Mark P

I've been having an issue on my DBW 1.8T for a while now, and after much debugging/logging/etc, I need some additional advice.

<b>HIGH LEVEL SUMMARY:</b>

Shortly after I start my car, it reports that it's running excessively rich, it then pulls fuel to compensate. At this same time, it will issue a fault code saying there's a Pressure Drop between the Turbo and Throttle Valve. Somewhere along the way, the car goes into Limp Mode, and then finally after pulling as much fuel as it can for a bit, it issues a fault code for System Too Rich.

The "catch" to this story is that after all of this happens, if I Clear Codes, then the car runs perfect. I can drive for hours without any of the above reoccurring. I can park the car, start again, no problem. Most times, I can park the car for hours at work, and start up again with no problems. However when I leave it for 12 hours overnight, then I have the issue again, every morning.


<b>HIGH LEVEL SUMMARY: (now with some tech detail)</b>

Shortly after I start my car, it reports that it's running excessively rich (VAG block 31:1 reports Lambda at the minimum(rich) possible value of 0.75), it then pulls fuel to compensate and gets to -25% LTFT (VAG block 32:2) shortly thereafter. At this same time, it will issue a fault code (17705) saying there's a pressure drop between the Turbo and Throttle Valve. Somewhere along the way, the car goes into Limp Mode, and then finally after being at -25% LTFT for a bit, it issues a fault code for System Too Rich (I forget exactly which #, so I'll have to check it again)

Also, if I clear codes very early in this process, the problem will still continue to occur. If I wait at least until the LTFT are almost to -25%, and clear codes then, the problem will not come back. That is, AF Lambda does not report rich, LTFT do not dramatically adjust, and I don't get either fault code.

It certainly seems like the problem is specific to how the car is operating when it's "cold" (ie off for a long period of time). I have had this issue in both summer and winter, so it's not specific to ambient weather temps.


<b>DETAIL:</b>

<b>the 17705 code:</b>

The exact code is:
17705 - Pressure Drop between Turbo and Throttle Valve (check D.V.!)

<b>I was getting this code when the car was Stock as well:</b>

I was getting the Pressure Drop code (17705) when the car was Stock and also now that I've installed bigger turbo parts.

When I was stock, I didn't think much about it, as I was soon going to replace the stock turbo with a GT28R, and replace most of the associated hoses at that time. Therefore, I had thought the code might go away. I didn't observe any problem caused by this code (didn't go into limp, etc), just the code itself. After changing the Turbo, the Throttle Body, and most of the hoses in between, I still had this code.

<b>replaced parts:</b>

When upgrading the turbo, I basically bought a used APR Stg3 turbo/manifold/hose kit, from a few different source cars. The injectors and software are different, but almost all of the connecting pieces are from a well made APR kit. The intercooler is a TyrolSport side mount.

In addition to changing the turbo kit parts, I have also changed:
- MAF sensor
- Throttle Body
- Primary O2 sensor
- DV
- ECU
- Coolant temp sensor
- plugs (changed and gap'd)

I would get this fault code with both the stock ecu, as well as on custom software. I also tried custom software on a 2nd ecu, and still had the issue. Therefore, I don't think it's specifically software or ecu related.


<b>Cold start details (O2 and LTFT):</b>

After the car has been sitting overnight, I start it and monitor various items in VAGCOM, I note the following:
Air/Fuel (Lambda) from the primary O2 sensor first reads 1.00 (block 31:1) when the car is started. It stays this way for some period of time. I believe this is a way that the ECU "masks" the real reading as the O2 sensor is warming up, in an effort to not react to incorrect readings. The desired AF Lambda is a bit richer at this time during startup. At about the same time the RPMs settle, the desired A/F goes to the expected 1.00. As I start to drive, this time period (of when Actual A/F is pegged at 1.00) expires, and then I see the lambda value start to drop. Within a few seconds, it now reads the minimum value of 0.75 (rich). [at this point, if I check for codes, I will already have the 17705]. Since the car is reporting very rich condition, I will then observe the short term fuel trims (block 33:1) will be at -25% as they are trying to remove fuel to run at stoich. While those short term trims stay at -25%, they then impact the LTFT. The (block 32:2) make their way to -25% as well. At/around this time, the car goes into limp mode and also throws the System Too Rich DTC.

