We were standing on the dock in Sitka, Alaska this past summer, chatting with our neighbors as we topped off the aft fresh water tank. We were getting ready to leave, and being almost full of fuel, we wanted to fill the water tanks as well to test motion in the big swells we expected to find in the Gulf of Alaska. The combination of extra heavy displacement, and long-period single-direction wave sets promised good engineering data.

The air vents for the water tanks empty into the forepeak and engine room. We rarely fill these tanks, but when we do, we just wait until the bilge pump starts to cycle, indicating the tanks are overflowing, to turn off the hose.

You can imagine our surprise when we heard the high water alarm sounding. We turned off the hose, and then checked the engine room. There was four inches (100mm) of water in our usually dry bilge. The big PAR diaphragm pump was running, but it wasn't pumping (and as a result was quieter than normal).

We used this as an opportunity to test our hydraulic damage control pump, and within a few seconds the water level was down to an acceptable level.

We've been using one form or another of these Jabsco PAR bilge pumps for 30 years. They have a large diaphragm, geared motor, and check valves. We have found them to be more reliable than the submersible pumps, and they don't back flood as will a submersible. These pumps depend on a pair of check valves - one of which is shown in the base above. The check valves are their weak point. If they are fouled by debris they will not work, so you have to use a large strainer between the pick up point and pump.

We knew from previous experience that the problem was apt to be in the check valves. Pulling the top of the pump we found that the rivet on the outlet check valve had failed, allowing the rubber flapper to drop into the body of the pump.

We have a spare pump and parts, but as the valve was in good shape, aside from the rivet, we decided to repair rather than replace it. You can see the rivet in the photo above. It would not come through the hole as the other end was enlarged, so we put the side cutters to work to cut off the head.

The hole in the valve was too small for any of the bolts we had aboard, so this was drilled out to accept a 4mm (3/16") bolt. We used a lock washer and double nutted the bolt to lock the assembly in place.

Here is the body, now with both check valves in place. We bolted the motor and diaphragm assembly back on top and the pump is now working fine. This was the first problem we'd had in two years with these pumps.