Preventing water leaks: Part III of III.

The pounding could be heard halfway across the plant. As the clock on the wall ticked toward the end of the shift, the repair shop was in high gear assembling several molds that had been pulled for scheduled maintenance. This is the time of the day, appropriately known by some as the "four o'clock slam," when shortcuts are taken and mistakes are made by repair technicians under the gun to get molds back out to production--and go home. No fiddling and fitting here--it's hammer time.

On one bench I could see the cavity plate lying face down and 32 cavities all ready with o-rings and a light coat of grease, perched in their bores and waiting to be installed. With a 5-1b rawhide hammer in hand, the repair tech quickly sighted down the dowel pin in each cavity to its locating slot in the plate. Satisfied that the alignment was "close enough," he let fly with the rawhide and drove the first cavity into the bore with one bell-ringing swing.

The cavity bottomed out in the bore, and without missing a beat he drove the others home in a random "whack-a-mole" fashion until all 32 cavities were seated. I was amazed that only four of the 32 locating dowels broke off.

Afterward, the retaining plate was bolted on and the plate hooked up to a portable water table to check for leaks. As I took cover well out of range, I was not surprised to see water seeping from several of the abused cavities' o-rings.

Like he had done a hundred times before, the tech marked the leaking cavities, mumbled something about the mold being "junk," disassembled the plates, and removed the affected cavities as brutally as he had installed them.

Upon inspection, every leaking o-ring had a nice half-moon gash cut out of it where the o-ring crossed over the internal water channel.

I politely inquired of the technician as to the supposed advantages of his method of installation. He informed me that he had been taught that cavities with o-rings must be installed quickly, so that the o-ring can "jump" past the water line before it has a chance to expand into the channel and be cut.

He appeared open to suggestions, since he was tired of repeating the procedure over and over until, simply by luck, the leaks just stop--at least on the bench, under no clamp or injection pressure. In fact, I did have a few suggestions.

A few rules of thumb

After many years of impromptu behind-the-bench debates with dozens of repair techs over the best way to perform particular jobs, certain methods have proved effective in troublesome installations:

1. Make certain the o-rings are the correct size. With this type of seal, o-ring cross-section should not exceed 15% of the total gland depth (including tolerances).

2. If the bore walls are rough and pitted enough to abrade the o-rings, try using o-rings made of a rubber with more abrasion resistance, like carboxylated versions of high-acrylonitrile butadiene copolymer (XNBR). Just keep in mind that polar solvents are bad for nitrile-rubber orings, causing softening, swelling, and degradation.

3. O-ring lead-ins (chamfers) should be 15[degrees] and as wide as the o-ring cross-section. Short, 45[degrees] chamfers will shave o-rings in half.

4. Break all sharp edges of intersecting waterlines, counter-bores and dowel alignment slots.

5. Prior to installing o-rings, apply ample grease to coat completely the o-ring and especially the bore walls. If not completely coated with grease, o-rings that are twisted during installation will remain twisted. But I have seen well-greased o-rings slowly untwist in their glands after installation.

6. Because of varying gland finishes, some o-rings will remain slightly twisted even though they are fully greased. When this happens you can help untwist them by carefully lifting them out of their gland with a small diameter punch (old ejector pins work nicely) and rotating the punch around the cavity several times. This will help coax an o-ring back into its original molded shape. Be careful not to overstretch the o ring--i.e., not beyond 75% of its original diameter.

(7.) Instead of "rolling" an o-ring over cooling ribs to get it into its gland, fabricate cones to go over these ribs, allowing much easier and quicker installation (see photo).

(8.) When installing cavities into bores, start by manually lining up any flats or dowels. This can be accomplished more accurately with a small square, but most people do it by eye. Make sure the cavities are straight and not on a bind.

