Oil Testing Report

1.      Background

It has long been claimed by the general boating community that Yamaha engines ‘make oil’, and in quantities that can be regarded as significantly substantial (up to 3 quarts have been reported).

To test this theory, Total Marine Developments designed a test to determine if, and how much, oil is being made, under specific operating conditions.  While it is not possible to make oil in a practical sense, for the purposes of this test it was assumed that any additional fluid volume created would be primarily due to fuel dilution, which could then be tested and quantified.

Research has shown that it is common for internal combustion engines to dilute the engine oil with fuel, as fuel and combustion products make their way past the piston rings and into the sump.  Taylor (2021) suggests that several factors can influence the amount of fuel that contaminates the engine oil, from “driving conditions, engine design, fuel type, and lubricant type”.

The test scope was also extended to include oil samples from an engine running a factory Yamaha ECU calibration and a Nizpro Marine ECU calibration, to assess oil quality differences and by inference, rates of engine component wear.  As the two ECU calibrations utilise a different fuel strategy, with one injecting additional fuel, it was proposed that an over-rich strategy would show as an increase in fuel dilution.

In addition to the general claim made (“making oil”), solutions to this problem have also been put forward by the boating community over time.  The testing, therefore, also aimed to supply a conclusion of plausibility (or not) to these proposed solutions.

One specific solution centred around a malfunctioning thermostat, and therefore a separate test was conducted to assess the effect of cold start fuel enrichment quantities for a standard (functioning) thermostat vs a thermostat that had failed and ‘stuck’ in the open position.  A thermostat that is stuck open will extend the time taken for the engine to reach operating temperature and therefore extend the duration of fuel enrichment.

To further understand oil dilution quantities, consideration was given to the expected evaporation rate of gasoline.  A review by Ian Taylor (2021) indicated fuel evaporation rates change relative to oil temperature, as shown in the graph below:

This illustrates that as engine and oil temperatures decrease, the amount of fuel that can effectively be evaporated also decreases.  An extreme example of this can be found in field trials, performed by Koleman et al. (1998), who found that short duration trips of less than 10km from a cold start in automotive engines can produce as much as 10-20% fuel dilution after travelling a total distance of 4000 – 8000km.

To put fuel dilution levels into context, Ljubas, et al. (2010) states that a fuel dilution quantity (by weight) of 4-5% is acceptable, while a fuel dilution level of 7-10% is unacceptable.

The remainder of this report outlines the details of each experiment undertaken, their results, and brief commentary on results interpretation.

2.      Aims

• To assess the claim that a faulty thermostat will cause the engine to ‘make oil’.
• To investigate claims that Yamaha engines ‘make oil’.
• To assess the claim that an engine must be run ‘hard’ to prevent ‘making oil’.
• To assess the claim that using synthetic oil will cause the engine to ‘make oil’.
• To assess the claim that a steel bore or plasma spray coated bore is the cause for excess oil.
• To assess if low speed running causes significant excess of oil.
• To assess if standard ring bedding in process is adequate.
• To quantify the level of fuel dilution over time for both Yamaha and Nizpro Marine calibrations.
• To assess oil quality over time for both Yamaha and Nizpro Marine calibrations, as it relates to oil viscosity and additive quantities.
• To identify any adverse effects on engine wear resulting from the use of a Nizpro Marine calibration.

3.      Equipment

The following equipment was used during testing:

• 2021 Yamaha F150XB 2.7L 4-stroke 4-cylinder engine with steel cylinder liners
• Castrol 10w-30 Full synthetic oil
• Ryco Oil Filters
• Vacuum pump for oil extraction via dipstick opening
• DynoLog Dynamometer with vertical mount water brake retarder
• 14,000 Litre Cooling water tower.
• Zeitronix ZT-4 Lambda sensor for calibration validation
• Yamaha diagnostic software (engine monitoring)

 

4.      Method

Test 1: Thermostat effect

The following steps were taken to obtain fuel compensation data during engine warm-up.  These steps were performed twice; the first with a standard thermostat, and the second with a thermostat mechanically held open.  A Nizpro Marine ECU Calibration was used for both rounds of testing.

