My first board has arrived!!

I received a confirmation e-mail from BatchPCB, on the 27th of September, alerting me that the board has shipped.  It then arrived today, the 30th.  It only took 19 days to receive the board!  Not bad for $25.86.  I was surprised to find that although I ordered only one board, they sent me two!!  I had actually wanted two but was broke at the time I placed the order.  An additional board would have cost $14.98 (dependent upon area of board).  Maybe I just overlooked that fact when reading through the site.... a mystery....

Board front and back

After about 15 minutes of searching around for the components I had on hand and another 15 minutes of actual soldering, I have a mostly populated board!  (It took longer to take these snazzy pictures)

Mostly Populated Board

I was lazy (mostly broke) and have not yet ordered the finishing parts; joystick and power plug from SparkFun, and the PIC microchip... probably from DigiKey.  I left the crystal out for now because it is close to the surface mount PIC and I didn't want to make soldering harder than it needs to be.  I'll try to get those ordered in the next week.  I really hope this works...

Project 3 - Detailed Design

I have spent a few more hours adding detail to the design.

Full XYZ Router Assembly

YZ Carriage

Most of the parts will probably be purchased from McMaster Carr.  Here is a list of the current parts (not the final parts list)

Bronze Flange Bearing 

6338K437 @ $2.80 each (8 needed) - 1" ID

6338K421 @ $1.01 each (4 needed) - 1/2"ID

Threaded Rod 1/2"-13

90034A033 @ $2.67 / 2' (2 needed)

90034A440 @ $1.89 / 1' (1 needed)

Weld Nut 1/2"-13

90611A500 @ $6.93 / 10 (3 needed)

Lag Screws

91478A628 @ $14.28/100 (? needed)

Precision Rod

To be determined - need 1" and 1/2"

Wood (to be purchased at local hardware store)

amount to be determined

Just with the items listed above, the cost will be about $55.00.  That is without the wood, another $20, and precision rod, about $30 for the X and Y (4 total).  I may have to go with regular bar-stock instead to reduce the cost.  I'm starting to think that 1" rod for the X and Y axes is too expensive (maybe 1/2" is better).

There are still a number of things to figure out, but I made a lot of progress today.  Another day or two like this, and I should be ready to buy the online parts (depending on my cash-flow).

Project 3 - XYZ Bench-top Router/Mill

Project 2 got me thinking... how am I going to cut the shapes?  I can design cool looking robot legs, but if I can't make them easily, why bother... enter Project 3.  I have been looking at a number of Blogs which discuss rapid prototyping machines and CNC routers and realized that the complexity of the machines and cost is relatively low.  I could probably build the XYZ stage for under $100, and buy the electronics for another $250.  With this tool, I could create any simple or complex shape I can model.

Take a look at the below sites for more building information.

Build Your CNC  a great step by step instructional site, who also sell everything you might need

OnShoulders a great blog which the writer has built his own rapid prototyping machine

HobbyCNC supplier of inexpensive driver electronics and kits

Probotix another supplier of inexpensive driver electronics and kits

After looking at the DIY as well as the turn-key versions, I started to design my own.  The first image uses the rollers shown at Build Your CNC (Rollers).  I was planning on building my own from replacement roller blade bearings, which are about $20 for 16.  In order to do a three axis machine, you would need 8 per axis.

XYZ router with roller bearings

This version has its drawbacks, mainly cost of bearings, and the volume each slide occupies.  In order to get the most out of any sliding system, the opposing bearings should be far apart as possible.  This reduces the amount of angular misalignment.  I found that the ball bearings on angle took up a lot of space.  Also, depending how big you want the bed size to be, there may be a lot of deflection between the bearing, which will add to the inaccuracy.

After looking around in Ace Hardware for a while, I found round bar stock and brass bushings.  This will allow me to design the router using the same methods as the professional machines, without the high cost of the linear bearings.  I will be loosing a bit of accuracy (compared to proper linear bearings) and gaining higher friction, but it should work pretty well.

XYZ router with brass busings

Once I have completed the design and built the first prototype, I will make it 'Open Source'.  Although I would like to sell the plans, legally I can not.  The version of Pro-E I have is only for student use, i.e. not for profit.  Bad for me, Good for you. But maybe I'll add a donate button for all you generous folks out there...

