs. What a revelation!

You've only stated the obvious facts that any bike mechanic knows. Using the big/big can suggest a chain that is 1 inch shorter than necessary. It will still work just fine, however. As far as that RD spring that you seem so worried about, it just sees the same tension is would with 4 teeth smaller. No big deal. It's what SRAM suggests as a setup method but there is no reason to use this method, exclusively. This method will work the same with any brand of RD. It merely suggests the shortest possible length and produces the maximum chain tension on the lower section of chain - regardless of the brand. If you use that method and switch from an 11-23 to an 11-27, you'll need a new chain that's 1 inch longer.

The little/little method should ALWAYS produces the maximum wrap capacity from any RD. It is NOT what Campy suggests, but it works perfectly. I can use any 11 speed cassette, 11-23 to 12-29 with this ONE chain length and my short cage RD, because it has enough wrap capacity.

If you choose to exceed the RD's wrap capacity, it's up to the user to decide if they want to risk problems if accidentally shifting to the big/big or put up with chain slack in the little ring and smaller cogs. Some users think nothing of not being able to shift to the big/big.

Personally, I always use a RD with enough wrap capacity for the drivetrain I'm using.

[regularguy412]

Could it be that the OP had the chain at the 'long limit' originally and that changing only the crank then allowed the resultant slippage?

( I didn't read every post. Sorry if this is redundant)

Mike in AR:beer:

[Velosmith]

Wow Andrew and Dave I learned a bunch from both of your experiences.

Just a suggestion, take the personal stuff out of the conversation. I'm pretty sure there is room for both opinions.

btw Andrew.... Dave is highly respected in this forum.

[Dave]

I agree about the personal stuff, but what I've posted are facts, not opinions.

There are several ways to adjust chain length and each has it's limitations. Fact number one is that all methods can be used with any brand of drivetrain.

The big/big plus 2-3 links (1-1/2 inches) is the shortest length that guarantees the chain will wrap the big/big but does not insure that the chain won't hang loose if you're using a RD with insufficient wrap capacity. It can also result in the need for a longer chain if you set the length with one cassette and later change to another having a larger biggest cog.

If you're using a normal setup, recommended by the manufacturer (having adequate wrap capacity), you can never go wrong with the little/little method. It will suggest a length that works with any cassette within the RD's wrap capacity without changing the chain length.

The big/little combination can angle the RD cage slightly forward or slightly backward and be confusing to those who might think anything but perfectly vertical is a problem. The slightly backward angle is usually the proper choice, since it suggests the longer chain, but you still end up checking the big/big and little/little combinations to verify the choice works in both cases.


All of my Campy 11 bikes angle the cage very close to vertical and that is with the maximum chain length, setup using the little/little combo. It may be a Shimano suggested method, but it works the same with Campy. If I use the SRAM suggested method (it's been around much longer than SRAM has existed) it will tell me that my chain is 1-inch too long. It's not too long, just 1-inch longer than the minimum length and it would work fine if it was 1 inch shorter. 1-inch of chain equal 4T of wrap. There is no point in making the chain shorter, when I sometimes use a 12-27 cassette and might want to try a 12-29. That's why I say that the big/big method is not the best choice. If I had followed that often-suggested method, I'd need a longer chain for a larger cog. The point here is that the method to use is not brand-exclusive.

[AndrewS]

Let's be real clear here. Dave posted a technique I don't agree with, so I said so and posted why. There was nothing insulting about my post (unless simply disagreeing with Dave is insulting). Dave responded by pulling the "I'm an ENGINEER" malarky, which has nothing to do with the question.

The question:
How do I determine chain length?

Answers:
Dave - Use a standard rule of thumb method that should work with every set up. Or use the OTHER rule of thumb method.

Andrew - Since the RD uses chain length to function properly, and since RDs have many different geometries, using the manufacture's recommended chain length formula is the BEST method.


Everything else, like Dave bringing up how much I was paid 18 years ago, or that he is an engineer but has no idea what I do, is immaterial to the OP getting a couple reasonable, fact based answers to his question. This is a blind forum, so facts and sound arguments are much more valuable than attempting to discredit posters based on irrelevancies. Respect comes from action, not reputation. Or pulling rank.


Dave, the example I gave of using a short cage rear derailleur for a mountain bike was totally in keeping with what Shimano thought people would do with their XT and DX short cage rear derailleurs. You need to avoid the extreme crossover gear, but otherwise is a nice setup if you want to get more clearance and minimize slap. And the correct way to set it up is with the Shimano manual method of making the pulleys vertical. Anything else would not have worked well. Extreme example? Not really, they were built to work that way.

