Scangauge fuel usage is right on, but MPG is way off?

[jimnms]

I've had my SG2 since 2008, and it has only proved to be an accurate fuel gauge. It's usually within a few tenths of a gallon every time I fill up, but the MPG it reports is never even close to what I actually get by calculating actual miles driven divided by gallons pumped.

I just got a new GPS last month (Garmin 2360LMT) which has mileage computer in it. You can buy a cable they call ecoRoute HD and plug it into the OBDII port so it can do what the SG2 does, but even without the cable it does a decent job of estimating fuel usage when planning a trip. I don't have the cable, so it's only giving estimates based on the city/highway averages I've told it I get.

I had to drive to Birmingham and back yesterday, so I used that as an opportunity to compare them. Before leaving, I filled up, reset the trip computer, the SG2 and the GPS log. I drove there and back, returning to the same station and the same pump. The car's ODO and trip show that I drove 575.0 miles, the GPS was showing 575.4 miles. The SG2 shows that I had driven 581 miles. The GPS estimated I did 41.9 using 13.9 gallons. The SG2 says my tank average was 45.4 MPG, and the fuel was only off by 0.2 gallons.

So how do they compare to what I really did? It took 13.473 gallons to fill up, and having driven 575 miles that is an average of 42.7 MPG. The GPS was only off by less than 1 MPG, and the SG2 was off by almost 3 MPG. Once again though, the SG2 was almost right on the mark with the fuel consumption, but way off on the MPG calculation.

I was surprised at how accurate the GPS was considering it's basically just guessing since it's not reading fuel flow from the car's computer. On this trip the GPS came out more accurate than the Scangauge. Is there some way of tweaking the SG2 to get the MPG calculation more accurate without messing up it's fuel accuracy?

 

[Jack Frost]

The car's ODO and trip show that I drove 575.0 miles, the GPS was showing 575.4 miles. The SG2 shows that I had driven 581 miles. The GPS estimated I did 41.9

Your SG mileage needs to be recalibrated. There is a procedure to adjust that using percentages (I can't remember the exact buttons). But the percentage need to be reduced by (581-575)/581*100 = 1%.

This is not very inaccurate but that is how you adjust the mileage of the SG.

You will never get perfect agreement between three digital devices estimating an analog value. Choose one and ignore the others. I just use the mileage from my odometer.

 

[UFO]

I agree, adjust the speed in the SG to get the SG to accurately measure distance. It doesn't matter if the SG speed matches the speedometer, only that the SG matches the odometer. The only way to do that is with the SG speed calibration.

 

[Ski in NC]

MPG has only two variables: Miles and gallons. If your gallons are spot on, then the miles must be off.

 

[jimnms]

Your SG mileage needs to be recalibrated. There is a procedure to adjust that using percentages (I can't remember the exact buttons). But the percentage need to be reduced by (581-575)/581*100 = 1%.

This is not very inaccurate but that is how you adjust the mileage of the SG.

You will never get perfect agreement between three digital devices estimating an analog value. Choose one and ignore the others. I just use the mileage from my odometer.

There is a speed calibration (More > Setup > Speed), but not a distance calibration. I calibrated it according to the recommendation in the manual. I set the GPS and SG to display KM/h. At 100 KM/h on the GPS, the SG also reports 100 KM/h. I checked it, again after getting new tires and once more after the glow plug recall and ECU flash. Even though the dash spedo is off by about 2.5 MPH, the speed reported by the ECU is correct and did not require me to adjust it.

So you suggest I forget about setting it to an accurate speed and adjust it to get the mileage correct?

MPG has only two variables: Miles and gallons. If your gallons are spot on, then the miles must be off.

The thing is, even if I take the mileage the SG reported and divide by the gallons the SG thought I used, it's still way off from what it said my average MPG was, and even further off of my actual calculated MPG.

 

[Jack Frost]

There is a speed calibration (More > Setup > Speed), but not a distance calibration. ....

