posted 12-16-2005 07:45 PM ET (US)         

"However, I'm curious to know on what basis the Mercury people think the SmartCraft fuel measurement system is more accurate than any other measurement system and even more curious why there appears to be in this particular case a much greater deviation in this particular test from the lows, highs and averages reported elsewhere than there is in the case of other manufacturers. Is there that much variability in the SmartCraft fuel flow measurement system? "

From what B& WB indicates about Smartcraft, and Seahorse here about E-tec, the fuel measurement systems appear similar, and one would expect, with same degree of accuracy.
Since Yamaha "Command Link" is also similar, I would think all three engines would similar, and comparable, fuel measurement. So I would think B&WB's comparative fuel economy statistics on the three engines would be valid.

Also, I noted in all these shootouts, 200, 225 and 250, the Optimax achieved the best economy, Yamaha in the middle, and E-tec at the bottom.

posted 12-16-2005 09:16 PM ET (US)            

I don't see any basis to make the giant leap of faith that because two systems measure fuel flow using similar indirect methods that they would be equally accurate. They might both be consistent in their measurement, but that could very likely be consistently inaccurate.

posted 12-17-2005 12:26 AM ET (US)            

------------
"The fuel flow system on the E-TEC gauges use the calculated flow from each injector pulse. Yes, the EMM "counts" every injector firing. It also re-calibrates each injector as it wears. It counts in multiples of 250,000 pulses.

I compared the EMM calculated fuel flow with a fuel flow meter and every time it was within 0.1 or 0.2 gallons per hour."
-------------

Seahorse has given us great information on the I-Command system from Bombardier. It sounds like it has the potential to be very accurate. Just to expand on this, the fuel injectors on an E-TEC engine are very precisely calibrated. Each injector has a measured coefficient which describes its characteristic fuel injection rate. This coefficient is stored in the engine management module, and it is used to calculate the precise electrical signal to send to each individual injector in order to control the fuel flow to each cylinder in a very precise manner. Because this fuel flow is related to injector pulses, total fuel flow rate can be deduced from total fuel injector pulse duration or counting.

I don't know where there is a source for definitive information about how the SmartCraft gauge fuel consumption measurements are made. It would be interesting to know how they do it. If (as apparently as been suggested elsewhere) the SmartCraft system simply deduces fuel flow from looking up stored values that correspond to some other parameter (perhaps engine RPM or throttle position sensor output), it may not be as accurate as the injector-pulse counting technique.

Also, an allusion is made above to some information in BASS & WALLEYE BOAT magazine about the SmartCraft fuel measurement techniques:

"From what B & WB indicates about Smartcraft.." (LHG)

In reading the article about the 200-HP outboard tests, I cannot find a single word on this topic. Are details about the measurement technique given elsewhere in that issue? If they explain how SmartCraft measures fuel and certify its accuracy, it would be good to get that information.

There is interest in how fuel consumption measurements are made because the price of fuel is becoming quite a factor in overall boating expense. In a comparative test it is vital that fuel consumption measurements be made in a uniform and accurate manner.

posted 12-17-2005 08:06 AM ET (US)            

Concerning the flow at the injector position: I have often wondered about the need for measuring the real-time flow of the fuel at the injector itself. Where does the EMM or ECM get its information from in order to know how much fuel it has to deliver in the first place? IMO, mainly from a very accurate, highly sensitive AMF meter (air mass flow meter). Together with the throttle position. If it knows exactly how much fuel the injectors have to deliver to the chamber can't it just deliver that same information to the fuel flow instrumentation? Just a thought.

posted 12-17-2005 09:48 AM ET (US)            

What I think Seahorse is saying is that the injector's capacity to inject fuel each time it is cycled changes over time as it wears (these injectors are little fuel pumps) and the E-TEC's EMM takes that into account as part of its injector control function and recalibrates its instructions to compensate for the wear. For example, when the injector is new, let's say that the EMM issues instructions to cycle an injector 100 times to provide X volume of fuel at a given point in time to the cylinder. Some time later, after so many cumulative cycles of the injector, the injector wears a bit and now needs to be cycled 101 times to provide the same X volume of fuel. (On the contrary, perhaps it "breaks in" and only requires 99 cycles to provide X volume of fuel.) Based on knowledge of how the injectors wear and how many cycles they have cumulated, the EMM adjusts or recalibrates to provide instructions to cycle the injector 101 or 99 times as predetermined, for example, instead of 100 times based on the known wear to get back closer to the X volume of fuel it wants to deliver to the cylinder and that it used to be able to deliver by cycling the injector 100 times when it was new.

posted 12-17-2005 09:48 AM ET (US)            

" Where does the EMM or ECM get its information from in order to know how much fuel it has to deliver in the first place?"