(9.) I am not a fan of using deadblow or soft hammers here or anywhere that you need to "feel" how tight the fit is and to gauge how much you are actually moving (or binding) the cavity or piece of tooling. I like to make hammers from CRS (cold-rolled steel) because it is softer than store-bought hammers and doesn't chip and flake like aluminum, brass, or bronze.

10. As a general rule, a cavity must be inserted at least one-fourth of its diameter into a bore before the bore walls will effectively guide it. (For example, a 2-in.-diam. cavity must be inserted at least 0.5 in. into its bore.) Until the cavity reaches that point, you run a higher risk of binding or cocking the cavity and creating a burr on the bore wall that can cut an o-ring or get trapped under it.

11. When started straight and properly greased, cavities of 4 in. or smaller diameter often can be installed with hand pressure alone. If you can't do it by hand, then use light persuasion gradually, allowing the o-ring time to compress in the gland. Moving slowly lets the rubber of the o-ring compress and slip past the edges of a water channel or other transitions instead of being cut by these edges in a single swift movement. As a bonus, it is much easier and less damaging to the mold plates and tooling when gradual installation allows the cavity to rotate slightly so the locating dowel can "seat" itself into the slot without binding.

What to do with pitted glands?

When glands become pitted and cause slow leaks, the right thing is to send the plate or tooling out and have the gland welded or plated and re-cut to print dimension. In some cases, you can re-cut pitted glands to take the next larger size o-ring. Unfortunately, when production is on the line, there is not always time to do this, so temporary "Band-Aid" measures are used to get by until the glands can be repaired properly. The challenge then remains: How best to fill these pits and stop the leak? I have used several methods with success. Though not a professional fix, they still work in a pinch.

Silicone rubber works well to fill pits and stop leaks. But don't get into the habit (as many do) of using it anywhere a mold springs a leak or a gland looks pitted. Silicone will effectively glue plates and tooling together, making disassembly and clean-up very difficult.

Before applying silicone, the glands must be thoroughly cleaned (ultrasonic works best) and free of grease or oil. If you apply too much silicone, it will squeeze out into bubbler (fountain) return channels, where it sets up and restricts water flow--or worse, breaks off and ends up lodged inside a fountain tube.

Wait until the silicone sets up (usually 4 to 8 hr) before applying water pressure, or you risk more leaks and having to clean out half-cured silicone. You can speed up curing by applying some heat (200 F max.) if working on the manifold side of the mold.

On the other hand, cured silicone is hard to remove from the glands if the process must be repeated, such as after removing a cavity for replacement or inspection. And you must remove all traces of old, cured silicone before new silicone can be applied.

The alternative is two-part epoxies (I prefer JB Weld), which work well for filling glands and are more permanent. But they are not trouble-free:

The glands must be glass-bead blasted to remove all traces of grease or contamination for good adhesion of the epoxy.

Make a miniature trowel from a small rod that fits the gland area so that you can spread the epoxy evenly in a thin, uniform coat.

Once dried, if the coating isn't as smooth as you want, you can spotface the epoxy in a mill or, if you have a surgeon's hands, use a Dremel and a small, fine wheel stone.

Finish the job right

Always perform a final check on a mold before releasing it to production. The best way to prevent unnecessary mold pulls is to discover the problem while the mold is still on the bench.

Apply water to the mold and take it to maximum line pressure, usually about 90 psi, then back down to standard operating pressure of 45 to 60 psi. Don't be alarmed if water initially begins to dribble out around a cavity. As the o-ring responds to water pressure, there will be initial seepage while the o-ring untwists, pushes out the grease, and finds a seat. Light and careful tapping on tooling with a small hammer will send vibrations through the tooling that will speed up this process.

Steven Johnson worked as a toolmaker for 26 years, rebuilding and repairing multicavity molds for Calmar Inc., and today is a mold-maintenance engineer for Hospira Inc., a medical device manufacturer. He also founded MoldTrax in Ashland, Ohio, which designs and sells software for managing mold maintenance (www.moldtrax.com). He can be reached at [e-mail] or [phone].