You’ll have to forgive the odd aspect ratio, but I decided that this photo was too much fun the way it is. This is the first (mostly) successful cutout from the new Shapeoko build. I burnt out a couple of stepper motors and upgraded the Z axis to a high power nema 23 motor.
This photo shows the worst of the issues. Mostly it’s just a combination of the bit walking combined with some play/backlash in the system. Educational, yes? I’m going to look for some gcode designed for tuning these issues out – and then we’ll be in business!
This week started with my trying my hand at making some items for a play my wife is working on. I got the shapeoko up and running and fed in the gcode…
After about an hour of cutting, things seemed ok, so I left it to run while I got a couple other things done. (It’s not pleasant to stand next to a robot that’s spewing saw dust!) After a while more, I found that machine behaving erratically. It turned out that I managed to fry the motor on the Z and X axis!
The Z motor actually melted my shop vac hose a bit, and the X was rather warm. I shut things down and let them cool off. The next day, I found that both motors were missing steps and not performing. So I dug up another motor and swapped out the X after some cursing and dropping lots of washers.
After that, I took a stab at the Z axis. I turned one of my couplers in my lathe to open it up for the bigger motor shaft. Then I realized that the bearing mount on the Z was failing. After some consideration, I ordered a Z axis Nema 23 kit that’s made by improbable construct. (I’m a fan of his work!) For $25 it was worth it!
While waiting for the parts, I added limit switches. I use a breakout board made by the same guy who made my stepper driver boards. It has handy inputs for limit switches. Based on that, I added switches to my machine. Each switch is wired to be normally open, and closes when they contact a limit. I fed a ground line to all of them, and the return line heads back to a terminal block. At the block, the output to the computer is held high with a pull-up resistor (the venerable 10k). This way, a short against the chassis will trigger the limit, but it won’t damage the pc input.
I haven’t run into noise on it yet, but I’ve read that a .01uf cap across the input will filter noise. (I have a few hundred of them, so I’m just waiting for it to happen!)
The server was rebuilt recently and we forgot that the post viewer here needs mod_rewrite. (doh). That explains the lack of new comments. We also added the adafruit resisty captcha.
In other news, the Shapoko build is running! It’s still getting a few tweaks (mostly CAM/CAD process changes).
Just a quick update about tuning the Shapeoko build.
The X and Y axis both use belt drive. Both are MXL Belts which have .08 inches per tooth. I used 18 tooth pulleys for each (shapeoko standard size). I’m using a .9 degree per step motor on X and a 1.8 degree per step motor on the Y axis. (Degree per step is usually labelled on the motor.)
To complicate it more, I’m using a microstepping driver. It manipulates the coils to create an additional 8 microsteps as the motor moves.
The math is simple once you have all the numbers. 1.8 degree/step equals 200 steps per revolution. Multiply that by the 8 microsteps to get 1600 steps per revolution.
The MXL belt has a pitch of .8″ and we have 18 tooth pulleys. Divide it out to get steps per inch. The .9 degree motor is double the steps, so we cheated and multiplied by two.
Finally, the Z axis. It uses a M8 threaded rod, which has 20.32 threads per inch. (TPI). We’ve got a 1.8 degree per step motor on it, so that’s 1600 steps per revolution.
We just multiply it: 1600*20.32 = 32512 steps/inch.
- X: 2500 steps/inch
- Y: 1250 steps/inch
- Z: 32512 steps/inch
Our new plates haven’t come in yet, but we did whip up some new belt mounts to match the new plates.
The belt mount was made from 1/2″ HDPE (Cutting board, actually – it’s cheap). The slots line up with the new plates (We got the dxf for the new plates from Edward and just imported the vectors for the slots.)
The plate is from the original shapeoko kit – we just added holes to match. There’s enough room in the slot to be compatible with single or double makerslide designs. Now, off to cut another on the mill.
We’ve suffered a bit of a snag on some upgraded parts for the shapeoko build, but we did order some good stuff this week. First up are some really inexpensive (not IP67 rated tho) snap action switched from Pololu.
These will serve as limit switches for the Shapeoko CNC bot. Edward included mounting holes on his latest generation of Motor plates and we’ve already run cables through our drag chain for them. Once the new plates show up, we’ll post an update with the machine’s ‘Hello World’.
Second up is this great infrared temperature detector. It has a built in laser point for clearly indicating the sample area. It’s a great deal if you have prime. (We love prime. love it.) Hit the link above and yes, we get a cut – it pays for our research so go for it. This unit seems relatively accurate and is better than burning your fingers!
We built a simple controller to drive the electric mountainboard. We dug out a Wii Nunchuck and grabbed an arduino to get things going. Power is supplied by a BEC switching converter and the speed controller is driven just like a servo. We’ll add a secondary switch input later from the brake system.
Our order of 1/4″ HDPE came in today, so we tossed some strips into the X2 mill and made the bearing mounts for the build.
A single part design was easiest, so both parts are the same Nema 23 dimensions. For the non-motor end, Iweattached it with M5-10mm sockets.
On the motor end, the bearing mount is held fast by the motor spacers themselves which are threaded for M5s. Some 30mm socket heads hold the entire assembly firm. A lovejoy coupler serves for now until we replace it to reduce backlash.
I cut out this initial bearing plate for the Shapeoko build. It was modeled in cam and then cut on the X2 Mini Mill. Not too bad. Tomorrow I should have some 1/4″ plate. This was made from some 1/2″ hdpe cutting board. The bearing was pressed in with a bearing press we like to call the Mallet of Knowledge.
You’ll have to forgive the photo quality, consider this a preview. This weekend we whipped up an initial prototype of an electric mountainboard. It’s a blast to ride, but it definitely needs some improved braking.
Board: MBS Core 95
8s 5ah Turnigy Lipo
5a 5v BEC (aka switching power supply)
Temporary Hobbyking transmitter/receiver.
6.18:1 Gear Ratio
#25 chain drive
3.2Kw Brushless Motor
Castle Creations Phoenix Ice2 HV 160 Speed Controller
Removed rear foot binding for improved safety. (Yeah, Will fell on his butt)
Hand built 6061 Aluminum Motor mount (Sawzall Special)
The biggest issue with building the board was mounting the sprocket. We pulled up the MBS brake installation manual and viewed a few tutorials on installing them. From that we hunted down the appropriate machine screws and promptly left them buried on the desk for a couple of months. Then we had to track down another sprocket to light up with the 5 screw mounting system. With a 5 spoke sprocket in hand, we created a drilling jig using half of one of the wheels, some hdpe and an aluminum dowel. The mount was finished off with some simple aluminum spacers (actually aluminum crimp ferrules made for aircraft cable).
The board is very ride-able, but needs a better braking system. RC Brakes just aren’t made to deal with human loads and feedback.