[sailorboy]

yea, I don't even care about the OP quetion at this point, the petty BS between Dave and Andrew is more entertaininig.

IMHO, Dave comes across as the bigger asshole since he started the personal attacks and continued to provide us with his resume throughout.

The OP could set the damned derailleur up either way and would work. If it wears out a couple thousand miles 'prematurely' cuz he chose the 'wrong' method then just spend 50 bucks in the classified section and get a new one.

[Dave]

It's winter. I have to do something for entertainment, at least once a week. Andrew is an easy target -thinks he knows far more than he does.

At least I provided some education. The topic is not complicated once it's explained and you've seen the results of using all three of the suggested methods. It still remains one of the most mysterious subjects even to some folks who have wrenched on bikes for 20 years.

One of the most common problems I read about is someone using the big/big method to set the chain length, then later wanting to switch from an 11-23 to an 11-27 or 12-27. Then they need a 1-inch longer chain and modern flush pin chains can't be lengthened safely. If you've got any notion of using different cassettes, be sure that your chain length is set to the longest possible length to start with.

[oldpotatoe]

AndrewS...

WOW! A bike PRO bike mechanic for 5 years. That's right up there with oil change technician. High school kids get paid to work on bikes so they are also PRO bike mechanics.

I love your statement "Campy I'm not sure. But the rear derailleur determines how long the chain is for the gearing you have." If that were true then the big/big method that bypasses the RD would not be valid for all RDs would it? In fact, it always suggests the shortest possible chain (that will wrap the big/big) for any brand of RD or any drivetrain for that matter.

I'm not a bike mechanic? I've been wrenching on them for 25 years, but only as a hobby. Thank goodness I never held a job that paid that little! I was rebuilding car engines when I was 16 and machined aircraft parts from age 18-23. Both are far more complex than any bicycle mechanic job. After I got my BSME (working part time as a machinst while in college), I was responsible for creating the manufacturing processes for nuclear weapons parts for most of my career.

You're making up new rules about the RD subject. First you mention what the manufacturers recommend, then you switch the subject to special cases. I'm speaking strictly in terms of RDs with enough wrap capacity to handle the cassette being used. That is also what the manufactuers speak of - not some setup that would never be manufacturer recommended. None of them would ever suggest using a RD with insufficient wrap capacity and a cog that was larger than recommended. If you deliberately use a RD with insufficient wrap capacity, then using the big/big method will prevent an over-extended RD in the case of a shift to the big/big. It will also produce chain slack in several of the smallest cogs. What a revelation!


The little/little method should ALWAYS produces the maximum wrap capacity from any RD. It is NOT what Campy suggests, but it works perfectly. I can use any 11 speed cassette, 11-23 to 12-29 with this ONE chain length and my short cage RD, because it has enough wrap capacity.

Personally, I always use a RD with enough wrap capacity for the drivetrain I'm using.

Actually, small-small and 10-15mm chain to der clearance IS what Campagnolo suggests...and what I use also, have for 25 years..never failed me yet.

Compact crank on a MTB and short cage RD on a MTB, with a 11-34? Yep, gotta use your head here and figure out wha's goin' on...not just use a rule.

[Dave]

oldpotatoe...

You're absolutely right.

Campy essentially suggests the little/little method. They just suggest a gap of 8-15mm to maintain. I'll accept even less, as long as there's obvious chain tension and no rub.

http://www.campagnolo.com/repository/documenti/en/11s_Chains_UK-03-09.pdf

[RPS]

One of the most common problems I read about is someone using the big/big method to set the chain length, then later wanting to switch from an 11-23 to an 11-27 or 12-27. Then they need a 1-inch longer chain and modern flush pin chains can't be lengthened safely. If you've got any notion of using different cassettes, be sure that your chain length is set to the longest possible length to start with.
That's exactly what happened to me when I bought my Co-Motion and the shop built it for me. They cut the chain based on the big-big which worked fine except the cassette was only an 11-21. The first time I went to install a 12-27 to take the bike to the Hill Country I ended up replacing the chain.

Had I installed the chain myself as I always do that would have not happened. I go with the longest chain -- and have also never exceeded derailleur capacity that I can recall. It's not that exceeding capacity can't be done or won't work with proper precautions and limits, I just prefer not doing it on my bikes.

