I'm not aware of an online calculator that does this for you, although one could certainly be produced or may be available. I did check
http://www.sengpielaudio.com/ which features a lot of other useful audio and acoustic calculators, but unfortunatley not this.
To clarify a bit on the geometry aspects, in addition to the math required to get the actual solution which wforwumbo sums up nicely above-
The direction in which each individual microphone of a stereo pair is pointed will have no influence. What does are two things: 1) The angular difference between the wavefront arrival of the sound in question and the axis along which the microphones are spaced, and 2) How far apart they are spaced.
Wavefronts arriving along the medial plane (from directly in front, directly behind, above, below..) will produce no time of arrival difference between channels, whereas wavefronts arriving from fully 90-degrees off to one side or the other will produce a maximum time-of-arrival difference between channels, determined by how far apart the spacing is. Wavefronts arriving somewhere between those two extremes will produce time-of-arrival differences that are dependent on and change with the angle of arrival. This is where the cosign function comes in.
All of the above is true in regards to time-of-arrival differences between a single stereo pair of two microphones. With two pairs it will also matter if both pairs share the same spacing axis or not, and their positions relative to each other.
In your case, lets assume the to pairs share the same spacing axis and are symmetrically arranged with respect to each other. If the spacing between the PAS hyper pair and the subcard NOS pair happens to be identical (PAS pair mounted directly above or below the NOS pair), the time of arrival differences will be identical for both pairs. Both pairs will produce no time difference between channels for wavefronts arriving from directly ahead. Both will produce increased timing differences the farther the source is off center.
And the timing differences produced by the two pairs for off center sources will be the same. If instead the PAS pair is spaced differently than NOS, there will still be no difference in time-of-arrival between the two pairs for any wavefronts arriving along the median plane, and as before, in each pair arrivals from further and further off the median plane will create increasing time-of-arrival differences. But in addition, the time-of-arrival differences
between the two pairs will grow increasingly different as the wavefront arrival moves increasingly farther and farther off center.
Lots more complexity generated by two pairs that are spaced differently.