Bad Mileage until unplugging battery, after 1 tank, bad again

I have an annoying problem that I cannot seem to figure out. I have an 04 BEW TDI w/ 5spd. The thing gets about 35mpg mixed, MAYBE 39 hwy (and I mean ONLY hwy). What I found was if I unplugged the battery, it would immediately go up to 41mpg mixed and 45hwy. After half a tank or maybe 1 full tank, the efficiency falls back down to 35mpg. I drive consistently, and am not particularly hard on it. Sometimes I'll step on it to get out of the way, or go 80+mph if I'm in a huge hurry, but that's rare. I have the Ross-Tech VGCD cable, and have confirmed that my timing is .5 - 1 (it seems to range with different engine speeds/loads but gravitates to .5).

I have checked for:
dragging brakes
performed cage mod
air filter
cleaned out maf w/ correct spray
had recent glow plug recall
have no ecu codes
no alignment issues
changed tb O-rings
snow screen (don't have one)
replaced temp unit w/ OEM green one, temp registers at 90C during operation

The airfilter and fuel filter have been changed recently.

Can anyone help me diagnose the problem? Like I said, I have the VGCD, but don't really know what to look for.

THANKS!

You definitely have a puzzle... and I'm interested in following this/your THREAD to a final conclusion/fix!
For my $0.02 worth, I think this is "software", NOT any physical component failure/problem!
IAW, look to your ECM/ECU's algorithm(?)...
because I think your unplugging the battery forces it to relearn the best MPG routine and in going through this learning process,
where it is finally settling on a routine that is not as good as its learning algorithm??
I hope I did not confuse you folks!
OK, you Guruh(s)... confirm in clearer terms or blow me out of the H2O!
Sam Ross
[edit]What is "VGCD"... the same as eBahn DVD software repair manuals??

Shouldn't the software be the same as what was programmed into it when they re-flashed it after my Glow Plug recall? Could there be multiple algorithms that exist within the ECU, perhaps a specific one that controls fuel delivery learning?

Are these actual tanks calculated after fill-ups, or are you just watching the MFD or some other Scan Gauge type devise?? One tank or so isn't meaningful anyways, but disconnecting the battery just resets MFD program.

Is your 04 still on the original battery? Maybe your alternator is working harder to maintain a weak battery. This is just a wild guess and wishful thinking on my part since I just replaced my OEM Varta. The new battery never goes below 12.7 volts after sitting for 18 or more hours, whereas the Varta would go down to 12.3. My rationale is that the alternator now puts less stress on the engine.

I have verified this with repeated fill-up calculations, many of which are at the same pump. I think someone has already Vent-ectomied the car, there is a little hole rather than a button at the bottom of the filler opening. I have verified good fuel efficiency on the highway using the VGCD data. It came in at about 45mpg, and was consistent with the calculation I made filling up after .5 tank driving all hwy.

I have several VGCD logs with different variables logged. I will use these and compare them with the same variables when I'm getting low efficiency again (which I don't doubt will be soon). Maybe if someone else out there has a VGCD, they can maybe send me a similar log? That way, I could compare it to a healthy TDI. This may help weed out trouble with something like a bad MAF...

On another note: for some reason, my VGCD data stops logging after 250 seconds, is this an adjustable time frame, or is it fixed to only log for 250 consecutive seconds?

PS: Battery is new as of this summer, and the old one never really gave me any trouble until it just failed.

JB05 said:

Is your 04 still on the original battery? Maybe your alternator is working harder to maintain a weak battery. This is just a wild guess and wishful thinking on my part since I just replaced my OEM Varta. The new battery never goes below 12.7 volts after sitting for 18 or more hours, whereas the Varta would go down to 12.3. My rationale is that the alternator now puts less stress on the engine.

Click to expand...

That would be very small and I doubt you could accurately chart it.

IDoSeaDoo said:

I have verified this with repeated fill-up calculations, many of which are at the same pump. I think someone has already Vent-ectomied the car, there is a little hole rather than a button at the bottom of the filler opening. I have verified good fuel efficiency on the highway using the VGCD data. It came in at about 45mpg, and was consistent with the calculation I made filling up after .5 tank driving all hwy.

I have several VGCD logs with different variables logged. I will use these and compare them with the same variables when I'm getting low efficiency again (which I don't doubt will be soon). Maybe if someone else out there has a VGCD, they can maybe send me a similar log? That way, I could compare it to a healthy TDI. This may help weed out trouble with something like a bad MAF...

On another note: for some reason, my VGCD data stops logging after 250 seconds, is this an adjustable time frame, or is it fixed to only log for 250 consecutive seconds?

Click to expand...

VGCD?? Explain please.

You said you started noticing the economy loss after the dealership did an ECU flash during a glow plug recall visit?

Sounds like a software issue.

Mileage has always sucked

By VGCD (or maybe it's VGDC) I refer to the VAG-COM Diagnostic Cable, which I use to plug into a laptop and read the vehicle's various component parameters in real time.

