All about Framebuilding Part 1


First off, this article series  (Book?) is about hobby building. It’s intended to share what I learned over a number of years of hobby frame building and some commercial framebuilding (that was really just hobby building with much higher costs.)


No one believes that their first frame will be trash. I didn’t when I built my first frame. Embrace the fact that your first frame will be trash. Give that trash a hug! Buy extra tubes ( or better yet, but straight gauge tubes). Copy an existing design.


Why copy an existing design?

If you’re enough of a bike nerd to want to build your own frame, you probably have an older steel bike that you like and are familiar with. Copy this frame. Copy this frame because you already know you like it. Frame design and geometry is a guessing game until you’ve made a number of frames in your size and then ridden them. You might also have a complete set of components for that old steel frame. If you copy it, you can move all of the components over to your new handbuilt frame and ride it. Chances are that you’ll love your new frame just because you made it, even though it’s a piece of crap. Embrace this bias and ride the wheels off of your new shitbox. Then break it down, build the original older steel frame back up, and realize that your first frame was inferior to the old production donor frame in every way.


Why a steel frame?

Steel is the easiest bike frame material. It’s light, strong, easily weldable, easy to braze, has a long fatigue life, and it’s inexpensive. If steel was invented tomorrow, the inventor would get a Nobel prize the next day.


But what about [some other material]?

New high end bike frames are made of carbon fiber. If you’re insistent that only carbon fiber will do, you should go to Dave Bohm’s school in Tucson. (Hat tip to Dave Bohm for getting )

Aluminum frames require welding, and post weld heat treatment. Also, welding aluminum is less forgiving than steel because of aluminum’s reactivity with oxygen and other contaminants, not to mention that you have to go a lot faster. It’s doable, but is a level 2 ( maybe 5?) skill. Heat treatment for 6000 series aluminum requires high temperatures that you can’t do at home and 7000 series can be hard to source for the little bits unless you also have a machine shop to make aluminum in any shape you want.

Titanium frames require professional quality welding and fastidious cleanliness. Titanium is also a much more expensive material. I’ve never built a titanium frame because I never had the money to buy the tubes and a TIG welder at the same time.


What kind of frame?

Your first frame should be as simple as possible. A track frame is ideal, even if track bikes are a little silly. A road frame also makes sense. A rigid mountain bike frame is also good. Single speed bikes are a great choice. Also, Single speed bikes are rad. You should make one so you own one and ride one.

DIY Air Bearings with Graphite

Hackaday recently posted about DIY air bearings here and here. Lots of good stuff in there, from lapping the graphite in to discussions of the stiffness of the bearings around the 3 minute mark in the next video.  Air bearings are nice because they give all of the benefits of hyrdrodynamic bearings without the oil mess, and they’re self cleaning because they push contamination out of the beaing.


I also found this nice explainer of types of air bearings at Specialty Components, and this nice breakdown of design considerations for air bearings at IBS Precision Engineering.

Side Pulls in the LMP Shop

2.810 uses a plastic gear that can be tough to source, and not all of them have a set screw, which makes them harder to use.  I wanted a mold making challenge and the ability to make gears at will, so I machined the molds.


My design goals were:

1. Mold a gear that meshes with existing RC car drivetrain components

2. Insert mold a set screw into the part.

Link to the CAD file.

The first goal was pretty easy. Fusion 360 has an add in gear generator. I knew the pitch of my gear and the number of teeth.  Find this generator under Design – Tools – Add-Ins – Scripts and Add Ins – Spur Gear.

When building the molds for the gear I only scaled the generated gear by 1%, as that’s a low estimate for polypropylene shrinkage. The mold could always be remachined to be bigger, so starting at the low side for part shrinkage made sense.


Creating the side pull is mostly about adding an extra parting line. I chose the centerline of this gear because it’s symmetrical, but any line that makes sense for your part is fine.  The side pull mounting holes and the hole that holds the set screw were machined first, then the side pull and rest of the core were machined as a set to avoid steps, mismatch and flash.

Die springs operate the side pull. These were found in a drawer of misc springs in the shop. They’re really strong – I can’t close it without a press. The side pull is closed in the injection molder with an angled plate. There’s a bronze plate behind the side pull to prevent aluminum to aluminum galling as the two parts slide past each other. The side pull has a section of allen wrench pressed into it, with a magnet behind it to retain the set screw.  It’s a little fiddly to put the set screw on the mold, but it stays on there well and releases well in the machine.

