Nov 18, 2016
One of the biggest challenges when I first started building book scanners was getting the glass panes for the platen. To minimize glare, the design called for a 100 degree angle between the two panes of the platen. And to ensure a tight mesh between them, I needed each pane to be mitred to a 50 degree angle.
But there are no glass shops who offer this kid of mitre. Most commercial machines that mitre glass are only adjustable up to 45 degrees. So in order to get the panes I needed, I built a jig to let me grind them to that angle myself.
Glass Grinder Jig
The core of the jig was the hobby glass grinder. This kind of grinder is typically used to smooth and shape the edges of stained glass tiles when making windows.
I cut and glued together an acrylic stand to support the glass at a 50 degree angle. The glass was held just above the grinder table and pressed against the grinding cylinder to make a mitred edge. The stand was on a rail so it could only move left and right. So even though the grinder head is a small cylinder, I could still grind a straight miter across the edge. The rail was mounted on a vise jaw which let me move the glass slightly closer to the grinder with each pass.
In practice, this jig had a lot of room for improvement. There were too many degrees of freedom and the pane wasn't held rigidly enough. I had to keep the whole thing stabilized by hand as I moved it back and forth. While my wife and I were able to make a couple of batches with this jig, we had to check each piece against a guide afterwards to make sure it turned out.
CNC Glass Grinder
I improved this setup over the course of the next year with new iterations of the jig. The latest iteration is now a motorized CNC device which fixes all of the flaws of the original.
The glass support is much thicker acrylic redesigned to avoid useless projecting faces and sharp corners. The rail is now double gantries running on solid C beams that are mounted to the table. There are rails on both sides of the glass grinder so that the weight of the glass and acrylic holder is fully supported.
Because it is computer controlled, it can move very slowly and consistently. This eliminates the need for moving the glass closer and further from the grinder. The quality is better and more consistent. And it requires a lot less work.
The grinder is a true 1-axis CNC device. I run a simple gcode program to do the grinding. It is probably overkill for such a simple task, but it allowed me to fine tune the speed easily by just tweaking code.
More photos and a video of the device working can be seen here.
I still have a few tweaks that I want to make to the grinder. I will be adding a touch interface at some point and using a relay to control the hobby grinder. But overall, it is feeling like a project that is close to complete. Now I just need to keep grinding glass.
Oct 31, 2016
On the left is a successful end cap for T-slot aluminum extrusion. On the right is a bowl full of failed trials. Each of the failures didn't fit correctly or couldn't handle repeated use.
Each end cap has to fit snugly. You don't want it to slip out if gravity or some other small force pulls on it. But it must also be removable and reusable. So you need a very slight snap fit. But this puts stress on the plastic. And the more you stress the plastic by reusing the cap, the weaker the snap fit is.
With my first designs, only a couple of uses would render the cap useless because the plastic would permanently deform. By tweaking the design, I was eventually able to get to a couple of dozen uses when 3D printing in ABS.
After reaching the limits of ABS, I switched to a plastic called Alloy 910. It took some time to figure out the best way to print with the new material and finalize the design that worked best with it. But Alloy 910 is less brittle than ABS which means that it can be used many more times without permanently deforming and losing the springiness which is necessary for a snug fit. The success on the left can be used more than fifty times and still work well.
I will now be able to make these end caps myself instead of ordering them. In addition to being cheaper, they are much higher quality. The injection molded end caps I was using before began to deform and become less useful after only a handful of uses.
This makes the many dozens of attempts and hours of tweaks and trials worthwhile. These end caps are a very small thing. But I now believe that I might be making the best end caps on the planet.
May 10, 2015
It has been a while since I have posted to the blog. If you follow along with this guide, you will see why.
Step 1: Select a Scanner Design
Step 2: Build All The Things
Step 3: Ship it
Hundreds of pounds of plywood, two car loads of bubble wrap, dozens of boxes...
Now I need a nap.
Jan 31, 2015
With the CNC Router fully operational, I turned my attention to stampmaking. The hope is that I will be able to draw custom stamps for my wife or others to help with various inkcraft projects.
After flailing around a bit with trying to use standard cam programs to make my stamp, I found F-engrave which can do magic. I have v-shaped engraving bits which can cut different sizes based on how deep the bit digs into the wood. With just the sharp point, you can score a thin line. With the whole bit in the wood, you can dig out a large channel. F-engrave understands this property and so can make a tool path to precisely cut out fine patterns.
There are four different cuts which need to be done. First a rough pass with an 3mm bit to clear away all the material not near a boundary. Second, a v-bit tracing of the outline which moves up and down to account for how thin the gaps must be. Third, a v-bit cleanup is done. Areas outside of the outline but too small for the roughing are cleaned away. And finally, the box around the design has to be cut out of the board. F-engrave can generate the first three tool paths and save them to different files. You need to use a standard CAM program to do the fourth one.
There are two gotchas here to be wary of. You want to scale the design before loading it into F-engrave. That will make it easier to ensure that the cut out tool path is correctly aligned to your engraved area. I screwed this up and ran a bit through my beautiful engraving. Also, make sure your art is already at the origin before importing it. If it is offset, the tool path F-engrave generates will be offset as well.
You can see my stamp attempt turned out pretty well before I accidentally carved across it later. The material is linoleum. I have read a few places that linoleum can work as a stampmaking material for CNC and you can find a few places that sell linoleum mounted to wood.