So you suggest I forget about setting it to an accurate speed and adjust it to get the mileage correct?

Speed is distance divided by time. Since time is a constant, if you adjust speed as the manual suggests, then you also adjust distance the same amount. The manual only hints at this.

However, your speed and/or mileage is only off by 1%, so this is not the source of your errors. I suspect that you are dealing with metric/imperial errors. My Scanguage incorrectly calculates fuel cost unless I set it up in liters, but when I do, it incorrectly calculates mileage. I emailed the company about this, but they never replied.

I will have to take a closer look at your first post tomorrow.

 

[jimnms]

Speed is distance divided by time.

I know that, and that makes me think the SG's clock is wrong. The SG reads the speed and fuel flow from the ECU. The speed and fuel flow have both been calibrated and are correct, so it should be able to accurately determine the MPG with those three variables (speed * time = distance, distance / fuel used = MPG). If speed and fuel flow are correct, the only thing left is time. Where does the SG get it's time from?

 

[Jack Frost]

Hi Again,

I checked your calculations and although you don't quote what the SG reported in fuel consumption, if it was different from the fuel pump by 0.2 gallon. Since you haven't indicated which way the difference is, it must have been either 13.273 or 13.673 gallons. Is that right?

If so, dividing your SG-reported distance (581 miles) by the SG-reported fuel consumed, it should have either reported for mileage either 43.8 mpg or 42.5 mpg. Your hand calculated mileage was 42.7 so it doesn't seem to me that your SG' reported mileage is off by 3 mpg. Perhaps you should check your figures again. It look all right to me.

BTW, how SG measures time had nothing to do with it accuracy in reporting speed or mileage. The reason why is mathematically obscure but trust me. But here is a hint.

If a Canadian car and an American car are travelling down the same road side by side at the same speed, does it matter if the Canadian's driver's odometer is calibrated in kph and the American in mph. Of course not, they are going at the same speed. If they are driving the same car in the same condition and started with a full tank of fuel, they will be both out of fuel at the same time in the same place. It is kinda like that. Speed and fuel efficiency can be expressed in different units of time and distance but that doesn't mean they can't be the same. For the same reason, it does not matter how SG measures time.

But if you really want to know, SG marks by how long it take to charge a transistor inside one of it chips to about 4 volts - usually a few billionths of a second. Like the system clock on your computer. The operating system will then translate that time interval to something more useful. It could be something like seconds but probably something more arbitrary. A conversion factor would be used to change that time interval into hours for display. Kinda like gears in a clock.

 

[jimnms]

I checked your calculations and although you don't quote what the SG reported in fuel consumption, if it was different from the fuel pump by 0.2 gallon. Since you haven't indicated which way the difference is, it must have been either 13.273 or 13.673 gallons. Is that right?

Sorry, I was trying to decrypt my writing on the back of the fuel receipt while using my leg as a desk. The SG thought I used less fuel and drove farther than I actually did. The SG only displays gallons down to tenths. It was actually reporting 13.2 gallons used, so if I round up the fuel on the receipt from 13.473 to 13.5, it would be 0.3 gallon off.

If so, dividing your SG-reported distance (581 miles) by the SG-reported fuel consumed, it should have either reported for mileage either 43.8 mpg or 42.5 mpg. Your hand calculated mileage was 42.7 so it doesn't seem to me that your SG' reported mileage is off by 3 mpg. Perhaps you should check your figures again. It look all right to me.

Yes, if the SG divided its reported distance by its reported fuel usage, it would be closer to my hand calculated mileage, but the SG must not be doing that to calculate MPG because it reported a tank average of 45.4 MPG, which is off by 3 MPG (2.7).