Shrinkage Factor for Polypropylene

TL:DR – 1.9%.


When injection molding parts, the mold sizes need to be slightly bigger than the final parts to allow for thermal contraction of the plastic. How much bigger? I wasn’t sure.

I wanted to create a lot of data fast, so I designed molds to create four parts with the same thickness, and I (arbitrarily) chose 0.100” as the thickness.

After molding around 60 parts, I measured them all in this scheme:



All of my measurement data is in this Google Sheets Link.

I used a scatterplot and linear regression because I wanted to see the effect of part size if it was present.  The data show a nice linear relationship between the mold size and the finished part size, with a shrinkage factor of 0.981.  A typical workflow for drawing molds includes scaling up the original part and subtracting it from the mold blanks. The amount to scale the part is 1/0.0981, or 1.019.

Design for Manufacturing – All about Axle Blocks.

TL:DR – if you need machined features, make the part a rectangular prism.

All about Axle Blocks:

In 2.810, many teams choose to make lots of parts out of waterjet aluminum plate, and then post machine those plates.  This is a pretty intuitive strategy, but isn’t the fastest. I created a couple sample parts and then filmed myself doing a time study for how long each one took.

Onshape Models and Drawings

The waterjet is still an amazing tool with great capabilities, but if you’re going to do any post machining (except maybe tapping waterjet holes), you’re better off machining the whole part.

Punch Press Tools – Lessons Learned

In 35-125, we do lots of flashlight trainings.  Once or twice a year I have to find another shop willing to do me a favor by laser cutting clear flashlight lenses.  I thought that making them with a press in our shop could be a fun project.


CAD Model Link.

Spring Housing CAD model link


Design Goals:

  • Make the lenses
  • Make all lens features in one hit, instead of reindexing blanks.
  • Strip both the lens and the material from the punch tools.
  • Minimize fingers between punch and die.


My initial design used a couple stacks of belleville washers around screws or shoulder bolts. After fighting with those washers, I don’t think I’d use them for much beyond indicating bolted joint or bearing preload. Stacked up inside a hole or around a screw they have a tendency to get stuck on screw threads or poor surface finish, and then I got a bunch of kinked belleville (former) springs.


The main takeaway from this project was about punch and die clearance – it seems like a tighter clearance would give better shearing, but it makes it really hard to get everything together and makes the parts much harder to strip off of their tools. After consulting Dayton Lamina I realized that the clearance I initial designed for was WAY too tight. I made everything to be a close slip fit, and I could have a great deal more clearance at 5% of the material thickness. On the second iteration, all the punch and die interfaces were given %5 of the material thickness (0.060”) on each side. That makes about .006” on diameter, which is easy to hit with an end mill or a reamer.



The parts for this tool were made mostly from Tight-Tolerance Hardened Multipurpose 4140 Alloy Steel Sheets and Bars. They are really nice, and at Rockwell C30 are soft enough to machine with high speed steel tools. (High Speed Steel tools have a hardness around Rockwell C60.)  We have nice sets of high speed steel reamers and drills in 35-125, and I didn’t want to order specific carbide tooling for all the holemaking on this project. Round punch blanks were made from drill rods. They can be purchased in most fractional sizes and are as hard as high speed steel tools.  They can also be machined into special drills if needed.



I wasn’t sure how long the drill rods as punches would last, so I made the punch holder out of three pieces with holes on the split lines between each part. That way I could remove and replace punch blanks without pressing them in and out of a housing.

All of the other parts ride on two quarter inch pins pressed into the die plates, 4 inches apart on the centerline of the tool. This allows all of the layers of the tool to be precisely located without a lot of fussy adjustments.



Punching through the material is only the first half of the process. Next you have to get the part off of the punch. Typical punch press tools do this in an orientation like this:



This design is pretty straightforward for a single punch, but gets harder to implement for a tool like mine because of the combined features. I tried a couple versions that were a round or hex button in the middle of the lens. These didn’t strip the lens well and tended to bend it like a diaphragm, making it bind on the hole punches even more.


I tried a number of urethane rubber sheets for the bottom stripper, and they worked well.  They initially got made as a byproduct of trying urethane sheets in the top cavity as a captive spring.  The 5 ton arbor press tended to squish those rubber sheets into useless stringy rubber bits.


The top stripper was where I found out all about throw, preload and durability of die springs. This is where I kinked and ruined my belleville washers. I ended up finding a spring with a 500lb/in spring constant, which turned out to be stiff enough to strip the lens out of the top cavity.