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

[WriConsult]

OK, I've only had my 2000 Golf TDI for 4 days now, but I'm already noticing an odd thing with the mpg on it. The (indicated) city mpg is spectacular, but the highway mpg is quite a bit lower. I am a fairly decent hypermiler so I am able to get better mpg than your average driver, but even so ... something is out of whack.

Around town I'm pretty consistently able to get mpg in the mid to high 40s (indicated), sometimes low 50s if I'm not having to stop too much. On Saturday I made a 15 mile round trip on suburban parkways and got an indicated 67 mpg.

But on the highway, no such luck. A 25 mile round trip the other day at very moderate (50-60mph) freeway speeds yielded about 43 mpg. Looking at the ScanGauge's instantaneous readouts, I'm seeing 1.1gph to 1.2gph fuel consumption even at a steady 50mph (which equates to substantially less than 50mpg, of course). Consumption rapidly spikes higher than that at higher speeds. And this was after swapping out the original high rolling resistance tires (Bridgestone G009) for my preferred LRR model (Toyo TourEvo) at 44psi.

I understand that a lot of factors are working against me: rainy roads, cool temperatures (30-40F), B50 fuel (4% less energy content), possible ScanGauge calibration error. But if any one of those was the culprit, then my city mpg should be impacted too. But it's not. As you can see above, it's awesome!

One key difference between my city driving and my highway driving is the revs. In city driving I'm almost always shifting between 1700-2000 rpm and only very briefly hitting the rpms I need to steadily cruise the highway. Is there something that could cause efficiency to dramatically drop once the revs get into the 2000 rpm range? ScanGauge shows no codes, by the way.

Things I can think of to investigate:

• Wheel bearings. This week I'll check each wheel to make sure it spins freely.
• Air filter.
• Fuel filter.
• [edit: added] Thermostat. Although the dash temp gauge gets up to 190F reasonably quickly, I've never seen the SG indicate anything higher than 167F. I read the thread about this, and apparently it's common for them not to agree, but in some cases people have found the cause to be a bad 'stat. I'll get that checked out too.

Anything else?

Later this week I'm going to take some controlled runs at various speeds to see exactly where the mpg drop is happening. I'll do it in both 4th and 5th gear too, to see if there's a sudden drop based on rpm.

 

[Bradtholomew]

I'm surprised you have your Scangauge calibrated properly after only 4 days - how many times have you filled up?

Anyways, before digging too deep into possible mechanical issues, I'd get your SG dialed in over a few tanks. Course, a little tune-up could never hurt..

Brad

 

[DFWDieselJet]

Hmm. Interesting.

The scangauge is known for being accurate - given a calibration based on a particular driving condition. For example, my all-city-driving calibration is -30%, but my all-highway driving calibration is around -22%. I've never heard a good explanation why, but this does seem to incdicate that the scangauge - or more accurately the ECU - is not always very accurate.

I would guess it is LEAST accurate in city driving because of the relatively low sample rate and rapid variations in RPM while city driving.

With only owning the car for four days, have you had a chance to get even one calibration done? Or even a real MPG based on mileage and gallons filled? This will help show WHICH scangauge reading is the more accurate one.

Also, you're right on taking a slight MPG hit from the B50, but you're in ideal temps for fuel economy (from a stock car at least) - assuming the car has time to warm up.

 

[40X40]

I would add that you want to use the 'current' average, not the instant MPG.

Bill

 

[WriConsult]

Just to clarify:

No, I have not yet calibrated the ScanGauge. I am well aware that it could be off by as much as 10-20% in either direction, and I'm well aware of the procedure.

Calibrating the SG will affect city and hwy mpg equally. All the SG does is add or subtract a flat percentage based on what you entered for actual fuel fuel vs. what the ECU reported. But DWFDieselJet, thanks for the info on the SG not being consistently accurate on hwy vs. city mpg. That is not something I've heard about or observed on my Subaru, but if there's something about the TDI that causes this to occur maybe that's the explanation.

With the exception of the one mention of instantaneous gph on the highway, all of the mpgs I'm reporting above are trip averages. After a number of both city and hwy trips, it's pretty consistently low to mid 40s on the highway at a very modest 50-60mph, upper 40s to low 50s in town, and 67 mpg on that specific parkway segment I mentioned.

