Lug Practice!
The water bottle bosses were actually not all that hard, to be honest. But it was good to get the torch running and start figuring out how to read the flux. Essentially, as the flux heats up it goes through a couple different stages:
- Pasty white goop. Looks like this when it comes out of the container and gets brushed on.
- White with active bubbles. As the water evaporates, the flux starts to bubble and get visibly dryer — more caked crust, less goop.
- Hard white with clear pockets. As the flux heats more, it develops little pockets where it’s clear.
- Clear, almost glassy. As most of the flux goes clear, I’d look for the spots that still had the little pockets and heat them until that area went clear.
- Clear liquid. At this point, the silver is ready to add.

Tube Mitered and Ready for Assembly
So on the second day we did additional practice on a more complicated piece: an actual lug. Before joining the tubes, one tube has to be mitered — essentially trimmed and filed so that its profile matches the intersection with the other tube. This maximizes the surface area between the tubes and lug that will get brazed and makes for a really strong joint. Once mitered, both tubes and the inside of the lug had to be given the rubdown with emory to clean the surfaces where the joint would form.
The lug joint is a much larger piece, requiring more silver. Essentially, you put the joint in the vice so that you can add silver at the top and let gravity/capillary action pull the molten silver all the way through. Once a ring of silver develops on the bottom part of the joint, you know you’re done. You have to add to the joint in the proper order to ensure that you don’t get too much warping. First, add along tube centerline on the larger angle, then add along the tube centerline on the smaller angle. Then add on the sides. Ideally you just add a small bit of silver in each place to tack the weld — then go back to fill the whole thing in.

First Practice Lug Done!
The lug required a lot more heat than the water bottle boss. It’s not enough to just heat the silver when the flux is up to temperature, you have to heat the area where you want the silver to go. So the entire lug has to be up to temperature. On my first lug, I used way more silver than I needed to — you could really see globs all over the place. But that would get better over time.
Getting Down to Brass Tacks
After practicing with a few lugs, we moved on to the brass brazing technique that’s typically used on the dropouts. Silver can really only fill in very small gaps, but brass can be globbed on to fill much larger ones. Brass also has a higher melting point that silver, so a dropout could be brazed into the chainstay using brass and then rack eyelets could be added on using silver brazing without melting the brass-brazed dropout.

Practice dropout all fluxed up.
But even with bigger gaps, the process was pretty similar. We slotted a tube and then cut notches in a fake dropout, then fluxed the area where heat would be applied. In this case we used the “blue” flux that’s specific to brass brazing’s higher temperatures.
For silver brazing, typically the torch is set pretty low — the oxygen tank at 5psi and the acetylene tank at 5psi. Once the torch is lit, an 8-10″ flame with a secondary cone about 2-3x the size of the primary cone is good enough to get the job done. Because the brass melts at a higher temperature, the acetele gets cranked up very slightly to 7-8psi and you use a little stronger flame. Bike building really only uses a 0 or a 1 size torch tip, but we did size up to the 1 for the brass brazing.
Heat is applied mainly to the dropout (or in this case the steel plate) — since it takes much more heat to get that up to temperature than the thin-walled tubing. Once up to temperature, we started feeding melted brass rod into the hole. It took a lot more than it did with the silver, since the gap we had to fill was significantly larger. The brass was much harder to work with — I got globs all over the place. And because I kept the heat on so long, some of the copper started separating out of the brass (the reddish-brown spots on the finished product). I’m sure practice makes perfect, and I did get better after a few attempts. But the challenges I had here were part of the reason I went with socket dropouts instead of brass-brazed ones. But more on that later.
Next up: Let’s draw pretty pictures.

Questionable brass brazing