Warning. Geeky stuff ahead, no wildlife pictures. Those without a deep interest in fuel systems or human suffering might consider skipping this post.
We are back in La Paz and I made arrangements for Rob Cross to help me get to the bottom of the fuel system issues. From the last post we know that the air leak that I see as bubbles forming in the Racor filter housing must be coming from somewhere between the tanks and the supply manifold. I had the opportunity to talk to a technical specialist at Racor since the last post and he agreed with my conclusion about the potential source of the leak. He also convinced me that the fuel levels in the filter housings have reached a steady state around halfway full, and that still provided adequate filtration. Any additional air coming in was passing through and going out as bubbles in the output line. He said that the bubbles in the output is common and not a concern. So, I eventually stopped refilling the housings every day and in fact have not even checked the levels since leaving Marina Puerto Escondido at the beginning of April. That has been 289 miles and 49 hours of engine run time over 14 voyages in the last month. I am pretty sure that this leak is not going to cause the engine to stall at an inopportune moment, but I am still determined to track it down.
Rob got to the boat and we started by pressure testing the fuel supply lines. To do this, we removed the supply line at the tank and plugged that end. Then we used the fuel transfer pump to create positive pressure in the fuel line. We shut all the other valves in the supply manifold, leaving open only the valve for the fuel pump input and the line to the tank. We turned on the pump and… no leaks. The pump is rated for 8-11 PSI of pressure, equivalent to 16-22” of mercury, which is easily 4-5x the amount of vacuum on the system when the engine is running. No leaks on either side.
The next step was to take the inspection plates off the tanks to inspect the fittings and dip tubes. As I observed before the thread sealant on the fitting that goes between the plate and the fuel lines was old and cracking. Bad thread sealant could be the source of the leak.
We labeled the lines and fittings. 
A good look at the dip tubes. 
Fuel in the tank.
On removing the inspection plate we saw the dip tubes for the supply and transfer circuits, both with stainless steel screens at the ends. The welds on the tubes looked good, as did the tubes themselves, and the screens were free of debris. Rob took them to his shop for pressure testing, and they are fine. He used a high quality thread sealant on the fittings, so that should be eliminated as a leak source.
The next step was to reprime the system and fire up the engine to look for the telltale bubbles. Before we did that, Rob suggest that we pressure test the Racor manifold and supply manifold, again using the transfer pump to create positive pressure. The pump is rated for 8-11 PSI or 16-22 inches of mecury, the unit of measure displayed on the Racor vacuum gauge. This is at least 4 times the normal vacuum level when the engine is running (2.5-4 inches of mercury). We found no leaks anywhere.
Next, we primed the transfer and supply dip tubes using the transfer pump, and topped off the racor filter housings. There was about 3″ of fuel in the housings before we topped up, one month after I last checked. The filter elements are 5″ tall, so we were about 60% full, as good as I have seen when I was measuring every day.
We selected the starboard tank for fuel supply and return because it has a shorter hose run and therefore lower vacuum in operation compared to the port tank. We selected the forward filter on the Racor manifold and then started the engine….
Disappointingly, there were still bubbles forming in the filter bowl. We could reasonably expect some bubbles from residual air trapped in the system as we disconnected and reconnected various lines. We used a rubber mallet to tap on the supply manifold and the filter manifold hoping to dislodge residual bubbles. Even after tapping for a while, we were still seeing a small but steady stream of bubbles, perhaps less than before, but the goal is zero bubbles (or, at most, tiny “champagne” bubbles). When we switched the selector to operate both filters, the bubbles disappeared (after some transient air bubbles in the aft filter bowl). What remained were champagne bubbles in both bowls. Progress, but I was not satisfied. At Rob’s suggestion, we checked the fuel tank vents to eliminate the unlikely possibility of blockage there. Then, just to be sure, we plumbed some clear line into the input port of the Racor manifold reasoning that if there was any air at all, we might see at least some sign of bubbles. Nope. None. The fuel going in was absolutely clear. There was nothing more we could do. I believe we addressed any and all possible leak sources, summarized on the table below.