Here are some graphs showing the O2 reporting 0.75 rich after cold start, and the follow-on LTFT going to -25%:
<img src="http://ctny.audiworld.com/mark/golf/misc_data/17705_coldstart_o2_ltft.gif">

and here's one right after I cleared the codes. Now everything is all fine and dandy - strange...
<img src="http://ctny.audiworld.com/mark/golf/misc_data/17705_codeclear_o2_ltft.gif">


<b>Cold start details (MAF Sensor, MAF Calculated, and MAP):</b>

One other item I looked at was the MAF values. I logged block 211 to compare the "MAF Sensor Readings" to the "MAF Calculated Values". What I found was that after a cold start, the Calculated values were about 1.8x higher than the Sensor reported values. After I cleared codes (and my problem went away), I would then find the Calculated values were only 1.3x higher than the Sensor reported values. (side note: I don't expect the Calculated and Sensor values to perfectly match as the software I'm running didn't alter the MAF curve for the larger housings, but rather adjusted the inj pulse as needed). However, the fact that it changes from a 1.8x relationship to a 1.3x makes me think somehow this is related. I swapped out the MAF sensor just to see if that was the issue, but no luck. So, it might be due to a part that deals with the Calculated MAF value.

<b><font color="red">Does anyone know what sensors/info the car uses to *calculate* MAF as shown in block 211:2? Maybe some combination of MAP and RPM, and maybe load? other?</font></b>

I logged MAF Sensor and MAF Calculated, both while the car was in this cold start condition, as well as much later when it was running fine with no issues (after I had cleared codes).

Here you can see the relationship between Sensor and Calculated is drastically different depending on whether the car is in that cold start condition, or later on after code clear:
<img src="http://ctny.audiworld.com/mark/golf/misc_data/golf_maf_act_vs_calc_cold_hot.gif">

(the log is from casual highway cruising)

As noted above, I did swap out the MAF Sensor, to ensure the difference wasn't due to the sensor. The results were identical afterwards as well.

I had wondered about the MAP's role in this, so I decided to log Boost(MAP), MAF Sensor, and MAF Calculated. I did so both while the car was in this cold start condition, as well as much later when it was running fine with no issues (after I had cleared codes).

Here is the graph of that data:
<img src="http://ctny.audiworld.com/mark/golf/misc_data/golf_maf_act_calc_vs_map_cold_hot.gif">

In the above graph, what I observe is that the relationship between the MAP and MAF Sensor is the same regardless if the car is in cold condition or hot (two blue graphs). Therefore, I don't think the MAP sensor is to blame as it is operating consistently throughout. However, sure enough, we notice that it is indeed the MAF Calculated that reports noticeably different between cold and hot conditions (two red graphs).

<b>What is column 3 in block 211?</b>

I have noticed that at the same time the ratios change between MAF Sensor and MAF Calculated, so does the value in 211:3. During the cold start problem (when the Calc/Sensor relationship is 1.8x), I note that 211:3 is around 0.98. Later on, when the car is hot, after I've cleared codes, this value transitions and becomes stable at about 0.72.

<b><font color="red">Does anyone know what the data in the 3rd column of block 211 represents?</font></b>

<img src="http://ctny.audiworld.com/mark/golf/misc_data/vagcom_block211.gif">


<b>OTHER INFO:</b>

- I have pressure tested and don't find any leaks.

- I suspected my Rear O2 sensor may have been malfunctioning due to some odd readings I was getting from it. I disconnected it from the exhaust to see if that would affect my above issues. It didn't appear to make any impact. I didn't think it would as the rear O2 should just be for emissions, but I wanted to ensure it wasn't related. I haven't changed out this rear O2 yet as it doesn't seem to be causing any problems.

<b>SUMMARY:</b>

It certainly seems that my problem is with some part failing, but hasn't failed completely. It seems to only be malfunctioning when "cold". I don't think, for example, I have a leaky hose as I would expect that to give this problem all the time, not just after it's been sitting for 12+ hours.

