abshipp 

This thread is inspired by a post I made in the ISO thread about wanting to replace a Pelissier track hub, better to use a separate thread so not to clutter up that one. (Thanks Unca_Sam )

So this is the first track hub I've ever owned, and being someone who was in college during the fixie boom of the 'aughts - but never had a chance to own one of my own - I wanted to learn a signature fixed gear trick: skidding the rear wheel.

Armed with my new wheels freshly built up, I headed out to a gravel patch since I figured it would be easier to get the body mechanics figured on a surface with less traction. A few tries in and I get the wheel to lock - only to feel the cog unscrew. WTF?! I had a lockring on there! And I'd ridden the wheel for a few hours beforehand so the cog should be tight! And I literally just made sure that the lockring was nice and tight against the cog!

So this is what greeted me back in my garage:







It appears that the hub shell failed at the lockring threads.

In trying to figure out why this happened, I discovered a few things:

• The distance between the flange that the cog bottoms out on when fully threaded on to the hub to the start of the lockring threads is approximately 5mm
• The cog I had on there is 7.5mm thick (17t, 3/32 Origin8)

So I'm guessing that I was trying to rely on too few threads to resist the force of the cog unscrewing. Given that this was my first ever experience with a rear track hub, I didn't think much about it when I only had a few threads recruited by the lockring.

So this begs the question, are modern cogs thicker than ones that would have been available at the time this hub was produced, or is this hub just poorly designed? For an additional data point, I also have a 16t 3/32 Dura-Ace cog that measures just shy of 8mm thick.

abshipp 

Here's an annotated picture for reference:

oneclick 

I just went and measuered a cheapo flip-flop hub, 5.47 from the first shoulder to the second - but the outer threaded section does not start as yours, it has a clearance band of about 1mm before the threads. The sprockets I measured were 6.8-something, and when they were spun on finger-tight the bottom of the lockring-section threads were all clear of the sprocket.

JohnDThompson 

Were you perhaps using the wrong lockring; e.g. an English thread 1.29" x 24tpi lockring on a metric thread 33mm x 1mm track hub?

In any case, there still appear to be a couple good threads left. I suspect you could get enough lockring engagement to safely use the hub, as long as you avoid skid-stops that put a lot of stress on the lockring. Install front and rear brakes, and use those to stop your bike instead.

63rickert

Yes, modern cogs are thicker. Also tend to be plain flat slabs of metal with some teeth around the outside. Old fixed cogs are machined to be thinner in the middle. To give the lock ring a place to go. No surprise that a Dura Ace cog would be one of first to do it the cheap way. Shimano always cut corners and still does. And always chooses the wrong corner to cut. And by putting a Dura Ace logo on it validates everyone else who wants to do it the junk way.

I can remember purchasing TDC cogs for 50 pence that were done the right way. Now it is an arm and a leg for a very simple piece of hardware and they do it wrong every time. This is why you buy vintage and run from anything Shimano.

63rickert

Just measured and compared a 2003 Suzue fixed hub against a 1930s FB and a 1960s Campagnolo. The lockring threading/spacing is set up to work with thicker cogs on the more recent hub. 30s FB and 60s Campy are of course identical.

The subtle erosion of every standard at every level is how Shimano creates obsolescence. A lot of it is not even intentional, they just started marketing before they knew squat about bikes. Good enough to sell but not good enough to ride. The purchasers of Shimano slime are equally to blame.

rustystrings61 

I hadn’t thought before about there being differences between vintage and current fixed-gears cogs but of course, it makes sense when I think about it! The only vintage track hub I ever used was a 70s Maillard that I ran with a SunTour 16T - and I bet when measured that SunTour will be thinner than current stuff. Is this the flip side of all the dire warnings against the use of “cheap, stamped cogs” - their dimensions are fine in vintage steel hubs but lousy for more contemporary hubs?

I’ve run Suzue Pro-Max, Kogswell, Formula/Iro/Origin8 and Surly track hubs through the years with Dura-Ace, Surly, SunTour and my personal favorite Euro-Asia Imports cogs with no issue. The next time I have wheels off perhaps I should pull some cogs and break out the digital calipers!

delicious

I agree with John. What cog and lockring were you using?

abshipp 

Thanks for the replies, everyone. I read all of them on mobile but wanted to wait until I got to a desktop to reply.

Interesting, are your cogs a recent production or are they older? Your cog is .7mm thicker than mine - not a significant difference in my eyes, so your modern hub must be designed around a thicker cog.

The lockring I am using was the one that came with the hub, it has no identifying marks and is made of steel. The lockring threaded on to the hub properly, so I doubt it was the wrong one. The cog I was using was a brand new Origin8, 7.5mm thick.