Also, I'm seeing this even when the car is fully warmed up.

 

[Brock_from_WI]

But when you say city are you really talking suburb? like cruising at 40 with a couple spread out lights? Or do you mean stop and go like every other block or so? I get the best mileage on my trips to and from work, mostly at 40 mph with 2 stop signs; I definitely get worse mileage on the highway compared to what I call “rural”.

For the record I would agree about SG being different if you drive all city or highway, it changes for sure, not sure why but it is different.

 

[WriConsult]

Mix of city and suburb. Even in city driving with lights every few blocks I'm getting mid 40s. But I'm not talking about primarily downtown driving where you're stopping every block or two, nor am I talking about stop and go. Although I'm still getting impressive trip averages when some of that kind of stuff is mixed in.

Some of the suburban roads out by where I work aren't any better for mpg because they're 40-45 mph. Even though the lights are 1/2 mile apart, you almost always get caught at them. And even though I tend to not always go all the way up to the speed limit, a lot of fuel ends up wasted getting up to higher speeds than in the city.

 

[WriConsult]

I did some steady state tests yesterday, recording trip average over the same stretch of parkway repeatedly at different steady state speeds. In 5th gear:

• 40 mph (~1500rpm): 0.59 gph, 67.7 mpg.
• 45 mph (~1650rpm): 0.80 gph, 56.0 mpg.
• 50 mph (~1850rpm): 1.11 gph, 45.2 mpg.
• 55 mph (~2030rpm): 1.45 gph, 37.8mpg.

4th gear:

• 40 mph (~1920rpm): 0.99 gph, 40.4 mpg.
• 45 mph (~2140rpm): 1.36 gph, 33 mpg.

So it's the revs that are clobbering me here. Of course you always want to keep revs down, but any engine -- even a diesel -- should have reasonable fuel consumption at 2000 rpm and light load. Downshifting from 5th to 4th even at 40 or 45mph is increasing my fuel consumption by 70%. ScanGauge calibrated or not, that shouldn't be happening. I would ordinarily expect something on the order of 30%.

Is it possible there's something wrong with the turbo? I've heard the variable vanes on these turbos can have problems sometimes -- if something in there is sticking, could it be causing fuel consumption to spike unnecessarily?

By the way, at these light loads it doesn't look like the turbo's doing very much, so it's not like it's overboosting. MAP reading was in the 15s and 16s throughout (baseline reading was 14.8 at idle). Will get up into the high 20s if I floor it, so the turbo's definitely working.

 

[darkscout]

What we have here is a solution in search of a problem. I don't see anything wrong with any of your numbers. Obviously all of your data was at 'light load'. Anything under 2500 on flat ground is a light load.

Most of anything you're going at those speeds is inertia and internal friction of the motor.

Here's a graph with all the data you just gave. Seems pretty linear to me.

 

[weedeater]

I would be less inclined to believe the ECU at light loads than at more moderate ones. This has to do with how much fuel is actually consumed versus pushed toward the injectors by the pump. Remember that all the excess fuel goes back to the tank, not true with the Suburu (gasser).

 

[WriConsult]

Well, I guess I'm going to have to hope that it's simply a matter of SG (or the ECU from which it's pulling data) not being consistent in its reporting of the TDI's fuel consumption under various conditions.

If true, this will be a bit of a disappointment to me. I've found the SG to be highly accurate on the Subaru. I could drive whole tank in town, then go on a road trip and switch back, and it was always accurate within 0.5mpg (<2%) and only needed recalibration when the temperature achanged dramatically from winter to summer. The only major inaccuracies I've heard of with respect to gas engines are that it doesn't properly factor in the fuel cut that happens while decelerating in gear on newer, and it doesn't handle lean burn mode properly.

Nothing I can't live with, and the SG will still be useful even if not truly accurate, but I sure have grown accustomed to being able to observe my actual fuel economy as I'm driving.

If it turns out the SG is accurate, and that I really am getting the same mpg at 60mph (as currently indicated) as in my Subaru, I'll come back and ask again.

