Scangauge accuracy

[dremd]

Works very well in vibe.
Works like crap in TDI.

[tomlai]

I have been using the ScangaugeII for about 9 months now, vehicle odometer and SG odometer match to within 1-2 km for every 400km which is very good.

However my tank fills are way off - even today it is reading as having used 33Litres of fuel whereas the actual pump fill-up reading is 42Litres ... my SG Fill up compensation is now nearly +150% and rising with every fill up !!!!!!!!!

What is the problem here?
I have been facing this problem since I have started P&G with extensive ICE-on coasting in Neutral whenever I can ... I can get "current trip" mileage of under 4L/100Km using this technique.

Although people say that the SG can be used as an indicator of Fuel efficiency, it influences my driving technique by indicating that if I P7G in this way I will get terrrific mileage. My question is should I believe the overoptimistic trip mileage figure, to guide my driving style?

[PaleMelanesian]

Check the Cutoff setting - that might be messing things up.

Is this a diesel? It's not very good for those, either.

[tomlai]

Its a petrol (gas) vehicle. Actually if cut-off was working then SG would show even lesser fuel consumption.

My problem is that SG shows lower fuel consumption than actual!

[MetroMPG]

I also found that the SG accuracy dropped when doing a lot of pulse and glide. It was fuel used that was off, not distance, and it was under-reporting fuel used. (But by about 8 or 9 percent, not to the extent you're seeing.)

I attributed it to the fact that the SG was guessing wrong at the fuel ratio when in open loop mode acceleration (which you pretty much need to do in the Metro to get good accel at low/moderate RPM).

When predominantly driving with load (avoiding open loop conditions) it is accurate within a couple of percent.

I had to re-do my calibration offset last year when I switched to DWL rather than P&G as a highway tactic.

[tomlai]

Under a normal tank cycle, there will be periods of DWL and periods of P&G.

Lets assume that the SG is accurate during the DWL cycles and way off during P&G.

What I want to know is whether I should DWL more (for just about OK mileage) or should I trust the way-off values (for fantastically good mileage figures) when I P&G in the hope that there is some truth in the figures and that I do improve mileage by P&Ging ?????

The difference is in the range of 60+ mpg (P&G) compared to values of 25-35mpg (DWL)

[PaleMelanesian]

I just did a tank with 570 miles of highway P&G. Scangauge said 12.62 gallons, I put in 12.60 gallons.

My car does not go into Open Loop during pulsing. Maybe that's the key. It also drops immediately into closed loop on a restart - not even one SG refresh cycle in open loop.

[MetroMPG]

tomlai: if your goal is maximum mileage through P&G, then you need to keep trying to calibrate the SG to your actual tank results.

It's a frustrating situation, to be sure. When my SG is calibrated for P&G, it gives pessimistic DWL readings (which can be good for motivation, but still annoying).

Your best option may be to go with an open source MPGuino, which measures actual fuel use through monitoring injector pulses, instead of extrapolating it from air intake (which is what the SG does). Once you get the MPGuino calibrated for your vehicle, it will be accurate regardless of technique used, open/closed loop, etc.

[tomlai]

[QUOTE=MetroMPG;88057] ... trying to calibrate the SG to your actual tank results... QUOTE]



I am getting these wierd fillup compensation.



The SG read 35.5 Litres used, whereas tank fillup was 43.4, and I upward adjust the fillup compensation at every tankup. And over the last 7-10 fillups the compensation has risen to 173%. It is going to rise even more at the next fill, I am sure.

What I am surprised at, is the consistent under-reporting of fuel consumption.

[winkosmosis]

Originally Posted by MetroMPG

tomlai: if your goal is maximum mileage through P&G, then you need to keep trying to calibrate the SG to your actual tank results.

It's a frustrating situation, to be sure. When my SG is calibrated for P&G, it gives pessimistic DWL readings (which can be good for motivation, but still annoying).

Your best option may be to go with an open source MPGuino, which measures actual fuel use through monitoring injector pulses, instead of extrapolating it from air intake (which is what the SG does). Once you get the MPGuino calibrated for your vehicle, it will be accurate regardless of technique used, open/closed loop, etc.

Why doesn't the Scanguage do that?