Hwy mpg substantially lower than city per ScanGauge - any ideas?

[WriConsult]

Eddy, thanks for the map. So can I presume from this that maximum efficiency will be obtained by heavy throttle between 1500 and 2100rpm? I'll assume I should do this both while accelerating from a stop and while P&Ging.

McBrew, I am using DIESELa and the numbers I am reporting are trip averages -- I know well the pitfalls of looking at the instantaneous mpg reading. I'll keep adjusting the calibration over the next few tanks, which should be all-city. Seems to me that calibrating for the all-city tanks is going to make the all-hwy tanks still read low, but we'll see.

Yes, believe it or not my average on the Subaru since last May really is just over 30mpg. I used to average 20-21mpg in my Subarus in winter (I've had 3 of them) and about 23-24mpg in the summer. Rarely any better on the highway because it was always either 70+mph or with bikes or a cargo box on the roof, all of which absolutely clobber fuel economy. Now that I've learned to hypermile (keep the speed down, anticipate conditions better, pulse and glide, nothing on the roof) I'm seeing about 28/33 city/hwy in the summer and 25/30 in the winter. This fall I finally managed to pull a 500 mile tank (35mpg) on a roadtrip, and that was with speeds of ~60mph.

 

[dremd]

I have just completed my semi-scientific test regarding this issue.

In a Nutshell
1) 6 Half tank fillups on a very normal route (low speed urban) yielded very repeatable Scan gauge readings. (within 2/10's of a gallon each time) (about 44~45 mpg actual)

2)Drove 100 Miles ALL highway
a) filled up pulled out and continued accelerating to interstate speed
b) Indicated SGII mpg 25~28 mpg (70~75 mph cruise)
c) Pulled off went straight to Fuel station (total non highway mileage less than .5)
d) sgII estimate MPG for tank 26.3 mpg
e) Pen and Paper mpg for trip 48.7

Obviously the 2 do not match

There ARE others with this issue as well.

Apparently Most TDI's don't have this issue.

So now the question is -> What IS the issue?
My Hypothesis include
1) Thermostat: Ecu may calculate fuel flow based on operating Temperature. This is on the top of the list because more than 1 person with this issue has a suspect thermostat.
2) Plugged injectors/ Pump
3) Faulty lift sensor in whichever nozzle it is that has the lift sensor
4) Ecu issue (don't know what)
5) Scan Gauge issue: My scan gauge would occasionally lock up in the past, hasn't done it in a while, but my unit is somewhat suspect.

Should I start a new thread?

 

[EddyKilowatt]

So can I presume from this that maximum efficiency will be obtained by heavy throttle between 1500 and 2100rpm? I'll assume I should do this both while accelerating from a stop and while P&Ging.

Well, that's the conclusion that I, and others wiser and more experienced than I, reach based on that map. Others here, with good mileage numbers to show, take the opposite tack and go for the light foot and never getting on the boost... it has been a topic of considerable debate here, and I don't claim to know the answer myself.

It is true that the map is most likely based on steady-state dyno performance, and perhaps the transient peformance (as when going up thru the gears) is somewhat different. Still, the shape of that map represents some basic thermodynamics that makes sense to me, so I put myself in the heavy foot/shift early camp ...with the occasional run to redline to keep things cleared out and the smile going.

Eddy

p.s. down around 1500 the turbo compressor is running along the edge of its surge (stall) line, so I believe it is advised not to get all the way into the pedal until 1750 or 1800 rpm...

 

[jettawreck]

I have just completed my semi-scientific test regarding this issue.

In a Nutshell

2)Drove 100 Miles ALL highway
a) filled up pulled out and continued accelerating to interstate speed
b) Indicated SGII mpg 25~28 mpg (70~75 mph cruise)
c) Pulled off went straight to Fuel station (total non highway mileage less than .5)
d) sgII estimate MPG for tank 26.3 mpg
e) Pen and Paper mpg for trip 48.7

Obviously the 2 do not match

Obviously not. In my opinion, half tank and even more so, 100 miles between fill-ups is not as accurate as full tank fills. The fill error (there is always some) is virtually the same for 15 gal. as it would be for 2 gal. giving wildly inaccurate calibrations/calculations.
If you can not get quite accurate readings after a couple of full tank fill-ups and calibrating either you have setup errors or the unit is suspect.

 

[dremd]

Ok JettaWreck In that case you come up with an experiment and I'll do that.

I'll be driving around town until easter then be taking a 800~900 mile trip (Houston Supra meet http://www.houstonperformancedriving.com/forums/forumdisplay.php?f=192(in a golf LOL)). Will that do?

