Jim,
I never heard of the 'old light bulb trick', even in those pre-SSB days...
I'll definitely try the diode-detected DC level method. But I still wonder how simple 2 or 3" CRTs would behave at airband frequencies. Wave envelope and trapezoidal patterns might be useful to check linearity. Bad modulation is a common problem. And dynamic or electret mike adaptation to the old standard carbon-mike input often seems to be the culprit. I never found a good publication addressing those interface problems.
Willy