These worked great on the main tubes. However, I need to fab up something smaller for the stays.

Time for the braze-ons. As you can tell, I like making tools. These are off the shelf Pony brand spring clamps with the rubber protectors removed. I took a some pieces of 1/8" welding wire and bent them into shapes and silver brazed them into the convenient groove in the clamp.

It's precarious, but it held. Again, joined with 56% silver.


Scribed for bridge placement. I checked with a wheel to make sure I wasn't being a moron.


It worked well though. Much easier than trying to clamp a 2" long piece of tubing in a tube block. The bridge is 1/2" x .028

Pretty straight forward. Nothing more than a stick with a hole and a pinch bolt/slot.

Here is a simple clamp I made for mitering the bridges. I made one for 5/8" and one for 1/2"

The tool stays on during brazing. I'll be attaching everything with 56% silver.



This is the tool that I made to help me miter and position the seatstays while I join them to the back of the seattube. It's very simple and does nothing more than clamp the stays at an angle and allow me to test fit them on the jig and then return them to the vice for some more filing.
It's a block of aluminum about .75" x 1.0" with 3 tapped holes in it. Above it is a clamp with corresponding holes and 1/4" cap screws. Sandwiched between are 2 mini V-blocks. They are attached from the back with 2 more 1/4" bolts so the angle can be adjusted independently of the clamp. Considering how many builders regard hand mitering S-bend fastback stays to be nearly impossible, this tool worked very well.

Ready for some stays.



This worked well. The pins are tight but can be removed easily with a pair of dike cutters. The bullets are a bit floppy but don't wiggle at all when the stays are installed.



It then used a transfer punch to mark the dropout and drilled it out as well. I initially tried using an undersize drill so that the spoke would have a press fit. My plan was to grind a taper on the pin so I could get it started. This didn't work so well. I discovered that the pins don't need to be tight to work in this application. Instead, I made the holes oversized and bent the pins slightly before tapping them in.


In order to keep the SS plugs on the dropouts while I was fitting/brazing the stays, I decided to pin them. This is a common technique for those building using tradition methods and lugs. The standard "pin" is a soft nail. Some builders even use square nails because of their natural taper.

I've never tried using pins myself so this was new territory for me. I chose to use an old 14ga stainless spoke as the pin since this area will not be painted. I first drilled an .080 dia hole in the bullet. Location really didn't matter.

1 pair of stays with S-bends. I was initially worried about making sure the bends were in plane to each other. I realized that you only have to get close. Checking them on flat surface is easy and minor corrections with a vice are also quick and easy. Also, do the larger bend first.


I did a couple of test bends. I got a piece of 3/4" x .035 to 25 deg before it started to crimp. Plenty good enough for S-bend stays. I can put the bends about 4" apart if I move one of the outside dies towards the center. This limits the bend to about 18 deg.
My old mill vice makes a decent press. It takes a couple of extra hands to get everything lined up but it's better than buying an arbor press.




This is my stupid simple tube bender. It's a blatant rip-off of the one I found at Steve Garro's blog. I modified only to fit the materials I had on hand. The "dies" are all machined from 7075 alum. The base is simply a piece of .120 wall scrap tubing with 2 pieces of mild steel bar stock stick welded on it. I did key the pieces of bar stock to make sure they ended up on there straight. I made all the dies on a vertical CNC mill and surfaced the curves.

Fit up and ready for some seatstays.


Radiusing the tops of the dropouts


Originally I was planning on using 3/4" x .035 for the seatstays as well as the chainstays. After getting the yokes all done, I was amazed at how stiff the chainstays were and realized that 3/4" was overkill for the seatstays, especially on a bike that I'm trying to keep light. I switched to 5/8" x .035 This meant that I needed to make new bullets. These are much lighter. I even added a little cutout on the side just to save those precious 2 grams.

A view from the bottom.


The cups worked. I also used my normal Harris white flux and tried to be a bit more patient with the heat. On the second of the 2 frames, with more BB warpage and therefor more gap in the joint, the silver flowed very easily. If I did this type of joint again, I would rethink a few things like designing an intentionally smaller radius on the yoke to create a larger gap with the shell.



And with the bike....


New cups installed on the frame fixture.



My solution to the heat control issue was to make some new BB fixturing cups. My standard setup is a stainless insert that is a light press fit into the BB shell. This slides on a 1" dia post on the fixture. The insert makes near full contact on the inside of the shell. These new cups only hold the outer .200 of the shell on each side. This will allow the center of the shell to reach brazing temp with less heat. I made these out of 6/4 titanium. Why? Because Ti is a crappy conductor. I wanted a material that would help contain the heat in the frame. ( I couldn't afford ceramic)


Well that didn't go as planned. It seems my fit between the yoke and the BB shell was a little too perfect. There wasn't enough gap to pull the silver into the joint. I also had heat control issues. I tried using black flux since I was anticipating some overheated spots. The BB post of the fixture acts as a massive heat sink and prohibited me from getting the shell hot enough.

