So when I first got my cruze, the average mpg and fuel used were about 10% off, consistently. I haven't been tracking it for a while, mostly out of laziness on my part. Anyway, we just drove to Niagara Falls (We drove seperately, because I had to work). It was raining a good portion of the way, so I had the defogger on (running the A/C compressor), and was also going 75mph for the first 4 1/2 hours of it. Beyond that I had to go much slower, because of snow on the thruway. Anayway, I travelled 297.5 miles on 9.22 miles of gas, which works out to a little over 32.2MPG. Not bad, I think considering the conditions. But the most surprising thing, was that for the first time ever, the DIC was spot on. It said 9.2 gallons burned, and average fuel economy 32.3mpg. I wonder if they become more accurate over time, because my mom's equinox did the same thing, (she average 24mpg at 75).
DIC More accurate over time
- Thread starter Nick R
- Start date
- Status
I think it's more of a function of comparing the inferred estimate (DIC) with a manual calculation, which has a lot of variables. The DIC calculation is simply a math equation, without any basis for error correcting. Ie, it doesn't know if it's correct or not. For example, in the ScanGauge device, you can set a multiplier to help calibrate the system. In the DIC, the multiplier (if there is one) is hard-coded into the software by GM. In other words, I doubt it fundamentally gets more accurate over time...it's probably a coincidence of some sort.
Right - the DIC has no feedback loop - it's a dumb system. It has no way of knowing whether you used 4.5 gallons or 45 gallons if it were to be off by that factor.
I do notice better accuracy when the tank is used "all at once" however. I think the computer struggles when it deals with a lot of open loop, start-stop driving over the course of a tank.
I do notice better accuracy when the tank is used "all at once" however. I think the computer struggles when it deals with a lot of open loop, start-stop driving over the course of a tank.
I think the least accurate part of the computer's mileage estimate is the distance travelled. That can be affected by everything from tread wear to tire inflation, to amount of continuous wheel slippage driving into a headwind vs. downwind (small, but present).
In contrast, the computer should, even in open loop mode, know rather precisely how much fuel is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential.
In contrast, the computer should, even in open loop mode, know rather precisely how much fuel is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential.
Except that whether hand-calculated or using the DIC, the trip distance used for calculation is the same.
Originally Posted By: 440Magnum
In contrast, the computer should, even in open loop mode, know rather precisely how much fuel is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential.
The computer does know, precisely, how much fuel it THINKS is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential. Variables here include injectors that may stick or leak, or pressure changes in the fuel system due to a lazy fuel pressure regulator, etc.
I agree that in theory, the computer's calculated number should be the most accurate possible number. But that assumes that every single mechanical item in the line operates at perfection all of the time, and that there is no allowable tolerance from one vehicle to another. That won't be achieved in the field.
Interesting related tidbit: our Acura MDX has a trip odometer as part of the gauge cluster, and also a trip odometer in the navigation system as part of the fuel economy data screen. The trip odometer in the gauge cluster gets its data from the vehicle's speed sensors. The trip odometer in the navigation system gets its data from the GPS receiver. In theory, the navigation system's trip odometer should be the most accurate, and there is a consistent delta between the navigation system's trip odometer and the gauge cluster's trip odometer, but it's EXTREMELY slight. The gauge cluster's trip odometer appears to read ever so slightly fast, but it's a percent difference. After just a few miles, they are exactly synchronized. But after 80 or 100 miles, you can watch the gauge cluster change 0.1 mile and quickly look over at the navigation screen and watch it change 0.1 mile. The most I've ever seen them different is exactly 0.1 mile, so that the trip odometer on the gauge cluster might read 246.7 miles and it'll read 246.6 miles on the navigation screen.
In contrast, the computer should, even in open loop mode, know rather precisely how much fuel is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential.
The computer does know, precisely, how much fuel it THINKS is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential. Variables here include injectors that may stick or leak, or pressure changes in the fuel system due to a lazy fuel pressure regulator, etc.
I agree that in theory, the computer's calculated number should be the most accurate possible number. But that assumes that every single mechanical item in the line operates at perfection all of the time, and that there is no allowable tolerance from one vehicle to another. That won't be achieved in the field.
Interesting related tidbit: our Acura MDX has a trip odometer as part of the gauge cluster, and also a trip odometer in the navigation system as part of the fuel economy data screen. The trip odometer in the gauge cluster gets its data from the vehicle's speed sensors. The trip odometer in the navigation system gets its data from the GPS receiver. In theory, the navigation system's trip odometer should be the most accurate, and there is a consistent delta between the navigation system's trip odometer and the gauge cluster's trip odometer, but it's EXTREMELY slight. The gauge cluster's trip odometer appears to read ever so slightly fast, but it's a percent difference. After just a few miles, they are exactly synchronized. But after 80 or 100 miles, you can watch the gauge cluster change 0.1 mile and quickly look over at the navigation screen and watch it change 0.1 mile. The most I've ever seen them different is exactly 0.1 mile, so that the trip odometer on the gauge cluster might read 246.7 miles and it'll read 246.6 miles on the navigation screen.