There are two ways that the engine can operate: open loop and closed loop. Initially the engine supplies fuel according to a stored map or look-up table. This is the open loop mode. If the engine has been designed with an exhaust sensor, it can measure the exhaust gas mixture and get feedback on the stoichiometry of the fuel burn equation, usually deduced from an oxygen sensor. The engine can use this information to make adjustments to the fuel mix. Closed loop systems have been used in vehicle four-stroke engines since at least 1993. It is not clear to me if any of the current four-stroke outboard motors or the E-TEC use a closed loop system. That is really another topic in itself.

The turbine transducer in a FLO-SCAN or other fuel measurement gauge is a direct measurement of fuel flow. By this I mean that it is the flow of fuel itself which causes the sensor to react. The sensor is directly operated by the flow of fuel.

In contrast I described the EMM based fuel flow systems as indirect, because they impute the amount of fuel flow based on something else happening. Basically they measure how long the valve was open that controls fuel flow (the injector). If the fuel pressure and flow rate is constant and well known, then you can calculate the fuel consumption by adding up all of these times that the valve was open. It is just that this technique is an indirect measurement.

ASIDE: The increase in computing power in an engine or vehicle has allowed for many new features to be added based on analysis of existing data. The fuel flow computation is a good example. Another example in vehicle applications is the deduction of air tire pressure differences from analysis of wheel rotation speeds. There is a federal regulation that requires vehicle manufacturers to be able to monitor tire inflation pressure. The direct method would need special sensors installed in each wheel and coupled via RF to the monitoring system. The indirect method compares the rotation speed of the wheels on a long term basis. It deduces a tire is under inflated by the difference in the rotation speed. This method saves money because separate sensors were not needed. (The wheel rotation speed sensors were already in place for use in the anti-lock brake system.) Another example: a misfire can be deduced by careful analysis of instantaneous crankcase velocity, which is deduced from the crankcase position sensor data.

The question still to be answered: how does a SmartCraft gauge compute fuel consumption? Is there a flow sensor in the engine? Or is the fuel consumption computed by an indirect method.

posted 12-17-2005 09:52 AM ET (US)            

Knowing exactly how much fuel is delivered to each cylinder is the E-TEC specialty and the heart of the system. Where other motors (Opti and HPDI) can only "guess" at the fuel delivered per the time the injector is open and depending on fuel rail pressure, the E-TEC injector and EMM are
"matched" with the calibration number assigned to each E-TEC injector. That is why each injector is "assigned" to a particular cylinder and can't be interchanged.

The EMM even compensates for wear during the millions of pulses each injector endures. The EMM knows precisely how much fuel is injected for each cylinder and the calculations on flow are very accurate.

posted 12-17-2005 01:04 PM ET (US)         

I have no idea whether any of this is more accurate than the Navman paddle wheel method, but I am now convinced that the Mercury Smartcraft system, being the innovator in this field and now obsolete compared to the newer "Command" technology used by all other brands, is being calibrated so that Mercury engines consistently record less fuel than actual consumption, allowing Optimax to win the "shootout" fuel economy categories. I am also convinced that the Evinrude system as defined by Seahorse is the least accurate measuring system, which is why the Evinrude, which is actually getting superior fuel economy to Mercury and Yamaha, recorded the highest fuel use.

posted 12-17-2005 01:44 PM ET (US)            

That's a bit like saying that your Cadillac and a Porsche that both use fuel injected reciprocating piston engines driving rear wheels with a transmisson are similar and therefore it should be safe to assume that they would get similar performance. LOL.

posted 12-17-2005 02:50 PM ET (US)            

If you have any evidence that computation of fuel consumption using the injector pulse counting and accrual method is something that was an invention of Mercury, could you please follow up with the patent number for us to read.

