One fly in the ointment that I probably didnt explain very well is that the affected circuit (the 4.5mc trap and peaking coil) probably doesnt affect anything at all when doing a conventional sweep alignment of the IF. In this scenario, a sweep at IF frequency is injected at the mixer grid, and the set's video detector acts as the demodulator probe. This is when I was using the direct probe in series with a 47k resistor at TP301. The only signal coming off the video detector is very low frequency, probably in the audio range.
In the sceond scenario, a carrier at IF frequency is modulated with roughly a DC-5mc sweep, and injected at the mixer grid. What comes off the video detector in this case is (hopefully) a chunk of the DC-5mc sweep as wide as the IF bandwidth. I then add a high-impedance video detector probe at TP301 to demodulate the sweep. The response should be a combination of the IF response, and the response of the parts following the video detector.
If I assume that the IF alignment is ok (because I just checked it with a conventional sweep), then the parts immediately following the video detector must be peaking up 4.08mc, a lot.
John