Maybe you filled up with the first accurate gas station pump
? On a serious note, I assume your DIC was reporting higher milage than calculated from fillups and you used different stations and still get consistent mpg disparity between DIC and from fillups?
Originally Posted By: 440Magnum
I think the least accurate part of the computer's mileage estimate is the distance travelled. That can be affected by everything from tread wear to tire inflation, to amount of continuous wheel slippage driving into a headwind vs. downwind (small, but present).
In contrast, the computer should, even in open loop mode, know rather precisely how much fuel is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential.
I agree. I don't think the injector pulse width is far from actual fuel flow. If PCM caculations were off that significantly it'd be exceeding the fuel trim allowed and setting codes. I think any significant optimism or dispcrepency of the DIC is intentional. With calcualting by odometer/fill ups you'd expect to see optimism there because usually speedometer/odometers are calibrated to read faster/further than actual.
I think the least accurate part of the computer's mileage estimate is the distance travelled. That can be affected by everything from tread wear to tire inflation, to amount of continuous wheel slippage driving into a headwind vs. downwind (small, but present).
In contrast, the computer should, even in open loop mode, know rather precisely how much fuel is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential.
I agree. I don't think the injector pulse width is far from actual fuel flow. If PCM caculations were off that significantly it'd be exceeding the fuel trim allowed and setting codes. I think any significant optimism or dispcrepency of the DIC is intentional. With calcualting by odometer/fill ups you'd expect to see optimism there because usually speedometer/odometers are calibrated to read faster/further than actual.
As usual everyone is ignoring the HUGE variations from manufacturer to manufacturer.
Just like 'dumb' gauges that are given a reading to display, DIC/EVIC's are wildly different from maker to maker.
Just like 'dumb' gauges that are given a reading to display, DIC/EVIC's are wildly different from maker to maker.
My Cruze's DIC has gotten pretty good, too. The most I've seen it off in the past month has been 0.5 gallons. Considering I fill at different gas stations all the time due to work, it's not bad. My ScanGauge is within a few tenths, but it's calibrated.
They can be right or wrong, from day to day.
Handy, but not to be trusted like a real measurement.
Handy, but not to be trusted like a real measurement.
Originally Posted By: SteveSRT8
As usual everyone is ignoring the HUGE variations from manufacturer to manufacturer.
Just like 'dumb' gauges that are given a reading to display, DIC/EVIC's are wildly different from maker to maker.
I've never had one that didn't match hand calculation quite well. But then I've only had them from one manufacturer... ;-)
As usual everyone is ignoring the HUGE variations from manufacturer to manufacturer.
Just like 'dumb' gauges that are given a reading to display, DIC/EVIC's are wildly different from maker to maker.
I've never had one that didn't match hand calculation quite well. But then I've only had them from one manufacturer... ;-)
Originally Posted By: Hokiefyd
Originally Posted By: 440Magnum
In contrast, the computer should, even in open loop mode, know rather precisely how much fuel is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential.
The computer does know, precisely, how much fuel it THINKS is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential. Variables here include injectors that may stick or leak, or pressure changes in the fuel system due to a lazy fuel pressure regulator, etc.
Oh, I know a malfunction can change fuel volume estimation greatly and manufacturing tolerances can skew it to a lesser degree. But given the types of systems, parts, and measurements involved I still think distance-over-ground is the single most error-prone part of the process.
Originally Posted By: 440Magnum
In contrast, the computer should, even in open loop mode, know rather precisely how much fuel is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential.
The computer does know, precisely, how much fuel it THINKS is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential. Variables here include injectors that may stick or leak, or pressure changes in the fuel system due to a lazy fuel pressure regulator, etc.
Oh, I know a malfunction can change fuel volume estimation greatly and manufacturing tolerances can skew it to a lesser degree. But given the types of systems, parts, and measurements involved I still think distance-over-ground is the single most error-prone part of the process.
Patman
Staff member
I think that unless you are manually calculating the MPG over a large number of fillups, it's almost impossible to get an accurate reading of MPG just by filling up the tank and checking how much gas you pumped in. With gas stations correcting the volume of gas based on temperature, along with the fact that the nozzle might stop flowing at a couple of tenths of a gallon different from one fillup to the next, shows how inaccurate it can be. For instance, let's say you went 300 miles and pumped in 10 gallons, but the pump stopped even a tenth of a gallon sooner or later. That tiny tenth of a gallon still results in a MPG variance of 0.3 MPG! So it's hard to truly know how far off your DIC is unless you are not resetting it for a very long time and are averaging out your calculated MPG for many many tankfuls.
This is the problem I'm running into right now while trying to calibrate my scangauge with the numbers I'm seeing manually. I even go to the exact same station and pump number every time, but I know for sure that when I'm filling up the car the pump stops flowing at a slightly different point each time. So I fill up the car a number of times before resetting and readjusting the numbers on my Scangauge. I'm almost there now, I'm about 0.3 MPG off. But I want the numbers to be DEAD ON, so even after having the Scangauge since July, I'm still fiddling with the calibration to this day.
