Going down a small hill, don't remember if I was gently "braking" or simply
letting it spin my legs, but it completely threw the chain. The road is mostly
smooth, maybe a small bump here and there, and I've ridden that rode hundreds
of times.

The chain has several hundred miles on it since I replaced it--no problems since.
I had to take the rear wheel off to untangle the chain, rode home . . . . . . . .????

Why, why, why?

That happened to me once when the drive-side rear wheel nut got a little loose and the cog side suddenly moved forward making the rear wheel want to turn left (or vice versa). Fortunately it wasnt bad because I was using a freewheel at the time... makes me wonder how it jumped off since it was a freewheel after all! On the bright side nothing broke and it only scratched the paint a bit.

Other than maybe some chain tension issues resulting in a chain being unhappy and/or kinking on itself I cant think of many reasons why this would happen.

I stretched a chain enough to do that in one ride. It was a POS KMC chain. I threw it in the middle of a 4 lane intersection. It was my fault for not checking it because I knew it was a brand new chain. It happens.

it happened to me goung up a hill. my drive side nut came loose. didnt have any tools and had to walk 2 miles home. same as prosperity red.

A freak road debris flew on the chain ring and derailed your chain.

all interesting . . . . . . I'm still wondering . . . . .

road kill......BUT it wasn't you

It's happened to me once or twice going over some especially bad railroad tracks.

Well presuming your cainline is Ok, this happens to me when I do not run my chain tight enough.

I thought about chainline, but I simply put my rear wheel back on and
centered it between the rear triangle just to get home after throwing it.

Is there a surefire way to check for a proper chainline other than to
eye it up from atop and the rear?

I thought about chainline, but I simply put my rear wheel back on and
centered it between the rear triangle just to get home after throwing it.

Is there a surefire way to check for a proper chainline other than to
eye it up from atop and the rear?



measure....centerline of cog and ring ...to centerline of frame....

Just some thoughts, assuming that there wasn’t a stick or stone pushing the chain off the teeth:

-For a fixed chain to derail, it has to be somewhat loose at the point of derailment and a lateral force (like a front derailleur would) has to act on it. Under power, the top run of the chain is under tension, so the probable point of derailment is the bottom of the rear cog. When back-pedaling, the bottom run of the chain is under tension, so the probable point of derailment is the top of the front chainring. During soft-pedaling, there’s little tension on the entire chain, so the bottom of the cog and the top of the chainring become probable points of derailment, increasing chances of derailment by 100%.

-Lateral force strong enough to derail a fixed chain can come from a bump, groove or hole in the road which kicks the rear wheel over sideways, even just a few centimeter. It’s also possible to put lateral force on a chain with a violent swerve.

-Back in the day, we’d put our track bike in a stand and spin it up. Then with a long tool, we’d lightly tap the side of the chain near the possible points of derailment to see if it stayed on the teeth. It’s crucial to use a long tool and not get your hands or any loose clothing anywhere near the chainring or the cog.

-To make shifting ever more idiot-proof, the teeth of road chainrings have gotten shorter and shorter over the years, plus shift gates were invented. If you run a road chainring, consider replacing it with a chainring that has higher teeth. You really don’t want those road chainring “easy derailment features” on a fixed gear.

Interesting write-up wim, got's me a thinkin' more of a bump combined while transitioning between pedaling w/force and then slowing back-pedaling prompted this. . . . ? ? hum.

Interesting write-up wim, got's me a thinkin' more of a bump combined while transitioning between pedaling w/force and then slowing back-pedaling prompted this. . . . ? ? hum.

Yes, that's exactly what I was thinking. At that transition point, all sorts of weird things happen to the chain. But because there's no way to be certain, I didn't want to say "here, that's what caused your chain to come off!"

FWIW, the last time my chain came off on my fixed gear was on a dirt trail at about 4 mph. I was soft-pedaling while rolling over a root at 45 degrees to the direction of travel. The bike gave a little lurch and the chain was on the rear axle. I put the chain back on and it has stayed on for months now. So it happens, but rarely. I don't worry about it.

there is an excellent article on chainline at http://sheldonbrown.com/chainline/index.html
another possible factor would be a crankset that is way out of round so that the chain tension varies considerably....road cranks are generally not nearly as round as quality track cranks since the dérailleur controls the chain tension...this is part of the reason that track cranksets are expensive.
Anyway, FWIW my bike hasn't dropped the chain once in 4 years of steady use with correct chainline (within 3mm) and a Sugino 75 crank that is almost perfectly round

Just some thoughts, assuming that there wasn’t a stick or stone pushing the chain off the teeth:

-For a fixed chain to derail, it has to be somewhat loose at the point of derailment and a lateral force (like a front derailleur would) has to act on it. Under power, the top run of the chain is under tension, so the probable point of derailment is the bottom of the rear cog. When back-pedaling, the bottom run of the chain is under tension, so the probable point of derailment is the top of the front chainring. During soft-pedaling, there’s little tension on the entire chain, so the bottom of the cog and the top of the chainring become probable points of derailment, increasing chances of derailment by 100%.

