The ShopBot CNC router hasn’t always been an object of desire for me- It’s easy to judge it as inferior because it doesn’t use linear rail bearings, ballscrews or servo motors in favor of ball bearings on rails, rack and pinion gears and gearhead steppers. I have learned more of the capabilities and strengths of the ShopBot by seeing the work BecauseWeCan do with theirs, as well as hearing Jeffrey talk about it at EatFoodTalkShop. Last thursday he described the some features of the machine in a way that really piqued my interest- mainly concerning the “Job Code” that is used by the ShopBot’s control software. As well as mentioning that they were planning to open and document the language they created and use (as opposed to Gcode)- which it appears they have now done, he said it is a much more comprehensive script-capable language than Gcode, with lots of built-in functionality- including built-in randomization.

That in particular really got me thinking- After quickly reviewing the documentation, it appears that the ShopBot code- “OpenSBP” supports user vairables, some math (although it is not described in the manual), basic flow control, BASIC style subroutines, windows message box triggering, file I/O. It’s a pretty good start- I can imagine some reasonably interesting possibilities of dynamic part definitions from within the SBP files. I am really attracted to the idea of being able to define parts based on random variables, such that every part will be “designed” on the spot, randomly. It is of course possible to write a seperate script to generate some Gcode or “dumb” SBP code, but the idea of having the randomness executed at machining time is sort of the ultimate in instilling organic ephemerality to the machining process.

It’s clear that the SBP standard is still under construction, as is its documentation. I believe that it looks better than Gcode, and has a lot of potential in future revisions. That said, the documentation is a bit vague on key issues at the moment, and there are still some complex processes that aren’t going to be efficiently implemented in SBP. You can’t define even simple data structures in SBP and I personally am much more familiar and happy with with C-style programming over BASIC style programming. Although it is not described in the current programming manual, I presume there is a random number function- this could easily be used to vary the parameters of some texture defined in SBP, like by varying the Z axis plunge depth at each point in the texture. I have thought of an application where I would like a randomly generated pattern of overlapping concentric circles- for a more logically complex operation like that I would want to work with an environment that would let me have data structures- arrays at least, and very well implemented subroutines, looping and flow control. A script to generate complex random toolpaths could be written in C, or Processing, or PHP, or whatever- but that abstracts it from the actual execution of the job, which as I mentioned earlier, is one of the coolest parts about scriptable job code. At least it is clear that the ShopBot control software is command-line accessable- that helps. Maybe there could be some way to include hooks to other scripting software from within a SBP file…

Anyways, I like the direction the ShopBot folks are headed with SBP. It’s already better than Gcode, and is only going to get better by documentation and standardization. I can imagine that I would enjoy very much working with a ShopBot machine and using the SBP- it looks like lots of fun. I was thinking before about wanting to build a CNC router- but now I find myself qute attracted to the ShopBot package- I don’t doubt the machine’s quality or capability to do what it’s designed to do, and their software/hardware package integration and support is very attractive. I feel like I could actually get somewhere right out of the box with the ShopBot system.

So why exactly do Iwanna one now? Because my mind has, as is inevetable, begun to imagine what I could do with it. Just like Jeffrey and Jillian at BecauseWeCan did, I am definitely dreaming about what my ShopBot business plan would be. I can think of about a half a dozen different relatively simple things to start with that I could design and be able to manufacture with something like a ShopBot. I am excited about that quickness from design to prototype through to manufacture- just like it has for BecauseWeCan, it’s a straight-fire way for someone with a interest in visual or functional design to turn cheap materials into complex prototypes that would be very nearly impossible to make any other way- and then redesign or manufacture those designs with just as little trouble. Seriously it is amazing the variety of unique furniture and interior pieces BecauseWeCan make with the CNC, good bits, good wood suppliers, sanders and a good finish spray setup. The emphasis is able to be focussed much more on design rather than the mechanics of manufacture.

So, maybe within an few years I’ll borrow some money, crank out some designs and nab a ShopBot- I think it may be the fastest route to some of my biggest goals.

I’ve been thinking more about how I would I’m going to build a CNC mill. My earlier ideas have changed a bit, some a lot.

