Lathes- metal spinning

[Curt Goodrich]

Not for the faint of heart nor excessive lathe accident researchers-

[Noteddy]

Thanks for that. I've always admired this particular craft. That fella has a callous from armpit to waist. Respect.

[Jake Rosenfeld]

awesome, I've spent some time looking at this and its just amazing what these guys build.

thanks

[bencooper]

Some other good videos from them too - metal spinning is wonderful to watch.

[11.4]

Y'all aware that Dura Ace 7600 high flange track hubs and the old Suntour Superbe Pro high flange track hubs are made by metal spinning? Story is, Shimano had one old guy who did all their high flange track hubs. He wanted to retire at something like 78 years of age, and was gone for a few years. The prices of the hubs skyrocketed, in the meantime Suzue went out of business, and riders were screaming for them. So they brought this guy back and now he just sits and spins hub shells each day, something he'd done for fifty years and always loved.

All the other hubs I'm familiar with are milled (or in a few cases forged). Spinning gives much nicer microcrystalline structure in the metal and allows the flanges to be as narrow as they are on those hubs. Campagnolo achieved reliability with their old Nuovo Record high flange hubs simply by going to an extremely soft alloy, at the expense of spoke holes that were too easy to ovalize. They tried to change it with the C-record series, both road and track, and just ended up with brittle flanges that fractured easily. Phil Wood combines tougher alloy and much thicker flanges plus a local hardening of the spoke hole by the method they use to create the spoke radius.

[11.4]

These hubs were actually spun in two pieces. The blank was a short thick round billet, clamped at the headstock on what would be the inside of the hub, on which the flange was raised all the way from about a 22 mm diameter solid form to an 82 mm flange diameter. The blank was held between a 4-jaw chuck and a 4-jaw tailpiece so it was completely rigid, and was spun from both sides. Once the flange was up, the core was bored through the headstock. Then the piece was parted from the tail and the cavity for the bearing cup was spun (this is why it needs a large plastic cover -- you can't spin that far into the blank and also get the cup in without having a sizable hole to work through. Finally it was parted off the headstock.

At that point two of them were checked for accuracy of the bearing cup seats and for general dimensions, and two of them were mounted with the center sections facing each other and with an expanding mandrel through the axis hole of both of them for support. The joint between the two flanges was made by spinning -- you can actually fuse metals without heat by spinning them. If you look at Suntour Superbe Pro track hubs, you can usually see the marks of the join in the midpoint of the hub. If you have one apart, you can see a short ridge where the metal is spun against the mandrel and then evened out. It's only occasionally visible on the Dura Ace because they used a more precise internal mandrel so the join formed more smoothly. The join is as strong as the rest of the hub and has never been a point for breakage.

We did a big study of flange breakage a few years ago, collecting hundreds of hubs with broken flanges and also cutting apart a lot of hubs (broken and not) to examine the metallurgy. When you see the benefit of a spun hub flange, it makes you want to train a bunch of apprentices to make hubs. I watched this old guy at Shimano playing around one afternoon -- he was spinning track chainring blanks that were amazing and was actually using spinning to shape and join alloy tubular rims -- post-forming the extrusion to relieve spoke hole cracking and push metal into the actual nipple bed. He could modify a rim every three or four minutes and sold them to keirin riders for four times the price of the un-spun rims -- they could take off 50 grams and yet have a much stiffer rim, all while leaving the NJS stamp in place. He said the keirin pros cheated every chance they could get and laughed.

[Blakey]

These hubs were actually spun in two pieces.


We did a big study of flange breakage a few years ago, collecting hundreds of hubs with broken flanges and also cutting apart a lot of hubs (broken and not) to examine the metallurgy.

I don't suppose you have any photos of spinning at shimano or micrographs of the hubs you examined? Interested materials engineer over here.

[11.4]

I do have videos of both Suntour and Shimano manufacturing. The details are quite different because Shimano has a separate traditional bearing cup on each side, while Suntour managed to sneak in the back door of NJS approval by using annular cartridge bearings but building them up with what looked like a regular cone on the axle. Suntour was very smart about this -- it meant that you can replace bearings on a Suntour Superbe Pro track hub and was simply a really smart piece of engineering. You wish Suntour were around today to design as smart a bottom bracket.

On the hub study, we have hundreds of micrographs and fracture data, plus we did dye perfusion and fracture analysis on broken hubs to see what had caused the breakage. Part of the images are at MIT, part at the Lamborghini test lab. We started with a focus on Campy C-Record track hubs, just to see whether they were unfairly targeted as fracture-prone and to see what might have caused breakages. One question was whether radial lacing was really an issue or not -- a question that a number of factory wheel manufacturers were trying to get a better answer on. We actually found, after 867 broken hubs, that C-Record hubs had no greater breakage rate than any of several other hubs, that radial lacing didn't appear to increase hub breakage, that changing spoke entry angle when relacing hubs with spoke hole set marks from preexisting lacings didn't increase breakage. We found that CNCd hubs had the highest breakage rates, except for Phil Woods, and that was because Phil didn't chamfer the spoke holes but actually used a method similar to spinning (one that Zipp copied for their 28-hole high flange track hubs) that relieved stresses on the spoke holes such that it didn't matter how the rest of the hub was made. There's a ton of info and a bunch of it went to some of the big factory wheel companies, but we don't really have anything coherent that could get posted here. Some of the later work was also under contract with wheel manufacturers, after we were able to show the interesting results we were getting, and it's their proprietary work. Most of the interesting data came from macroscopic studies relating to basic hub design, like that Chris King Classic hubs had their holes so close that many failures consisted not of breakout of a spoke hole bridge, but instead the breakage pattern zippered from spoke hole to spoke hole until the whole area simply tore free. That C-Record hub breakages that were publicized so broadly actually involved something like a wheel driven over by a vehicle -- vectors of critical deformation force were not remotely in line with spoke tension. That Nuovo Record hubs only broke because of corrosion in the flange and if hubs were honed by hand, it sealed the surfaces to corrosion and breakage rates dropped to the lowest of any hub we ever tested. And so on. We wanted to do the same kind of analysis on custom hand built steel frames so we could see how frames might break, but we heard that eRichie was going to put out a contract on us, so we demurred. Anyway, Vivalo was imploding (literally and also in a corporate sense) and we did a study in Japan to understand what had gone wrong, and then took some of that information to a certain very large US frame manufacturer who was having alloy steer failures. They needed better metallurgists.

Anyway, hope I've filled in a couple little gaps. If I can get permission to release more of this stuff, and then assemble it, we could have an interesting discussion about frame failures. And learn to make our own high flange hubs at the same time.