A Long Story of an InterestingProblem with a 2003 Mercury 115 4-Stroke Engine

[GoldenDaze]

During summer 2023 and right through to the end of the year, I had a very perplexing problem with my 2003 Mercury 115 4-stroke (on my 160 Dauntless). I didn't post here looking for help since it wasn't strictly Whaler-related; instead I had a running thread on thehulltruth.com where I got lots of input. Some of it was even useful! But I wanted to relate the story here just in case it proves helpful to anyone else.

Starting in the early summer, my engine would occasionally "cut out" at cruise speeds. At first it was just a little blip, almost too quick to notice. I'd describe it as if I'd very quickly turned off the ignition and turned it back on. As summer went on, the problem became more pronounced, with the engine sometimes running very roughly for a few seconds before smoothing out again. The problem didn't really seem to be related to anything in particular, other than I was always running at planing speed. I checked the ignition switch and the kill switch, but both seemed to work fine.

Although intuitively it seemed more like an ignition problem to me, I replaced both the large cartridge fuel filter as well as the small filter that this engine has under the cowling right next to the low-pressure fuel pump, but to no effect. During September the engine would suddenly no longer run above a fast idle. Pushing the throttle up would simple cause it to quite abruptly die at around 2500 RPM.

I started my trouble-shooting in earnest and tried to check all the sensors with an ohmmeter. I did find that the throttle position sensor (TPS) was very noisy (that is, the change in resistance was not smooth) and that seemed like a very plausible culprit. But while the TPS was demonstrably bad, replacing it didn't help. At this point I also checked the fuel rail pressure which read a steady 42 PSI, which eliminated the high-pressure fuel pump and the vapor separator tank as a likely problem.

This engine actually has the Yamaha F115 powerhead (during the early 2000s Mercury bought 75-115 and 225 HP powerheads from Yamaha while they worked on developing their own) and there's a diagnostic connector where you can hook up an indicator lamp. The intent is that the lamp will blink codes at you. All I ever got was one long flash when the engine died, which wasn't even a documented code. I did have to jury-rig my own connector (the factory connector is no longer available new) so I thought perhaps I didn't have the pinouts correct. At that point it didn't seem like there was really anything else I could do, so I took the boat to a local shop. They had the boat for 3 weeks, then told me all it needed was some new spark plugs. They also managed to scratch my freshly-repainted cowling which was quite annoying. Naturally, when I got the boat back it still didn't run. Clearly they're not the top mechanics here at the lake, but the experience did reinforce my desire to fix it myself.

Given that the engine was now starting but then shutting down within 10-20 seconds, my suspicion next went to the manifold absolute pressure (MAP) sensor. This engine doesn't have a mass airflow (MAF) sensor, instead it estimates the intake airflow based on the TPS, the MAP pressure reading, and the intake air temperature sensor. It seemed like if the sensor was bad, then perhaps the engine control module (ECM) would let the engine start but then shut it down when the manifold pressure wasn't seen to be dropping. The MAP sensor was a bit pricey, and you can probably anticipate that it didn't fix the problem either.

One important thing I did learn in the thehulltruth.com discussion was that the Yamaha computer-based diagnostic system does work on this engine, so I bought it on eBay and received the software as well as an OBD-II reader and an adapter cable to fit the engine's diagnostic connector. The diagnostic system turned out to be critically important, just not in the way that I was expecting. What I learned was that when the engine died, the connection to the ECM died with it. This led me to conclude that the ECM itself was shutting down, though I had no idea why. The main relay was a possible cause, so I temporarily jumpered the connectors to simulate the relay connections staying closed, but to no effect.

I brought in a neighbor who's a classic sports car guy, and his thoughts went right to the voltage regulator. I'd replaced the voltage regulator around mid-summer when I noticed the system voltage was hanging around 12 V and there seemed to be no charging current. Given that I had a brand-new OEM voltage regulator, I thought the odds of that being a problem were just about zero. But as a first test, I simply unplugged the voltage regulator. Well what do you know, the engine ran perfectly. I bench-tested the voltage regulator and determined that the rectifier diodes were working properly, but I couldn't do a very good test of the voltage output since the engine wouldn't run for more than a few seconds at a time. With the regulator disconnected, I checked continuity and voltage output from the stator coils and they were within spec. My hypothesis was that perhaps some AC voltage was leaking through the voltage regulator and freaking out the ECM.

