The Experiment I board, for the most part, is working. Unfortunately, my old oscilloscope (a Tektronix 2465BDM) died during this experiment and has temporarily left me without a measuring instrument. After some research, Jon and I decided to purchase a Rigol DS11032 (Rigol's web site is out of date. Try here for a description). This is a 100MHz DSO in the $800 range. It will be delivered mid-November. I also plan on getting a 20MHz to 50MHz analog scope at some point. Probably used via eBay.

Here is what is working:

1. Power. All voltages are up and working (+24V, +13V. +5V). The two power regulators each had their shutdown pins wired wrong. They should have been left floating. To correct this, I cut off those pins. Also, the DC and AC input terminals didn't have the right size holes in the board. To correct this, I just drilled them out and fixed the wiring by hand.
2. Zero-Crossing Detectors. Both zero-crossing detectors are working fine. The zero-crossing detector recommended by the Insteon reference design works best. It has a crisp transition right at the zero voltage point.
3. Transmit Path. The schematic had the wrong resistor value in the emitter (is was set to 15K ohms when it should have been just 15 ohms). I lowered this value but blew it up. I only had quarter watt resistors on hand. So, I just put in a 100 ohm resistor which drew far less current from the source. Works fine for now.
4. Receive Path. I was able to see the Insteon BPSK signals near the zero-crossing of the AC power lines on the output of the receiver. I generated the signals by asking a working Insteon device plugged into a AC outlet to do an 'all-link'. This makes it send Insteon signals for a few minutes.
5. Charge Pump. Charge pump is up and running. It generates enough voltage to clip on the zener diode.

Unfortunately, we were not able to get the actual Insteon chip to work. We figure that this is because the chips sent to us by Insteon were not actually programmed with modem software (Insteon told us this might be the case). So, we were not able to get a fully functioning modem.

To fix this problem, we decide to suck the program out of a working modem and then reprogram the chips ourselve. However, this turned out to be more difficult than we thought because we didn't include the proper header pins in our design to program our own chip (we needed a program voltage, clock and data). After burning up a couple chips, we gave up. We might try and fix this at a later time when we have a rework station (for getting a working Insteon chip out of a factory modem) and a working oscilloscope.