Everything fit up (minus the front end)


Checking fit. I took this whole assembly and installed it on a wheel to make sure I wasn't doing anything stupid.


Checking for square.

Trimming the chainstay to length with a file.

Small .050 vent holes in the yoke were an after thought.


Trimming the chainstays to length. This isn't that critical as my yoke design allows for lots of slop.


Checking to make sure the new block/dummy axle is square.



Here is my scabbed on block to extend my dummy axle.


I ran into a couple of snags on the frame jig. This is really the first complete bike I've built on this jig. First off, it isn't adjustable enough. If you build your own jig, make sure it can go well beyond any norms for geometry and size. This bike has such a low BB, that the headtube adjustment is maxed out and the dropout is well out of range. Another issue was the headtube cups. The originals were stepped. This made setup easier and more accurate, but when the headtube ovalized slightly during brazing, the cup was stuck. Combined with the cooled flux, it was a chore to get it out. Tapered cones are the answer. I will still use the other ones where I can since they are easier. Another lesson I learned is to allow more slop. My old headtube cups had .001 of clearance on the ID. This would have been plenty if it hadn't gotten hot. The new ones have .003. Same with the BB post.

Tabs removed.



Second side complete but with the tabs still attached.


Part bolted to the fixture. I start out with just 3 bolts and machine around them.


Simple fixture for the second side. It's just a piece of scrap with the bolt pattern drilled in it to match the part.



First side complete. The holes are "tabs" for holding the part for the second side. 2 of the holes are reamed for a .250 dowel pin. This ensures the features on the second side are in perfect relation to the features machined on the first side.


Into the mill. The first side can be done in a vice.


The part starts as a 1" thick piece of 4140. Cut to length, and well deburred.

The chainstay yoke: This will be machined from a solid chuck of 4140 plate. Modeling done in Solidworks. Programming is done in Mastercam.

I learned a few thing about brazing over the last few days. Some that people have told me before and have finally clicked. Others that I figured out on my own. I'm not a master by any means. However, half the motive for this blog is for my personal notes.

1.) Always work uphill with the torch pointing in the direction of travel. This gets said over and over but if you're not doing it, you need to be.

2.) Preheat: I realized that getting the whole joint dull red, including the backside of the mitered to tube, makes everything easier. It helps with distortion and it helps with heat control.

3.) Heat the inside of the tubes. I've decided that heat sinks do more harm than good because they cause cold spots and prevent you from getting the torch down the inside of the tubes.

4.) Keep the filler rod close to the flame. If the filler rod is 4" away, when you add it to the puddle it takes too long to melt. Keeping 1/2" means that when you add filler you get molten brass right away.

5.) After you add filler, make sure the puddle wets out to the root of the joint. I found myself concentrating on the shoreline of the joint. I noticed the puddle would wet out on the sides a moment before it wetted out to the front (the root). If I moved ahead and added more filler, I would create pocket and seal it when I added more filler. This may have been the reason I was getting poor penetration on my practice joints and went to a wet first fillet second technique. I may try some practice joints with one pass and see how it looks.



When I first started building this bike, I tried this joint in brass. I couldn't get it flow so I ended up using silver. I was concerned about the later joint and the higher brazing temps. You can see a bit of silver bleed out, but overall everything is fine.



Looking better still. Since this bike is being built to be ridden, this one probably won't see much filing. My goal is to get good enough so that I never have to file fillets.
As a machinist I was taught that a good machined finish is harder to achieve and therefor looks better to the trained eye than a sanded or scotchbrite finish. I feel the same way about fillets. Anybody can lay on a bunch of brass and spend an afternoon filing it all off. It takes a lot more work to make a fillet that can never see a file and still look good under paint.



BB went way smoother than I could have expected. I did have quite a hard time keeping the ST/DT intersection from overheating, but it flowed with a bit of patients.



I got the whole front end brazed up today. This is my first joint. It's not into the realm of the "no file fillet" yet, but I'm getting there. I learned a lot and things started clicking.
The homemade headtubes with the extra thick ends really paid off on this joint. I had zero issues with heat control even though the brass went right up to the edge. My TT is only 7mm from the top edge of the headtube.

All jigged up. The black paint on the HT/DT was just me playing with fillet filing. Paint helps show weird spots in the fillets.


Test fit.


This is the downtube after I cut the relief for the seattube by hand. This was much easier than I was expecting.


Do the math and move the gage up. I know, I need to buy a real indicator clamp.



