Sunday, April 20, 2008

Last weeks stretch

If anyone has ever worked on a project where you design as you go you will certainly understand what it is like to be on the last week before a project is due. All he things that did not get enough attention during design are now screaming back at you "see I told you , but you never listen...."

Actually it is not that bad but we can feel the pressure. Last week we accomplish some big mile stones. The electronics where completed and the mayor machining of the casting was delivered. The one thing that can be a problem is a single draw bar (it holds the tool on the spindle) which is being shipped overnight on Monday. I don't know what overnight to Colombia form the US is like, so I am crossing my fingers.

Now some cool pictures

Here you can see that we have all three axis on them machine....
this is me smiling after finally assembling the Z axis with the spindle and drive pulley.

Interior of the machine being fitted.. We will later paint this with urethane paint (car paint). I actually got the MFG green and blue colors ready mixed so we have only to decide which parts we paint blue and which parts green.

Thursday, April 3, 2008


I’ve been kind of silent lately….and a few mayor things happen since then. First. I failed to deliver the project on time. There where many reasons for that, they run the gamit, designs that didn’t work out, mistakes, supplier delays, even local holidays. I am not used to failing in this way so it was hard to write or even focus on what next. But after some consideration I feel it’s time to focus on what we have learned and what is coming up.

So I will focus the discussion on what we have learned: First lets talk about castings…they are key to a CNC machine, at least in small quantities. Not only do you have to get the molds right you have to take into consideration the time it takes to get every step done and what happens when there are complications. Here are some tips, get your carpentry gear out , sharpen your scraper, measure twice, cut once and start gluing MDF. Mind this is the cheapest way to get it done. We used spray and paint on wood sealer on the MDF, otherwise it will be hard to paint since it will absorbed to much paint and distort itself. The sandable kind was nice, its finish after a light hand of sanding can be almost ideal. We finish off our molds with enable paint. This helps with the de-molding. I also advice reading a little theory about how to designing cast parts –constant wall thickness, draft, material flow and part cooling. There is a good description of this in Slocum’s book “Precion Machine Design” and on his course posting.

Keep in mind that when creating tall enclosures you may have porosity problems, the material creates gases that can be trapped if the casting acts like a bell. So make you you leave openings for the piece to breath. Also put some thought into how you want the material to flow. The person doing the casting is probably better at this so make sure you have this discussion if you are making a complicated piece. I will post all the molding instructions for our part, which took two modification to get it right. Also make sure you talk to your caster about none compliant parts, make sure you have a clear understanding of how he will redo pieces that did not come out right the first time.
Motor mountings: Painted MDF mold after releasing - Cast motor monut top view - motor mount next to Y axis servo motor

The other part that is important to note is that this pays off if it’s a local effort. Many time large foundries won’t take your work and overall castings are heavy to shipping so cost can be high. In Colombia we are where being quoted between 4500 to 9000 peso a kilo which is the equivalent of $2.43-5/kg or $1.10-2.20/lb. this may sounds cheap but remember that there are post machining operations. For small parts we paid about $100 for all three motor mounts, for larger pieces it wasn’t so simple, we have gotten quotes for $3500 for squaring and machining the Z axis and it X axis interface. Clearly that is way too much so we are finding ways around this. I will complement this post once I’ve perfected a method for replicating flat/parallel surfaces with $50 of metallic resin.

Making the Zcasting mold. MDF mold views

Modeling the Zcasting in SolidWorks

Lina and Camilo working on the Zcasting Mold

The Making of the Zcasting: Mold released -Mold sealed - Two views of the casting once out and cleaned.

Wednesday, April 2, 2008

What is Open Source Hardware?
It’s an initiative to lower the barrier to access of production machinery, by reducing cost and enhancing availability. In its own way it similar to Linux, it doesn’t cost much and you can get it pretty much anywhere. With hardware the hope is that with the proper open design users can source components locally, make their own and assemble production machinery them selves.

In other words we are trying to design a machine equivalent to the entree level VMC (vertical machining center) in the states that runs you about 50k, and we are shooting for BOM cost of 20k or less. The machine will also be design so that you can build it your self, in your shop/garage. This can have an impact not only for American machine shops and hobbiest but also in parts of the world where the cost of acquisition of such machine is actually much higher.

Why is MFG doing this?
Various trains of thought have molded the idea; initially Mitch envisioned the possibility for MFG to be the catalysis that enables open source hardware to happen. Like any initiative of this kind it has to have an organizing body that sets the direction. We are hopping to leverage the rage of suppliers in our network to have access to components and sourcing information world wide

Mitch also envisioned a day when trade school students, instead of making practice parts and discarding them would instead practice making parts that could be assemble into a working machine. There are many ways in which the information embedded in the design package can be used but most importantly we want our user community to decide how they use this information. The is a radical departure with the way things have been traditionally. Manufactures of equipment tell you exactly how a piece of equipment should be used, and we are giving user ultimate modification power over the design so they can do what ever they dream up.

We have also considered that such a design can be a gift to our customers and that it would also broaden manufacturing capabilities around the world creating a larger market for MFG to act as a broker.

Why is this important?
Personally, I share a vision where personal fabrication and mass customization could become a reality, and I see this as the first step. Think computer main frames and printers, initially large expensive complicated machines that later became house hold items. Well I envision a day where you buy design or instruction sets for a personal or local machine to make something “custom” for you. In a way MFG is already doing that but we are not sharing designs. To go full circle we need not a one to many but a many to many relationship. Where a design embodied as an RFQ can be quoted by many suppliers and then purchased by many users. Today it’s one to many because of confidentiality agreements but under a creative common license we could do away with that requirement. At the very begging we have to prove that there is compelling opens source content to merit a business case and the infrastructure to deliver the information.

What are the challenges faced?
First we are trying to make this very low cost, initially 30-50% of retail value. And we want this to be something that a small, basically equipped shop can make. So we don’t have any economies of scale nor we use more complicated machines to make precision parts. In fact that challenge is two fold, it’s like buying a car by going to AutoZone, piece by piece it’s a lot more expensive, so we have to choose the components wisely. This has been hard especially on the controls areas, GE and other Taiwanese and German companies control this market and to buy a controller alone would blow our budget. So we are using Linux a PC and piecing together our electronics. And similar to a car that is designed to be put together in an assembly line, with special welding and assembly tools, most machines are made out of large castings and we have to do ways with that replacing castings with structures that 1-2 person crew can put together themselves.

We have found very innovative solutions to these challenges, many are our own inventions. We still have a few more challenges to solve but eventually sharing all this information and receiving the feedback is what I am the most worried about. Only when a critical number of users collaborate on the design, will it reach its full potential in cost, assembly and performance. Only then will open source machinery become a reality, right now we are trying to kick start it with an enticing project/proposition.

Where is the project today?
The project is currently under-way, we keep a blog and a facebook group that tracks our progress, feel free to join Open Source Machinery at:
Read my posts at:
Or review the projects wiki: