louky

I do a lot of charity bikes for unhoused people in my town. Most of the donated bikes are run of the mill box store "mountain bikes" but I sometimes get cruiser bikes donated. I usually take the wheels off to check for rim tape and a rudimentary wheel truing. When reinstalling, I notice that there ALWAYS seems to be a place in the chain where it is too "tight". I understand that there needs to be some slack in the chain when adjusted properly, but it seems like if I leave any amount of slack at the tight point it makes the chain too slack at the slackest point. Is this something they are doing on purpose in these bikes? It seems that there should be at least once in a while that they are consistent but I haven't seen it yet. Needless to say, these are all cheap bikes.

FBinNY 

What do you consider to much slack, and why? Not looking for answer, just want you to consider it.

Understand that both sprockets are eccentric. They're usually pretty round but not mounted perfectly centered, especially on low end bikes.

So there'll always be some degree of change in chain slack. I've yet to see one bad enough to be problematic, though some are pretty bad.

55murray

You just make sure the tightest spot is not too tight, and don't even consider what it is like elsewhere in the rotation. Too tight will rapidly start to destroy things, too loose never will.

If you do nothing else with a cruiser, new or used, ensure that chain is never too tight.

Andrew R Stewart 

One could try to change the rotational "clocking" of the rear cog and/or the chainring to see if other rotational arrangements offer less eccentricity. Often there's some slop between the ring and the RH crank arm which can also be played with if the two are bolted together.

As always the "correct chain tension" is zero tension. Andy

Jeff Wills

The fit between the one-piece crank and the chainring is so sloppy that the chainring will never be centered and, even if you manage to re-center it, it won’t stay that way.

As others said, make sure it doesn’t get too tight and call it good.

FBinNY 

To clarify,

References to being too tight may lead to a logical error, implying that some tension is OK. A wee bit tight is already too tight.

Chain drive must be slack at all times. Ideally, you want to minimize slackness, but never want it to be zero.

FBinNY 

One thing to keep in mind is that you have 2 eccentrics turning at different speeds, producing ever changing chain slack.

Imagine 2 eggs. The chain will be slackest when both point toward each other, and least slack when they point away. Other arrangements produce slack someplace within that range.

The thing to keep in mind is that the extremes are relatively rare and might not happen within the few crank turns while you're checking. So you need to leave some extra margin for error.

alcjphil

I would imagine that eccentricity is more pronounced at the chainring than at the rear cog. I am prepared to be called out on this.
If the tightness occurs every time the chain reaches a certain point on the pedal stroke, think chainring

FBinNY 

Yes, that's generally the case for a number of reasons. But, regardless of which or why, you need to maintain slack.

grumpus

You'll find it varies in time with both the wheel and the chainring, so hold the crank in its tightest position and stop the wheel in it's tightest position, use those locations to set the chain's minimum slack. If you only allow for the chainring you've only done half of the job. And when you're done setting the slack, check again that the chain doesn't go tight.

Bill Kapaun

Depending on the number of teeth of the crank & cog, it may take a LOT of crank rotations to achieve that EXACT degree of tightness/looseness.
IF BOTH are PRIME numbers, it's probably worse. a 37:13 for example, would take 37*13 or 481 revolutions to get back to the exact combination of engagement.

FBinNY 

This approach is impractical, or even impossible since the only way to identify the high spots is by looping the chain. Thereupon the relationship is set. The chainring variance is once per crank turn, that for the hub is once per wheel turn, but they only line up at the max/min once per many crank turns. As Bill, who did the math, it could be as rare as once per 480 crank turns or worse.

Fortunately the crank variance dominates, so set chain length based on that leaving some margin for the effect of hub variance.

The message is simple, do not try to optimize anything, just leave some slack at the tightest spot.

GamblerGORD53

Actually, my 1973 CCM 1 piece crank is the BEST for roundness. I tried filing my other bike's $150 Sugino 5 arm spiders and it only improved a hair.
The OP has the right idea for sure, that's what I've always done. And NO I don't have to turn it a hundred times to get it lined up. More like 3 or 4 chain revs to get it to repeat. Cheap bike has nothing to do with it. My CCM cups are pitted and so I've had to set it loose, been that way 7,000 miles so far.
My White Industry 46/48T ones weren't perfect either and the buggers didn't even offer full 1/8".

FBinNY 

In this discussion, it might help to understand how bike chainrings and sprockets are made.

Generally, they're stamped, a process not unlike using a cookie cutter. As a practical matter, you cannot stamp the teeth and the inside cutouts at the same time. So, it involves a minimum of two operations.

That means, while it's easy to make a round sprocket with well formed teeth, it's not so easy to ensure that internal cutouts are exactly centered.

Some are much better than others, and some are way off, but none are perfect. Fortunately, chain drive systems can handle that without issues.

Bill Kapaun

I should have gone for a 53:17!
To test, check every 1/4 mile?

I notice only one has to be a Prime if that number isn't a LCD of the other factor.
I'm just drawn to Prime Numbers.

tcs

The thread title contains "generic coaster hub". If the meaning was coaster brake hub, note that more (excessive) slack = delayed braking. Other posters have included the word 'minimum'. Subtle advice not to be overlooked where generic coaster brake hubs are involved.

FBinNY 

Any delay caused by chain slack is going to be in the range of micro seconds, representing an inch at most of added pedal travel. This is smaller than the motion lost to the Bendex, or time possibly lost because of pedal position, which is something we've always accepted, along with very long braking distance.

The real hazard from an overly slack chain is much more straightforward, namely the chain falling off, and total loss of brakes.

Fortunately, there's a wide band between too tight and too loose.