I feel like the issue is related to whatever *calculates* air pressure and likely also the MAF. I think some sort of sensor is not operating properly when cold, and this is leading to an incorrect air flow calculation. This incorrect value is then the root cause for two issues which happen in parallel. First, it causes the 17705 code as the calculated pressure doesn't match to the MAP sensor. And Second, and separate from that fault code, it also screws up the car into thinking that more air is entering than actually is. This causes more fuel to be added, and thus the car runs rich.

If the above paragraph is true, that also means that the car must be using the *calculated* MAF values, at least when the car is cold. Maybe the system does that for a while at startup to ensure the MAF sensor is warmed up before using it for MAF data - similar to how the O2 Sensor is "set" to 1.00 at startup. ??

This 17705 code seems to suggest that there is a pressure drop between the turbo and the throttle body. Since there are NOT two pressure sensors, the car must be doing some other sort of calculation to determine that the pressure is so much at one spot, and something different at a different spot. My guess is that it's using the MAP sensor for one of those, and perhaps the MAF (along some sort of other info) as the 2nd. I tried changing out the MAF, and that didn't help. I haven't changed the MAP, but haven't noticed anything else in the logs that would suggest the MAP is failing. Perhaps there are some other sensors that come into play here, but I don't know what they are. Maybe whatever is providing input to the Calculated MAF Values (block 211) might also be an input to the calculation that causes this 17705 DTC?

If anyone has any other thoughts to what I should be considering, please let me know. I'd prefer to not randomly purchase more parts if I don't know mine are bad, but I really want to get to the bottom of this. I don't think it's a leak or torn hose as I would expect that to happen every time I started the car (not just when the car has sat for 12+ hours), and I also would expect it to happen after code clear as well, which this does not.

Also, if anyone has suggestions of other things to log to help point to the problem, please do share.

Maybe I'll give a prize to the winner ;-)

Thanks for reading,

AVANT S4 KING

I found a leak, that I didnt find when I pressure tested the car warm only a few days earlier, by letting my car sit overnight and testing the next morning. I could have been a 'new' leak, but with as few miles as I put on my car between tests, I really doubt it.

Mark P

Thank you - there's a good idea. 1 point for you! ;-)

VR_Life

STFA for Breeze CT Clamps

Ochlockonee

"Does anyone know what sensors/info the car uses to *calculate* MAF as shown in block 211:2? Maybe some combination of MAP and RPM, and maybe load? other?"

Many ecus have volumetric efficiency tables in them in some form or another. With a measured baro (it can sample map at key on engine off and maybe even at idle if the sensor is upstream of the throttle body)at a certain rpm with a given throttle angle and a general neighborhood of ignition timing the ecu should expect to see a certain amount of airflow. The ecu doesn't seem to be seeing what it wants at the maf and it's setting this code with a somewhat misleading definition.

Cars with maf sensors but without baro or map sensors actually use these tables to calculate barometric pressure using the maf sensor signal at WOT.

There may be a problem with idle air control at cold starts. This may just be a case of a throttle body malfunction. I'm sure vag-com has some measuring blocks for idle speed, airflow and air control that may help check to see if the throttle body or IAC valve is keeping up at cold starts.

[badmoon]

"I've been having an issue on my <b>DBW</b> 1.8T for a while now"

-SlowSilverTIPJason-

No misfires though huh? I'd expect a good number of misfires if it is in fact running that rich. Could also be some water in fuel return line that is freezing with our cold weather lately, and causing your FP to be higher than it should. After running for a while and getting warm fuel returned the tank the line defrosts and FP returns to normal. Do you have a way to check your fuel pressure?

Mark P

FYI - I have logged/watched the baro sensor (block 113:4, I believe) and that reads as expected (980-990mbar).

When I searched for this code a while back (seems nobody has any solid root causes for it), I found someone that did associate it with a plastic one-way valve of sorts. I don't recall which one off hand, but perhaps this is similar to what you mention as an IAC Valve?

Thanks again for the suggestions

Mark P

I updated that after he asked ;-)

Mark P

I haven't had any during the first few minutes that the car is like this before I clear codes. Of course, during that time, I'm not getting on it much at all, which is when I would think misfires would be more likely. But in general, no I haven't had misfire codes during these startup conditions (first ~5 minutes of driving)