• The engine was left overnight to allow it to cool to ambient air temperature, which was recorded as 13°C. Engine cooling water was also recorded as 13°C.
• The engine was started and a load of 30HP applied using the dynamometer at an engine speed of 1800RPM and a timer started.
• Engine runtime was recorded as the engine temperature increased, with duration measurements taken for every 5°C increase in engine temperature, from 15°C to an operating temperature of 65°C.

 

Test 2: Fuel dilution

Each test ran for a total of 20 hours, which is 1/5^th of the prescribed engine service interval.  The test was carried out twice; first with a Yamaha ECU calibration, and secondly with a Nizpro Marine ECU Calibration.  During each test, the engine was monitored for temperature, load, and Air/Fuel ratio, with 60ml oil samples extracted at 6.66 hours, 13.33 hours, and 20 hours.

The following steps were taken to obtain the required oil samples to assess fuel dilution levels.

• The engine was left overnight to allow it to cool to ambient air temperature, which was recorded as 15-16°C. Engine cooling water was also recorded as 16°C.
• All engine oil was drained from the engine and a new RYCO Oil filter fitted.
• The engine was filled with a measured amount (4.5L) of Castrol 10w-30 Full synthetic oil.
• The engine was started and a load of 30HP applied, using the dynamometer, at an engine speed of 1800RPM and a timer started.
• Using the vacuum pump, 60ml of oil was extracted via the dipstick tube at 6.66-hour and 13.33-hour marks. The engine remained running during each of these extractions.
• The engine was shut down after 20 hours of running and the final oil sample extracted.

Once testing was complete and oil samples obtained, the oil samples were sent to a specialist oil sample laboratory for analysis.  An additional 60ml of oil in the ‘new’ condition was also supplied for baseline comparison.

 

5.      Results

Test 1: Thermostat effect

Table 1 shows the results of the first test, with a standard thermostat allowing the engine to reach an operating temperature of 65°C in 485 seconds.  The same engine warmed up with a thermostat that is stuck in the open position took 876 seconds.

 

Table 1: Effect of open thermostat on fuel compensation

Engine Temperature	13	15	20	25	30	35	40	45	50	55	60	65
Std Thermostat (s)	0	30	60	82	103	123	149	180	220	278	363	485
Stuck open Thermostat (s)	0	37	68	90	118	149	188	246	323	421	585	876
Fuel compensation (%)	18	15	10	8	6	4	2	1	0	0	0	0

 

Using available data from the ECU calibration, the percentage of fuel enrichment at each engine temperature was calculated and plotted.  For the standard thermostat, a period of 220 seconds was recorded before the fuel compensation had reduced to 0%.  With the thermostat stuck open this duration extended to 323 seconds.

 

Figure 1: Engine warmup duration and fuel compensation effects

 

Test 2: Fuel dilution

During each of the test performed, one for each ECU calibration, the engine maintained an equilibrium temperature of ~75°C.  Lambda was also steady throughout testing, with a measurement of 0.68 for the stock calibration and 1.03 for the Nizpro Marine calibration.

The table that follows shows the results of fuel dilution at the various sampling points of each of the two tests.

At the 6.66-hour point, the Stock ECU Calibration test showed 4.27% of the sample was fuel, whereas the Nizpro Marine ECU calibration sample had just 3.38%.  Fuel dilution at the 13.33-hour point had the Stock calibration at 4.25% and the Nizpro calibration at 3.71%.  Finally, after 20 hours of running, the Stock calibration measured 4.98% of fuel dilution, with the Nizpro calibration reaching 3.98%, a 20.1% reduction.

 

Table 2: Fuel dilution over a 20-hour period; Stock ECU vs Nizpro Marine

	Fuel %
Sample Time (hrs)	Nizpro ECU	Stock ECU
6.66	3.38	4.27
13.33	3.71	4.25
20.00	3.98	4.98

 

These results were also plotted graphically (Figure 2), such that a general trend may be investigated over a period of 100 hours.  From this trend, after 100 hours the anticipated fuel dilution levels will increase to ~5.75% for the Stock calibration and 5% for the Nizpro calibration.

