3D printing electronics: solder extruder version 3

Impatient Inventor and I have continued work on the solder extruder.  To dispense the solder more consistently, we decided to try using a pump to pulse air into the solder chamber and hopefully thus pulse solder out the bottom of our extruder.  This depends on an amount of pressure not possible with our previous nozzle.  So, for this version, we opted to use a pipe tee:  we will feed solder into a barbed fitting connected to the top, use an aquarium pump to pulse air in into a barbed fitting on the side, and extrude solder through a barbed fitting on the bottom.

Given that we had a bigger extruder head, we needed more heat or more patience.  We opted for more heat:  we decided to connect four soldering irons to the extruder.  To do this, we got the help of a talented welder friend of Impatient Inventor’s.  He welded four 1/4″ (the inner diameter of the soldering iron with its tip removed) aluminium posts on the bottom of the extruder.

With the extruder made, we hooked everything up.  We loosened the set screws on the solder tips of four irons and fastened them to the posts.  We connected an aquarium pump to the side barbed fitting.  We used some wood to shim the irons such that the extruder was level.

With it assembled, we fed solder into the top of the extruder:

It turned out well–better than version 2, but not yet good enough.

For the next version, we have a couple of improvements:

  • Using a band saw, we shortened the soldering iron mounting posts to promote faster heating.
  • We flipped the extruder upside down so the posts wouldn’t get in the way.  To do this, we had to switch the fitting on the bottom and the fitting on the top.  This was easier said than done.

We had to use a torch to get the fittings loose, because they had been left on while welding the posts.

They were done for, so we picked up another pair:

We had to re-tap the tee given the mess we made of the threads.

  • To find the right pressure and pulse frequency for the air, we wanted to find a device that allowed us to vary both.  We will try connecting the tube to a speaker.  Using software, we can vary the sound frequency and thus the pulse frequency.  By increasing the volume, we increase the pressure.  Through experimentation, we hope to discover a pulse frequency and pressure that yields the most continuous solder flow.

3D printing electronics: solder extruder versions 1 & 2

One of the projects that came out of our 3D printer project brainstorming event was the idea of making a machine that can 3D print electronics.  Impatient Inventor and I have decided to pursue it.  If we are successful, then new possibilities are opened up in design and manufacturing.   Instead of figuring out how to the electronics and mechanical parts separately and assembling them, you can make a single object that integrates both.  Imagine being able to make anything from multilayer circuit boards to car doors with integrated wiring on a single machine.

Ideally, the 3D printer would be able to put down plastic, wiring, and components.  To start out, though, we will make a machine that just puts down wiring and plastic.  CNC gantries and plastic extrusion are solved problems, so we began by prototyping an electrical wire extruder.  We picked up a four axis router for the gantry and are still looking around for a good plastic extruder.

For our first wiring compound, we chose solder due to its low melting point and strong conductive properties.  To ensure that ABS and solder played well together, we jammed solder into a soldering iron and letting it drip on a raft left over from a 3D print job.

It turned out all right, so our idea of using regular solder for the leads seems viable.

For our next version, we wanted to see if we could do it in a more controlled way.  Pulling this off took a couple of advancements:

  • To see if we could precisely control the solder application, we 3D printed a block with a channel instead of using the left over raft.
  • To prevent the solder from flowing everywhere, we opted for flux-free solder instead of solder with flux.
  • To better control the flow rate, we put the solder in a nozzle that we hose-clamped to the soldering iron tip, instead of putting solder directly on the iron.

3D printing a block with a channel
We wanted to see if we could get the solder to fill out a 3D printed channel.  So, we started out by drawing up a 3D model in CAD software (SolidWorks).

We then used 3D printer CAM software (ReplicatorG), to convert the the 3D model into instructions for a 3D printer (gcode).

Finally, we sent the instructions to the 3D printer (a MakerBot Replicator).

The result was our 3D printed block with a channel in it.

Turning the solder nozzle
We then made the nozzle that would put down the solder.  We started with a small aluminium rod that Steve generously turned down for us on a lathe.

Our solder nozzle:

Using a hose clamp, we fastened our nozzle to the end of a soldering iron and the next version of our extruder was born.

Testing it out
We pushed the flux-free solder into the nozzle and to test our newly minted extruder on our 3D printed block.

This is what we got.  Not the most handsome wire, but…

It is functional!

In the next version, to further improve the flow control, we will use a pump to sputter the solder out at a defined frequency.

The State of Industrial 3D Printing

This is my impression of the state of 3D printing technology based on my visit to RAPID 2012 (http://rapid.sme.org/2012/public/enter.aspx) and reading some articles on the industry.  The post includes the applications of additive manufacturing, the additive manufacturing process, a comparison against subtractive manufacturing technologies, and a comparison between the different additive manufacturing technologies.

It is the basis for a talk I will give to set the context for a 3D printer design project at a hackerspace, Pumping Station:  One.  At the event, Jeremy, from TinyWorkshop, will discuss what’s going on in consumer 3D printing.