The poor mileage I noticed it on day 1 of getting the car. This is my second PD TDI, and on the drive back from Atlanta, on my second fillup, I noticed it was only getting 40mpg. Sad was this little green-machine driver , I was hoping that my previous TDI had something wrong with it, and that this one would get closer to 50mpg....

The PD BEW doesn't get the same MPGs that its older brother the ALH gets. Sad, but true.

Have you checked your MAF accuracy and logged your turbo boost requested v.s. actual?

Have you checked your engine temps? You may just need a new thermostat, or coolant temp sensor.

I saw a 4mpg increase when after I installed a new thermostat, the old one was opening up to early and the engine was not getting hot enough. Ignore the dash thermostat, use VCDS

IMHO, I think you should put the battery issue out of mind and check the items above.

The PD BEW doesn't get the same MPGs that its older brother the ALH gets. Sad, but true.

Have you checked your MAF accuracy and logged your turbo boost requested v.s. actual?

Have you checked your engine temps? You may just need a new thermostat, or coolant temp sensor.

Click to expand...

Yea, the ALH does get a bit more bang for the buck, but since I use mine for a bit of towing, I figured the extra bit of power would come in handy (and it has). However, my previous car was exactly the same as this one and with the same driver, consistently got 41mpg or higher.

I have not checked the thermostat itself, but have changed out the temp sensor to an OE green sensor from IDparts. No change. VCDS logs operating temp at 90C, is that what it should be?

Will check the MAF requested boost vs actual tonight. Thanks for the tip.

Drive it harder. Sounds to me like your easy driving trends the ECU's fuzzy logic to the far end, and when you mess with the battery you zero it. You could do the same thing in VCDS (just go to fault codes, and clear them, even if none are present...it should do the same thing).

I'd also check the smoke load before and after, see if it changes, by doing a floorboarded acceleration from a standstill in 1st gear to redline (at engine operating temp).

It is quite amazing not only how much better these cars will run, but how much less fuel they will use overall, if you keep 'em blown out!

I'd also check the smoke load before and after, see if it changes, by doing a floorboarded acceleration from a standstill in 1st gear to redline (at engine operating temp).

It is quite amazing not only how much better these cars will run, but how much less fuel they will use overall, if you keep 'em blown out!

Click to expand...

Heh, I've never known about this concept before... guess after only a year, I'm still a noob at diesels Is there a way I can reference what the smoke value should be? (haven't really explored those parameters with VCDS).

How often would you say I should red-line the smoke out?

If it belches out a cloud you can see from your rearview mirror, then does NOT after you reset the ECU, you've been driving it too easy.

My TDIs tag the redline every single day, at least once.

" IDoSeaDoo ",
For a moment let me go back to your Orig.POST ("OP") and
ask you to clarify some of what you initially said in your necessarily lengthy "OP":

IDoSeaDoo said:

I have an annoying problem that I cannot seem to figure out.
I have an 04 BEW TDI w/ 5spd.
The thing gets about 35mpg mixed, MAYBE 39 hwy (and I mean ONLY hwy).
What I found was if I unplugged the battery, it would immediately go up to 41mpg mixed and 45hwy.
After half a tank or maybe 1 full tank, the efficiency falls back down to 35mpg.
I drive consistently, and am not particularly hard on it.
Sometimes I'll step on it to get out of the way, or go 80+mph if I'm in a huge hurry, but that's rare.
I have the Ross-Tech VGCD cable, and have confirmed that my timing is .5 - 1 (it seems to range with different engine speeds/loads but gravitates to .5).

Click to expand...

This is what made me immediately jump on the likelihood that this is "software" driven!
For further clarification please answer my seemingly dumb Q(s) that follow:
==============================
Q1 - In your own words...
what led you to "unplug the battery"...
IOW why did you do this...
what were you hoping to accomplish when you first did this?
Q2 - How long typically have you left the battery "unplugged"?
Q3 - Have you ever done this by disconnecting both battery terminal/cables...
and grounding them together for however long you have been doing this?
==============================
Regards,
Sam Ross

Q1 - In your own words...
what led you to "unplug the battery"...
IOW why did you do this...
what were you hoping to accomplish when you first did this?
Q2 - How long typically have you left the battery "unplugged"?
Q3 - Have you ever done this by disconnecting both battery terminal/cables...
and grounding them together for however long you have been doing this?

Click to expand...

A1:After the glow plug recall, the dealer said they "reflashed" my ecu. I noticed an immediate increase in MPG's and thought they had fixed something in the software. Then I noticed that it had gone back down and stayed down. Knowing that the car has "adaptive" software subroutines built-in, I thought that maybe clearing those would get me back to default settings. From previous experience, the only thing I knew that cleared re-writeable ECU parameters is disconnecting the battery.

A2:at least 30 seconds

A3: both cables have been disconnected when battery was removed, however, when I do it, I just pull the negative cable off. never tried grounding them together. would doing so do something special?

IDoSeaDoo said:

================clip==
A3: both cables have been disconnected when battery was removed, however, when I do it, I just pull the negative cable off. never tried grounding them together. would doing so do something special?