Setting the headtube height on my jig using a height gage. My BB post is exactly 6" off the table so I use a 6" mic standard to set zero.

I just use a good quality 24" machinists scale to set it.



This is the tool that I use to measure tube length. It's basically a fancy trammel bar, but instead of a point one both ends, I built it with a 1/2" dowel pin. The dowel pin sits in the valley of the first miter and the scribe indicates the position of the second miter.

Cutting the miter: I use a 3/4" dia serrated roughing endmill to cut the miter. I program the mill to cut a radius .010" bigger than my mating tube. This extra seems just the right amount to limit hand finishing.
The serrated tools are really good because the tool pressure is up instead of in. This means it is less likely to fold a tube. They also last forever. I've never replaced this tool. They are expensive at about $60, but they pay for themselves.



Finding the centerline of the tube with an indicator. Sweep one side, then sweep the other, when they read zero, you're on center.


Here is my setup in the mill. The extra tall vice jaws are only required when doing steep miters like the DT. You can just make out the angle block on the fixed jaw.
This one is easy cause it's only 2 deg.



To set the angle of the cut, I use angle blocks. These go down to 1/4 degree. That seems to be accurate enough.



Top tube clamped with matching blocks. By rocking the blocks on the surface plate, I can tell if they are aligned. With a little trial and error, I can get them near perfect.



I finally got some pictures my tube mitering technique. I use tube blocks made of aluminum. They are matched sets. For most tubes, I only need to use one at a time. When I cut the top tube, I use both so I can keep the miters timed to each other. If I made these again I would make them shorter.

Some Glamour Shots of my dropouts. I think they look pretty hot in satin. I might leave them and for go the polish.




I also got tired of wrapping utility cloth around random stuff so I made some sticks for the job. I also made one out of Delrin. Again, it works better than the oak.




I didn't have a 3/4" tube block so I made a new one. This time I thought Delrin might make a good material. It's plenty tough but it won't mare metal. It's reasonably cheap, though hard to find in small quantities. It works well with wood or metal working materials. You can cut it on the table saw and bore it on the mill. You can't say that for maple or aluminum. It worked so much better than my wood ones that I'm going to make a whole new set and resign my old ones to brazing only. You really don't want to see what happens when you get Delrin hot.

I forgot to take any pictures as I was putting this together. These were done with 56% silver. I drilled two 1/8" feeder holes in the top and bottom of the chainstay so I could get silver into the joint. I only fed from one hole. The second hole is just so I could see that my silver went all the way around. These will be filed flush and never seen. It was a little tough getting the dropout hot enough without cooking the flux on the bullet clevis. Next time I will use black flux.


The fixture is symetrical with the slot in the dropout being on centerline. Reversing the v-block that holds the cs makes the fixture work for the opposite side. Measuring from the edge to the block gets the angle close to the same. By moving the block to a different slot, I could build a bike with a different cs angle.


I realized that I needed a way to fixture the stays to the bullets to the dropouts all at the same time as they would all be brazed at once. It didn't have to be super accurate, just repeatable so the left side would be the mirror of the right. If I was going to take any time making a fixture, I might as well make something modular that I could use in the future.


Like so.


I filed a radius on the tab of the dropout to fit snug in the slot of the bullet.

Cutting the slot for the dropout in the bullet.


Attaching the dropouts to the stays presented an issue. I didn't want to use an off the shelf stay for several reasons. One is the chainstay yoke. I wanted to make the yoke from 3/4" plate for cost reasons but I couldn't find a stay that was the right size that would fit cleanly on the dropouts. I decided to make the chainstays from 3/4" x .035 tube and join them to the dropouts with "bullets". Paragon makes these in Ti, but not steel. I made my own from 17-4. I lack a CNC lathe so mine had to be tapered vs bullet shaped.


These are the Paragon Machine Works dropouts I'm using. These are becoming the dropout of choice for builders and riders when it comes to singlespeeds. They appealed to me because it allows me to run SS or geared without running tugs. They are also much easier to build around than an eccentric. I could have made my own, but at $90, they are a bargain.



My breakfast looked so good I had to take a picture. How much better is this than work?


Yeah! Snowday. This much snow is a big deal in Seattle. It's not that we're soft, it's that MOST of us don't know how to drive. Hey, any accuse to stay home from work.

They look much better after the flux is soaked off.

Joint done. I'm not the best brazer, but I'm getting better and I'm confident in their strength.



Crappy picture of my tacking sequence. Obtuse angle, acute angle, left side, right side.
Then I do a "tinning pass" to make sure the brass gets good penetration and builds the internal fillet. If I don't fully consume the tacks in this pass, I get voids.


Fluxed and ready for the torch. Notice the heat sink/weight used to keep everything where It should be.