NMEA-2000 has been in development since 1994. Let me say that again. NMEA-2000 has been in development since 1994. Much of the underlying technology in NMEA-200) is based on CAN-BUS, and this is a much older standard which is very mature and widely used in real-time automotive data systems. It was really a very intelligent decision to piggy-back on CAN-BUS because there are already hundreds of devices available at very low cost to implement CAN-BUS. There is no way to know much about what SmartCraft is or how it works, because it is a closed system. I would bet they also used CAN-BUS technology.

As for the shape of the gauges, rectangular digital gauges have been very common in marine instrumentation for decades. We used to have some on our sailboat back in the 1980's. I don't think there is anyone who really thinks that Mercury invented the rectangular digital marine instrument in 2002.

We have quite good information provided about how the E-TEC engine computes fuel consumption, but so far no information at all about how the SmartCraft system works. It is hard for me to make the inference about one being a copy of the other when we don't really have any information about the SmartCraft system.

Once we finally discover how SmartCraft computes fuel flow, we can compare it to other techniques that are known.

To relate back to the initial discussion about fuel measurement techniques used by the boating magazine, the bottom line question is whether all the measurements were made using the same instruments. If a magazine presented speed results and used three different systems of measurement, there would be the same interest. For example, if one boat's speed was measured by a stop-watch, one by GPS, and one by a paddlewheel hooked to its own gauge system, I think people would be asking the same type of questions.

The real concern is did the fuel consumption data all come from the same measurement technique? I have not seen the details of the test technique used. I am curious if it was the same for all motors.

Whether SmartCraft fuel consumption is accurate is a separate question.

There is no doubt in my mind, or probably in anyone else's, that if the magazine used the I-Command data to measure the fuel consumption of the E-TEC and declared it the winner, there would be a similar inquiry.

I also think that there will be little reason for Mercury to modify their SmartCraft instruments to become compliant with NMEA-2000. It goes against their way of thinking. Mercury, through Brunswick, likes to think in vertical integration. They like to own everything in the boating business, from the boat builder, the engine maker, the dealer, the finance company, the bilge pump maker, and even perhaps the marina and slip where you keep the boat. They don't want to let anyone else in. So they probably will never make their system open. They probably looked at NMEA-2000 and intentionally decided not to participate in it so as to exclude other vendors from being able to compete with them for connecting to Mercury engines.

On the other hand, the following companies are producing and shipping NMEA-200 compliant devices:

ENGINE MAKERS:
--Honda
--Yamaha
--Bombardier
--Volvo-Penta

posted 12-17-2005 03:20 PM ET (US)         

All Mercury's Smartcraft instruments are made by VDO, a highly respected instrument maker.

posted 12-17-2005 04:13 PM ET (US)            

http://www.boatingindustry.ca/BIC0604/0604Honda.htm

I assume the system used by Honda is similar to the ones used by Yamaha and others outboard makers. Like jimh said earlier, a test using the traditional paddlewheel type with the [direct] method is probably needed to make any accuracy claims valid.

posted 12-17-2005 10:33 PM ET (US)            

I seem to recall said technician had very little difficulty dialing in his Yamaha fuel management gauge (1997 2-Stroke engine), or his Floscan. With the floscan, he did have to re-orient the flow meter to face UP since trapping bubbles in the turbine path tends to mess up your readings. (the previous owner or installer apparently disregarded the note about pointing the flow sensor vertically!)

posted 12-19-2005 10:59 PM ET (US)            

But the service manual does explain how the EFI system delivers fuel through the injectors. The ECM (engine control module) monitors the air temperature, manifold absolute pressure (MAP sensor), engine head temperature, and throttle position (TPS- throttle positioning sensor).

In the MAP sensor description it says "The MAP sensor is functioning through the full RPM range and is continually signaling induction manifold readings to the ECM. The ECM in turn determines fuel flow as signals are received."

I would think it is not a stretch to assume if the ECM makes the decision on how much fuel is to be delivered by the injectors, it can easily calculate the sum total of fuel sent, convert that to gallons per hour and send an instantaneous fuel consumption figure to a simple display.