I bought a new voltage regulator (this time a cheap after-market version for 1/5 the price) but to my surprise the engine still started up and then died. It ran for a bit longer, maybe 30-40 seconds, but overall the result was the same. I was glad I only spent $30! With the engine running just a little bit longer, I was able to see that the rest voltage at the engine was about 12.5 V, dropping to 11.0 or so during the start, then continually climbing until the engine stopped. The maximum that I was able to measure was 15.0-15.5 V. That's unusually high for a nominally 12 V system, which should have a charging voltage of around 13.7-14.2 V. Perhaps the ECM was shutting down due to over-voltage protection?

While examining the SELOC manual for the 1000th time, I came across a description of the voltage regulator that said something like "when the battery voltage is at the normal charging voltage, the voltage regulator shunts excess power to ground." I know this is an overly-simplistic explanation, but I sat bolt upright -- if the voltage regulator ground was bad, it wouldn't be able to "shunt excess power" which could result in an over-voltage condition! I pulled off the voltage regulator ground connector (which was shared with the ECM ground) and it looked fine. This may be a 21-year-old engine, but it's primarily freshwater and it's super-clean. Nevertheless I polished up the engine block land, both sides of the ring terminals, and the bolt itself with 300-grit sandpaper and put it all back together. The engine idled perfectly. I put the boat back in the water and she ran great from idle all the way up to wide-open throttle.

I spent the better part of 4 months chasing this problem, spent $500 on parts I didn't really need, and ultimately solved it in 5 minutes with a piece of sandpaper. Lesson learned: check your ground connections.

-Bob

[jimh]

BOB--thanks for posting your narrative of how you found the problem with the intermittent cutting-out of your 2003 Mercury 115-HP four-stroke-power-cycle. Finding the cause of that electrical problem was quite an accomplishment.

Regarding electrical connections in a low-voltage system like a 12-Volt system, your description of the appearance of the ground circuit from the voltage regulator to the engine chassis was very interesting.

I have commented before that in working with low-voltage circuits, the appearance of just a very thin layer of oxidation at a connection can interrupt the flow of current. At low voltage, an electrical discontinuity can be created and be so thin that you cannot notice it by visual inspection. A connection may have very normal appearance, but not be connected at all electrically.

The initial circumstance when this problem appeared were, in hindsight, also indicating the root cause: as the engine speed increased, the voltage being produced by the alternator was also increasing. The voltage regulator was then trying to operate as intended, and regulate the now too-high voltage output by shunting current to ground. When the path to ground was no longer available, the engine control module must have decided it had to protect itself and shut off the engine. However, it seems odd that eventually the problem worsened to the point that the engine would even sustain operation at lower engine speeds. This could be explained by a change in the resistance at the poor ground connection. At first, perhaps the connection was not completely an open circuit, just a connection whose resistance was much higher than it should have been. At lower speeds, and thus at lower alternator voltage output, the connection was just good enough to maintain the needed shunting current for the voltage regulator to operate normally. Over time, the connection worsened, and eventually became such a high resistance that even at engine idle speeds the voltage regulator could not keep the alternator voltage under control.

[GoldenDaze]

initial circumstance when this problem appeared were, in hindsight, also indicating the root cause

Exactly! All along I was searching for something that tied all the progressing symptoms together, and I was never really satisfied that I had the right answer with the TPS or MAP sensors. The over-voltage finally tied the whole year together.

By the way, the new voltage regulator that I installed earlier in the year (an expensive OEM one) turned out to be just fine, so it's back on the engine. The "problem" with that one was that it raised the voltage too fast for me to see the pattern with my voltmeter. The cheap aftermarket one brought the voltage up more slowly, so in a way -- like the diagnostic system -- it was critical to the solution, just not in the way I expected.

[jimh]

...The "problem" with [the OEM voltage regulator] was that it raised the voltage too fast for me to see the pattern with my voltmeter. The cheap aftermarket one brought the voltage up more slowly, so in a way -- like the diagnostic system -- it was critical to the solution, just not in the way I expected.

That is quite a twist in the plot.