 

[RZedBoy]

Seems like a dumb question but how does the scan gauge measure the fuel in the tank? Based on the amount you put in, you probably did the tank vent modification. If so does the fuel level sensor get stuck at full for a long period of time? How does scangauge actually measure fuel used? Does it go by how long injectors are open or does it use the fuel level sender info? I am very interested in getting a scan gauge to help me get better milage but without my undertanding of how that above is handled by the gauge I am hesitating.

Seems it is just easier to use a notebook and get the real numbers but then I won't have real time numbers........

 

[Jack Frost]

The SG does not measure how much fuel there is in the tank. It calculates how much should be remaining.

There are two ways it can do this. Initially before your first fill up, you enter the size of the fuel tank, then you tell it that you have filled it up. After that SG subtracts how much you used in real time. The cumulative amount of fuel that has been used is then subtracted from the tank size to estimate the fuel remaining in the tank.

After that, every time you fill up the tank, you enter in SG the amount of fuel that you have added and the process starts again. SG subtracts how much you used in real time. The cumulative amount of fuel that has been used is then subtracted from the tank size to estimate the fuel remaining in the tank.

There is no data used from the fuel level sensor or the injectors. Somewhere on the fuel line, there must be a little flow sensor that sends digital data to the engine's computer. Every time the shaft on the sensor rotates a full turn, it then sends a voltage "blip" to the engine's computer. Every "blip" might represent a consistent but arbitrary amount of fuel. I don't know how much, but the smaller the amount, the more "blips" the engine's computer gets. The engine's computer reports the number of blips as they occur to the SG. The SG takes these blips and multiplies them by some number to get the liters or gallons of fuel consumed. Every car is different.

Part of getting a SG accurate is teaching it how to convert "blips" to gallons or liters. That requires that you enter in it accurately how much fuel you have pumped into the tank upon every fill up.

 

[Ski in NC]

Engine ECU's have data available to compute fuel burn rate. ALH use QA position and rpm, PD's use injector solenoid on-time and rpm, cr's use injector open time, rail pressure and rpm. These data are used by the ecu and scangauge and vcds to estimate burn rate. No flowmeter in system. Accuracy varies, but with scangauge it can be tuned over repeated tank fills. Tank level sensor not in system.

 

[RZedBoy]

I like your explanation. While 'blips' is not incredibly technical it replaces alot of blah blah and means the same. Nice.....

I am glad to hear that it is not relying on the fuel level sender. Now, if I do the vent mod, I really do not know how large my tank it so it is just based off me telling it how much I put in and eventually it will converge on the true tank size as I run it from completely full to empty over a few tanks?

Also, does the SG use Canadian/IMP gallons or US gallons? I would really hope that it uses the US ones as I fill in the US a fair amount.

 

[Jack Frost]

The SG only displays gallons down to tenths. It was actually reporting 13.2 gallons used, so if I round up the fuel on the receipt from 13.473 to 13.5, it would be 0.3 gallon off.

Yes, if the SG divided its reported distance by its reported fuel usage, it would be closer to my hand calculated mileage, but the SG must not be doing that to calculate MPG because it reported a tank average of 45.4 MPG, which is off by 3 MPG (2.7).

Ok, the SG reported 581 miles and during those miles, SG reported that you consumed 13.2 gallons of fuel. SG should have displayed 44.0 mpg, but you are telling me it displayed was 45.4 mpg.

Does not compute!

Even if you entered your fuel pump figures wrong and even if you haven't got the distance calibrated properly, SG should be displaying a mathematically accurate mileage.

Perhaps you were looking at one of the figures for current, today's, or previous days' trip. Those figures would be close but not exact. Why don't you check all your figures again at the next fill up. I am going to do the same for my SG at the next fill up and report back to this thread. To keep things simple, leave your GPS at home.

 

[Jack Frost]

Now, if I do the vent mod, I really do not know how large my tank it so it is just based off me telling it how much I put in and eventually it will converge on the true tank size as I run it from completely full to empty over a few tanks?