I can thread the lockring on and get it reasonably snug, but I really don't think it's doing much. I do run front and rear brakes, I want the levers on the bars for aesthetic and ergonomic reasons, so I might as well run both brakes as well I rode it a few days ago without using the brakes and the cog is secure enough to do that here in flatland IL, but I haven't yet found the breakaway torque for the cog when I'm braking with my legs. I assume it's a bit higher than I think.

Very interesting, thanks. My cogs are machined, not stamped, but they are fairly thick. I do not know the age of the DA cog, but the Origin8 I bought new earlier this year. I searched for vintage track cogs, I never noticed before that they were machined thinner in the middle. There might be enough good threads that I could still use the hub with lockring if I picked up a vintage cog.

That's an interesting idea, it kind of pains me to think of using a stamped cog but maybe it would be advantageous here.

My other thought, which might sound a bit blasphemous, is to pick up a Miche carrier system. It might be possible to permanently or semi-permanently install the carrier onto the hub with loctite - then all the lockring needs to do is hold the cog in plane tight against the carrier. There shouldn't be any significant side loads, and the threadlocker could deal with the rotation loads.

Not sure if I want to go down that road though, kind of feels inappropriate and a bit of a kludge.

repechage

not a Shimano fan but my hunch is that there came an NJS standard that was adopted- even if not marked NJS they migrated- but no tech mention -
I also suspect a small chance the lock ring may not be 100% compatible.

if mine, I would machine the backside of the cog to insure 100% engagement of the threads left on the lock ring threading of the hub.
and as Mr Thompson notes- no skid braking.

abshipp 

I did think about that. Taking 1-2mm or so off the back side would give some more room for the lockring, and moving the chainline in that tiny bit wouldn't be terrible especially since I run 3/32 drivetrains at the moment and road chains can deal with the slight misalignment.

Now if I only had a lathe...

juvela

-----

Sorry to read of this event.

From your very good description it sounds to me like it may not be a mechanical failure so much as a spanner twisting one.

After that first climb or sprint with a newly mounted cog it is necessary to retighten the lockring as the cog is certain to have cinched down a bit.

By not doing so you left space/play between the lockring and cog such that the first time you placed reverse pressure on the pedals you had this event.


-----

63rickert

A few weeks back was just riding along with a much slower friend when a fast group came by. And then I felt a familiar hand on my backside giving a big solid madison push. Was instantly midfield in a downhill/uphill sprint. Won the sprint. Second place told me (definitely no onboard electronics on the ‘58 Rivetts) the max speed downhill had been 36mph and his speed at top was 28mph. I was in a fixed gear of 48x19. So about
180rpm downhill and powering over top at 140 or better. Was very glad to have good chainline and good chain tension/slack. Chain would most definitely have come off with bad chainline. Messing with chainline on a fixed is a very bad idea. Don’t do it.

Old fixed cogs were machined on the front face. This left an attractive shelf underneath the teeth and a recess for the lock ring. Some lesser cogs were made of thinner metal and had a dish pressed into them. Again leaving a recess for the lockring. The reason and the only reason for flat slab cogs is cost cutting.

Juvela makes an excellent point. My recommendation would be put the hook spanner in your back pocket on first ride. Dismount and tighten the lockring after first mile and again at top of first hill. Continuing to use that fractured hubshell might or might not work, seems like asking for trouble to try and keep it in service.

My experience with older and better Japanese parts is limited. What I have seen is that the anglophilia and francophilia were extreme. They did not miss much and they knew how to make a copy. Shimano, on the other hand incessantly messed with absolutely everything, usually with no understanding of why the original version worked.

juvela

-----

BTW -

in case it should prove of assistance Outillage VAR offers hub dies

here are two pages from an early 1980's catalogue showing what was offered at that time
you would of course want to check current offerings...





[pages posted as example only]

-----

JohnDThompson 

Unfortunately, none for track hub lockrings.VAR-39 is for bottom bracket cups and VAR-40M is for hub spocket and freewheel thread. A thread file could clean up the damaged area on the OP's hub.

clubman 

Nothing like a BSA #3 track cog. I've still got a couple of 18s.

Oh and a Madison push is equally delightful, even in a road peleton

juvela

-----

Yes indeed!

Have no idea what VAR offers today; hence the disclaimer at the bottom o' me post.


-----

clubman 

I have a couple of excellent condition 16 tooth Brit cogs, and probably a lockring if you want to talk. I can measure.
1/8th of course.

juvela

-----

abshipp -

the groove visible just inboard of the hub's threads and outboard of the drive side flange indicates that Perrin made it of BSC threading. this means easier cog and lockring availability.


-----

JohnDThompson 

Like this:

63rickert

Ah, 18T cogs. We have a traditionalist. I would like to, but never seem to get warmed up starting on that.