First 6 fillups were ~450 miles each
Last was Just over 100 (just looked; 106)
Edit:
Indicated Fill 4.0 gallons
Actual fill 2.167 Gallons
So for the Scangauge to be correct I would have to have been nearly 2 gallons off on my fill ups (venectomied, fill all the way to the brim after all foam is gone is my normal fill)

Edited:

 

[jettawreck]

So now the question is -> What IS the issue?
My Hypothesis include
1) Thermostat: Ecu may calculate fuel flow based on operating Temperature. This is on the top of the list because more than 1 person with this issue has a suspect thermostat.
2) Plugged injectors/ Pump
3) Faulty lift sensor in whichever nozzle it is that has the lift sensor
4) Ecu issue (don't know what)
5) Scan Gauge issue: My scan gauge would occasionally lock up in the past, hasn't done it in a while, but my unit is somewhat suspect.

Should I start a new thread?

1) I dont belive SG calculates that way.
2) Doubtful. Your actual mpg numbers are not bad. Runs good??
3) #3-but also doubtful.
4) no clue me either (see #2 and 3)
5) best chance of the 5

yes-and/or contact Linear Logic about defect or setup errors.

 

[dremd]

1) I totally don't know
2) runs pretty darn good I smoke a little at WOT but otherwise no complaints
3) I totally don't know
4) Same here
5) How about a Trade with the Unit in My Vibe? It is always within .1 gallon at fillup?

 

[McBrew]

dremd, I think I am missing something. You are saying that you filled up, drove around, and filled up again... and that the SG os off quite a bit. Did you correct the SG after the fillup? I have found that it is best not to correct it all the way, but to correct about halfway between SG's numbers and actual tank fillup. That way you get less of a swing from tank to tank. It took about 6 or 8 500+ mile tanks to get mine to be pretty accurate.

 

[dremd]

I've had my SGII for several months I can pretty much tell what the FCF will be based on what type of driving I do. (not hard proof but by the end of a tank I can guess on the FCF and be pretty darn close. I started the 6 tank "calibration" pre-calibrated. all of the 6 fill ups were with in 2/10's of a gallon (which in an ~9 gallon fill up gives an accuracy of around 97% (Very acceptable) Compare this to the ~50% Accuracy on the highway and you can see there is some issue.

Your 1/2 correction basically just Mask the issue, it gives a rounded off fuel consumption, probably great if your gauge is fairly accurate, but if it is off by a factor of nearly 50%, like mine is, it is useless.

 

[DFWDieselJet]

dremd - Thanks for doing the experiment to confirm what many people have reported. It's interesting that you got the opposite of what many people report. Usually in city driving the SG shows lower than actual FE, and in highway it shows higher.

As for your thermostat theory, it's possible, but if your dash temp gauge shows normal temp (190) once warmed up, and it takes a reasonable time for your car to warm up, I doubt you have a problem. I think the reason thermostats have come up a few times is that there is a known (but not commonly known) problem with the scangauge reporting the trans temp as the coolant temp. Read back a ways.

I like McBrews binary division mode of calibrating the SG, especially after a good basline calibration is established. It's a good hysteresis algorith.

 

[DFWDieselJet]

Ron's a great guy

I sent an email to Ron at Scangauge about this issue, here is my questions and his reposonse:

Gene,

I would have to speculate too.

We determine the fuel being used by reading the injector pulse-width. The fuel flow may be affected by the following:

1) A difference in highway fuel pressure vs city fuel pressure.
2) The difference in manifold pressure between highway and city.
3) The different fuel temperature from the higher under-hood temperatures and longer time the fuel is in the lines at the engine in city driving.
4) The higher heating of the fuel in the tank during city driving.

We are going to have to look at ways to compensate for these and see which ones work.

Ron

----- Original Message -----
Sent: Saturday, February 09, 2008 6:50 AM
Subject: Re: Question about some readings


Hi Ron,

Thanks for the quick help on this last issue with the Scangauge. There's another thread going on on the TDI Club forums right now, and I'd like to ask a question. We all know that the MPG calibration will differ based on driving conditions. For example my "all city" calibration is about -30%, but my "all highway" calibration can be as low as -17%. It is my belief that the innacuracly lies in the ECU, not the scangauge, but there is a lot of speculation going on on the forum. Would you mind me asking which values you're reading for the fuel consumption measurments?

Thanks

​

 

[jettawreck]

Ok JettaWreck In that case you come up with an experiment and I'll do that.

I'll be driving around town until easter then be taking a 800~900 mile trip (Houston Supra meet http://www.houstonperformancedriving.com/forums/forumdisplay.php?f=192(in a golf LOL)). Will that do?