Figure 2: Percentage of fuel diluted in the engine oil throughout the test, and 100-hour trend

 

In addition to the fuel dilution quantity, another important part of the test was to quantify any changes in viscosity.

The overall results for V100 (Viscosity at 100°C), V40 (Viscosity at 40°C), and the Viscosity Index (VI) are shown in Table 3.  Of note are the V100 data after 20 hours of running.  Compared to a baseline value of 10.11 centistokes, the Nizpro Marine calibration reduced 7.2% to 9.383, while the Stock calibration reduced to 8.938, an 11.6% reduction.

 

Table 3: Viscosity degradation over time raw data

Hours	0	6.66	13.33	20	6.66	13.33	20
	Baseline	Niz 6.66	Niz 13.33	Niz 20	Stk 6.66	Stk 13.33	Stk 20
V100	10.11	9.679	9.625	9.383	9.441	9.356	8.938
V40	55.29	58.04	56.03	53.64	55.26	54.05	51.02
VI	172	152	156	160	153	158	155

 

Viscosity at 100°C was selected for trending analysis as this most closely relates to the engine and oil operating temperatures when compared to the 40°C Viscosity figures.  Figure 3 shows the V100 results graphically and has also had a trend applied out to 100 hours.  This shows an anticipated reduction in viscosity to ~7.2 centistokes for the Nizpro Marine calibration, while the Stock calibration would have oil viscosity reduced to ~5.7 centistokes, almost half the baseline figure.

Figure 3: Viscosity degradation over time @100°C, and 100-hour trend

 

Further to the fuel dilution and viscosity data, a breakdown of the individual elements in the oil was also quantified from the baseline and testing samples (Table 4).  Comparing the two 20-hour samples, both showed very few of the elements associated with engine wear (Iron, Aluminium, Tin, Nickel, Manganese), with all wear elements less than 4 Parts Per Million (PPM). These levels are considered typical and are indicative of a healthy engine.

Anti-wear agents and detergents such as Molybdenum, Calcium, Phosphorus, Zinc and Magnesium were also comparable between the Nizpro Marine and Stock samples, with the Nizpro Marine samples generally maintaining a higher PPM count than the Stock samples.

A large drop in Molybdenum was observed across both sets of samples, by a factor of eight, although after 20 hours of running the Nizpro Marine sample contained a higher PPM count.

Both Zinc and Phosphorus levels were reasonably well maintained over the sampling period when compared to the baseline, however the Calcium levels increased significantly for both the Nizpro Marine and Stock samples.

Although not typically seen, the Titanium levels in the oil samples were elevated.  In this case, the Castrol oil contained Fluid Titanium technology and is therefore not of concern.

 

 

 

Table 4: Elemental component quantities (PPM); Baseline vs Nizpro Marine vs Stock

Element	Baseline	Niz 6.66	Niz 13.33	Niz 20	Stk 6.66	Stk 13.33	Stk 20
Copper	0	0	0	0	0	0	0
Iron	0	1	1	3	1	1	2
Chromium	0	0	0	0	0	0	0
Aluminium	0	0	0	1	0	0	1
Lead	0	0	0	0	0	0	0
Tin	0	0	1	2	0	0	1
Silicon	4	5	6	8	5	5	7
Sodium	0	0	0	0	0	0	0
Potassium	0	0	0	0	0	0	0
Molybdenum	824	94	97	107	96	99	103
Nickel	3	1	1	1	1	0	1
Silver	0	0	0	0	0	0	0
Titanium	30	44	43	47	44	44	46
Vanadium	0	0	0	0	0	0	0
Manganese	0	0	0	1	0	0	1
Cadmium	0	0	0	0	0	0	0
Calcium	1973	3064	3056	3036	3100	3105	3009
Phosphorus	647	650	665	686	670	688	668
Zinc	812	831	824	852	840	838	843
Magnesium	7	21	21	23	21	22	27
Barium	0	0	0	0	0	0	0
Boron	9	68	63	64	69	65	65

 

6.      Discussion

Thermostat testing results showed that while there was an extended period of fuel enrichment of 103 seconds, the engine was still able to reach a temperature whereby fuel enrichment reduced to zero with only light (30HP) load on the engine and at an engine speed typical of slow motoring.  During these 103 seconds fuel enrichment is less than 2%, and so the net effect of this additional fuel becomes less and less significant.  A quick look at the area under the curve for enrichment also supports this.