[RPS]

The problem with setting it up for small/small is that it will not produce enough chain in some wide gear set ups to prevent the chain from over tensioning in accidental Big/Big shift.
Setting engineering backgrounds aside, you do agree that "ROAD" groups are sold by all major manufacturers so that the rear derailleur can handle the "entire" range of gears, right?

The listed capacity of rear derailleurs includes the entire range of small-small to big-big; which when combined with the smallest and largest rings and cogs gives the total capacity. I have never seen a "ROAD" group where the manufacturer suggests using less than the right equipment.

Don't get me wrong, I know it can work if done correctly provided the rider limits which gears he uses. But from a liability standpoint I see no way a major manufacturer will ever suggest using a RD that has less capacity than necessary to cover the entire range. If they do that seems stupid on their part to put it in writing.

[AndrewS]

RPS,

We're not talking about installing a new group. The OP changed part of his equipment (chainrings and cassette), and kept the derailleurs. So he could well be exceeding or coming close pretty close to the capacity of the RD. And even if he's still within the RD capacity, it is going to require the right amount of chain to center the pulleys within the desired range.

If the capacity is 28, that means that a 53x39 with a 12-26 puts the RD at maximum. That means the small/small crossover is going to have minimum tension and the Big/Big crossover is going to stretch the RD as far as it was ever designed to go. If you don't get the chain exactly the right length - centered between those two extremes, the performance at the two extremes, and the gears near those two extremes becomes sketchy. That's why I don't like setting up chain length from ONE extreme.

And if you have a 28 capacity RD and you want to use 53x39 with a 12-28, do you HAVE TO buy a new RD? Technically, yes, but the reality is that correct chain length will get you through as long as you avoid the crossover pairs.

Of interest, compact gearing set ups aren't just lighter: They require a lower RD capacity for a given gear ratio range. Which is a cool advantage when you want a bigger range but don't want a triple.

[Dave]

The manufacturer would never suggest setting the chain length so short that the big/big had to be avoided to prevent damage to the drivetrain.

If I understand Shimano's instructions (which are a bit vague), they suggest setting the chain length on MTBs using the big/big plus 2 links method. Like most instructions, they forget to mention what to do when the ends of the chain don't match up (add one more link).

http://bike.shimano.com/media/techdocs/content/cycle/SI/SI_5W70A_005/SI_5W70A_005_En_v1_m56577569830655487.pdf

Not being an MTB rider, my understanding is that it's intended for a short cage MTB RD to not be useable with the little ring and the 5 smallest cogs, since the chain will hang loose. The chain length IS supposed to be adequate to wrap the big/big.

http://forums.mtbr.com/showthread.php?t=284688

Road bike riders can do the same thing, using a short cage RD with a triple, but they would be restricted to using only the larger cogs with the little ring. Setting the chain to a length that will wrap the big/big, you should still be able to use the largest 4 cogs with the little ring, assuming a typical 30T little ring and 52-53T big ring, but the chain will hang loose in the smaller cogs. I wouldn't care for a setup like that, but it's doable.

[Mike748]

I've learned a lot from this discussion. Question - all things considered, is it better to have a little more derailleur tension (shorter chain) or less tension (longer chain)? My guess is shorter chain shifts better, longer chain wears less.

[AndrewS]

Again, this depends on RD design. Derailleurs with pulley cage pivots mounted well below the upper pulley (like many of the Shimanos and Campys) will often shift worse with more tension because the top pulley ends up further out of contact with the cassette and with less cog wrap.

With a SRAM derailleur the upper pulley and pulley cage pivot are co-located, so it doesn't matter as much. The upper pulley stays in pretty much the same position regardless of chain tension.

If you are between two chain lengths, I would set the chain length based on which one puts the upper pulley closest to the cassette in the most used gears. That will give you the best shifting. But since cassette profile and chainring differences vary, I don't think making a blanket statement like "less tension always better" is a good idea.


Topic change: I don't think I stated this earlier, but the OTHER reason not to use chain length formulas the don't reference the RD is that those methods make no reference to the cage length of the derailleur. A long cage derailleur is designed to work with a longer chain than a short cage. If you use the same chain for both you're automatically increasing the chain tension of the long cage RD, and that is not how that RD was intended to be used.