This is the problem I'm running into right now while trying to calibrate my scangauge with the numbers I'm seeing manually. I even go to the exact same station and pump number every time, but I know for sure that when I'm filling up the car the pump stops flowing at a slightly different point each time. So I fill up the car a number of times before resetting and readjusting the numbers on my Scangauge. I'm almost there now, I'm about 0.3 MPG off. But I want the numbers to be DEAD ON, so even after having the Scangauge since July, I'm still fiddling with the calibration to this day.
Originally Posted By: 440Magnum
Originally Posted By: Hokiefyd
Originally Posted By: 440Magnum
In contrast, the computer should, even in open loop mode, know rather precisely how much fuel is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential.
The computer does know, precisely, how much fuel it THINKS is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential. Variables here include injectors that may stick or leak, or pressure changes in the fuel system due to a lazy fuel pressure regulator, etc.
Oh, I know a malfunction can change fuel volume estimation greatly and manufacturing tolerances can skew it to a lesser degree. But given the types of systems, parts, and measurements involved I still think distance-over-ground is the single most error-prone part of the process.
See my previous post.
Originally Posted By: Hokiefyd
Originally Posted By: 440Magnum
In contrast, the computer should, even in open loop mode, know rather precisely how much fuel is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential.
The computer does know, precisely, how much fuel it THINKS is passing through the injectors, since flow rate is a very precise function of duty cycle and pressure differential. Variables here include injectors that may stick or leak, or pressure changes in the fuel system due to a lazy fuel pressure regulator, etc.
Oh, I know a malfunction can change fuel volume estimation greatly and manufacturing tolerances can skew it to a lesser degree. But given the types of systems, parts, and measurements involved I still think distance-over-ground is the single most error-prone part of the process.
See my previous post.
Originally Posted By: cchase
See my previous post.
Agreed that the same source data for distance is used in both, but its apples to oranges. In the manual calculation, the varying amount of fuel put into the tank at each fill-up FAR swamps every other error.
See my previous post.
Agreed that the same source data for distance is used in both, but its apples to oranges. In the manual calculation, the varying amount of fuel put into the tank at each fill-up FAR swamps every other error.
Originally Posted By: 440Magnum
Originally Posted By: cchase
See my previous post.
Agreed that the same source data for distance is used in both, but its apples to oranges. In the manual calculation, the varying amount of fuel put into the tank at each fill-up FAR swamps every other error.
Okay, I can agree with that aspect.
I read
Originally Posted By: 440Magnum
I still think distance-over-ground is the single most error-prone part of the process.
and took it out of context to mean that there was an implication that the trip meter was the source of error.
On the whole, I don't have much luck on many cars with getting consistent pump shut-off.
Originally Posted By: cchase
See my previous post.
Agreed that the same source data for distance is used in both, but its apples to oranges. In the manual calculation, the varying amount of fuel put into the tank at each fill-up FAR swamps every other error.
Okay, I can agree with that aspect.
I read
Originally Posted By: 440Magnum
I still think distance-over-ground is the single most error-prone part of the process.
and took it out of context to mean that there was an implication that the trip meter was the source of error.
On the whole, I don't have much luck on many cars with getting consistent pump shut-off.
Originally Posted By: 440Magnum
Originally Posted By: cchase
See my previous post.
Agreed that the same source data for distance is used in both, but its apples to oranges. In the manual calculation, the varying amount of fuel put into the tank at each fill-up FAR swamps every other error.
There could be intentional optimism in the speedometer/odometer of say 5% that isn't in the vehicle speed sensor. So you would get a more optimistic mpg calculated by odometer/fuel fill.
That's not the case in this thread as the DIC is optimistic and I don't think the Cruze is even using distance/divided fuel flow. I'm just saying I agree with you that there could be and is several factors that could put mpg claculation off by 5-10%. I tend to think VSS/injector pulse width is probably pretty accurate.
Originally Posted By: cchase
See my previous post.
Agreed that the same source data for distance is used in both, but its apples to oranges. In the manual calculation, the varying amount of fuel put into the tank at each fill-up FAR swamps every other error.
There could be intentional optimism in the speedometer/odometer of say 5% that isn't in the vehicle speed sensor. So you would get a more optimistic mpg calculated by odometer/fuel fill.
That's not the case in this thread as the DIC is optimistic and I don't think the Cruze is even using distance/divided fuel flow. I'm just saying I agree with you that there could be and is several factors that could put mpg claculation off by 5-10%. I tend to think VSS/injector pulse width is probably pretty accurate.
I know that the one on my 2004 Monte Carlo SS is always off. I filled up today and manually figured my mileage and it was at 25 mpg. The DIC said 27 mpg. It pretty much always shows about 2 mpg over what I manually figure.
Wayne
Wayne
- Status
Similar threads
