Then, about two days later, the joint on the right hose started to leak. Not as bad, but it was noticeable.
The engine coolant runs through a series of hoses and pipes down to the keel cooler. Below is one section where pipes make a complex turn to get from the engine bed to the keel cooler. This is below a small hatch in the engine room floor, and although not impossible to access, it's not the easiest spot either. Those blue shop towels are my tell-tales to locate leaks. When they get wet it shows really well against the blue - much better than white paper towels. With a few of these strategically placed, it makes it much easier to identify the origin of a leak.
Here's an example of two hose segments coupling together pipes, or in this case a pipe and the keel cooler fittings. After closer inspection, the upper connection to the keel cooler was dripping very slowly as well.
Now the question was what to do about all this. If the leaks stayed very small, we could just keep going. But if they got worse, it could incapacitate the main engine which would be a real pain. Embarking on any repairs also carries the risk of complications, some of which could make the situation much worse. And, just to add more complexity, we had guests arriving shortly to cruise with us for 10 days or so. The last thing I wanted to greet them with is an incapacitated boat. And even if I did want to attempt a repair, what would that be? It was completely unclear why these joints were leaking. And any sort of repair would necessitate draining all the coolant out of the system, and if only it actually did drain. You can drain it out of the engine, but the contents of those pipes and the keel cooler have to be pumped out somehow. There is somewhere between 10 and 15 gal to be removed.
Research time! The only thing I could think to do was to separate the hose joints, apply sealant, and reinstall them. But my experience with hoses like this is that they seal very well on their own as long as everything is the right size, etc.
I also noticed that on the keel cooler fittings you could see a bulge in the hose where there was a bell at the end of the nipple. The bell is just a short section at the end of the fitting that's wider than the rest of the pipe. It helps provide a better seal, and ensures that the clamped hose can't pull off the end of the pipe. The two fittings coming off the engine had similar bells visible, but I could not see any on the various pipe ends. This raised my first suspicion which was that the hoses were just clamped over straight pipe. That would provide less effective sealing, and might have allowed the hoses to shift due to vibration. If this were the case, sealant was probably the best solution.
Then I started looking into the different kinds of hose and learned a bunch of important stuff. The hose used is wire reinforced, and commonly called hard walled hose. It is used for long runs, runs with bends, and anything with suction. The wire ensures the hose retains its shape in these situations. It is arguably the most rugged hose one can use on a boat. The down side is that the same wire that ensures the hose holds its shape also resists being compressed to make a tight joint connection. I learned that the fittings need to be the exact right size, and that sealant is often required. This was sounding like a likely cause of the leaks.
In contrast to hard walled hose, there is also soft wall hose which is the same stuff but without the wire reinforcement. Soft walled hose is more pliable and seals much better at connections, especially where the fitting size isn't exactly right. Since all these hose runs are straight, short, and involve no suction, soft walled hose seemed like a better choice for this application. So one option was to replace the hose segments with soft walled hose.
Then, someone told me they had cured leaks by using Silicon hose which is every bit as tough, but much more pliable at the same time. It seals even better at connections, and is much easier to muscle into place. This was starting to sound like the way to go. One reason I was reluctant to tear things apart is that I expected it would be very difficult to get the old hose and pipe segments apart and back together given all the tight spaces. Silicon hose sounded like it would go back together much easier. Either way I needed to get hose, so figured silicon was the way to go. One thing I knew for certain was that if I was going to do a repair, I only wanted to do it once. So I ordered some hose from Seattle to be flow in ($$ ouch $$).
Then I procrastinated for a day, pondering the pros and cons of leaving the system as-is and risking a problem in some far away place, or make the repair while in Juneau and risk making matters worse.
Yesterday morning, I finally screwed up the courage to attempt a repair. The first step was draining the coolant which took about 3 hrs. First I drained the engine block, and then the heat exchanger That was easy. Next I had to disconnect the highest pipe at the engine and use a wet vac duct taped to a hose to suck out the coolant. I worked the hose down the pipe as far as I could, essentially intubating the cooling system. Once I had extracted as much as possible, I started to separate the joint that had been leaking the worst, and as I pulled it apart, used the vac to suck up the coolant that was still in there. Next I tried to remove the coupling to the keel cooler, and sure enough there was not enough space to pull either end of the connection off. So I ended up cutting into the hose and again used the vac to suck up the coolant as it came out. Finally, I used a sawsall to just cut the hose between the two fittings. What a messy, pain in the ass job. But at this point I was passed the point of no return.
Once all the joints were apart, a number of new things surfaced. First, I discovered residual epoxy or something like it on the keel cooler fitting. It looks like this got on there while the boat was being built and hardened like a rock. Then later the hose just got put on over it. Based on how much there was and how irregular a surface it created, it's no wonder this joint was leaking. The other fitting also had some of this crap on it, but not nearly as much. A little emery cloth and elbow grease and it cleaned right up.
The next discovery was a pleasant one. It turns out that all the pipes have nice stainless barbed fittings on their ends. The barbs are not large, but enough to create a good seal and keep the hose from pulling off. So you might be wondering why those joints hadn't sealed well given the barbed fittings. The reason became obvious once everything was apart. Most of the hoses had been clamped over the smooth pipe and not over the barbs. The clamp on the pipe in the picture below was up near the line that I've drawn. So the barbs were serving no purpose at all.
So at this point I was actually quite encouraged that between cleaning the crap off the keel cooler fitting surface, clamping directly over the barbed ends of the pipes, and the use of silicon hose would solve the problem.
But before starting to reassemble everything, I needed to mark all the pipe ends so I could be sure where to locate the clamps to get them centered on the barbed surface. The next two pictures show examples of this. I marked a line somewhere that wouldn't be covered by the hose once it was reinstalled, then measured the distance from the line to where the center of the clamp should be.
Now it's time to put it all back together. First are the pipes under the floor. In the picture below, the hoses are in place and the clamps are ready to be tightened, but I'll wait until I have everything connected and lined up before tightening anything.
Second are the couplings between the engine and the main pipes. You can clearly see the bulge in the lower hose where it goes over the bell in the end of that pipe. There also are two brackets that secure the pipes to the engine, and they need to be reinstalled at this point so the pipes are held in the correct position.
Now we are left with the final coupling between the long pipes and the s-shaped pipes under the floor.
Today I got all the pipes aligned and the clamps tightened up, then fired up the engine to test things out. The only load I could put on the engine was the hydraulics for the stabilizers (I didn't want to leave the engine in gear when I was alone on the boat and down in the engine room), so it took a while to come up to temp. But no leaks! I kept it running for about an hour to monitor, but we will really have to be underway to complete the testing. With the engine fully warmed up, the pipe going into the keel cooler was hot, but the return pipe was still very cold indicating very little coolant flow required to maintain proper engine temp. But given what I found, I'm pretty confident that this problem is now resolved.