 

[AudiLikeA4]

Pulse and glide on the highway, I can maintain about 47-53mpg on the hwy at speeds of 70-80mph using an on the accel off the accel technique. You speed up and coast in gear, generally trying to coast on any downgrades if possible. In gear coasting is free I'm driving an 04 PD so you should be able to do better, ymmv. I also power up hills, the coasting on the downgrade usually recoupes about double what it cost uphill.

 

[WriConsult]

mobe, I can't speak for the TDI specifically yet, but in most cases pulsing and gliding (or at least pulsing and coasting) has HUGE benefits if you do it right. I use it quite a bit on the Subie, and even with its 0.2-0.3gph idle I can get a 10-20% improvement over steady state driving. Although (perhaps due in part to the Sube's low gearing) I usually don't do it at speeds over 50-55 mph (that's 2400-2700 rpm) unless the terrain is rolling. As an extreme example, for me to get to the western or southern suburbs here requires a 500-600' climb (depending on which freeway) over the West Hills. If I climb the hill at 50mph and coast down the other side (typically a 2-3 mile coast at 55-65mph), I can easily hit 37-38mpg for the trip in the Subaru, easily ~4mpg better than I could get cruising steady state at the same speeds. That's an unusual case, but it works equally well for shorter glides. I've always found the SG to be consistently accurate on that car so I don't believe it's overstating the benefit. Feel free to hit cleanmpg.com for lots of testimonials on how well it works.

The reason it works is that an engine is not particularly efficient at light load. Although it uses more fuel, you are getting more work per ounce of fuel if you make your engine work a bit. Pulsing judiciously lets you get the engine to put out power at maximum efficiency, and then coast for almost free in between the pulses. I say "judiciously" because if you're stomping on it in the pulses you'll be worse off than steady cruising. Just nice moderate (not light, not heavy) acceleration.

I didn't have time yesterday, but in the next few days I hope to take the Golf out for some more runs including steady vs Pulse/Glide and Pulse/Coast. Although if the SG turns out to be as inaccurate on this car as I fear, maybe it won't tell me anything.

 

[rotarykid]

I'm surprised more don't ask this same question . In reality we do actually achieve higher mpgs on city runs but most of those gains are lost by most stopping & starting . If you can avoid the great loses imposed by stopping & starting city runs are more efficient . I have learned the techniques that save fuel in city driving so I almost always achieve higher mpgs in city only loops .

With much effort I've hit around 65 mpgs a couple of times on low speed city runs in a 97 Passat TDI . Mid 50s are more the norm in city runs if complete stops can be avoided , over acceleration is avoided & the brakes are used as little as conditions will allow . Driving other than rush hours helps to avoid stop & go which kills mpgs . You throw in using gravity to start the engine avoiding the starter battery drain . And then add killing the engine a few times while coasting in "N" using gravity to restart the the engine when power is needed . If all of the above is done the high 50s to low 60s mpg US are possible on summer fuel .


The winterized ULSD has really put a hit in winter mpgs over what was achievable with LSD . I've seen as much as a 28-30 % drop in winter mpgs with winterized ULSD . I have close to 28 years of records that back this up . I've never seen such low mpgs in any of my diesels as I have seen with winterized ULSD .

 

[Brock_from_WI]

WriConsult while P&G does net some gain on a TDI a diesel engine doesn't suffer as much with a low load as gassers do so the gain isn't nearly as great. If I P&G I can go from 85 to 90 on the same trip, in our van that same trip running P&G I go from 24 to 30. I just don't have the patience to do it consistently unless I am really bored.

 

[jettawreck]

I'm surprised more don't ask this same question . In reality we do actually achieve higher mpgs on city runs but most of those gains are lost by most stopping & starting . If you can avoid the great loses imposed by stopping & starting city runs are more efficient . I have learned the techniques that save fuel in city driving so I almost always achieve higher mpgs in city only loops .


The winterized ULSD has really put a hit in winter mpgs over what was achievable with LSD . I've seen as much as a 28-30 % drop in winter mpgs with winterized ULSD . I have close to 28 years of records that back this up . I've never seen such low mpgs in any of my diesels as I have seen with winterized ULSD .

I agree, especially with the second part.

