Is DOE Always Necessary

[brentb]

Say you do the "4 Corner" or Appearance Process Window and all the visually parts are "good" dimensionally, giving a Cpk in excess of 2. Would doing a "DOE" be necessary?
I have a couple of molds that we are validating that are so tight, we can't get them to flash, even after more than doubling the pack pressures required to produce a visually good part.
The customer also gave us super-wide tolerances.

KOM
brent

[Suhas]

Hi Brent,
Good question buddy! I hope a molding damager, I mean manager reads this :)
4 corner = 2 variables, 2 levels = 2 factor, 2 level DOE . So you already did a DOE ! Should you vary another variable? I think yes. Let us say you used hold pressure and melt temp for your 4 corners. What is the effect of cooling time? If it is an olefin, what is the effect of mold temp? I think these would be good to know. May be at the cool time and mold temp you have the process is robust. But you should also know the 'weaknesses' of your process and hence the answer is yes.
Hope this helps.
Suhas

[brentb]

What muddies the waters is the fact that the mold has hand-loaded inserts, and molded part is hand removed. Cooling study is rather pointless. Mold and parts are not too hot, but comfortable to the touch. Insert weights vary, hence no gate freeze study, but we put plenty of 2nd stage time on. Of the six step, we do the viscosity curve,pressure drop,cavity balance ( with a very short shot and then remove insert) and 4 corner.
With the tolerances involved, things are actually a molder's dream,will do PQ next and some longer-term SPC

KOM
brent

[skyhew]

Scientific moulding is more important then DOE in Injection Moulding. Setting the most stable parameter is better, rather then to find the parameter windows.