As with any mechanical thing, use your head and look at what you're doing. Don't wait until you're riding to figure out whether you set the bike up right or not. Universal links are nice because you can start with the chain on the long side and trim it down pretty easily without much work or problems with some of the fancy pins now in use.

[Mike748]

Derailleurs with pulley cage pivots mounted well below the upper pulley (like many of the Shimanos and Campys) will often shift worse with more tension because the top pulley ends up further out of contact with the cassette and with less cog wrap.

Can't I adjust the A tension or B tension to accomodate this?

[Dave]

I'll provide the counter argument and say that the length choice is not affected by the RD's brand or design.

Shimano suggests using the same big/big method of setting the chain length for both long or short cage RDs, at least in some cases, like the MTB example I posted previously.

The reason the big/big method works regardless of the cage length is that the cage will be almost fully extended in the big/big and cage length differences will not affect the required length in that position. If the setup uses all of the RD's wrap capacity, the little/little method suggests the same length as the big/big plus 2 link method. I've read postings from bike shop mechanics who never use any method other than the big/big and think any other method is stupid - probably because it sells more chains when customers make unnecessary chain length changes.

If you read the latest instructions for the Shimano 6700 RD, they suggest the vertical cage method unless using the new 11-28 cassette. Then they switch to the big/big plus 2-link method. Once again, the little/little would suggest that same longest chain.

http://bike.shimano.com/media/techdocs/content/cycle/SI/SI_5X90A_002/SI_5X90A_001/SI-5X90A-001-ENG_v1_m56577569830671285.pdf

Chain length should have no effect on chain wear since the high tension that affects wear is all along the top section of the chain. On a road bike, users of cassettes with large cogs in the 27-29T range only have one choice - the longest chain length. That length will be the same whether the big/big or little/little methods are used. In most cases, the shortest length is only 1-inch shorter. That's why Campy suggests the little/little method. One chain length will then handle any cassette within the RD's wrap capacity.

Chain length has no significant effect on shifting precision either, since the shifting is done at the upper pulley. The lower one just swings to take up chain slack after the shift is complete. That slack does not have to be taken up before the chain can drive the bike. The lower section can hang loose and still drive the bike. If using a cassette with the largest cog and greatest wrap, you have no choice on the chain length.

[AndrewS]

Can't I adjust the A tension or B tension to accomodate this? To a certain extent, but adjusting the location of the whole derailleur via spring tension in an attempt to make up for chain length is kind of treating the symptoms, not the problem.The a or b tension adjustment are there primarily to adjust the derailleur to shift over larger cogs, like a 28t. As you increase b tension the whole derailleur moves back and pulley cage tension goes up even higher. If you decrease b tension that counteracts a too short chain, but now the upper pulley will be bumping into your biggest cassette cog.

Sometimes a large adjustment to b tension will be accompanied by adding or removing links from the chain. I would say that chain length, along with b tension, high/low stops and cable tension are all essential adjustments to setting up a derailleur optimally. Some of them interact, but that isn't the same as saying that they are interchangeable.

[John M]

Time for a couple of guys to go ride their bikes....

http://www.googlefight.com/index.php?lang=en_GB&word1=Dave&word2=AndrewS

[AndrewS]

Chain length has no significant effect on shifting precision either, since the shifting is done at the upper pulley. Sure it does, depending on the derailleur. SRAM - none. Shimano - some. '80s Campy - a lot. A Nuovo Record RD had the pivot centered between the pulleys - so the top pulley moved just as much as the bottom.

It depends where the cage pivot point is located. The further it is from the upper pulley center the more the upper pulley location changes with chain tension changes.

I'm using the word "tension", but really this whole discussion is about location. Greater chain tension means that the pulley cage is angled more forward, possibly affecting upper pulley location in some RDs. Greater chain tension also means that the spring that counters it in the RD pulley cage is either working within its desired range, or is being compressed more than designed.

Sound complicated? Maybe, but using the recommended chain set up instructions from the RD manufacturer makes it a non-issue.

[RPS]

RPS,

We're not talking about installing a new group. The OP changed part of his equipment (chainrings and cassette), and kept the derailleurs. So he could well be exceeding or coming close pretty close to the capacity of the RD. And even if he's still within the RD capacity, it is going to require the right amount of chain to center the pulleys within the desired range.

In the OP's case exceeding the RD capacity is not likely. If he has a 50/34 compact with a 12-23 his total capacity requirement is 16 + 11 = 27. Most all road derailleurs should handle that.