As far as accuracy, I have not tested it against another metering device. I don't think you could really test instananeous consumption accurately by known gallons of fuel burned- you would have to run the boat at an absolute fixed speed uninterupted over a period of time, then accurately determine the exact amount of fuel consumed to be precise. You could certainly calculate an average using that method, but I think all test results we are talking about display instantaneous burn at a snapshot moment-rpm and speed at that instant.

A well calibrated paddle wheel system, which produces accurate and consistant readings I think is easier for us to accept as accurate because we actually saw the fuel level drop a certain amount after running for a while during the calibration process. Fuel level actually droped "x" amount, and the meter said it consumed that amount. A little wheel spinning is easier for us to accept than a computer matrix calculating very minute amounts of fuel dispersed from multiple injectors. It appears that you calibrate the paddle wheel system by getting the system to accurately display a fixed amount of fuel gallons consumed over a period of time. If it will accurately display fuel consumption over a period of time, it is safe to assume that the instantaneous reading is also accurate. That is a bit of a leap of faith too I think.

My vehicle displays miles per gallon. It is wrong. It is not instantaneous, but it does seem to give a progressive miles per gallon over a tank full or even longer period. If you divide the displayed miles per gallon by 1.12 after it calculates the figure over a tank of fuel, the result is consistant with the tried and true fill the tank, reset the trip odometer to zero, run 300 miles, fill the tank, calculate miles per gallon. 320 miles, 20 gallons used, week after week tells me I get 16 miles per gallon. Computer says 18 miles per gallon. This is a recent Toyota 4Runner, design is by a Lexus Engineer!

I think it is logical to build some error into all methods used to calculate fuel consumption, be it instantaneous or averaged. Five to ten percent, in my opinion would not be unreasonable when using the data. Any study of data always has some degree of variation built in to take into consideration variations of all sorts in test conditions. Especially when testing boats, there is just too much variation built in from weather, sea conditions, salinity, horsepower variations of motors, hull differences.

Rich

posted 12-20-2005 11:03 PM ET (US)            

I would bet that any of the digitally calculated flowrate measurements would be inherently more accurate than any of the paddlewheel types,and any calibration you do is only going to be as good as your measuring equipment.

The only accurate way is to have a graduated volume of fuel that you can switch into and out of the fuel system at various speeds and accurately measure distance travelled.
The speed side is relatively easy with today's GPS units but
you'll need a pretty specialized fuel system installation to get better accuracy results than any given manufacturers engineered system.

posted 12-21-2005 01:00 AM ET (US)            

Perry--that is a very good point you made. The accuracy of the speed data of course would directly affect the accuracy of the computed miles-per-gallon data!

I would assume that there must be some setting in the SmartCraft instruments which allows one to pick the source of the speed input. A GPS would probably be more accurate than a paddlewheel. In my NAVMAN 3100 Fuel instrument, you can select the data source for speed information to be either a paddlewheel or GPS.

Even among GPS data, however, there may be some difference in accuracy depending on the Horizontal Dilution of Position (HDOP) of the GPS measurement. Units with Differential GPS or WAAS capability would be more accurate. And I would not be surprised to learn that there might be some variation among units in the accuracy at which they deduced the speed from comparison of successive position plots.

One minor point: GPS speed data includes the effect of current, as it measures speed over bottom. A paddlewheel will measure speed through the water. When operating in water with a current, there will be a big difference.

Bthom--good information on the comparative accuracy of the displacement pump versus the turbine blade type of measurement. I would tend to accept that as true. I don't think the issue has ever been that the injector pulse counting and accumulation method of measuring the fuel is not accurate.

The fundamental issue that really began this discussion was the lack of information on what sort of technique was being used in the SmartCraft system on Mercury motors. I think we are still cloudy on that.

posted 12-21-2005 02:39 AM ET (US)            

First, we don't care about GPS speed in calculating MPG,
just distance, measured as successive positions. And unless
we are doing some really convoluted course, the total
distance will be VERY precise. Assume GPS position is
accurate to 45. The distance will be accurate to a MAX of
90' over many miles. Any errors cancel each other out,
and don't add up. If one segment is short, the next will be
long, zero net error.

Second, even if we did care about speed, the same thing would
be true: any errors would cancel each other out. If one
segment is fast, the next will be slow, zero net error.

