If you haven’t already considered pack pressure, pack/hold time and cushion size, you might want to take a look at it. Most non-filled, semi-crystalline materials will have warp problems if they are over-packed. Of course, part geometry, wall thickness, etc. will play a part in warping of the part. But I would start with checking the fundamentals of the setup; the pack/hold dynamics would be a good place to start.

I have summarized the advice I got with a few remarks of my own based on the experience gathered:

1. Perform a mold flow analysis on the part especially concerning residual stresses, gate size and location and cooling channel location. This process is still on going so I really can’t comment on the results.
2. Heating the mold helped considerably.
3. Differential heating also helped (20C temperature difference between mold sides).
4. Low pack and hold pressures.
5. Use of fillers actually made things worse.
6. Cooling on fixtures helped.
7. Use of MDPE helped.
8. Foaming didn’t help much.
9. Making sure mold temperature is constant; this was never an issue so I can’t comment.
10. Making sure wall thickness is constant; this was never an issue so I can’t comment.
11. Annealing helped.
12. Increase temperatures in the central nozzles were never tried.
13. Nucleating agents didn’t help much.