I run a 40 mile, one way, commute most every day with a mix of about 50% @ 65 mph, the other 40% @ 55mph and the remaining 10% @ 35/45mph. My average, when not running blended fuel in the cold (May-Sept here), is 51+. My "current average per SG" on the 65mph steady nonstop part is in the upper 40's. At 55/60 mph its around that 50mpg mark and the "slow" part nets over 55mpg. I haven't done all the math to see if those "averages" average out to my tank average that I keep records on but, the SG is very accurate after calibrating it a few times on the tank to tank numbers so I have no reason to suspect it to error in different types of driving situations. I think the often quoted "rule of 120" is a bit over-estimated. I get pretty good mileage, but there is no way I can achieve 60 mpg @ 60 mph w/o a BIG tailwind. I think those that do (think or say so) perhaps have a fair amount of miles at lower speed in that equation that they dont recall or account for, IMO. For the original poster however, his "uncalibrated current averages" seem rather low at the speeds he reported. If they don't go up after a few fill-ups/calibrations, I would suggest he start to look at some basic items-snowscreen, airfilter, nozzles, timing, etc. I don't know the extent of fuel blending where he lives to even guess how much, if any, that may be a factor. Done rambling, out of time, back to work.

 

[DFWDieselJet]

• [edit: added] Thermostat. Although the dash temp gauge gets up to 190F reasonably quickly, I've never seen the SG indicate anything higher than 167F. I read the thread about this, and apparently it's common for them not to agree, but in some cases people have found the cause to be a bad 'stat. I'll get that checked out too.

The scangauge is probably showing you the transmission temp. I had a problem with the values "jumping", but after contacting Linear Logic, we determined that the temp gauge is showing trans temp most of the time. The gauge on the dash is actually pretty accurate. Check out this thread:

http://forums.tdiclub.com/showthread.php?t=193859&highlight=jumping

 

[dremd]

I have the Same issue with Mine.
Let me dig Up my thread and Post a link here.
I've never resolved the issue; I'd Contempleating partially clogged injectors; but I honestly don't know

 

[AZGolf]

In the Scanguage setup page you did select Diesel A as fuel type right?

 

[dremd]

http://forums.tdiclub.com/showthread.php?t=191061
To frame that thread I should state that I was contacted privately by multiple other members who had the same issue; but were afraid to post publicly (for a variety of reasons).
And on another note I suspect that I have a sticky thermostat (cold) as well.

 

[DFWDieselJet]

Another thing to watch for is intake air temps. I've noticed (especially in these chillier winter months) that my fuel economy immproves as the intake temps fall. Best observed FE is with intake temps in the 40s. This appears to be emissions control related - the ECU pulls injection timing to prevent NOX emissions with high IATs.

The reason I mention this is the testing done by WriConsult and Dremd (thank you, btw, it's nice to see people post actual data!). IATs are highest right after acceleration, then can take several miles of steady-state cruise to cool off. To best measure highway fuel consumption, you might need to start your measurment after the intercooler has reached it's minimum temp.

You might add IAT to your scanguge set and see if you notice the same effect. The effect might be more pronounced on my '04.

 

[DFWDieselJet]

Before changing your thermostat, add these XGAUGES and see if your Coolant Temp doesn't look better:

GAUGE TXD RXF RXD MTH NAM Notes
Trans Temperature 686AF10105 031A04410505 2808 00090005FFD8 TFT F
Coolant Temperature 686AF10105 031004410505 2808 00090005FFD8 WTR F

From -
http://www.scangauge.com/support/pdfs/XGAUGE.pdf

 

[AudiLikeA4]

Err... sounds like you got a perpetual motion machine on your hands. That's what it might appear like on the SGII but how does it translate into real life/mpg?

It translates to driving 70-80mph and averaging 45+mpg on the tank, as opposed to driving 55-65mph and getting 45mpg on the tank. I get places faster and it doesn't cost more money, it's also a much more enjoyable commute due to the torque giggles. The cost of the SGII was worth the knowledge gained about how my engine operates.

