After becoming owner of a Yamaha-Garmin combination, NMEA 2000 connected, I started wondering if all the available information on the Garmin is reliable and useful.
So one day with calm sea and blue skies I decided to investigate fuel economy readings on the Garmin. The boat is a Finnmaster 5300 Br with a Yamaha 100 and a Garmin GPSMAP 526s.
The purpose of the trip was to find out:
- Are fuel related readings reliable?
- How do I optimize fuel economy?
To find out if fuel readings are reliable, I tested if the reported fuel consumption matches the actual consumption (measured at the pump). To make sure I got the actual consumption as correct as possible, first step was to fill the tank to a well-known level.
To document the filling, I took a picture of fuel level both before and after the test (click any picture to see full size version). Before sailing, the Garmin GPSMAP 526s trip was reset. The trip data includes fuel consumption, which is what is what I need for this test.
Next step was sailing, which also included taking some screenshots for later analysis.
The trip started and ended in Rungsted. The trip was non-stop, except for a lunch-break.
After completing the trip, I took a screenshot of the trip data. The trip distance was 74 nautical miles, and the fuel consumption according to the Garmin was 49.2 liter.
Next stop was the gas station, where the tank was filled to the same level as before. You can compare the pictures of both fuel levels. I estimate that fuel levels has a vertical difference of no more than 3 cm, which equals 0.02 liter or 0.04%.
The pump indicated 44.76 liter. The conclusion is that the Garmin deviates from the actual consumption with +9.9%
The first thing to note is that Garmin reports the fuel consumption higher than actual, which means that is will never report remaining fuel, when the tank is actually empty. This is a safe approach which is certainly better than the opposite. This deviation might be “by design”, like speedometers in cars usually indicate the speed to high. Or it might derive from inaccuracy in the measurement, precision on flow reporting from the engine. Anyways, a deviation on total fuel consumption below 10% is OK for me.
The second part of purpose was to find out if this can used for practical purposes, like saving fuel. Readouts were logged at different RPMs, after fuel flow and speed stabilized. This produced data for plotting speed and fuel economy versus RPM.
The red graph shows speed in knots, scale on the right. The blue graph is fuel economy in liter per nautical mile, scale on the left. The “step” on the red graph represents the threshold to planning. The boat is fully planning at approximately 20 knots. The really useful information is the optimal fuel economy RPM/speed, the minimum value on the blue graph.. The optimal fuel economy is just below 4.000 RPM, which results in 0.6 liter/nm at 22 knots.
“Taking it easy” at 7-8 knots results in twice the fuel consumption. Even though the fuel economy is displayed in real time on the Garmin, this graph stills adds some understanding, which is hard to get while in action on the water.
Most sailing was at around 4.000 RPM, only interrupted by some harbour manoeuvering and the tests described above. The readout on the Garmin was 0.6 l/nm most of the time. The overall actual trip economy is 44.76 liter / 74 nm = 0.605 liter/nm, which matches the readout.
Comments on method, results and conclusion:
- The comparison of indicated and actual fuel consumption does not uncover possible RPM dependent deviations, as only totals are compared.
- The precision of the gas pump is unknown
- Engine trim angle was in the middle position
- There might be another optimal (economy) speed in the idle range
- The boat was loaded with 2 persons (~ 160 kg)
- The engine/propeller installation is original from yard, 21″ pitch alu propeller
- The hull is not painted below the waterline
- The Garmin GPSMAP 526s firmware version is 4.20
- Engine firmware is 11006D9-03_ENG_D18_P00
- The total engine hours were 40
Less informative pictures from the test: