The Troubleshooter, Part 26: Disk gates can be trouble

I opened one of the boxes of parts sent in daily by molders who need help with something that is causing them problems. I could see cosmetic problems on the outside of this part-some sort of cover-not just at the gate but in the center of the part and at the end of fill areas. This part had burn marks, short shot areas, and cosmetic problems all over it.

The part was being fed by what looked like a sprue gate into the center of the part and the material appeared to be a glass-filled nylon of some type. The wall thicknesses were pretty uniform at .090 inch with ribs of .070 inch. The glass-filled nylon is a low shrinkage material, so the ribs can be upwards to 70 percent as thick as the walls they are attached to.

Sink where the ribs attach to the walls doesn't seem to be the issue, so I'll leave the rib thickness for later. I could see generous radiuses in the areas where ribs or other projections attach to the nominal walls. This is very important to parts molded out of glass-filled nylon because those sharp corners can cause a lot of post molding cracks to appear.

Gate Design
After a closer look, I could see that what I thought was a sprue gate was actually a disk gate into a 1/4 inch hole on the part. When I saw how the disk gate was feeding from the sprue into the wall of the part, I could understand why I was getting a chance to help this molder out. The area where the material transitions from the sprue into the part is very restricted.

Glass-filled 6/6 nylon usually processes at a 540F melt temperature, and I could imagine the molding technicians working on this job had to work a lot higher than that, probably 580 or 590F, just to get the nylon to flow through the restricted gate area. Not only that, but I could also see that the sprue O diameter was too small for running glass-filled nylon.

The sprue O diameter is .170 inch and the nozzle orifice is .125 inch-much too small for this type of material and a part this size (approximately 71/2 by 41/2 by 31/2 inches). I am sure the nozzle is freezing off a lot, and the molding people are probably putting pieces of cardboard in between the nozzle and the sprue bushing to keep the nozzle open.

The sprue has a step in it that jumps from .210 inch to .280 inch in the middle of the sprue. Just eliminating the step would add .070 inch to the O diameter and increase the nozzle orifice to .195 inch or maybe a little more. I would prefer a sprue with an O diameter of .275 inch so I could use a 1/4 inch orifice in my nozzle.

Cosmetic Signs of Problems
The flow marks that are originating at the disk gate and running pretty much straight out from the gate are being caused by the high barrel heats, set in an attempt to overcome the flow restriction at the gate. The small amount of sink where the ribs attach to the part wall can probably be packed out when the gate is opened up.

The disk gate depth appears to be .050 inch now, and it should be opened up to a minimum of .090 inch. If a little roughness on the top side of the gate is not a problem, the depth could even go more than the .090 inch wall thickness and fill the part both into the wall and on the top surface, around the 1/4 inch diameter hole. The combination of gating into the side of the 1/4 inch hole and on the top wall will create a wide open area to dump material into, and this material can use all the help it can get.

You have to remember that glass-filled nylons flow pretty well until it comes to small nozzle orifices and small gates. It is a volume thing in most cases. Unfilled nylon flows like water and isn't bothered by small gates or nozzle orifices, but glass-filled material is different.

The Usual Culprit
The only thing left to figure out was why we are getting the roughness at the end of fill areas. I was sure it was just a venting problem, and after looking the part over with a magnifying glass, I was sure of it. We sometimes forget that trapping air in the corners of the part is going to cause the trapped air to compress and heat up to the point where it burns the material we are trying to shove into those last little corners.

It shows up as short shots or burns in those last places to fill. On this part, it is showing up as both, short shots or non-fills in some of the corners and actual burning of the material, which is giving the part a grainy appearance, in some of the other corners. Of course, venting the mold is the answer. I can't believe we still have moldmakers who don't know venting is so important to our being able to mold a part that is visually and dimensionally acceptable to our customers.

We do not have a runner to vent this time, and because we are gated into the center of the part it would be appropriate to use perimeter venting around the entire parting line of the part to get rid of as much of the air as possible. The vent depth for glass-filled nylon is somewhere around .0010 inch to .0015 inch, depending on the barrel heats, injection speed, and pressures being used.

Personally, for perimeter venting of glass-filled nylons, I typically use a vent depth of .001 inch. Go out .060 inch from the parting line at .001 inch deep, then drop into a .040 inch deep channel to atmosphere, and finish up by polishing the vent lip to an A1 finish in the direction of air flow. I generally figure, if I do a good job of perimeter venting, I will probably not have any trapped air in the corners of the projections on the parts. This isn't always true of course, but if I do have the problem I can always use blind vent pockets or vacuum venting to get rid of the extra air.

I called the molder and gave him the good news about how little it was going to take to correct the cosmetic problems on his part. I also gave him a little extra good news by suggesting that after he opens up the flow restrictions and gets rid of the trapped air, he will be able to lower the barrel heats to the 540F melt temperatures I mentioned previously. Then, depending on how well the mold cooling is designed, he might be able to cycle a little bit faster than he is currently running. I told him I would expect a part like this to run somewhere around a 30- to 35-second cycle.

There was silence for a few seconds; then the molder came back with the comment that he could use a faster cycle because he was currently running a 54-second cycle and had quoted 45 seconds. I told him to check his dimensions when he brings the barrel heats down to be sure he isn't going to get into trouble with his customer, but we speed most cycles up an average of 25 percent when we do this type of optimizing. We can't get faster cycles all the time, but this one looks pretty good for a faster cycle to go along with all the cosmetic improvements.