Monday, January 30, 2012

CNC Router

Another project I'm thinking about is building a CNC router.  There are several designs out there already, but they seem to either fall into the "you don't have any tools or skills" or the "you have a lot of money to spend" categories.  Since I'm in neither, I'm looking around at some of the more popular designs and am going to design something that fits my budget and skills.  I'm going to be using this blog to track some of my build notes while I'm sketching things out.

One thing that gives me a big edge is that I already have a fairly well equipped wood and metalworking shop.  This gives me the ability to fabricate with metal, and most designs seem to shy away from that.  Many of the low-end designs use a lot of MDF as well.  I'm not against MDF, but the idea of tapping threads in it and using it without significant structural support bothers me.

So where am I?  Eventually, I think I'd like to be able to build a system that will work for plastic extrusion like RepRap, but for now I'm looking wood routing systems.  I have the capability to mill metal right now (manually), and so I wanted to focus on something that would actually expand my shop capabilities.  I don't need a huge bed, but something around 2' x 4' would be helpful.  Ideally, I'd like to be able to use the finished product to help me build musical instruments, including things like banjo pots, necks, etc.

I may end up discarding this idea, but at the moment I want this system to be able to support my full size Porter Cable router.  It's not huge, but it's fairly heavy, and so the gantry and related stuff will need to be beefy.  This will likely increase the motor torque requirements as well.  If this fails, I'll look at smaller routers, like this Bosch.

At the moment, I following in many footsteps and using the "JGro" design as a reference for the key components of a CNC design and looking at what components I can improve piecemeal.  The first major change is that I'm building out of primarily birch plywood and not MDF, and using screws and glue to do most of the joins.  The other change is that I'm looking to make the alignment blocks for the pipes made out of aluminum.

I'm just now starting to get the initial components purchased, cut, and assembled, and I expect I'll have some progress to report in the next couple of weeks.  I plan to get all of the mechanical stuff done first, before I look at electronics.  Particularly with a full size router, I need to make sure I understand the loads involved before I size the motors, motor drivers, etc.

I'll try to get some pictures up here as I make some progress, because what's a blog without pictures?

Tuesday, January 24, 2012

Hall Effect Pickups

As part of this project, I need to find a way to capture the individual tones from each string, and do it in a way that doesn't interfere with the "natural" way the banjo is played.  As I mentioned in my last post, this is done by a pickup, but I need one for each individual string.  Normally these seem to be coils, but I'm going to try to use a device called a hall effect sensor that, with the help of a strong magnet underneath, should detect the resonating string.  Many hall effect sensors only tell you "on" and "off," but I'm looking for a radiometric hall effect sensor that will give me an analog signal that will represent the string movement.  The ones I'm looking at right now are the Allegro A132X series sensors, which I believe will do the job.  Some experimentation is needed however, since I don't know how strong of a magnet will be needed or what the exact placement will need to be.

My next step is to mock up a couple of strings, magnets, and sensors to see how it behaves.  Why two?  Because one of the main goals of having a single string pickup is isolation, and I need to know how much noise bleeds into adjacent sensors.

Sunday, January 8, 2012

Solid Body MIDI Banjo Project

I'm starting a new series of projects relating to music, electronics, art, and other random hobby things.  I'm going to start with a project I've started recently, with a goal of designing and building a solid body banjo with the capability to output both a normal analog output as well as polyphonic MIDI.

Since you're looking at this, you probably have a sense of what I'm looking to build already, but just in case it's still fuzzy, I need to explain what's normal and commonly available before I can explain my particular version of fun.  Most electric instruments, and in particular a solid body guitar, use inductive pickups.  The signal is them amplified and then driven out of a speaker or recorded.  Since the coil used for the pickup spans all of the strings, this coil picks up a blend of all of the strings that are being played.

Now MIDI is a completely different beast.  That is a digital protocol that's used to communicate between computers and various digital music devices.  It can signal things like when notes start and end, how loud they are, and so on.  There's no audio in MIDI - it's like a computer program that needs to be interpreted by a synth of something similar to turn it into something we can hear and enjoy.

Where things get interesting is when you want to take a live non-digital instrument and generate the equivalent MIDI instructions.  In some cases, like a piano this is fairly straightforward - you track what keys are pressed, when and how hard, and you can get a pretty good approximation of the original performance.  With stringed instruments, it's generally been more difficult, since these don't have keys.

One solution is to replace the strings with the equivalent of keys, which is what some designers have done, such as the SynthAxe and the more recent You Rock Guitar.  In this case, you're using the guitar (or whatever) as a digital input, but you potentially lose a lot of the look and feel of the original instrument.

The other route is to take the audio signals that the instrument normally produces from a mic or a pickup, and convert that into the note information.  This simplifies the use of the instrument, but you lose some of the expression information.  It can also be fairly complex to extract discrete pitch information for each string what it's all blended together.

Next time I'll talk about some of the specifics that I'm working on, including a version of a hexphonic pickup for a banjo (which makes it pentaphonic!), and some planned experiments with hall effect sensors instead of the traditional coil pickup.