[SOLVED] Very innacurate readings on Hyundai Elantra Touring

[pherthyl]

Hi all,

I recently got a Kiwi Wifi and Dashcommand for my 2011 Hyundai Elantra Touring automatic transmission (http://www.hyundaiusa.com/elantra-touring/specifications.aspx)
I put in a new vehicle and set the engine displacement to 2L.  Everything connects and starts displaying data.

However, the fuel consumption data is wildly innacurate, like 30-50% too high.  For example, I've tracked fuel consumption manually for the last 5000km, and the average is 8.7L/100KM.  Dashcommand's average currently shows 15.9 (i've reset it recently, but on a 200km trip recently it was 14.3 L/100km instead of 8.4 that I calculated from a fillup before/after.

I suppose I could do the fillup correction, but the manual states that it is optional, and everything I've read about systems like this is that they should be within a few percent of the correct value without calibration (my brother has a Scanguage and his correction was 2% from fillups).  To verify that it is not the Kiwi, I plugged it into a friend's Toyota RAV4 and verified the fuel consumption in DashCommand against his onboard fuel meter and it was dead on (within 0.1 L/100km).

So something on my Elantra is reading incorrectly, but what?
I noticed that my fuel consumption does not go down when coasting, which might explain the discrepancy.  For example, earlier I was coasting down a hill, no gas, 2500RPM and it was in top gear.  The fuel flow gauge said 10.5L/hour.  That does not make any sense, since if I'm coasting down the hill it should be using zero or close to it.  

So can someone give me the formula for how fuel consumption is calculated?  Then I could watch the individual PIDs and see which one is wrong, or debug further.  

Otherwise, any other suggestions for what to try??

Thanks.

[pherthyl]

Ok it seems that the only actual sensor being read for fuel flow calculation is MAF, the rest are constants.  I'll check the PID's to see what that number reads and if the car has that.  

Have you come across cars that just report MAF incorrectly?  If so, I wonder how I can explain that to hyundai warranty ..   I'll do some more checking.

[pherthyl]

More testing, more info.

Looks like I don't have an actual MAF sensor on this car.  The console says: "MAF PID is undefined or invalid - Using CALC.MAF_A"

So I had a look at the two estimated mass air flow PIDs and they do get quite different values.

For example, at idle I get ~4.1g/s with Method A, and 3.2g/s with Method B.

Given that the Method B produces significantly lower values, my instinct would be to use that value instead.  Can I just enter CALC.MAF_B into the MAF PID field in the vehicle manager to force it to use that one?

Based on the reading, I gather that the other option is to tweak the volumetric efficiency until I get more reasonable values for fuel consumption.  However given that my fuel consumption is so far off, I think switching to the method B of MAF calculation might be a better first step.  Does that make sense?

[pherthyl]

After reading all 117 posts in this thread (http://www.mp3car.com/engine-management-obd-ii-engine-diagnostics-etc/75138-calculating-mpg-from-vss-and-maf-from-obd2.html), I'm beginning to appreciate the complexities of calculating fuel flow.  Have the car companies never heard of a simple flow meter??

[pherthyl]

I switched over to using CALC.MAF_B and the numbers seem a lot better.  I will give it a go for a tank and see how close my average is to the actual numbers.

Well I've been a lot of help, thanks!

[John@PPE]

Nice detective work. As far as the fillup page, I generally don't recommend using that page. Like you say the correction is pretty minor. And DashCommand has to be constantly monitoring the tank or you'll really mess up your reading.

[Weston@PPE]

It sounds like you've done a bit of looking into this. If you go to the CALC PID Reference sheet, on our downloads page, you can see what the differences are between MAF_A and MAF_B. I would trust MAF_B if its available -- MAF_A uses absolute load and RPM to calculate airflow, and MAF_B uses manifold pressure, IAT, and rpm. You are correct, to force it to use one or the other you would enter it into the MAF PID vehicle setting as CALC.MAF_B (I do believe this is the first time I've ever recommended someone fill in that setting).

Let us know how it looks after a tank. If you've done the fillup procedure before it will have saved some adjustments, and you might need to reset your fuel and trip stats.

[pherthyl]

Thanks guys.  It was very interesting to read up on the full picture.  I never imagined that measuring fuel usage would be so complex.

I haven't used the fillup screen yet for the reasons you mention.  I'll monitor a tank or two and then see if it's worth trying to tweak VE at all, but I suspect the current values are pretty close now.

[pherthyl]

Ran a tank (yeah I don't drive much) and it was pretty accurate.  8.6L/100KM indicated VS 8.4L/100km calculated manually.

I think the remaining discrepancy comes from the fact that Dashcommand thinks the engine is burning fuel when I'm coasting.  For example, coasting down a hill at 90km/h, the display says I'm using about 4L/hour when it should pretty close to zero.  I guess it isn't possible to detect this without the MAF sensor?

[Weston@PPE]

Even with a MAF we can't detect situations where the ECU cuts fuel. Air will still be moving through the engine, and just won't have fuel to go with it, so the MAF would still read higher than 0.

Its good to hear thats its accurate though, you can always ask if you have any other questions.