At this point, we called my technical contact at Racor and reviewed all of the findings. He had no suggestions for additional tests, agreeing that we had covered all the possible sources. He said that the bubbles we were seeing were due to cavitation, which, in his experience occurs when the filter is undersized compared to the delivery demands of the engine. However, he confirmed (what I already knew) that my filter unit was well within spec for the engine, and also confirmed that the vacuum levels were well within the normal range. His one suggestion in this regard was that I could replace my filter manifold with the next size up, whose filters were twice the size. The other area we touched on was the fuel supply and return to the tanks. When we told him that there was not a return dip tube into the tank he speculated that the return fuel dropping from the top of tank could be aerating the remaining fuel in the tank, which he called the “aquarium effect”. He suggested that adding a dip tube returning the fuel to the bottom of the tank could negate this effect. In my opinion, neither of these suggestions are worth the time/effort/expense to implement at present.
As the last step, we removed both sections of clear hose from the Racor input and output ports and fired up the engine again. This time, we were seeing some small amount of bubbling when running the front filter, no bubbling at all when running both, and, surprisingly no bubbling at all when running the back filter. I suppose it is possible that it took a fairly long time of engine run time to clear all of the residual air out of the system, but this was quite encouraging. We observed this running the engine at normal cruise RPM, but at dockside. We will need to do a sea trial to be certain of the results.
Sea trial and videos
We got out of Marina CostaBaja on a warm, sunny Saturday morning. After we got everything stowed and up to cruising speed, I went down to the engine room to check on the filters. I decided to run the Racor on the aft filter and was drawing from the starboard tank. The first video shows me checking for bubbles selecting the aft, then both filters, then the forward filter. The results were pretty encouraging. Very little bubbling from the aft filter alone, some bubbling from the forward filter alone, and still less when both were selected. Pretty good, but not perfect.
I continued to run on the aft filter for the 22 mile, 2.5 hour run up to Caleta Partida. When I checked the fuel level in the housing, it was down to about 2″ or so of the 5″ height of the filter element… lower than I’d like to see.
On the way back from Caleta Partida, I decided to run in tandem filter mode, after having refilled the aft bowl. Here is the video with the results.
Again, better, but by no means perfect. There is still a little bit of bubbling even running in this mode, although less than I was seeing before. I’d REALLY like it if there were NO bubbles at all. However, I remain pretty convinced from following this all season that this amount of bubbling is not going to lead to an engine stall at an inopportune moment.
Summary and my conclusions
Here is everything we did to test the system:
- Vacuum and pressure test Racor filter manifold. No leaks.
- Vacuum and pressure test fuel supply manifold. No leaks.
- Check/tighten all fuel fittings – hoses, supply manifold, Racor manifold.
- Check/tighten all valve assemblies on the supply manifold.
- Pressure tested supply lines – manifold back to tank. No leaks.
- Inspected/pressure tested dip tube assemblies in port and starboard tank inspection plates. No leaks.
- Resealed NPT to JIC fittings on the inspection plates.
- Reinstalled inspection plates, tightened all fittings.
- Checked all fuel tank vents. Clear.
- Observed fuel entering the Racor manifold using clear tubing. No bubbles.
I can’t think of any part of the fuel system that we didn’t look at and/or test, and I am as certain as I can be that there are any extraneous leaks in the fuel system. I now believe that the residual bubbling that I see is normal for the filtration system. In fact, a Racor Technical Bulletin discusses air separation in diesel fuel, and starts by listing these facts:
Fact #1: There is AIR entrained in diesel fuel.
Fact #2: A very slight pressure drop can cause air to form visible bubbles.
Fact #3: Air can cause problems.
Racor, Products Parts, Service
and Technical Information, 7480F
I love how understated they are with fact #3. In another Racor document, “Turbine Series Rebuild”, they state in the troubleshooting section that “It is normal for fuel level inside housing to be about 1/2 full after lid removal“. They also mention that if the fuel level gets too low, the engine will stall, and that excessive bubbles indicate either a system restriction (high vacuum) or an upstream air leak.
Going back to the very beginning, I did have engine stalls on two separate occasions last year. I am certain that both of those stalls were due to leaks within the Racor manifold itself, which I replaced back in January. From then until now, I have still seen some degree of bubbling, and have seen the fuel level in the housing consistently down to half full, but not lower. Until now, I have not been able to rule out an upstream air leak as the source of the bubbles. After this week, I conclude that there is no upstream air leak. The final question – is the bubbling that you can see in the videos excessive. I have decided, because it has never caused the engine to stall, that it is NOT.
Done. Really. Finally.
MV Miss Miranda 