Unknown reason for boost cut in high gear at 5000 rpm

[fivebanger12453]

Hi there,

I'm in trouble with some boost cut on WOT, I do not understand where it's coming from.
I attached a log with a WOT pull in 4th gear.

At 5000 rpm N75 drops to zero for a short moment and raises again afterwards, while N75 request stays at 90% without interruption

Seems to be reproducible, every time I'm in high gear at the same rpm.

Any suggestions where to look at?

[Crasher]

Also log the charge air temperature

[fivebanger12453]

IAT is logged, see attached file.

IAT starts at 30°C and ends up at 52°C at the end of the WOT pull.
When N75 drops to zero, IAT is at 44°C.

[Crasher]

Is that the Air Mass Meter sensor or the one on the manifold?

[fivebanger12453]

It's the sensor in the manifold.

[Crasher]

Over 40 degrees the ECU pulls the timing, from Self study program 111; it could explain it.

Ersatzfunction: erkennt die eine diagnose kurzschluss oder unterbrechung werden 40 degrees als ersatzwert angenommen der zundwinkel wid dadurch zeruckgennomen

Replacement function: detects a diagnosis short circuit or interruption 40 degrees are assumed as a replacement value the ignition angle is thereby reduced

[fivebanger12453]

From prjs definition I can see that the ECU pulls 1.5° timing from 30°C to 50°C IAT.
Load is also in relation to the IAT, but not that much that N75 should drop to zero.

You can see in the log that MAP request and N75 request are above actual. So why does N75 drop to zero?

[tekstil]

Try lowering "Base WGDC - pilot map" 5000 rpm and 87/112 cells value, maybe even at 4500rpm. I had same problem at 5200 rpm, N75 dropped to zero for a second and then came back up, but I never bothered to figure it out because I fitted RS2 turbo and remapped WGDC map. I think this is where it comes from, or it might be a bug in the files.

[fivebanger12453]

At the weekend I will install new knock sensors because I notice a lot of knock. If that does not help, I will follow your advice to lower the WGDC pilot map.

Thank you for your input guys!

[fivebanger12453]

Seemed to be a problem due to knock. I replaced the rear knock sensor, the front sensor was not reachable without dismantling the IM and I had no time for this yet. Noticed a lot of knock before, now cyl. 4 and 5 have no knock and cyl. 1 - 3 only some decent knock.

I did some 1/4 mile runs at the weekend and also some high gear pulls on the Autobahn. The car runs fine now without cutting boost.

[-ice-]

The reason was simply you hit the "map limit" threshold and this is why the N75 will drop to 0 then, after short timeout.

In other words your boost overshooted and it was limited.
One reason could be that the N75 base duty cycle was calibrated too high. This is why tekstil hinted to lower the N75 values. Lower boost will not trigger the threshold anymore.
Of course it would be also an option to increase the "map limit" map. You can do this, if you are sure the the boost peak does not bend your rods, turbo etc.

The knock sensors don't have anything to do with the issue (ok as far as i know, there is a safety feature which will totally disable boost control if the knock detection is not working at all). But this is a different behaviour.

The issue happens of course in higher gears... this is where you generate lots of exhaust gas to spin the turbo and generate high boost.

BR

[fivebanger12453]

This would make sense, but actual MAP was always below MAP request. This was the first thing I checked.

[-ice-]

Hi,

I talked about the map limit map, which has nothing to do with the map-request (this is your target map value)
As you can see in the stock "map limit" map for 5000 rpm, 270kPa are allowed. In your picture above (where unfortunately no rpm was plotted) we can see that you touched the 270kPa at the beginning.

There is no doubt if your symptoms were: driving in low load, maybe 3000 rpm, then hitting WOT in 4th or 5th gear, where boost suddenly builds up (then its likely to overshoot), gas pedal became "light" for short period (N75 =0) and then accelerates again (because boost was lower meanwhile).

In this particular case you have a options to fix:

- check if wastegate spring was tightend too much
- increase map limit map (only if safe for your engine!)
- lower N75 base map, so boost will not overshoot
- tune boost PID to not overshoot

BR

[fivebanger12453]

Indeed, at 3200 rpm I hit the MAP limit with a value of 271.
But dropping N75 to zero nearly 2000 rpm later does not make any sense for me.

Nevertheless all your explanations are right, thank you for that!

[-ice-]

Got your point. I also think this would be very late (too late, because the 270 kPa threshold is applicable for 5000 rpm not 3200).

The ECU will only drop N75 to 0, if it detects a safety issue. Too much boost is clearly one of them.
To reset this safe state it needs to leave the "boost mode" at least once. This is why after short period of time with N75=0 and dropped MAP, boost can raise again. This explains this typical gap in the plot for the MAP signal.

continuous knock can definitely lower boost. I mean lower it, not totally kill it with N75=0. This is why I think N75=0 is not a reaction to continuous knock. Correct me if I am wrong.

But, what we didn't discuss is: the ECU can also detect too low boost. Maybe it cuts N75 as well in this case. Never had this issue, can't tell you from experience. But in your first plot we see, that actual MAP is much lower then requested MAP. ECU might think sth. is wrong and protect the turbo from overspinning.
But this is just guessing. In this case you should have at least any fault codes about to less boost. A fix would be then to lower requested boost or to fix the engine to (nearly) deliver the requested boost.

As said, in most cases I have seen the overshooting issue with too much boost.