The promise of 3D printing is its ability to support general purpose making.  There is already a huge range of applications for which 3D printing has been used, piloted, or envisioned: Continue reading

World Makerfaire 2012 – New York

I attended World Maker Faire 2012 (http://makerfaire.com/newyork/2012/index.html) in New York.  The most impressive technological advances were in 3D printers and embedded systems.

3D printers
At the 2011 Makerfaire, it seemed that everyone was showing off their Makerbot Thing-O-Matic 3D printers (http://wiki.makerbot.com/thingomatic).  These allow folks to go from a 3D model to a physical thing by using software to slice the model into layers and then printing each layer in melted plastic.  This process is called FDM, or Fused Deposition Modeling.

At the 2012 Makerfaire, the main advances beyond the Thing-O-Matic were: Continue reading

Autonomous car – obstacle avoidance with embedded vision

I recently gave a talk at embedded systems night at a Chicago hackerspace called Pumping Station One (www.pumpingstationone.org).  In it, I explained why I chose to go with embedded vision for obstacle avoidance in my autonomous car and discussed options for overcoming the performance issues with embedded systems. Continue reading

Project ideas from IMTS 2012

I recently gave a talk at automated manufacturing night at a Chicago hackerspace called Pumping Station One (www.pumpingstationone.org).  In it, I discussed project ideas that came from IMTS (a 100,000 attendee manufacturing show, http://www.imts.com/) as well as technical highlights (http://jeffsinventions.com/?p=1177).

Some of these projects would result in original products.  While others may have been done before, they would improve on my current rapid prototyping resources.

  • Combine a 3D printer with the milling machine quick change tooling idea.  Instead of switching out cutting tools, the 3D printer could print a variety of materials.  Consider, for example, making a tire by alternating between printing in rubber and aluminum.
  • Use a pair of robots to perform a variety of manufacturing activities.  One robot might fixture the part and the other would perform a programmed series of manufacturing tasks on the part (machining, welding, finishing, inspecting).
  • Having spoken with a number of machinists at the show, one of the most challenging aspects of the profession is coming up with ways to support the part while it is being machined (fixturing).  Currently, folks use a wide variety of approaches for doing this—anything from vises, clamps, t-nuts, to vacuum tables.  It would be valuable to come up with a general-purpose fixturing device.
  • There were a number of demonstrations where a robot would pick up one of many arbitrarily arranged, uniform parts.  In discussing the technology with vendors, it sounded like a hassle to install the capacity for picking up a new type of part.  If this process were simplified, then these robots could be used for picking by distributors that carry a wide array of parts.
  • Currently, software exists, called CAM, which translates a 3D model of a part (a CAD drawing) into instructions (gcode) for a cutting tool to machine the part.  An ambitious project would be to generalize this idea:  to write software that would figure out how to coordinate all of the manufacturing operations that go into making the part—anything from machining, welding, finishing to inspection.  Imagine being able to draw a part, load a blank, press a button, and the part is gets made.
  • Add a fourth and fifth axis to a three axis, vertical mill using rotary tables.
  • Add quick-change tooling to a three axis, vertical mill.
  • Come up with a way to quickly load and unload parts on a three axis, vertical milling machine.
  • Make a pick and place machine to automatically load electronic parts into printed circuit boards.

Get in touch (jeff@jeffsinventions.com) if you are interested in collaborating on any of these.

The Vision Show – Boston 2012

This is my impression of the state of machine vision technology based on attending The Vision Show (http://www.visiononline.org/events/event.cfm?id=76).

There are typically two problems machine vision is involved in solving.  The primary one is quality assurance.  For example, a camera might evaluate whether components were soldered onto a printed circuit board properly.  The secondary one is guidance.  For example, an automated guided vehicle might be tasked with transporting material from one location to another.  Along the way, it must be able to identify and navigate around obstacles.

The end point of most machine vision systems is mechanical motion.  For example, a defective product might be placed in a defective parts bin.  Or, an AGV might brake when a person walks in front of it.

To get there, the basic steps tend to be:

  • A scene is illuminated with a light
  • An object moves past
  • A camera outfitted with a lens and sometimes a filter takes a picture
  • The camera transmits the picture over an interface
  • Analytics are performed on the picture
  • A decision is transmitted to a PLC
  • The PLC sends a signal to a machine
  • The machine moves

Based on the context, there are a variety of approaches available for accomplishing each step.  For a run-down of these approaches: Continue reading

Bob Dorf – How to build a great company

One long term possibility I am considering is starting a company.  When I was in Boston for The Vision Show, I attended a lecture at MIT by Bob Dorf, a man who has started a number of companies (http://legatum.mit.edu/content/1189).  He was promoting his book “The Startup Owner’s Manual:  The Step-by-Step Guide for Building a Great Company.”

I found three of his points particularly interesting: Continue reading