Caliper

I've got a bike that's set up fixed which doesn't have track dropouts. I can't get the right chain tension with my desired gearing and the gearing that gets good tension just bugs me... So, while pondering how make a chain tensioner for a fixed gear bike, I come across this video... Everything appears to work. There would be a bit more stress on the chain, but I can't see anything going catastrophically long. Has anyone run this solution?

Casualcoolguy

I’ve seen this exact video! Tried it one time just for sh*ts and giggles and it worked.... for a short period of time. Personally no matter how much tension I had on the chain with the sprocket, the second I started hitting bumps and rough road the whole sprocket would dislodge itself and go flying down the road ahead of me. You’d be better of making one of those stripped down Derailer chain tensioning things!

79pmooney

My question: does it allow you to keep proper slack in the chain? The chain going tight is a bearing killer for hubs and BBs. i"m guessing that with a high quality crankset and hub (ie velodrome worthy hardware) this could work really well, but with adopted mediocre road equipment, this could be excessively exciting on a fast downhill.

Casualcoolguy, no, on a fix gear,you cannot use any sort of derailleur style tensioner because you will simply fold it up and destroy it the first time you forget you are on a fix gear and try to coast.

I"d rather go with an eccentric BB or hub to do vertical dropout and fix gear. (I do not know if there is an eccentric BB that will work in a standard BB shelll. White Industries makes a hub that works well.) But I"d search out an old 1980s Japanese sport bike or equivalent with horizontal dropouts and just live life the easy way. My winter/rain;city fix gear is just that. Started life as a Peugeot UO-8. Frame was replaced by a Japanese built Schwinn sport frame, then a Sekine, then a Miyata 610 and now a Trek 400 (?). All with sport geometry and horizontal dropouts. (Early-mid '80s were the golden years for bikes. The Japanese had standardized the thin tubed steel frame and threads that everyone around the world had to copy if they wanted sales. You could pull the parts off one bike and put (almost all) of them onto your replacement frame. Might need seatpost and headset. Everything else just dropped on.

Ben

caloso

That looks hella sketchy.

The other thing you can try is a "magic gear combination." Basically a chainring x cog combination that allows for a chain length that perfectly fits the BB--dropout distance.

ThermionicScott 

As is often the case, Sheldon Brown documented the idea long ago: https://www.sheldonbrown.com/tandem-build2.html#okeefe

Casualcoolguy

oh good point I totally blanked on the fixed part hahah I assumed it was a single speed thing

AlmostTrick

I once fixed a V-dropout bike with a half link. Worked well with my preferred chainring and sprocket. I also wouldn't be afraid to file the lead or trail edge of the dropout a tiny bit if it was close.

Caliper

So my bike actually is a mid 80's Schwinn tourer with angled dropouts. It just doesn't quite work right with my desired gear ratios though. Either the chain is loose at the furthest rearward position or I'm nearly kissing the chain stays when I remove a link. Frustrating.

I'm trying to run this with the Wabi 16/18 dingle cog for a gravel fixed gear, so that limits things a bit.

How did I not know that a half link was a thing? That will do it, even if it doesn't look as cool.

hobbitman

Something I've done a few times is use a quick release axle and just cut the axle so it doesn't protrude beyond the lock nuts. This gives you a little wiggle room for chain tension. Note: I wouldn't attempt this unless you're using a good quality, closed cam skewer and it's done up tight. Sheldon Brown offered it up as an option, and I've had good success with it. YMMV, and I'm not responsible if you die a painful death. You can also get a similar result by filing a solid axle on two sides so you can rotate it to give you a little more tension. I've tried the ghost ring option and frankly, I'm just not a fan.

MAK

I was going to suggest the half link. I've never had to use them, but they seem to be a good way to fine tune the chain length. Good to see that they do work.

Mikefule

This idea with the floating chain ring: we used to use that in the Tandem Club as a way of taking the slack out of the chain that connects the two bottom brackets, rather than the final drive chain. It worked.

79pmooney

Half links work very well. I clean the threads (on half links with threaded pins) and use Loctite red. From then on, I break the chain away from the half link. Other half links use a cotter pin like a sailboat shackle. You have to treat them like a sailboat shackle but they are very secure if you do so. (I've trusted my life to those cotter pins many times as an ex-sailor.)

