I found a proper model for a 7.5mm resistor (silly me, should have known). I'm leaving the diodes as 5mm since they are small, but may move them to 7.5mm too depending on if anyone builds this version and gives feedback (I will probably change it to be honest just in case people wind up finding longer diodes). I'm including the Eagle files as well in case anyone wants to muck about with them (Eagle is the CAD program I used to put this together; it is really nice for a "trialware" product;
So here is what to do (it really is a bit of a pain; unless you really want to build one and give feedback, you may want to wait). Also recommend you see other PCB tutorials (there are some decent ones on Instructables.com and others).
Print out the "test4_brd.pdf" as a reference. You will need these parts (with their names on the board):
1 2.2k resistor (R3)
2 68 ohm resistors (R1, R2)
2 3.6v zener diodes (D1, D2)
1 DB9 male ("plug"), through-hole mountable, 90-degree bend (DB91 -- look at the photo to get the idea; I actually had a hard time finding this)
1 USB B socket, through-hole mountable (USB1)
1 20-pin IC socket (you technically could do without this, but it will be harder to drill, risk heat damage to your IC, and only costs a few cents). Width is I think 10mm on these.
1 MAX233CPP serial driver chip (this is the expensive part, at $5-$7; we could use a cheaper one but would have to add capacitors to the board). Fits in the socket.
1 DIP switch, 6-position (SW-1). In reality, we only need a switch or jumper on the TOP of the six pairs of contacts. So you can do without SW1 for the moment as long as you have some switch or jumper that you can remove between pin 0 of the Arduino and pin pin 2 of the MAX233CPP. These switches come in different varietes; the only reason I used a 6-position is to be able to disable pin 2 if necessary (to talk TO the orb) and to have four user-selectable pins to change options without reflashing the device.
Bits of "strip board header". This is a bit funny--this is soldered into the rows of holes and then plugs into the Arduino below. The reason it's tricky is that it's really designed so the short bit plugs all the way through the board and is soldered on top, which we can't do on a single-sided board. The solution is either to find extra-long header, plug it through from the top, and solder the bottom, or just plug it somewhat into the bottom and try to solder it there, which is trickier than it sounds but worked just fine.
0.8mm drill bits. Get a few spares as these are very thin.
A drill (or drill press if you are lucky)
Okay, on construction (assuming you've never made a PCB before):
You will need:
1 laser printer
1 sheet glossy magazine paper (or glossy photo paper, etc).
1 copper-clad single-sided PCB
1 nonreactive (glass or plastic) tray
1 nonreactive (glass or plastic) container (jar, whatever)
Etchant. You can use ferric chloride crystals (from an electronics place) or two parts hydrogen peroxide to one part muriatic (hydrochloric) acid (you can get muriatic acid from a hardware store).
Acetone (I used fingernail polish remover since it comes in smaller bottles).
Print "test4_bottom.pdf" onto the glossy magazine paper; we use glossy so that the toner comes off easily.
Clean the PCB well (I used soap/water; other folks recommend alcohol). Take a "green scrubby pad" and scuff it gently in both directions so that it looks a little scratched, which will help the toner adhere.
Now (trial and error time) place the toner side of your printed circuit onto the copper side of the PCB and iron it there. I used the "silk" setting of my iron. Hold it in place without moving it for the first bit (about a minute in my case) and then you can gently move it to get coverage. You're melting the toner onto the PCB.
After that, the paper will be glued to the PCB by way of the toner. You want to get the paper off, but don't just pull it off or you may pull off the toner. Let it soak in water for about 5-10 minutes and then gently peel off the paper and try to rub any extra bits off with your fingers. By the end of this you should have a nice clear circuit with toner (and some paper fibers). If there are small breaks in traces (compare with test4_bottom.pdf) you can try to connect them with a sharpie (which didn't work so well for me) or maybe other thick pen or paint.
Take a good look at the circuit. In particular, I found that the nice big pads that I used worked great almost everywhere but didn't have enough clearance on the four pins of the USB plug. You may want to try and clean that area a bit so the four pins are really separated (I had to do it after etching, which was a bit more of a pain).
Time to etch. Mix your etchant in the container and pour a bit (enough to cover your board by a little bit) into the tray--use plastic forceps or nitrile gloves to handle the board! Drop in your board and gently agitate; this takes a long time and for a while I didn't think anything was happening. Eventually the copper will etch away and you will be left with a circuit. Rinse the board off (if you have a metal sink, KEEP LOTS OF WATER FLOWING as you don't want to etch it!) and dry.
You now will have a board covered in toner. Use the acetone and a rag or cotton to clean off the toner.
Drill out all the holes with your 0.8mm drill. Use a VERY light touch. I got lucky and did all holes by hand without breaking a bit, but evidently it's very easy to snap them.
Then, stuff the board (components on the top, which should just be clean plastic by now) and solder on the bottom. Look at test4_top.pdf; the three traces there are going to be wires that you make and solder on the bottom (I used 22AWG insulated wire).
Also, on the side headers: if you look at the diagram you can see what headers actually NEED to be connected with the Arduino. I went ahead and drilled all of them, and for the production boards will probably put "through" headers so there are sockets on the top of the board in case people want to stack other shields on top, wire jumpers and LEDs in, or whatever.
Software to follow. While this isn't a particularly devilish build, there can be a few gotchas (I had to spend quite a bit of time testing traces with a multimeter to find where I'd gone wrong, as the serial side worked fine but the USB side had a broken trace so it just said "one of your USB devices is malfunctioning"--great).
Good luck to any who try; would like to hear feedback!