Repairing water leaks: Part II of III.

Last month we discussed how to find water leaks at the press and at the bench. Now we will examine probable causes and repair methods with an eye on preventive measures to prevent leaks from recurring--or at least to manage them better through specific corrective actions at predetermined cycle counts.

Every corrective action must be looked at from two perspectives. First, determine the immediate action required to repair the leak and get the mold back into production. Second, identify the root cause of the leak, so that it may be eliminated or controlled through scheduled downtime.

Leaks are generally related to failure of o-ring seals. Fortunately for most injection molds, seals are not required to withstand extremely high water pressures (usually only 45 to 60 psi), frictional heat, or abrasive wear. The seals are simple, and for the most part easy to get at, although I have worked on molds where o-rings were tucked into a groove inside a 0.125-in. bore and could be removed only with the help of a microscope.

Two types of seals

Static seals are o-rings placed between stationary tooling or fittings, such as cores trapped and sealed between plates, or cavities housed in a bore with external o-rings that seal against the I.D or walls of the bore.

Dynamic (rotating or reciprocating) seals are those where tooling is in motion against an o-ring or quad seal, such as unscrewing cores and cavities. Also in this class are hot-runner molds with pneumatic or hydraulic valve gates. The valve-pin pistons contain o-rings that undergo higher temperatures (600 F with some resins) and high pressures (500 psi for hydraulic systems).

Typical o-rings are molded from any of a half-dozen different elastomers available in hundreds of different compounds. The majority of o-rings used in plastic injection molds are made of nitrile rubber or a fluoroelastomer such as DuPont's Viton. In some cases silicone gum rubber or LSR is used when the application requires exceptionally pliable or soft o-rings of 50 Shore A or lower hardness. Typical o-rings in molds are 70 to 90 A durometer. Softer o-rings work well in molds with pitted glands and plates.

If you need more specific data on elastomer properties for meeting heat, chemical, or pressure requirements, several online resources publish troubleshooting and selection guides for tooling engineers or mold designers. (Two examples are www.all-orings.com and www.applerubber.com.)

Common o-ring failures

There are four probable causes of static and dynamic o-ring failure:

Material failure: Causes are compression set, thermal and chemical degradation, incorrect hardness, or wrong choice of elastomer.

Incorrect mold assembly: O-rings may be of the wrong size, twisted, pinched, cut, or missing altogether. Inadequate lubrication may also be at fault.

Tooling and plate condition:Common culprits are cracked tooling and worn or pitted glands.

Tooling/plate design: Incorrect stack dimensions, wrong gland depth or surface finish, and inadequate bolt patterns all can lead to o-ring failure.

Compression set causes most water leaks that are not related to poor installation. This means that an o-ring has resided in a compressed state in the mold for so long that it has lost its ability to spring back and thus cannot apply adequate force in the gland to overcome water, air, or hydraulic pressure.

The bad news is that compression set usually occurs with tooling that is difficult or time-consuming to remove, such as A- and B-side cavities, or tooling that requires a complete mold disassembly to reach. In many shops, if there is no reason to remove this tooling (like a compulsory preventive-maintenance program), then it is overlooked in favor of speedier repairs to get molds back into production. Thus, compressed o-rings are left in place too long.

Thermal degradation will accelerate compression set, such as when a hot-runner system is fired up without turning on the water to the cavities. The o-ring is then baked in a compressed condition--basically being "remolded" in the bore and reducing its ability to seal.

Such degradation is another in a long list of reasons why it is critical to follow correct mold start-up procedures and change o-rings at regular intervals based on cycle counts. The cycle count is established by visually inspecting the rings on a definite schedule to establish a baseline seal life.

Chemical degradation normally occurs only when tooling is cleaned with the o-rings in place. Some elastomers, such as nitrile and silicone rubber, are susceptible to swelling and softening if exposed to certain cleaning solutions like acetone or MEK (methylethylketone). Caustic cleaners such as sodium hydroxide, commonly used in ultrasonic cleaning tanks, will have the same effect on fluorocarbon rubbers, so it is important to follow the o-ring manufacturer's recommendation for cleaning solutions.

Wrong size o-rings

Another common occurrence in fast-paced repair shops is to grab any o-ring that looks like it can be stretched or compressed enough to fit, regardless of the correct size. This usually happens when o-rings are removed from tooling before first checking to see if new rings are available from the storeroom. If the old o-rings are damaged during removal (and they usually are) and you don't have enough of the correct size in stock, then you'll have to make due with a less-than-ideal size just to get the mold back into production.

The problem is that the substitute o-ring might appear to work on the bench, but it will begin to leak water slowly under the cycling of clamp and injection pressures. And if the next repair technician doesn't know that the new o-ring is not the correct size, troubleshooting this slow leak will be time-consuming and frustrating.

To avoid this problem, every maintenance manual (electronic or hardcopy) should contain the specific sizes of critical o-rings in a mold. But be advised that attempting to include all o-ring info in every manual of every mold can be overwhelming. Make it a habit to record the o-ring sizes of the molds that get repaired for leaks. Before you know it, you will have collected accurate data for the molds that really need it.

Short "cuts"

Many repair technicians (and production supervisors) do not like to disturb o-rings until they begin to leak--especially on a mold where the plates' water channels intersect cavity bores. Older molds (and newer "low-priced" tools) are often designed without stepped bore I.D's, which creates a razor-sharp edge at this intersection.

The more popular quick fix is to change out the o-rings only after they begin to leak, and then only the one that is leaking, leaving the others for another day. This is firefighting at its worst. While waiting for the rest of the o-rings to leak eventually, the rust that collects around the o-rings will speed erosion and pitting of the bore walls. Then you will have real problems.

The correct approach is to handwork these sharp edges (5 to 10 min per cavity using a Dremel with a medium stone and finishing with a medium Cratex wheel) until they no longer "bite" the o-ring as it passes over. This is a time-consuming procedure, but once done, all the o-rings can be changed more frequently, quickly, and safely with no more worries about random leaks shutting you down. Keeping the o-rings fresh and the bores clean will reduce short- and long-term damage.

Next month we'll discuss methods you can use at the bench during o-ring installation to prevent leakage.

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].