Today I tried Dave Swingle's suggestion to get start the car by pushing down on the defector plate:
ALSO - take off the air cleaner. Have an assistant try to start the car. While cranking press the air flap with your finger. This will put more gas in the engine. If it runs like this you probably have the CO set way too lean to run.
To this end, I put the cold start valve back in and got ready to start the engine. I immediately discovered how much fuel is sprayed around when you forget to tighten down the bolt on the CSV. After I let the gas evaporate for a bit, I was able to get to work.
Starting the Engine
The first trick was that I don't have an assistant. I needed a way to crank the engine from the engine bay.
Luckily, I had a simple solution. Years ago I installed a relay in the engine bay to work around a low-voltage starting issue. It's basically service bulletin ST-05-2/82, which I thought I'd already had but didn't. My solution was to mount a relay that, when triggered by the starter circuit, would run full battery power from the positive jump post to the starter solenoid.
Since I had this handy relay socket, I could easily remove the relay and jump two pins to turn the starter. The problem is that I need to hold the air cleaner down as well, and I needed two hands to insert a jumper.
To resolve this problem I went to Advanced Auto and bought a momentary pushbutton. They called it a "starter switch". It has a strong spring so that I wouldn't accidentally hit it. I attached two wires with blade-style connectors to it, and plugged it into the relay socket.
Testing the Mechanical Advance
To test the mechanical advance, I first needed to disconnect the vacuum advance from the distributor. This was easy to find, as I'd taken the engine cover off a few days prior and had a much easier view of the back of the engine.
Next I turned the key to the "run" position, held down the air deflector plate a bit and pushed the button. The car cranked and started right up. I then let go of the plate, and it died. I repeated this a few times, attempting to switch from holding down the plate to pulling the throttle arm back to rev the engine.
After a few tries I finally got it revving and was able to get a timing light on it. The red mark had moved far to the left (counterclockwise); I don't know the exact RPMs or angle, but it was probably 40-50 degrees from 0. This implies that the mechanical advance is indeed working correctly.
After that test, I released the throttle and waited for the car to stall. It actually kept running for almost a minute, although it did run rough. It finally stalled out, but at least I'd determined that the mechanical advance was OK.
I did notice a couple other interesting things while the car was running. First, there was a hissing sound coming from around the fuel distributor area. I wasn't sure if this was from the air deflector plate (which was partially opened as the car idled) or something else. I'm fairly sure there isn't a vacuum leak, as I thoroughly tested for those last year when tracking down a high idle that wound up being a bent throttle arm. That's not to say that one hasn't appeared since then, though.
I also heard a few popping noises while I was playing with the throttle. They sounded like the noises I'd previously attributed to the exhaust while I was driving to the garage, but this sounded more like it was coming from the air metering unit. It's possible that it was just the deflector plate banging against the top of the metering unit as it closed, but I'm not totally sure.
After some discussion with Dave Swingle and others, I've decided that it's not worth spending money on new K-Jet parts (a new cold start valve, possibly new injectors, and maybe even a new fuel distributor) when I plan on doing an EFI conversion in the fall. Instead, I'll be moving right to EFI.