Detecting, repairing & preventing water leaks: Part I of III.

* In the molding business, water leaks come with the territory, but they cost molders thousands of dollars in labor hours and lost product. Most root causes of leaks are simply oversights, failure to exercise caution during mold assembly and installation, and a lack of proper final check procedures. However, much can be done to more accurately identify, repair, and prevent this costly source of downtime.

In plants where downtime is categorized and tracked, internal and external water leaks have been shown to account for 45% of all maintenance-related downtime. Eighty percent of those leaks are due to missing or improper o-ring or tooling installation on the bench or to mold-set practices at the press. The other 20% is split between cracked tooling, worn or pitted o-ring glands in plates and tooling, and bad mold design.

Troubleshooting leaky molds is akin to troubleshooting leaky roofs. The location of the drip is not necessarily the source of the leak. This article (the first of three) will deal with improving leak detection.

In-press detection

The worst thing to do (but usually what's done first) is to immediately pull the mold and send it to the shop to have the leak repaired. Tooling and o-rings often leak only when under clamp or injection pressure, conditions that cannot be replicated on the bench. Other times, set-up technicians will notice water dripping from the bottom of a mold and pull it without first determining which half, quadrant, or general area the leak is coming from. This greatly lengthens the time it takes the shop to find the leak. And if the leak originates externally--from a loose or improperly installed pipe or snap fitting--much shop time could be wasted in looking inside the mold for an external problem that should have been discovered at the press. (A side effect would be added strain on the often delicate relations between the toolroom and processing sides of the business.)

When the inevitable water leak occurs, several steps should be taken--in the proper order--to help isolate the leak and determine exactly where it is coming from before the decision is made to send the mold to the repair shop.

Check out the history

Before you even walk out to the press, take a quick look (less than 5 minutes) at the mold history to answer three questions:

* Does the mold have a history of leaking and if so, where?

* Was the mold recently worked on, and were any o-rings changed or disturbed? (What was the last cleaning level performed?)

* Did the mold receive a proper final check before being sent to production? If it didn't leak on the bench, then it could be clamp or injection pressure related and will need to be run or dry-cycled to make it leak.

This knowledge will clarify puzzling visual clues you might see at the press.

Dry it out

Verify that all water lines on the suspect side of the mold are shut off. Unless the leak's location is obvious, dry the mold or tooling with compressed air and towels. If the mold faces or cavities are drenched, experienced technicians know that blowing air across the face of tooling is the preferred method and will actually pull water out of cavities, acting as a venturi. Blowing directly against the face will force water deep into tooling clearances and recesses, only to migrate out or cause other issues later.

Alcohol and in extreme cases, acetone, are sometimes used as drying agents to eliminate water quicker. I've tried this method, but I've never had consistent results nor been satisfied that it wasn't causing accelerated corrosion within the mold--not to mention being inherently dangerous.

So I won't endorse the technique, though I know it is standard practice with some molders. I am just not in favor of forcing any type of spray or liquid into cavity tooling recesses and clearances.

After the initial blow-dry is complete, actuate ejector stripper plates to extend the tooling and plates for better internal access. Be sure to blow out thoroughly all external water-fitting counter-bores and other cracks, slots, and holes where water can lie and leak out.

Usually you can now reopen the water valves and watch for the leak or drip to reappear. When a static leak occurs around an A- or B-side cavity, where the o-rings seal against a bore I.D, the water leak can drip down the face, making it easy to see. Then the mold--or just the affected cavity or plate, if possible--can be removed and taken to the toolroom for repair.

If the o-ring leaks at the back of the cavity, gravity carries the water to the bottom of a plate, so determining the exact location is much more difficult.

Also difficult are internal dynamic leaks around rotating shafts, which cause water to gravitate one or two plates away from the source of the leak. The object now becomes to find the cooling circuit that is feeding the leak. There are a couple of different ways to do this.

One circuit at a time

Redry the mold and turn on just one circuit at a time, allowing ample time for water to reappear at the bottom of the plate. If you are using jumpers (shame on you!) they must be removed and each circuit hooked up individually. By locating the circuit that is feeding the leak, you can narrow down the potential causes significantly, which is critical with multi-cavity molds.

Another method is using air pressure to test the individual circuits. Simply blow out the waterlines and then plumb an air fitting and pressure gauge with the snap-on water fitting and pressurize the circuit with standard line pressure (usually about 90 psi). The severity of the leak will determine how quickly you observe a pressure loss.

Once you locate the guilty circuit, the next step is to try and determine which component within this circuit is leaking.

Using dye

A method with which I've had moderate success both at the bench and at the press is using food coloring to identify where the leak is coming from. Adding several food colors (red, blue, yellow, green) to different circuits can speed up the process of determining which circuit, and sometimes which component, is leaking. Be advised that some preparation is needed to make this procedure accurate. Don't make the mistake of simply putting a few drops of the dye directly into the water "In" line, shutting off the outboard water line and waiting for the magic to happen. It won't. An adequate concentration of dye must be distributed throughout the mold circuit before you pressurize it.

The best method is to use several clear water lines equipped with the appropriate fittings, so you can install them coming right off the mold and into outboard or return lines.

Depressurize the system, leave the water lines full, and simply follow these steps:

* Unhook the inboard water line of the circuits you want to test. Remove just enough water from the line (using a towel as a sponge) to add the dye.

* Pour the dye (I use the 1-oz bottles) into the water-line hose or, if you use snap (quick-disconnect) fittings, pour it into the mold line and reconnect the hose.

* Be sure the outboard line is shut off and turn the water on to pressurize the circuit.

* Slowly open the outboard valve until you see the dyed water appear, then close it. The water will darken quickly--so be ready. This is the critical part. Wait too long and the dye will simply flow through the circuit. Shut it off too soon and there might not be enough dye at the point of the leak to notice it.

You can now dry-cycle the mold for a time, or even shoot a couple of shots if you suspect the leak is related to injection pressure. As a bonus, the dye may even show up on the part, revealing precisely where the tooling is cracked.

If nothing can be done at the press to fix the leak, open the valves until the lines run clean, then blow air through all the mold circuits to remove as much water as possible. Do this again when the mold gets into the shop. This will prevent the water from pouring out between plates when the mold is disassembled, which will dilute the dye trail left by the leak.

On the bench

Most water leaks are detected on the bench through the typical procedure of running water through the mold circuit and then pressurizing with 90 psi of air. With the mold disassembled, cores, cavities, sleeves and other components can be seen better while still retained in the plates, allowing the technician to jiggle, tap or otherwise side-load tooling looking for cracks or bad o-ring seals.

Occasionally, it is necessary to remove the retaining plate to expose the backs of tooling for a clearer view. A steel or aluminum bar of sufficient size (at least 2 x 2 in. if clamped across a line of eight cavities totaling about 3 ft) can be used to hold one circuit (line) of tooling in the plate, allowing moderate (under 75-psi) pressurization. The danger is using bar stock too small for the o-ring diameter and jacking up the pressures to high on tooling where the back o-ring is larger than front. I have seen 1-in. steel bars bend far enough (1/4 in.) under pressure to let the back o-ring escape its bore, drenching everyone around before the valve can be turned off. Also, if the clamps don't squarely engage the bar stock, they can slip off under pressure and all eight cavities will shoot backwards ... not fun.

Next month we will discuss repairing different types of water leaks.

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 De reached at [e-mail] or [phone].