Click to expand...

touching the disconnected cables together will discharge certain capacitors, resetting some sensors/devices (not sure which ones). This has been known to cure certain odd issues.
There have been a lot of threads about the dealers installing the wrong flash when performing the recall

It is quite amazing not only how much better these cars will run, but how much less fuel they will use overall, if you keep 'em blown out!

Click to expand...

Oilhammer, I like your smoke concept, and will def. try it, but how do you figure that is affected by the ECU and unplugging the battery?

IDoSeaDoo said:

A1:
After the glow plug recall, the dealer said they "re-flashed" my ecu.
I noticed an immediate increase in MPG's and thought they had fixed something in the software.
Then I noticed that it had gone back down and stayed down.
Knowing that the car has "adaptive" software subroutines built-in,
I thought that maybe clearing those would get me back to default settings.
From previous experience, the only thing I knew that cleared re-writeable ECU parameters is disconnecting the battery.
...

Click to expand...

"IDoSeaDoo",
I think you/we are dancing around the answer to this puzzle...
I think the dealer might have "botch" their "re-flash" of your "EPROM" that stores your ECM/ECU's adaptive algorithm...
or something that sounds like this in "GEEK" language!?
-----------

BobnOH said:

Touching the disconnected cables together will discharge certain capacitors, resetting some sensors/devices
(not sure which ones).
This has been known to cure certain odd issues.
There have been a lot of threads about the dealers installing the wrong flash when performing the recall.

Click to expand...

Bob - My highlights in your POST above help to substantiate my thinking on this THREAD's most interesting subject/issue!!
[edit]... I think it's time to:
(a) ask the dealer to conduct the re-flash recall of the car's ECU/ECM... what could this hurt?
(b) also [but first ] to run a diagnostic on the ECU/ECM if such can be done!?
(c) Q- Can a Ross-Tech VAG-COM do a general diagnostic on the ECU/ECM's central processor chip?
Sam Ross

An ECU reflash either works or it doesn't. The software check happens several times. If it starts and runs and there are no faults, the flash is taken fine. It is not like setting the gap on breaker points.

Since they recently updated the software for these cars, there will not be anything newer. They can only check the software version number, and if it is already the same, that's it, nothing more can be done. The dealer cannot reflash an ECU with the same software (although the tuners are able to wipe the ECU clean and start over).

Sometimes newer software DOES change things in such a manner that the fuel economy goes down, if the driver continues to drive in the same manner as before. Remember, emissions are not the same as fuel consumption. A different EGR map to improve driveability could just as easily knock 2 MPG off the car. Different shift patterns in cars that have autoboxes can also do similar things. It happens more than you think. I've flashed more cars than I can remember, Toyotas, Ford/Mazdas, GM, ChryCo, MB, VAG, etc. Common practice nowadays.

what are we saying here?

Are we saying that the re-flash screwed up my mileage on purpose? 6mpg's for driveability is really hard to believe. Why hasn't anyone else had the same problem? No, I do not believe this is the way it's supposed to be, for emissions or any other purpose. Something is wrong with this car, it's getting 10mpg less than it should. I'm afraid to take it to the dealer b/c of how shotty their workmanship is. I've seen how they did the glow plug harness, and they left out parts when they did the egr cooler... has noone out there had a similar problem??

I'd bet ya my first born if I drove that car for a month, it would get consistently right where the EPA said it should. Tank after tank after tank.

I've taken you up on your advice and started driving it harder. Due for a fill-up today. Would you say I should "reset" the ECU again, or should I just continue and see what it is going to be? I reset it at the beach and drove about 150 hwy miles, then I filled up again (getting 46mpg) and am about to fill up again. I expect no higher than 38 this time.

I agree with the hammer, if there is a problem, it' likely one of various devices sending bad input to the ECU.
Do not attempt to reset or load default values using VCDS.
If condition persists, do monitor various setting with VCDS, like IQ, timing, temps. etc. before and after the battery removal and see if something is changing

Filled up today and surprisingly got 39. Have been driving it a bit more spiritedly for about 100 miles. This was pretty much all highway, so I'm not thrilled. I shift around 2.5k or more, and cruise at around 1.8-2k. Will try to drive it like that, taching out once per drive in 1st gear and note the results. When I get some time, I also want to check my MAF readings. Still eagerly welcoming suggestions from all you TDI techheads and gurus

"I shift around 2.5k or more"

If that is the way you drive it I must be pounding the crap out of mine. Still get city/hyw 40.5 mpg. I shift over 3 90% of the time, over 4 alot.

39mpg

Driving harder these days, last fill up got 39mpg. Though better, I'm not convinced that that's what it's supposed to be getting. Can anyone let me know what the maf reading should be at level cruise? (give some other readings like engine speed, torque demand, fuel economy, etc.

I'm starting to suspect the MAF...

FWIW I'll record some data for you tomorrow on the way to work; engine speed, torque demand, fuel economy. Let me know if there's anything else. I have an auto so you'll have to account for that somehow. It's 95% highway, ~65 mph.

You also might want to take a look at this:
<http://forums.tdiclub.com/showthread.php?t=123766&highlight=deoxit>