Your tank capacity is what you tell SG it is. It is one of those setup parameters that have to be entered before your first fill up and it remains constant. The tank size for most of us is what the car's manufacturer says it is. But if you have modified it, you might want to enter your estimate of the tank size is. However, even if you are wrong, this will not affect the mileage of fuel consumption figures. It will only affect the SG estimate of your car's remaining range and total fuel remaining.

Your mileage figures may become more accurate with successive fill up providing that your reported fuel consumption to SG is accurate, but I don't know for sure. It could be that your mileage figures are only as good as your fuel consumed figures entered after your last fill up. It is possible this may be a source of errors the OP has. We can only tell by tracking all the relevant figures properly fill up after fillup and reverse engineering mathematically what SG is thinking.

Also, does the SG use Canadian/IMP gallons or US gallons? I would really hope that it uses the US ones as I fill in the US a fair amount.

The SG's gallons are whatever gallon the fuel vendor sells it in. Since the only gallons that fuel is sold in is American gallons that would be the only gallon that anyone would report.

I find my SG does not convert gallons to liters. If you enter tank capacity in liters and report fuel consumption in liters, then change the setup to display fuel in gallons, the fuel mileage figures are way way way off trying to display mileage in mpg. I haven't got it figured out yet what SG is thinking, but my suspicion right now is that they don't convert between units at all.

All I know is that they don't answer my emails.

 

[jimnms]

Ok, the SG reported 581 miles and during those miles, SG reported that you consumed 13.2 gallons of fuel. SG should have displayed 44.0 mpg, but you are telling me it displayed was 45.4 mpg.

That is correct. After I fill up, I write on the receipt the odometer reading, the trip odometer, then flip to the tank page on the SG and write down the SG's mileage. I just dug out 7 more receipts on my desk, and the SG's reported mileage is always 2-3 MPG high. I don't write down the SG's trip distance, but I've always noticed that it records more miles than I actually drive. I only wrote it down this time to document it.

When I fill up again, I will take pictures of the SG. I did dial the speed down one notch to see if it corrects distance. I may have to make that trip again next week, so I will post back next week if I do, otherwise it will be 3 weeks before I have to fill up again.

 

[Jack Frost]

Sounds like a plan. I will be filling up this Friday and will do the same. I only got my SG recently so I haven't figured it out completely yet. The last time I filled up, it significantly underestimated my fuel consumption. As a result the mileage it displays is very low.

We should write down for every ScanGaugeII fillup:

1. vendor's fuel pump volume
2. vendor's fuel price
3. Car's trip odometer (assuming it is reset at every fill up; mathematically this is irrelevant at this stage, but it makes a good common standard reference)
4. Hand calculated MPG
5. SG TANK GAL (for SG estimated fuel consumed)
6. SG TANK MILES (for estimated SG distance)
7. SG TANK MPG (for SG calculated average mileage)

(if I think of anything else, I will edit this post)

Don't get TRIP figures mixed up with TANK. They tend to look similar.

It may take a few fill-ups to answer the questions we have.

 

[LethalMindNinja]

I'm curious what you all set your tank size to on your SG2. I have a 2005 jetta tdi with a BEW 1.9L PD engine and have heard that if it's run to empty people are filling it up with around 17 gallons, but manufacturer lists the tank as around 14.5 gallons (i think). Would you all recommend the manufacture tank size or reported tank size?

 

[Jack Frost]

I would say that if you aren't sure, enter the tank size spec the car manufacturer provides. You can't rely on stories from others. On the other hand, if you have run your tank dry and fill it up the same way each time, then enter the fuel volume that you used to fill it.

Tank size only affect the SG's estimates of the remaining fuel, mileage, and driving time. It does not affect mileage at all. If you enter a figure that is larger than the actual tank size and you rely on those figures, you could end up stranded with no fuel in your tank. That's all. But that could be serious. It is a personal safety issue and it probably doesn't do your motor any good. So always error on a lower tank size.