Finally: My '87 Vette has an MPG meter. I don't know whether
it has a sensor in the fuel line or counts injector pulses,
but it's spot on vs. odometer, gas pump, and pocket calculator.

posted 12-21-2005 07:49 AM ET (US)            

Old school method--Started with a full tank, and filled up at the conclusion of the trip. Checked my GPS trip log (all the trip was under power, either cruise or trolling-no drifting)and divided by gallons burned==2.7 nmg.

Smartcraft==2.6 nmg.

Isn't there some velocity made good reading on the GPS, showing your actual speed over the surface? That would be a better gauge of miles. Maybe I should just pick a windless day and make a measured run on flat water at slack tide to better compare?

posted 12-21-2005 08:02 AM ET (US)            

Another item that should be published is the ICOMIA fuel milage in average gallons per hour. A motor is run at the same duty cycle as the emission tests and the result would be the fuel comsumption for an average boater. I don't have all the "test points" handy at the moment, but it is 6% at full throttle, 14% at 80% throttle, on down to 40% at idle. This is the same standard that all outboard manufacturers use to compute fuel usage.

The auto makers use a standard procedure also to compute the fuel milage that you see on the window sticker of a new car. Every car company uses that standard.

The only problem with outboards is that many "flo-scan" type units cannot accurately read flows less than 1 gallon per hour. Some E-TECs idle at 0.2 gph and many 4 strokes at 0,5 gph. You would need almost lab grade instruments to be accurate.

posted 12-21-2005 08:41 AM ET (US)            

With respect to the low flow rate accuracy of paddle wheel sensors, my Fichts idle around 0.3 GPH as I understand from Evinrude test reports and the Navman 3100 I have shows 0.3 GPH at idle for each motor.

posted 12-21-2005 09:45 AM ET (US)            

As far as know, a GPS does not transmit "distance" data. It can transmit position and speed data. However, if the GPS just supplied the fuel computer with the position, the fuel computer would have to retain the position data in memory and calculate the distance. I think it works on a much simpler basis. It gets the current speed from the GPS (miles-per-hour), and it also gets the current fuel consumption rate (gallons-per-hour) from its own measurement of flow. From these two it computers the miles-per-gallon.

Your hypothesis would require that the fuel computer perform the flowing:

--get a position fix (1) from GPS and remember it;
--note the time of this fix
--start measuring the amount of fuel consumed
--get a new position fix (2) from the GPS and remember it
--stop measuring the amount of fuel consumed
--calculate the distance between positions 1 and 2 (which would require the fuel computer to have a great deal of knowledge of spherical trigonometry it otherwise would never need to know)
--divide this distance by the amount of fuel used in the same interval

I really do not think it works like that. I think it works like this:

--measure fuel flow over predefined intervals and extrapolate this to a flow rate in gallons per hour (A)
--accept data from the GPS about vessel speed in miles per hour (B)
--divide A into B to compute miles-per-gallon

posted 12-21-2005 10:20 AM ET (US)            

And WAAS GPS (essentially all of them these days) are reputed
to be accurate to .1 MPH. I'd believe that. I've run a
a measured mile on cruise control in my truck, and a stopwatch
and the GPS agreed.

posted 12-21-2005 08:40 PM ET (US)            

Now, I have not read this entire thread - but Jim's concept of the mpg calculation is right on - with the possible exception that the conversion is probably based on a shorter period of time - seconds or minutes rather an hour. All GPS systems receive a very!! accurate (probably micro or nano seconds) time from the satellites and it is this time that is used in the speed and triangulation position calculations. Speed calculations/display could be a short term moving average - or something else - I just don't know.

Another comment; I have noticed where some have mentioned about being provided little or lack of information regarding how a device works. That is one of my major pet-peeves and will turn me off faster than anything else - someone not telling me how his/their device works. And I will accept "...I don't know, but I will find out and let you know..." I have been known to refuse to buy sophisticated instrumentation systems from a vendor who refused to tell me how their systems perform certain operations. And saying that the entire system is proprietary doesn't cut it. ---- Jerry/Idaho

posted 12-22-2005 10:21 PM ET (US)            

Are you saying that when I plug my GPS into my Navman 3100, it doesn't use the speed input to calculate MPG?