Keep an eye out for a new 'Designs' page.

Project 2 - RoboSpider

While waiting for the board to complete Project 1, I decided to start designing the parts for Project 2.  For this project, I decided to build a spider robot leg, with 3 degrees of freedom.  Using my student version of Pro-E, I modeled the below.

Prototype RoboSpider Leg

The servos shown are going to be Futaba S3003 or a servo of equivalent size.  The blue parts were going to be Delrin by DuPont, until I noticed the cost (I was thinking of using Delrin because it is a great engineering plastic and I wanted to see how workable it is).  I will probably use 1/4" thick wood instead (at least for the first prototype leg).

The plan is to build one simple leg to use to develop the code needed to drive 6 legs.  I am not a programmer and have had zero formal education.  All I know has been learned from a few books but mostly from trial and error.  So there will be a lot of time between the completion of this project and the start of the 6 legged creature.  But, it should be simple geometry, how hard can it be...?

DIY Workbench and Storage

Five months ago, my wife and I moved into our first house, but the exciting part is that it has a three car garage!  Today, I was finally able to invest some time (and money) setting up my workspace.  I modeled the workbench and storage in Pro-Engineer, a 3D modeling program (a Professional program like SolidWorks).  Here is the rendering of that layout.

Pro-E Rendering of Workbench and Storage

Pretty sexy, ain't it?  The workbench frame is made from industrial shelving I purchase from Lowes.  I got the idea from Craftsman's Gorilla Garage line (which is pretty expensive).  Here is the cheap version...

Actual Workbench and Storage

The shelving and wood cost about $160 (24"x96").  It will cost about another $150 to have the storage section added (the rack to the right).  That will have to wait another month or two.

But on to the workbench surface.  I did a little shopping around to see the prices of workbench tops.  You can find one online for about $90 to $225 depending on the material (24"x96").  Craftsman actually seemed to be one of the cheapest suppliers, take a look - Craftsman .  But, I know I am going to destroy the surface pretty quickly so I wanted a cheaper solution.  I decided to buy a few sheets of Plywood and Particle board and glue them together.  The cement was used to add weight to press the three layers together (there were still a few gaps around the edges.

Gluing In-Progress

Three layers of the workbench

The cost for this work surface was about $40 after two sheets of wood (3/4 of which was used), and a container of wood glue.  After the glue had dried, I sanded the surface and added a nice rounded edge.  I then sealed the wood with a rub-on seal/polyurethane coat.  It ended up looking pretty nice.

Wood surface protection

Completed workbench

Project 1 - Joystick Controlled Servos

If your anything like me, you'll understand the difficulty in completing projects.  I want to do so much, nothing gets done.  I call it Technological Attention Deficit Disorder (TADD).  I have been playing around with electronics for over 4 years now, and I have little to show for it.  (apart from hundreds of dollars worth in parts).

So to rid myself of TADD, I decided to complete my first project.  I have forced myself to start and finish one project before moving on to another.  This first project, I have cleverly named Project 1, consists of controlling two RC servos with a joystick.

As you can see from below (or not), I have a PIC16F877A from Microchip, the big black thing in the center.  Connected to that, there are two potentiometers (POT) (the poor man's joystick) on the far left.  Then there are connections for the two RC servos, on the lower left.  Finally, there are lots of other connections to support the PIC's needs; crystal for timing, power, ground, and the black wires which connect my programmer.

I use the PICKit 2 from Microchip which has three interfaces directly to the PIC and another two for power. This is a great programmer in my opinion (but I am only a novice at programming, so what do I know...).

Now I hope to have some video of the above breadboard assembly in action, but for now, you'll have to bear with my description.  You know when you do something for the first time... and everything goes just as you hoped?  This was not one of those times, it sort of worked.  I guess you could say it was a success, I turned the POT, and the servo rotated....

There was a lot of noise coming from..... somewhere, and the servos were not at all smooth.  I did try to move a few things around, but nothing seemed to help.  I had a look on-line and noticed that when designing printed wiring boards (PWBs) a ground plane is typically added to reduce/eliminate noise.

I'm not saying this was the right thing to do, but I jumped into designing the schematic and then layed out a board (for the first time).  So here they are, on order from BatchPCB.  All the parts will be ordered from SparkFun.