First 6 fillups were ~450 miles each
Last was Just over 100 (just looked; 106)
Edit:
Indicated Fill 4.0 gallons
Actual fill 2.167 Gallons
So for the Scangauge to be correct I would have to have been nearly 2 gallons off on my fill ups (venectomied, fill all the way to the brim after all foam is gone is my normal fill)

Edited:

How was your reported SG tank averages on the 5th and 6th "calibration" tanks?
Our vehicles are only 1 model year apart (Golf/Jetta but both VE) so no big difference there. I have never run a full tank just "city" so I cann't say with 100% certainty that my SG would be accurate that way but I do run a mix of conditions every day (same route pretty much) and get very accurate readings in day/current/tank options. I would think that if it were erring big time in the city portion as much as yours seems to it would alter my day and tank averages enough to throw the accuacy out the window.
The only time accuracy has been an issue, after getting calibrated, was when I installed the new nozzles. It made the average look better than it actually was for a few tanks while it "calibrated" again. Also there have been some erratic fills by me-out of town "Pay at the pump" pumps that shut off at $50 automaticly (short filled tank) or lack of patience filling when its -25F with a good/bad wind (sit in car and wait for pump to shut off and forget it).
Do you have any mods/chips not in your sig. line that may be affecting it?? If not, my hunch would be to try the other unit and see if the results are better/different. Perhaps it is faulty.
I need to fill tonite on the way home. Aprox 750+ miles (light came on mid way in) this tank with LOTS of -0F temp, so it'll be down around 48mpg. I'll post up the results with the actual numbers after.

 

[jettawreck]

I need to fill tonite on the way home. Aprox 750+ miles (light came on mid way in) this tank with LOTS of -0F temp, so it'll be down around 48mpg. I'll post up the results with the actual numbers after.

2/16/2008
Scan Gauge before fill-up: Tank average 49.5 mpg@ 15.1 gal. (forgot to look at SG miles, it always clocks about 10 more to a 700 mile tank)

Actual fill-up: 737 miles on tripmeter/15.072 gal. =48.9 mpg. Less than 1.5% error. No correction calibration needed. Current correction percentage read 8.6%.

Per that email from "Ron" maybe there is something up with the fuel temp sensor in yours since that is something that appears to enter into SG's calculations. Just a wild guess though, I'm not a tech. Hope you can get some help to get it more accurate for your driving needs.

 

[EddyKilowatt]

I sent an email to Ron at Scangauge about this issue, here is my questions and his reposonse:

Gene,

I would have to speculate too.

We determine the fuel being used by reading the injector pulse-width. The fuel flow may be affected by the following:

1) A difference in highway fuel pressure vs city fuel pressure.
2) The difference in manifold pressure between highway and city.
3) The different fuel temperature from the higher under-hood temperatures and longer time the fuel is in the lines at the engine in city driving.
4) The higher heating of the fuel in the tank during city driving.

We are going to have to look at ways to compensate for these and see which ones work.

Er, um, that there's gas-engine talk... at least the part about injector pulsewidth is... probably the bit about manifold pressure too (affects gas engine injector flow). I have a hunch Ron didn't realize you're talking Diesel.

But I hope you hear back from him again, because I've always been really curious where the Scanguage picks up its fuel-flow data when it is set to Diesel mode...

Eddy

 

[DFWDieselJet]

Er, um, that there's gas-engine talk... at least the part about injector pulsewidth is... probably the bit about manifold pressure too (affects gas engine injector flow). I have a hunch Ron didn't realize you're talking Diesel.

But I hope you hear back from him again, because I've always been really curious where the Scanguage picks up its fuel-flow data when it is set to Diesel mode...

Eddy

Good point, and I think you're right because I did some experimentation with the custom commands with the scangauge this weekend, and no fuel pressure commands (PIDS) are supported, and I couldn't find one for injector pulse width. I replied to Ron asking specifically what he's reading. I'm beginning to suspect it's NOT and OBDII PID, but that he's requesting "proprietary" info like vag-com. I'll let you know what he says...but I might be asking for information he's not willing to share.

 

[dremd]

As an experiment I swapped Scangauges yesterday, Started them both at "guessed" FCF's but the Golf one will require a month or more worth of driving before I have anything I would consider Data.

 

[McBrew]

Your 1/2 correction basically just Mask the issue, it gives a rounded off fuel consumption, probably great if your gauge is fairly accurate, but if it is off by a factor of nearly 50%, like mine is, it is useless.

Well, I don't know what the issue really is, since I can't actually see how the SGII plugs the correction factor into the FE math. I know how I would do it, but I don't know if they do it the same way.

All I know is that my FE was swinging from high to low, getting only a little closer to reality each time I filled up and made the correction. Once I went to my half-way approach, it took only 3 tanks to me off by 1%.