I watch the current trip gauge while driving as it is more accurate over each trip, but changes fast enough to allow me to change my habits to the terrain. On an uphill with an equally graded downhill I will lose say .2-.4mpg, but P&Ging downhill I end up .4-.8mpg better than I started. I also use the downhill for some free accelleration to try and keep the momentum for a bit longer. I get about 1.5-2x the fuel on the downgrade, constant fuel usage is 1.3gph accell uphill at 2-2.3gph, now coast downhil at ogph continuing with the momentum further than you have to power up. If you stayed constant you'd lose speed and use more fuel. Hill is (A+B) the 2 halfs, then you have distance divided by speed to find time(C/D), so (1.3gph+1.3gph)=2.6gph*(1 mile/65mph)=.015, so 2.6*.015=.039gal for a 65mph average, (2gph+0gph)=2gph*(1 mile/70mph)=.014, so 2*.014=.028gal for a 70mph average over the same bridge....Now yes this is a very loose equation as the number of actual factors involved would take too much of my free time to put together and compute. But you can see how this can work, it depends on the variables but the math can work. The speeds are different because I am powering uphill...maybe not jack flash but enough to keep me accelerating as opposed to using the same amount of fuel. I'm just using less fuel on the downhill than I used on the uphill as compared to if I stayed constant the whole time. I know there are points that can be picked apart by the vultures but this is a hypothetical to explain how the extra mpg come to be, the short version.

I feel like I didn't explain this as well as I'd hoped so don't put me on the firing line if I screwed up.

 

[AudiLikeA4]

I understand pulsing and gliding.
But not how someone may "recoup double the cost" of powering up a hill by coasting down it.
There is no give and take when it comes to fuel consumption, only give. We give the engine fuel, it never gives any back.

The engine gives you distance traveled, you give it fuel it gives you distance, when you give it less fuel over the same distance you can recoup the cost, see my explaination in the post above, hope it makes sense.

 

[WriConsult]

Thanks for all the posts guys. I did a hwy trip this weekend and now have a couple tanks of reality against which to verify the SG.

First tank (all city driving) yielded 41mpg actual, vs. 47.0 indicated.

Rather than adjust the SG correction to +14%, I adjusted it halfway. This is a trick (referenced in dremd's thread above) that I learned on the Subaru to avoid having the correction bounce wildly back and forth from tank to tank. So now I'm at +7% correction, which means it adjusts consumption up by 7% and mpg down by 7%.

As expected, hwy driving yielded even worse mpg readings than before. A 480 mile trip to north of Seattle and back gave an indicated 35.6mpg trip average on ScanGauge. Actual mpg was 47.2 for this tank. We filled up at the same tank at the very beginning and end of the roadtrip so that we could clearly bracket the highway mileage.

This leads me to believe that there is nothing wrong with the car, and I'm not going to tear my hair out trying to figure out what the "problem" is. That 47mpg tank was at 65-70mph, in heavy rain with lots of standing water northbound, and into a strong wind southbound. In those conditions, and on B50, 47mpg is as much I could hope for even from a TDI. Dremd, I'm thinking you're dead-on in your assessment, and it sounds like we're not the only ones. On this vehicle, the SGII appears to over-report fuel consumption on the highway and under-report it in city driving, and the swing between the two appears to be on the order of 30% or more.

Since most of my driving is urban interspersed with occasional all-highway roadtrips, I'm going to adjust the correction to report as accurately as possible for local driving. On roadtrips I'll just have to make the appropriate mental correction as I drive.

 

[WriConsult]

Before changing your thermostat, add these XGAUGES and see if your Coolant Temp doesn't look better:

GAUGE TXD RXF RXD MTH NAM Notes
Trans Temperature 686AF10105 031A04410505 2808 00090005FFD8 TFT F
Coolant Temperature 686AF10105 031004410505 2808 00090005FFD8 WTR F

From -
http://www.scangauge.com/support/pdfs/XGAUGE.pdf

Thanks for the tips. My SGII is pre-XGauge but I plan to upgrade it soon, and I'll add those once I get the chance. Meanwhile, I'll watch IAT and see if I can figure out a relationship.

Wow, IAT in the 40s yields best MPG? I don't doubt you -- it's probably just that diesels are such different animals than gas engines. Lots of things for me to unlearn here. From what I hear most gasoline engines get better mpg at higher temps (at least up to the point that you have to turn on the A/C!), and that's what I've observed on the Subaru. My mpg goes in the toilet as soon as it starts getting below 50-60 degrees out, and this makes a bigger difference than winter E10 fuel or snow tires.

