Black wallnut 

Chain tension is on the return or slack side of the loop not the drive or power side. The tension applied on the drive side between the rear cog and the front chain rig is the same regardless of tooth count of either. Force required will be the same if the ratio is the same. However there may be measurable differences in friction losses with a less straight chain line. Highly doubtful a mere mortal would ever notice.

You are overthinking this. If there is more tension because of greater rotation then at the same time the sum of the aggregate would be lower because of less rollers contacting the cog.

HTupolev

No. The chainring is basically a lever. If your power and rpm is the same but the ring is bigger, there's more chain speed and lower chain force.

Rollers contacting the cog isn't the problem. It's when things articulate and when they rub. Although frequency of these interactions is lower in the smaller-cogs case, both the size and the force of the interactions if higher, and this combines as a net drawback.

Friction advantages of larger sprockets have been known for well over a century (see the discussion on chain gearing in Bicycles and Tricycles by Archibald Sharp, 1896), and it's confirmed by modern drivetrain power loss testing (see page 9 in this report).

Black wallnut 

Ya might want to review the document you linked to. Page 3, item 3 in bold says that I am correct with regards to big ring and lowest three cogs is less efficient than the same ratio of small ring. So after all the exercises to measure such things my intuition prevails. Granted that I may be in error as to the why that I am right. Thanks for the link by the way.

HTupolev

Page 3 is including the effects of chain misalignment. My post that you responded to was commenting purely on chain tension and cog sizes. That is why I directed you to Page 9 of the report.

I agree that low gears while in the small ring is usually mechanically better-running than big-big crosschain, at least within the scope of typical road drivetrains.

grumpus

How does the crank turn the wheel if the chain doesn't apply forward tensile force?
That's clearly incorrect - halve the tooth count and double the force transmitted by the chain, twice the force at half the speed equals the same work done per unit time in each case because work = force * distance, but there's a higher load on the components and subsequent risk of e.g. breakdown of lubricant film.
Force applied at the pedal will be the same if you ignore the issue I raised i.e. losses caused by factors affecting drivetrain efficiency including but not limited to "cross chaining". But assuming all these losses are negligible then yes, the force required at the pedal to ride at a certain speed remains the same if the sprocket ratio remains the same.
That sounds suspiciously like a point someone was trying to make earlier, before this thread descended into word games and name calling.
"Fewer". I was playing devil's advocate - I don't know the answer nor do I have any particular desire to investigate further; I was simply pointing out that it is potentially more complex than just "more straight chainline always better than less straight chainline". But I think we have pretty much agreed that it doesn't really matter unless you're going for the hour record.

79pmooney

I guess I don't understand how gears work either because I have no idea what you are talking about.

smd4

I guess not.

If you don’t understand by now that the biggest cog in back is the easiest, and therefore slowest gear, while the large chainring is the hardest, but fastest gear, and that using them at the same time is dumb just for that reason, then I really can’t help you.

Best of luck in your cross-chained journeys.

Black wallnut 

What are we picking nits now? Playing devil's advocate?, so in essence you agree with me mostly but you just have to extend the conversation? I don't apparently know the cause but I do seem to know the answer however the answer, in regards to this thread, is that big+big cross chaining should be avoided and that it is in fact less efficient. Everything can be potentially more complex, is there really time and interest in an internet forum to dive into fine details? The linked document details some of what you are saying and although it seems you are right about the friction their charts show the losses to be almost nothing, a watt or two. For what it is worth the hour record is done on a track, with a track bike and those are single speed so I fail to see your point in bringing that up.

My answer didn't consider mechanical advantage. You win on that one.

I was thinking of the tension that the derailleur added, you were thinking of the tension the crank applies under force... guess we were both confused as to what each other was referring.

Sorry replying with a different order than your objections.

Further the linked document conclude that what @smd4 and others have been saying is factual. Cross chaining is to be avoided for several reasons.

Trakhak

In a Tour de France video from around 2005, the riders are on a slight rise just before the stage finish. A motorcycle cameraman zooms in on Lance Armstrong's drivetrain, showing that the chain is on the large cassette sprocket and the large chainring and prompting Phil Liggett to exclaim, "He's in the correct gear." He goes on to explain that under those circumstances, shifting down the cassette to get to his sprint gear is a more reliable move than shifting to the larger chainring from the smaller chainring.

cpach

The bickering about cross chaining is hilarious. Regardless, this drivetrain shouldn't do that in this gear if everything is well adjusted.

I'd suspect hanger alignment(checked), bent derailleur or excessive pivot play, excessive play in guide pulley(replaced), poor match of chain/pulley wear, and/or incorrect chainline.

Kapusta

Yep. But some people prefer to drag up worn out arguments rather than to contribute something useful (likely because they have nothing useful to contribute).

Camilo

Not really. I'm a big "don't cross chain" person, but it's not dumb to use the big/big combo every once in a while. Personally I don't make a habit of it, having "grown up" when cross chaining was taboo, but use it when the situation warrants - several of which have been described. No harm, sometimes expedient, not dumb.

AndreyT

Despite the scientific insignificance of that endearingly barn-sciency report on "modern testing", it does mention an interesting point: top-span chain loads are much lower in big-big than in an equivalent small-smaller combination. This is another point that helps to breaks the myths about "wear" and "lower efficiency" of big-big.

People have to understand the simple thing: all that mythology about purported "disadvantages of big-big" actually originates from marketing departments of bicycle drivetrain manufacturers. It all dates back to the initial transitional period from friction-based shifters to indexed ones. While no one can deny the advantages of the latter, the manufacturers still had to combat the uproar of disapproval from the cycling community, caused by the fact that they effectively blocked big-big when it was well within reach. While these companies frantically worked on their big-big solutions (which gave us "trim" feature in indexed systems, as I mentioned previously), their marketing branches generated massive amounts of "you don't need big-big, because..." propaganda to persuade people into buying their initial big-big-incapable offerings. And when trim became a thing in indexed shifters, it took some time to trickle from TOL groupsets to more mere-mortal-oriented ones.

What we are fighting today, with our excellent big-big educational effort here, is not mechanics. It is the residual effect of that massive brainwashing campaign that managed to take hold in some feeble minds.

smd4

"Trim" existed in index shifting systems starting in 1985. As I mentioned previously. It's not "new." It didn't "trickle down." Some feeble minds who have been brainwashed still don't understand this.

HTupolev

Earlier in the thread, you suggested that actual running characteristics of older chains was a contributing factor to the avoidance of big-big:
Now you're claiming that it arose purely as a way to deflect from the issues with un-trimmable indexed front shifting. So which is it?

At any rate, your new argument is false, as evidenced by admonitions toward cross-chaining from prior to that period. For example, here's a page from a 1978 guide by Huret on derailleur usage, strongly recommending users against riding "with a crossed chain."