Ben

Caliper

Well, the floating ring didn't end up working for my bikes geometry. Too bad. With the dingle cog in there, the outer gearing combo meant that even a tiny 24T ring would rub the inside of the chainstay and at the same time not give enough help to make the chain tension work.
It's now sitting pretty with a half link and the axle is still far enough back to clear full fenders so it's my ultimate dirt road bike for wet days.

BicycleBicycle

You think you've seen everything until something like this pops up.
It looks like it would work for general straight line a to b commuting, but throw any kind of variation in there, or stick it in an urban environment and I would suddenly feel very sketched out.
Some people suggest half-link links, but I'm surprised that a half-link chain is rarely mentioned.
Try a half-link chain.

ThermionicScott 

Half-link chains elongate faster, that's why they aren't recommended much.

JohnDThompson 

There's nothing that a half-link chain can provide that using a single half-link can't provide.

yelro

thank you for the tip and sharing your first hand experience of this type of tensioner. Iw about to install a ghost ring for my single speed conversion. now im im. convinced of the rear derailer as tensioner. and i saw many infos vidwos who choose it over attaching a ring. although it looks edgy and fun still.

yelro

im using half link and a rd rd tensioner. i imagined the wear and tear of tye chain would one day it would loosen atleast i have both solutions attached and ready

JohnDThompson 

I've used single half-links for decades and never had a problem.

AlmostTrick

Sure there is... half link chains provide goofy aesthetics.

TejanoTrackie 

Trying to set up a fixed gear with vertical dropouts is just plain dumb.

son_of_clyde

Here's a somewhat automated way to see what might work. I wrote the function in matlab, but it should work in octave as well (haven't tested there). Enter distance in inches between crank axis and rear hub axis and the number of chainring teeth. Function returns chain length for given chainring with a range of rear cogs (12 - 22 teeth). For example, P = chain_length(15.5, 45) returns: P =

45.6960
45.9193
46.1435
46.3684
46.5942
46.8209
47.0483
47.2766
47.5057
47.7356
47.9664

Possible "working" combinations would be those close (<0.06" maybe) to an integer value. In the above example, 47.0483 (18 tooth cog) and 47.9664 (22 tooth cog) look promising.

___________________________________________________________________________________________

function [P] = chain_length(L, Tf);

% Returns chain length based on L (distance between crank axis and hub axis, inches)
% front sprocket size (Tf, # of teeth), and an array of rear sprocket sizes
% (12 - 22 teeth). A vector of 11 chain lengths is returned, each element corresponding
% to a rear sprocket size.

Tr = [12 13 14 15 16 17 18 19 20 21 22]'; % # teeth on rear sprocket

r1 = 0.25/sin(pi/Tf); % (PD = pitch/sin(pi/Nt)), pitch = 0.5"
r1 = r1*ones(length(Tr), 1);
L_t2 = length(Tr);
r2 = zeros(L_t2, 1);

for i = 1:L_t2;
r2(i) = 0.25/sin(pi/Tr(i));
end

theta = asin((r1 - r2)/L);
x = L*cos(theta);

% P = 2*x + (pi + 2*theta)*r1 + (pi - 2*theta)*r2 % "continuous" length,
% scalar

P1 = 2*x;
P2 = (Tf*0.5)*((pi*ones(L_t2, 1) + 2*theta)/(2*pi));
P3 = (0.5*Tr).*((pi*ones(L_t2, 1) - 2*theta)/(2*pi));
P = P1 + P2 + P3;

crankholio

If you want to be able to lock up the rear wheel reliably for a skid stop, you need an eccentric rear hub. With vertical drop outs, I'd either get an eccentric rear hub, or I'd just ride single speed.

yelro

yeah I guess eccentric hub for the rear must be real solution if you wont resort to the other proposed solution i.e ghost ring, rear derailleur, or filing the dropouts..

ofajen

Yeah, I’m running a Schwinn Sprint frame from about 85 or 86 and it works great for this. This was the bottom of the line bike, so i assume it has straight tubes. Still, it’s a sturdy, lugged frame made for Schwinn in Taiwan and still holding together after about 35 years.

Mind you, I’m just running SS, not fixed, but the semi-horizontal dropout makes it a cinch either way. I have two spots on the dropout where my 42/16 could work, depending on chain length.





Otto