Although a separate 20-hour test was not conducted with the thermostat stuck open, the fact that the engine was able to reach a temperature of 65°C in ~15 minutes suggests that the equilibrium temperature of the engine during such a test would not have been far from the 75°C observed during the two tests performed. Therefore, the rate of fuel evaporation would be close, if not the same, as for a fully functional thermostat.

Considering an engine at idle, warm up times would certainly be extended further, and with this, the amount of fuel enrichment.  Offsetting this effect, when compared to the engine running at 1800RPM and 30HP, are lower injector opening times, of which fuel enrichment is a percentage of, and a lower mass flow rate of water through the pump which is linked to engine RPM.

From the above it was concluded that a stuck open thermostat would not cause sufficient over-fuelling to significantly impact upon fuel dilution.

Turning our attention to fuel dilution levels there are several considerations when reviewing the results.  Firstly, the Stock ECU calibration reached a mass percentage of 4.98 over the 20-hour period, which is still considered an acceptable level according to published research.

Secondly, the proposed trend has limitations in that it assumes that the engine would continue to be run for 100-hours at 1800RPM and 30HP, which would not be the case.  Typical use would also include periods of higher RPM and power levels which would increase the evaporation rate of the fuel in the oil as temperature equilibrium levels increase.  Considering this, the proposed 5.75% fuel dilution at 100-hours would be expected to be on the high side, and yet remains below the 7% threshold of what would be considered unacceptable.  It should be noted that even at a level of 5.75%, the total volume of oil would only increase by ~260ml (0.068 US Gal).

Thirdly, the 4.25% data point for the Stock calibration, taken at 13.33-hours of engine running, is likely to be an outlier when compared to the other five available data points and the typical trend of increased fuel dilution over time.  Confirming this would require additional data points over an extended period, which was beyond the scope of this test, however the trend used considered this and therefore remains appropriate for the data collected.

Finally, the results obtained from the Nizpro Marine calibration showed the most significant impact to fuel dilution levels.  From these results there is a clear correlation between the percentage of fuel diluted into the engine oil and the fuel injected into the engine, commanded by the ECU.  In these tests the Nizpro Marine calibration targeted an Air/Fuel ratio of 1.03, typical for best economy at light loads, while the stock calibration was set to 0.68, which would be considered extremely rich for this operating point.  The 35% reduction in commanded fuel injection resulted in a 20.1% reduction in fuel in the oil.  With a stock calibration there is more fuel being injected, which in turn means that more fuel will condense out of the fuel/air mixture on the cylinder walls, which will make its way into the engine oil.  By running the engine near lambda 1.0, almost all the available fuel in the fuel/air mixture gets used by the engine to develop torque, leaving considerably less to condense on the cylinder walls.

Reviewing the results of oil viscosity degradation over time it was obvious that the reduction in fuel dilution using the Nizpro Marine calibration had a positive effect on oil viscosity, with all three data points maintaining a consistently higher retention of viscosity over the 20-hour duration, and a 5.4% improvement over the Stock calibration at the 20-hour mark.  This is important for a number of reasons. Oil viscosity has a direct impact on oil pressure, with lower viscosity equating to lower oil pressure.  Reducing both viscosity and oil pressure affects the lubrication system, and particularly the journal bearings in the engine, which operate using a hydrodynamic fluid film to prevent metal-to-metal contact and to support the large loads required during operation.  From the data, and associated projections, the rate of degradation of the Stock calibration was higher, meaning that over time the Nizpro Marine calibration is shown to maintain superior oil viscosity properties.

