[Algorithms] Geometrical insight into K matrix for point to point constraints?

[<Pau...@sc...>]

Hi guys,

Are there any physics gurus out there who can give any geometrical 
insights into whats happening inside the K mass matrix for a point to 
point constraint? The K matrix is composed of a sum of a number of other 
matrices: point of application cross product matrices for each object 
multiplied by their respective inertia tensors and a identity scaled by 
the sum of inverse masses. This is then inverted to produce a matrix which 
will convert a relative velocity (at points of application) into an 
impulse which will zero said relative velocity on both rigid bodies.

Now, this is not immediately intuitive from a geometrical point of view 
(at least to me anyway), not in the same way as an impulse computed for a 
contact (which is basically a reflection vector with mass bias)... And my 
attempts to derive a point to point constraint using the same methodology 
as the normal impulse have resulted in a system not as stable as when 
using the K matrix, principally due to the fact that there is no contact 
normal for point to point constraint, so you use the application points 
delta vector instead, but this goes to 0 pretty quickly - and i notice the 
K matrix doesn't even feature this, so i must be doing something wildly 
wrong? :) 

Anyway, sorry for the long post!

Cheers, Paul.
**********************************************************************
This email and any files transmitted with it are confidential and intended 
solely for the use of the individual or entity to whom they are addressed. 
If you have received this email in error please notify pos...@sc...
This footnote also confirms that this email message has been checked for 
all known viruses.
Sony Computer Entertainment Europe Limited
Registered Office: 10 Great Marlborough Street, London W1F 7LP, United 
Kingdom
Registered in England: 3277793
**********************************************************************
P Please consider the environment before printing this e-mail