I’m not going to try to use servos, it is worth more to me to get the machine usable quickly, and the HobbyCNC stepper boards are pretty nice. It would also be an easy upgrade to hook it up to the step/direction signals from the KCam MaxStepper or DeskCNC board. These seem all like nicely developed, reasonably priced and reasonably effective solutions. I plan to use a big Chinese switchmode powersupply to power the HobbyCNC driver board, rather than the transformer/bridge setup they describe in the kits. You can pick them up for dirt cheap, and they are very easy to use.

I would still like to use a “CNC on a Budget” Z axis, I’m going to inquire about getting one of those, hopefully soon so I can use its design as inspiration for the other axes. I am pretty sure I want a gantry Y axis setup, but further than that, I’m really not so sure. I suppose I should probably start I just spent some time looking around at other’s machines on Google and CNCZone- it looks like bunches of people do the gantry arrangement with dual leadscrews at either side of the X axis (that moves the gantry up and down the table) I suppose that wouldn’t be too bad, just have to get a hold of some timing belt and pulleys. Then you really do have to have good thrust support bearings on the leadscrews also, since there won’t be any other thrust support from a directly connected motor- though I would avoid relying on that anyhow. I guess then I should continue shopping the eBay- see if any good deals come up on rails and trucks or rods and bearings. I think either would work fine. I like the idea of not having any important parts under the cutting table- I think it could increase the versitility of the mill greatly in some situations to be able to set it up on any sort of table material.

I found this site today and read through a bunch of their (excellent) content. Makes me think that using a single speed Porter Cable router, a variac, and their fancy chucks would make a pretty good spindle. There’s also lots of great info on there about bits in general and how to best use them.

I should probably hop on Sketchup again… see if I like it any better than I did last time I tried it. Or maybe I’ll test out SolidWorks. I’m in a class about AutoCAD right now, but I rather detest it- maybe I’ll draw this up in one of the aformentioned programs and then draw it in AutoCAD for my class project.

Anything’s better than just napkin sketches.

Drill Press, Laney College Machine Shop

Building a small CNC mill has been one of those projects I have thought about dozens of times over the years but have never initiated. It’s come up again now, and maybe this is the right time.

I would really like to be able to cut parts out of acrylic and aluminum sheet, as well as engrave text and mill out shapes. I can imagine many applications for milled acrylic as an enclosure for PCB based projects- potentially an asthetically pleasing, compact, cheap way of packaging a device.

So I’ve started shopping the eBay for linear motion parts, I would like to build the machine out of aluminum and use recirculating ball linear bearings and hardened rods, with Acme leadscrews. Great deals on new old stock parts can be found on eBay, but NOS bearings and rods are still pretty expensive, and are generally kind of a compromise between what you were actually looking for and what happens to be for sale. There are some parts on eBay now that I think would make a very nice set of axes, I have to decide whether they’re worth the (bit more than I’d like) price.

I am envisioning a gantry-type mill, using 24″ long rods, for a travel length of about 18″. An “Instant Z Axis” from CNC on a Budget is pretty tempting, as is their 3 speed spindle. I think his prices are very fair for the amount of material and work that goes into his products. So, a couple hundred dollars here, a couple hundred dollars on aluminum plate and linear bearings and you still haven’t got a CNC machine!

The most practical solution to simply getting a machine up and running would be to purchase a proven stepper motor based motion kit, like those from Hobby CNC. I would really like to go a different route though, I think stepper motors are a bit less than optimal- slow, noisy, generally not that elegant. Optimally, I would use any old brushed PMDC motors to drive the leadscrews, then attach a linear scale to each axis and run it closed loop servo-style. Capacative linear encoders are cheap and plentiful on eBay- search for “digital quill”. The data output on these is pretty well documented here and here, that’s not to say it wouldn’t be challenging to implement a system to read them and use them in a closed loop system with the motors. I bet it would work, though, possibly work well. I think it would be fun to build a custom control board with motor drivers, inputs for those scales and an onboard G-Code interpreter, then make a complementary app for the host computer that would work as a DRO and to feed the controller lines of G-Code.

Ok, not a small project. I should wait until I have a bit more machine shop experience then concentrate on building the machine and then pick up some steppers to get it running- the custom servo setup would be fun, though.

And, of course, I’ve got to finish the battery regulator system and the electric car first. But I’ve got a little cash burning a hole in my PayPal account, maybe it’s time to pick up some linear motion parts. Or maybe, as with so many ideas, it would be wiser to resist the impulse purchase and not regret it later- just like there are deals on eBay now, there will be deals later. Not an unusual quandary for me.