If we "assume" (big assumption on anyone's part because the OP wasn’t completely specific) that his chain was cut long enough for the previous big-big and that he went from 53/39 to 50/34 rings (also assumptions), we can take some educated guesses:

Previously: 53/39 and 11-25 stretches range from 78 to 50, or a capacity of 28 teeth.

Present: 50/34 and 12-23 stretches range from 73 to 46, or a capacity of 27 teeth.


I think the 27 capacity in itself should be OK, but the longer-than-necessary chain "might" just throw the RD beyond its limit when in the 34/12 combination (assuming a standard derailleur and variations in chainstay lengths).

Fortunately, this is so simple to test that it’s not worth arguing. Not with you or anyone else. If hairylegs wants to know for sure, just go out and try it. If OK, leave it alone. If not, shorten the chain.

[AndrewS]

I didn't mean that the OP was exceeding capacity in this particular instance, but that you can in any instance when you change out your gearing. Like in my example of going from a 12-26 to a 12-28.

But I did not post in the first place out of concern for his RD capacity. I posted because the move he made to all smaller cogs probably meant shortening the chain to keep everything working optimally. And I was merely suggesting that reading the RD manual would provide a good check on that.

[AndrewS]

All this interesting talk got me thinking about why Shimano chose the method they did for many of their derailleurs. The answer is interesting and involves geometry.

A pulley cage is most effective when it is 90 degrees from the returning chain. This is because it is pulling the chain through an arc, and as the pulley angle approachs being parallel to the chain they stop pulling any chain at all. In fact, 2/3 of the derailleur's chain rap capacity lives between 45 degrees forward and 45 degrees behind that 90 degree angle.

So for a derailleur to be working at optimum, the total capacity range should be centered around a 90 degree pulley angle from the chain. We can figure out where that should be with a little subtraction. Consider the following drivetrains: 53x39 12-26, 50x34 11-25, 52x42 12-21

Big/Big ........ Center ........ small/small
53+26=79 .... 53+12=65 .... 39+12=51
50+25=75 .... 50+11=61 .... 34+11=45
52+21=73 .... 52+12=64 .... 42+12=54

So if we seek the center of the derailleurs range (90 degrees from chain), we will end up with the minimum and maximum chain useage centered around the most effective part of the pulley cage arc, requiring the least amount of total angle change for the capacity needed.

As it should happen, that center point is usually within one tooth of the "high" gear, whether its 53x12 or 50x11. Which is why the Shimano method works.

So another rule of thumb would be: Find your drivetrain capacity's center, and using that gear combo set the pulleys at 90 degrees to the chain. Unlike other methods, this rule of thumb takes cage length into account. Other rules of thumb could have your derailleur working in the arc ranges where it is least effective or over-taxing the pulley spring.

You can also use this method to see if your chain is long or short enough after a major gearing change: Find the center gear combo and see if it is close to that magic 90 degrees.

Why bother? If people are going to spend $50 just on fancy shift cables, it seems like setting up shifting a precisely as possible might be of value, too.

(Note - the above method works with modern derailleurs. Old slant parallelogram derailleurs, like Nuovo Record, were not designed to have as much forward range to their pulley cages and you'll run out of spring.)

[oldpotatoe]

I've learned a lot from this discussion. Question - all things considered, is it better to have a little more derailleur tension (shorter chain) or less tension (longer chain)? My guess is shorter chain shifts better, longer chain wears less.

A shorter chain doesn't shift any better than a short cage RD vs a medium cage one. But if you have to error in one direction or another(and no reason to error at all), have the chain be a little long. Broken chains when going to big-big w/o enough chain always breaks lots of things, including frames sometimes.

[RPS]

Good point. A longer chain also makes on-the-road repairs easier to live with until you get back home. ;)

[Velosmith]

My wife is pretty small. 50cm X 50CM frame and a pretty short chain stay. She rides Campy 10 speed with a compact crankset 50x34 and 13X29 cassette. We really struggled with FD chain drops with a medium cage RD. It was really hard to control the front chain drop from the 50 to the 34 and she was constantly dropping chains on the front downshift.

Solution: I had a short cage DR we tried. Used Dave's method. I explained to her she didn't want to use the Big X Big combination as the rear DR was really pretty maxed.

Result: No more dropped chains and crisper rear shifts.

Go figure

[Velosmith]

BTW.... I've been an engineer for 21 years... Does Information Technology count? :)

Sorry .... I don't know what gets in to me sometimes... :crap: