I am looking at a steel dart falling onto a piece of glass that's bonded into a plastic housing by 4 spots of glue.

I started off by modelling the glass, the glue dots and the dart, making the glue dots built in on the surfaces where they would contact the plastic mount in real life.

So, I start the analysis with the dart in contact with the glass and moving at 1.4 m/s to simulate a drop from 100mm, no penetration contact between dart and glass. I managed to get the problem to solve but found it very unstable. Every time I changed a variable and then ran the solver I got an error message. If I then shut down SW and then re started it the problem would solve. No idea why, at 21000 nodes the problem wasn't huge. I was also surprised to find that when adding gravity SW defaulted to setting gravity to 0 at t=0 and increasing it linearly to 1 (X9.81) at the end time, but I guess those guts in SW are just making sure you're paying attention .

Anyway, I began to realise that the stiffness of the plastic housing might have a significant effect on the result so I added the mount to the model, see below.

Nothing will induce this model to run. I did a frequency analysis to check the duration covered about 3 periods of the fundamental frequency, I've run a static contact analysis to check everything is properly stuck together and messed about with the damping but still no solution.

Is this model just too complex for SW to handle or am I missing something obvious?