I decided to try a cheaper solution. I bought a linoleum tile from the local hardware store that had glue on the back. I just attached it to a small piece of wood and thought I was good to go. The real problem came when I tried to cut it out. The glue was not strong enough to hold a small section of linoleum on so when my cutout chipped off the linoleum around a stamp, the insides started shifting with the stress of cutting and I had to hit the emergency stop.
In order to make more progress, I will have to find a stronger adhesive to attach the tile to the wood. Or I will have to break down and buy pre-mounted stock online. So far I have learned a lot but have no stamps yet. Hopefully I will succeed soon.
Jan 27, 2015
After a month of tinkering, upgrading, and building my router is truly useful. In this picture, it has just finished carving a pattern for a stamp from linoleum. It will be interesting to see what distributed fabrication will do to the economy over the coming years.
Among the essential upgrades are homing switches that let me resume a job even after an emergency stop, a dust collection system with a dust shoe to save my lungs and sanity, and an upgraded spindle because the generic one that came with it failed.
Dec 20, 2014
The box contained a Shapeoko 2 which is a CNC router. This will let me cut out many kinds of wooden or even plastic parts and designs with a computer. I am not that great at classic woodworking skills like drilling straight holes and making sure things are square. This machine will let me do all of that by simply drawing them out at my computer which is considerably easier. The default kit will have a work area of 12"x12" or so. This is too small to cut furniture or other large projects, but there are still a lot of things I can make. Later on, I can upgrade it to do bigger pieces.
The box came absolutely stuffed with screws and plates and pieces:
Here are some of the pieces laid out on the new workbench:
The first big piece to be assembled was the z-axis. Then I rolled it onto the x-axis rails. The tower in the middle is built around a threaded rod which rotates to move it up and down. The tool itself is attached to it.
Here is a closer view of the main gantry. The x-axis is belt-driven and will move the tower left and right along the rails.
And here is the frame completely assembled. The x-axis assembly is mounted on two rails which form the y-axis. These are also belt driven and move the whole x-axis back and forth.
All of this is electronically controlled by an Arduino board wired to stepper motors via a shield. My current status is that everything is wired up, belts are installed, and I can use my laptop to control it moving it slowly around. But I am running into issues where the z-axis is binding up and not turning with the stepper motor sometimes.
The working machine is so close, but so far away...
Dec 06, 2014
The workbench is complete. Buying a separate bit that only did the countersink part was a very good decision. The hardboard top looks very good and promises splinter-free usefulness. I only screwed up one drill hole.
But, I finished the workbench just in time. For, a new and very heavy parcel has arrived. And I've got something new to put together using the new workbench:
Dec 04, 2014
At last, I am making good progress on the workbench. With luck I might even finish it tomorrow. After finally getting the legs cut to the right size, I made the two side frames:
And then I connected the two sides into a complete frame. There was only one major mistake on this. I had to unscrew and redrill the holes for one of the cross members because it was on crooked. Now that the basic structure is complete, I have only to secure the tabletop and shelf top boards.
After some experimentation, I have come to the conclusion that the cheap countersink bits I bought are completely worthless. So I will have to go buy a more expensive countersink in order to make the screws flush with the top when I screw it in.
And I can celebrate a sort-of milestone now. For the first time since I bought my drill earlier this year, I managed to run it out of juice in one sitting drilling pilot holes and driving screws.
Dec 02, 2014
This workbench business is harder than it looks. So far, I've managed to make almost no progress. It is especially hard to create a workbench without a workbench to start with because it is very difficult to clamp things properly for cutting. In order to try to escape from this catch-22, I've bought a mini-workbench to help me create my full-sized workbench.
Just a bit more work to do, then hopefully things will progress much faster.
I've also been trying to create a jig which will help me cut straight with a circular saw. In theory, I can clamp it across a board and then just saw along the line of the board to make a straight line. You can see the latest iteration here:
But I have now failed to make the jig twice. The first time, I made it too narrow so that when it was clamped, the clamp interfered with the saw blade. The second time I didn't cut the jig itself straight enough and so using it made all subsequent cuts crooked. So I will have to try again. Third time's a charm, right?
I managed to get four straight legs of the same size, but, alas, when I measured them, they had become too short. This was partly bad planning and partly mutliple miscuts. So I had to buy two more beams from the hardware store and try again. Perhaps the rather short legs will become something else in a future project.
Hopefully some real progress soon...
Nov 29, 2014
Earlier today, I bought a bunch of lumber and a pound of screws.
In theory, this is enough supplies to make a workbench from scratch. The 2x4s will provide the legs and cross supports. The OSB (a kind of plywood I just learned about) will form the tabletop and shelf underneath. Hardboard, a dense fiber composite, will be the top layer which will provide a durable and replacable surface that should be completely splinter free.
So far, I have managed to create a jig that should allow me to cut a straight edge with my saw very easily and do a lot of cleaning to make a space for the workbench. Tomorrow, I will cut everything to size and assemble it.
If everything works out, I will have a new table for a fraction of the price of purchasing it from a furniture store. It will be a lot more solid than my current desk and dining room table for constructing and clamping things together. But, above all, it will be replaceable. So I won't have to worry too much if I drip glue or scratch it damage it. In the worst case scenario, I will be able to spend ten bucks and get a new surface board.
My basic plan comes from here.