 

[DFWDieselJet]

Wow, IAT in the 40s yields best MPG? I don't doubt you -- it's probably just that diesels are such different animals than gas engines. Lots of things for me to unlearn here. From what I hear most gasoline engines get better mpg at higher temps (at least up to the point that you have to turn on the A/C!), and that's what I've observed on the Subaru. My mpg goes in the toilet as soon as it starts getting below 50-60 degrees out, and this makes a bigger difference than winter E10 fuel or snow tires.

That seems to be related more to emissions control instead of that conditions inherently provide the best economy. I suspect the effect is stronger on late-model TDIs. Apparently some of the aftermarket tunes add a lot of that timing back in, and provide FE that better mathces what you're seeing with your gasser. I'm planning on an Aligator tune soon, I'll let you know if that changes.

 

[EddyKilowatt]

WriConsult -- My ScanGauge has been calibrated over a couple dozen tanks and its predicted MPG vs actual is always within a few percent. But I can report the same thing as you... I can easily show 50 mpg around town (boulevards and suburban streets, not dense gridlock traffic), then 40 mpg on the freeway at 70 mph.

My initial thought was something like yours, a non-linearity in the reported vs actual fuel curve, but after a few all-highway trips I have to say I don't think ScanGauge is mis-reporting highway fuel flow rate. (I've never run a full tank strictly around town so can't comment there.) The ECU needs to keep pretty accurate track of injected quantity, so I think the big non-linearities must already be compensated for inside the ECU... I doubt they could meet emissions rules with a 30% mismatch between commanded and actual fuel flow rates.

I think we're really seeing the difference between good efficient in-town driving, where drag is low and there's good opportunity to recapture kinetic energy (by coasting a couple blocks before every stop), versus highway driving where it is just you versus rho-vee-squared (aka aero drag) for hours on end, and there's not much you can do about it from the driver's seat. I'd be happy to be proven wrong on that, however.

BTW, if you haven't come across it yet, here's the TDI engine fuel consumption map, courtesy of TDIClubb'r "Uponblocks":

Eddy

 

[jettawreck]

I agree with Eddy. My SG is accurate to within a few percent versus actual fill amounts usually. Even though I average over 50 mpg on my work commute (except now when its -20F half the time) the average is enhanced by some great mpg thru "town" slow speed driving. But my "current average" driving with resets between hiway/town seem to be right on.

 

[McBrew]

WriConsult, you really will need a few tanks on the SG to bring the accuracy into a realistic range. I am assuming you set the engine size and the fuel reporting to DieselA, right?

Anyway, what you report is what most of us experience. I can drive on 30-40 MPH roads to the grocery store and get somewhere around 70 MPG. On the highway, this drops off dramatically. I see a big drop between 40-60 MPH, but not as much of a drop between 60-80 MPH. I think maybe the engine is just running a bit more efficiently in the upper power range.

My "city" driving is a lot different than someone in NYC. The two "cities" I drive in the most are Annapolis, MD and York, PA. Neither of these towns see much stop-and-go traffic. Speeds are generally int he 30-40 MPH range with relatively few traffic lights (compared to NYC, Washington DC, or Baltimore, for example). I can easily keep my MPG numbers above 50 in town.

On my daily commute, my mileage constantly goes up to around 55-65 MPG (depending on traffic), until I hit the highway (65 MPH), where it steadily drops to around 50 MPG. Then I get off onto a 50 MPH highway for about 4 miles. I go through about 5 traffic lights, and it has a number of hills. Even so, my mileage actually goes up during this portion of the commute.

I don't go through much stop-and-go traffic... but we had a number of bridges closed (for about 5 hours) the other evening, and I spent a LOT of time in slowly creaping traffic. I really wished people would have just stopped... there wasn't anywhere to go until the bridges re-opened! Anyway... I watched my mileage steadily drop to a low of 26 MPG!

Oh, another thing... are you watching the real-time mileage, or the trip average? If you have a new SG, set up an xGauge that shows your trip average along side the real-time numbers. I ahve noticed that I subconciously look at the real-time numbers when I am getting low mileage (going up a hill, accelerating, etc). The trip average is evened out quite a bit.

If you were getting over 30 MPG ina Subaru, my hat is off to you! A few friends of mine have Subarus (legacy, forester, and impreza) and they all get around 20 MPG.