Elemental analysis of the oil samples generally showed that the Nizpro Marine calibration retained more of the oil additives and detergents (Calcium, Phosphorus, and Zinc), added by oil manufacturers to reduce wear and maintain a healthy lubrication system.  The large decrease in Molybdenum is typical, as this element is added with the express purpose of coating engine components to reduce friction and improve wear characteristics.  This was shown in the data as both calibrations reduced detectable Molybdenum levels by similar amounts.  At the elemental level, then, the Nizpro Marine calibration showed an improvement over the Stock calibration, which is also attributed to the reduced fuel dilution.

Assessing the claim that an engine must be run ‘hard’ to prevent ‘making oil’ really speaks to anticipated evaporation rates, as an engine that is being run hard will typically have a higher equilibrium temperature from the increased engine load and higher combustion pressures and temperatures.  Increasing oil temperatures to 100°C would increase the evaporation rate from ~40% to 55%, an increase of 15% from those expected during testing.  Taking 15% from 260ml would be an expected fuel mass of 4.9% which equates to 221ml in volume terms.  In isolation this is a reduction in line with that observed using the Nizpro Marine calibration and could be considered significant, however, running an engine at high load and higher RPM also increases the total mass of fuel being injected into the engine, and lambda targets at these loads are also reduced (richer mixtures), which would offset much of the supposed 15% gains.  The expected net effect would be fuel dilution levels that remain within acceptable levels.

Based on the above, it is our opinion that running the engine hard, while having an impact owing to increased evaporation, would not ‘fix’ an engine that is reportedly ‘making oil’.

The test engine used in this experiment started with a total of 5 hours of run time and had been run in from new following the standard Yamaha procedure as per the appropriate documentation.  At the completion of testing, the engine had been run for a total period of 94 hours, with over 90 of these at or below 2000 RPM and 30HP of load. Therefore, it can also be concluded that running an engine at low speeds and light loads for extended periods will not cause excessive (unacceptable) levels of fuel in the oil, nor will ring sealing be a cause for excessive fuel if following the recommended engine run-in process.

As all the testing performed used synthetic oil and an engine with steel bores, and in consideration of the results which were all within acceptable limits, the proposal that synthetic oil or steel bores is the cause of excessive fuel dilution was rejected.  In relation to the plasma sprayed bores, such as those found in the 4.2L V6, future testing will be carried out to determine its effect, if any.

 

7.      Conclusions

In summary, the following conclusions were made from the testing performed:

1. While the Yamaha engine did ‘make oil’, it was found to be well within the acceptable and expected limits as determined by research.
2. Running an engine hard will not, in and of itself, solve the proposed problem.
3. A faulty thermostat which is ‘stuck open’ will not cause a significant increase in the fuel dilution mass percentage, even at low engine speeds/loads.
4. Using synthetic oil will not cause significant amounts of fuel dilution beyond published acceptable levels.
5. Steel bores will not cause significant amounts of fuel dilution beyond published acceptable levels.
6. Low speed and load running will not cause fuel dilution beyond published acceptable levels.
7. Excessive fuel dilution was not observed for this engine, which has followed the recommended engine run-in procedure.
8. Lambda (or Air/Fuel Ratio) targets significantly impact the levels of fuel dilution of engine oil over time.
9. A Nizpro Marine calibration can reduce fuel dilution levels up to 20.0%
10. Oil viscosity is affected by fuel dilution and can be directly linked to the ECU calibration.
11. A Nizpro Marine calibration will maintain a higher level of oil viscosity than a Stock calibration and will degrade at a lessor rate.
12. A Nizpro Marine calibration will maintain higher levels of additives and detergents when compared to a Stock calibration, which relates directly to lubrication system and engine health.

 

 

 

 

8.      References

Kollman, K. et al., 1998. Extended Oil Drain Intervals—Conservation of Resources or Reduction of Engine Life (Part II). SAE Trans, Volume 107, pp. 738-758.

Ljubas, D., Krpan, H. & Matanovic, I., 2010. Influence of engine oils dilution by fuels on their viscosity, flash point and fire point, s.l.: s.n.

Taylor, I., 2021. Lubricants. 14 09, Volume 9, p. 92.