how not to design injection molds

Some time back, I designed a set of letter and number plates for our large Precious Plastic design injection machine. In hindsight, and to put it politely, I made a few choices that I would do differently were I to do it again. To be blunt- I screwed 'em up.

We have a lot more experience with molds than when we had these cut, both with in-house designs, as well as out-of-house designs. I like to think that everything's an experiment, and any experiment can fail. Right now, these molds are falling into the "fail" category. But there's almost always a fix, and that's what I'm in the process of now. But first, let's have a look at these molds:

A photo showing the problem- when the die plates flex, plastic flows in between them, creating a huge cleanup problem

The screenshot above seems pretty good at first glance. We specified aluminum for these plates in an attempt to keep the weight manageable. We are making more than one letter or number at a time, which is good from a productivity standpoint. And the layout has a central injection point and corner bolt holes that are all in the same position. This means these dies can be stacked- creating four or even six letters or numbers at a time. We arranged the letter combinations to minimize "closed" letter shapes, like A, P, Q, R, O, D, B, 6, 8, and 0. These letters need a dedicated backplate with the centers either welded, or in our case, riveted into place. The other letters can use a plain backplate, reducing the total number of plates required. This, in turn, reduces the cost.

But the thing that I've learned is that injecting molten plastic creates a huge amount of hydraulic pressure... enough to deform 1/4-inch aluminum plates if they're only held at the corners. When that happens, plastic squeezes out in between the plates, creating a huge cleanup problem. To prevent this, we need more bolt holes.

And here's where my layout mistakes become apparent. If we simply add four more bolt holes in the center spaces between the existing holes... well, that's OK for some of the dies, but not all of the dies.

For example, take a look at the 4&5 die. Adding four more bolt holes in between the existing holes is no issue at all, and should fix the problem. This remains true for most of the dies. But a few are going to be a problem. The most obvious example is the I&M die, where I was trying to squeeze two letters in a space that, really, should have been left as a single letter. The cutout for the I running across the top of the die presents a real problem. Other trouble spots include the K&L, E&F, and W dies.

These could be corrected by adding two new holes across the top & bottom, for a total of 10 new bolt holes. But this won't work for all of the dies... it will interfere with some of the other letter cutouts. Again, I was trying to squeeze letters and numbers that were too large into a small, compact space. The dies have to be a certain size in order to spin onto the machine, Plus, smaller dies are easier to handle that larger dies.

Two of these dies are a real problem (even more frustrating because it's a problem of my own making.)... the S&T die and the M&I die. As far as I can see it, I have solutions, but none of them are really good. 

One fix is making a custom top and bottom die plate with a custom bolt pattern... expensive and a bit of a pain to keep up with. Another fix is to recut the letter plates for individual letters... expensive, and a little painful since we've already spent a lot of money on these dies. A third fix would be to make larger steel top and bottom plates, with bolt holes that fall outside the outer edge of the aluminum. This might be the cheapest option, but it would result in a heavy & bulky die stack.

So for now, I'm adding four new bolt holes to the existing dies. That should solve the problem with 15 of the dies, leaving 5 unresolved. I think I'll sleep on it for a few nights before I decide how to fix the others.

---