There are a couple ways to optimize this.

As mentioned here, I really think this thing could be fit into a DB9 hood/backshell:
	http://paperlined.org/hardware/microcontrollers/serial/serial_sensor.html


	SN75155		RS232 transceiver		SOIC-8		$1
	12F683		microcontroller			SOIC-8		$1.50
			SMD ceramic 4mhz resonator
	LT1121		5v LDO regulator		SOT-223		$3.50
	GRM188R60J334KA01D	0.33yF cap		30mils		$0.20
	?		DB9 backshell					$3
	?		PCB				?		?

Also needed would be a $10 SOIC-8 clip for programming these.

Ultimately, it's probably just a PITA to put these together.

I don't know if it would be any easier to assemble this dead-bug style, against the inside of each shell half?
Here are the wires that'd be needed:

	1, DB9 for PIC=>DB9
	1, DB9 for ground
	2, DB9 for +12v and -12v tscvr power
	1, DB9 for voltage regulator

	3, voltage regulator
	1 regulator pin free/tied

	2, capacitor for regulator

	2, 12F683 for power
	1, 12F683 for PIC=>DB9
	1, 12F683 for VSS=>PIC
	4 12F683 pins free/tied

	2, tscvr for power
	2, tscvr for PIC=>DB9
	2, tscvr for VSS=>PIC
	2 tscvr pins free/tied

I believe this is 10 wires.  17 of 22 pins used, not counting DB9 connector.








Even further, since really ultimately the core part is reading things through a RS232 transceiver, and recording
timing info...  it may actually be possible for a PC to do this?

Hook the VSS line into a status line that triggers an interrupt in the computer.

In the computer, write a tiny interrupt-service-routine that does the work that the PIC is doing.  100Hz shouldn't be that bad.

As far as I can tell though, this requires writing a separate device-driver in windows:
	http://www.google.com/search?q=site:microsoft.com+IoConnectInterruptEx+OR+IoConnectInterrupt&filter=0
Which makes this a much less portable solution.

Also, it puts the serial port at greater risk to damage.