Then I hit it again with the cartridge roll. Then a quick wipe of alcohol before the flux goes on. Dawn works so well at removing oils, that the tubes corrode as they dry.

Here is the over-engineered tool I made to clean the inside of the tubes. Just a piece of brazing rod bent around half a sponge. Worked quite well combined with Dawn dish washing liquid. If you don't wash all the way down the tube, you get a cool fireball as the oil burns off.




Using cartridge rolls on a die grinder to remove the mill scale from the ID of the downtube.

Good enough for me


Woops. I goofed on the drawing and didn't realize it until after I had cut the HT miter on the DT. I'm off by about .75 deg. A little hand work required to get it where it needs to be.


Turning the headtubes. This started as a piece of 1.625 x .156 4130. The outside was turned to 1.600 for cleanup and the inside was bored for a headset plus .010. The center was relieved for weight. Wall thickness was left at .060.
The logic on this was to provide a nice large tube to attach the 1.5" DT. This made the fillet braze much easier. The integral "reinforcing rings" left on the headtube meant I could braze closer to the edge with less distortion and fear of overheating. It worked well. Heat control was much easier than with normal headtube stock. Since my headtube is on 4.25 long, I only left 10mm below the DT and 7mm above.

Post braze and still hot. A little saggy but not bad for doing the whole joint without rotating it.





Tubes loaded (minus washer and nut) and ready for flux



Checking the distance from the ST to the support. Checking at the top and the bottom ensures that the tube is indeed square.


The ST/bottom bracket area of my frame fixture

Thoroughly deburred


Using a #2 center drill for the vent hole.


Scribing the centerline of the shell for the vent hole location. I don't do this with my good calipers.




The bottom bracket shells are from Henry James. I really wanted to use the heavy walled ones from Paragon, but they were out of stock. Since the HJ ones come out of a piece of tube, they need quite a bit of cleanup. I sanded them in the lathe.

Two assemblies done. One in 56% silver, the other in brass. 1/4" vent holes drilled and well deburred both sides.


Well, I ran into a bit of an issue while brazing the ST sleeve. I was initially going to use brass because I was worried that if I used silver, I would run into issues later when doing the toptube joint. (It will be fillet brazed brass)
The issue was that I couldn't get the brass to flow very far into the joint. There just wasn't enough gap to suck the brass in.
I tried another joint in silver. (I'm actually making 2 bikes at once)
This went a little better but I ran into another issue. My original plan was to suck the filler from one end to the other. 4" proved to be just too far so I had to feed from both ends. However, I had neglected to drill a vent hole and when I wetted out the second end, the silver wouldn't flow. In a bit of a rush, I quickly drilled a vent hole and continued the joint. I did end up getting a lot of silver in the joint.
I also realized a drawback to my tapered sleeve design. While I like the aesthetics, the lack of a "shelf" where it meets the ST makes it hard to feed the silver without it running out onto the sleeve itself making cleanup difficult. If I did this again, I would either do the traditional design, or braze the sleeve and THEN cut the taper. And definitely drill vent holes first.



The sleeve and ST assembly in the lathe for a little trimming and cleanup.

Ready for the torch. The flux is a bit lumpy cause it's dry. I didn't feel like walking all the way to the house to get water. It works all the same. I preheat the tubes with a blowtorch before applying flux. Just enough so the tubes "sweat". This way the flux dries instead of boiling and falling on the floor. Both the flux and the brass are from Henry James.

Here is the seattube sleeve all cleaned up and ready to braze. I'm leaving about 1/4" of ST sticking out so I have somewhere to feed the brass.

Here is my seattube sleeve. It starts as a piece of 1.25 x .058 4130. I turned what will be the bottom end to a taper to help distribute stress. I like the look better than the more standard scallop cut.

Got the main tubes all checked and marked.

Before I start cutting tubes, I wanted to make sure I knew exactly where the butts were, so I made this little tool. It's a simple piece of 3/4" SS with a 1/4-20 hole tapped crosswise in one end. I filed a nice radius on the end of a set screw and loc-tited it in the hole. With a travel indicator positioned correctly, it will tell me the wall thickness about 8" in on either end of the tube. This will ensure that I don't accidentally cut too much off one end of a tube. It also serves as a redundant quality control on the tubing manufacture. Second picture shows the tool in use with my butt start and stops marked with a sharpie. I like to label the wall thickness just to help keep thing straight.

The frame will be a hardtail mtb with very aggressive geometry. It will have a very low BB and very short chainstays.
Tubing will be a mix of Zona and True Temper HT. The stays will be straight 4130 with an intricate machined chainstay yoke.
Paragon sliding dropouts will allow for easy SS or geared use.
I drew the frame in Solidworks.