-Lateral force strong enough to derail a fixed chain can come from a bump, groove or hole in the road which kicks the rear wheel over sideways, even just a few centimeter. It’s also possible to put lateral force on a chain with a violent swerve.

-Back in the day, we’d put our track bike in a stand and spin it up. Then with a long tool, we’d lightly tap the side of the chain near the possible points of derailment to see if it stayed on the teeth. It’s crucial to use a long tool and not get your hands or any loose clothing anywhere near the chainring or the cog.

-To make shifting ever more idiot-proof, the teeth of road chainrings have gotten shorter and shorter over the years, plus shift gates were invented. If you run a road chainring, consider replacing it with a chainring that has higher teeth. You really don’t want those road chainring “easy derailment features” on a fixed gear.


See I told ya Wim was smart.....making mental note...never stick your fingers in moving drivetrain....check...

My cog and ring are VERY true on my langster...so I can keep my chain very near "bar taut" The same as you would on a chainsaw.
I spin the pedal and if I get 10 spins or so I figure it is loose enough.

See I told ya Wim was smart.....making mental note...never stick your fingers in moving drivetrain....check....

Nothing to do with smarts—even a cow learns to keep its dumb nose off the electric fence wire after a few zaps . . .:)

Many years ago, I got my fingers bloodied two or three times, stupidly trying to wipe a moving chain on a fixed gear. Learned my lesson, sort of. Then one day I stumbled on Sheldon Brown's page about fixed gear dangers—the one with the gruesome photograph of a completely severed fingertip. That drove the point really home, believe me . . . :D

Nothing to do with smarts—even a cow learns to keep its dumb nose off the electric fence wire after a few zaps . . .:)

Many years ago, I got my fingers bloodied two or three times, stupidly trying to wipe a moving chain on a fixed gear. Learned my lesson, sort of. Then one day I stumbled on Sheldon Brown's page about fixed gear dangers—the one with the gruesome photograph of a completely severed fingertip. That drove the point really home, believe me . . . :D


I was crawling UNDER an electric fence one day, wet ground....drizzling...The lower wire caught the small of my back as my shirt rode up....BAM....face down in the dirt......there are some serious nerves running through the spine!!!!!!!!!

I saw that picture on Mr. Browns site....before I owned a fixed...

it's gross but he makes his point.

That's where a good slack chain would well be to your advantage!

Can I hi-jack my own thread and ask, how do I know WHAT my cahin tension
SHOULD be? It has a bit of slack in it and it's not too tight.

I think it's OK, but is there a sure-fire way to know for sure? Thx, good
informative thread.

Can I hi-jack my own thread and ask, how do I know WHAT my cahin tension
SHOULD be? It has a bit of slack in it and it's not too tight.

I think it's OK, but is there a sure-fire way to know for sure? Thx, good
informative thread.

Good question, but there's no defining answer other than the vague "tight enough to stay on, loose enough to run freely." For one, people disagree on what's right. And as ukiahb pointed out, chain tension often varies within one revolution of the crank. Sounds to me like you have yours just right.

Here's my test: I'm satisfied when my spun-up rear wheel comes to its initial stop, then moves in the reverse direction for just a second or so before coming to its final stop. I've never cared for chain deflection numbers because they depend on how hard you push on the chain. But that's just me. Hopefully, others will post their equally valid ways of setting chain tension.

a figure I've seen a lot (and use) is 1/2" up-and-down play halfway between the cog and chainring.....check this with the crank in various positions, if the slack varies the crank is out of round and you will have to compromise

a figure I've seen a lot (and use) is 1/2" up-and-down play halfway between the cog and chainring.....check this with the crank in various positions, if the slack varies the crank is out of round and you will have to compromise

Good advice..

JetSpeed: it occured to me that where you're riding should also figure in setting chain tension. Trying to win the 1,000 meter time trial ("kilo") on a smooth track, I'd run a droopy chain—with no no back-pedaling, soft-pedaling, swerves or bumps during that painful minute, I'd want as few frictional loss as possible. On the other hand, shredding cross-country on rooty trails, I'd tighten the chain within an inch of its life to keep it on. These are extremes, but you get the idea.

All good pointers, thanks.