The effect (on comb filtering) of adding a 30ms delay to one speaker in a stereo pair.
Caveats: room reflections are not considered here. Assumption is a mono sound source reproduced in both speakers. Speakers are 10m apart. 200Hz and above assumes a 60-degree directional driver. This assumes that single frequencies are being played for a significant length of time (much greater than 30ms, for example).

Note: 30ms diagrams assume speaker #1 has a 0 ms delay and speaker #2 has a 30ms delay.

40Hz, 0ms

40 Hz 30 ms.

50 Hz 0 ms

50 Hz 30 ms

60 Hz 0 ms

60 Hz 30 ms

70 Hz 0 ms

70 Hz 30 ms

80 Hz 0 ms

80 Hz 30 ms

Note: As we go to 200 Hz, a single wavelength is 5 ms. So a 30ms delay is six full cycles at 200 Hz, so there really won't be any difference in the patterns for the nulls and lobes. Any multiple of 200 Hz (400, 800, 1000, 2000, 4000, 8000, etc.) will exhibit this same behavior. For this reason I've included some additional frequencies at random points to avoid this problem.

200 Hz 0 ms Wavelength = approx 5.6'

200 Hz 30 ms

222Hz 0 ms

222Hz 30 ms

400 Hz 0 ms Wavelength = approx 2.8'

400 Hz 30 ms

431 Hz 0 ms

431 Hz 30 ms

800 Hz 0 ms Wavelength = approx 1.4'

800 Hz 30 ms

1000 Hz 0 ms Wavelength = approx 8.5"

1000 Hz 30 ms

1234 Hz 0 ms

1234 Hz 30 ms

2000 Hz 0 ms Wavelength = approx 4"

2000 Hz 30 ms

2531 Hz 0 ms

2531 Hz 30 ms

4000 Hz 0 ms Wavelength - approx 2"

4000 Hz 30 ms

8000 Hz 0 ms Wavelength = approx 1"

8000 Hz 30 ms