 

[DFWDieselJet]

Drat, we'll get no details from Scangauge on how they get fuel consumption...

Gene,

There is a way to get fuel injector pulse width, but I can't go into detail. It is one of the things we consider proprietary.

Ron


I hope he realizes, this means war! No, not really.

 

[DPM]

Well, IQ is roughly interchangeable with injector duty cycle/ pulse width. You could even go with mass airflow too if it wasn't for the EGR...

 

[DFWDieselJet]

Well, IQ is roughly interchangeable with injector duty cycle/ pulse width. You could even go with mass airflow too if it wasn't for the EGR...

Are you talking about the vag-com reading? Or is this an OBD PID or field?

If it's a vag-com reading, I ran some serial sniffer traces with vag-com on my laptop, and I've found the codes needed to request the monitor groups. But when I tried them as custom commands in SG, I got no response. I think I need to log into the controller before requesting monitors, but I'm not sure if SG can do that.

Another option would be to tap the serial cable between SG and the ECU and sniff that way. That'll tell us quickly if SG is using an undocumented OBD PID or some other proprietary way of reading data. I'll try that sometime soon.

 

[DFWDieselJet]

Tap cable worked.

Scangauge is reading:
0C - RPM
04 - Load
05 - Engine coolant temp
0B - MAP
0F - Intake air temp

Then using the engine displacement, the "Diesel A" formula, and the correction factor, SG calculates fuel consumption.

There are no other codes being read, so (after this and other testing) I'm forced to come to the conclusion that there is no way to read any directly related fuel info (like inector pulse width, or VAG monitor group 15, zone 3) through the OBDII protocol on Diesel VWs.

There is one other possibility...I'm monitoring the ISO K/L lines only. It's possible Scangauge is using the CAN lines to gather additional data, but I doubt it. If so, there would be no need to read these OBDII values. An easy test would be to remove these lines from SG and see if it can still display FE data.

This calculation will, and does, work well - as many Scangauge owners can attest. It does not, however, take into account certain factors like EGR activity, injector timing to control emissions, and changes in fuel quantity based on fuel temp. That's partly what the correction factor is for. But I see now why there is such variability beween the highway and city driving.

 

[DFWDieselJet]

It's confirmed. Scangauge is NOT using the CAN bus. The only pins required for the SG to work properly are ground (signal and chasis), power, and the ISO K (and possibly L) lines. CAN lines are not needed, so SG is not reading info off the CAN bus. The five codes above, plus vehicle speed, is ALL the data SG is getting from the ECU to calculate fuel consumption and MPG.

 

[dremd]

UPDATE: Scan gauge swap APPEARS unsuccessful.
I was still seeing heavy swings (doing 4 gallon is fill ups) in the neighborhood of %50 depending on where/ how I drove. I say appears, because I did not continue the experiment, long enough, to obtain enough data to come to a conclusion with any certainty.

I DID however (yesterday) swap out my thermostat. I do not know if it will solve the issue or not, but it did make a HUGE difference in my indicated mpg.

More data to come in the following weeks, lots of highway driving will be included .

 

[DPM]

You'll not get great accuracy if you're only doing 4gal top-ups...

 

[dremd]

You'll not get great accuracy if you're only doing 4gal top-ups...

Normally I would agree with you, however my error between City and highway is approximately 50%. I think it is quite safe to assume that fill ups are within 1 gallon each (25%). And I think it is mostly safe to say that they are accurate with in 1/3 of a gallon which would be 8.3% margin. Compare that to ~50% (margin of error on my previous SGII) and you'll see why I was / am using such small fill ups (more data points)

 

[jlallison]

Any new data to add anyone?

I've been noticing the same thing in my quest for fuel economy.

Much better mpg 50+ when driving under 45mph and much worse (below 35mpg) mpg when driving over 50mph. My averages are close per tank fillup but the ave rates indicated while driving seem to suggest keeping my rpm around 1700 to achieve the best economy.

 

[darkscout]

I just got the new firmware update and after a weekend on the road and in the city it was about dead on.

 

[Honeydew]

I just got the new firmware update and after a weekend on the road and in the city it was about dead on.

what version are they up to now? (select "more" until "version" is an option and select that) Thanks

 

[dremd]

Update1: Thermostat did make it closer, but far from correct.
Update2: Installed a set of pp520's and had them balanced last week. MPG read out now "appears" to be either accurate or close. I've got 400 miles on them so far, I will keep everyone posted. FYI Old nozzles did feel worn to my untrained hand; some sticking; little loose; etc

Edit: also see http://forums.tdiclub.com/showthread.php?t=216604