I have this strange problem on my 1999 Volvo V40 1.9T T4 (B4194T engine, no VVT, ~150k miles), I'll try to provide a detailed description.

I noticed that only when the engine is warm, whenever I drive at some RPM and release the throttle along with pushing in the clutch to get the engine to idle, often the RPM quickly drops down and below it's normal idle of 750 rpm and then jumps a bit and recovers to the normal value. The other times it drops more slowly and smoothly and there is no dip. This actually caused the engine to die in a couple of occasions (days were hot, as far as I remember), but it's extremely rare. The dip in idle is accompanied by slight dimming in lights and a lower noise of fuel pump, which are probably caused by a drop in voltage.

After opening the hood and operating the throttle by hand I noticed that the engine pretty much always does this after revving and the dip in engine speed is always accompanied by knocking a few times. The number of knocks varies from day to day, sometimes it's a couple, sometimes it's closer to 10. After connecting a scan tool to check the ignition timing values I noticed that the knocks cause ignition timing advance value to go negative for a moment (~-10%), while usually they are kept positive. I assume this is what causes the revs to dip, as detonation is detected and timing is retarded (NOTE: I do not have enough experience in engine diagnostics to know whether this is normal or not, maybe all engines have their timing retarded when RPMs drop). So basically it goes like this:

  • I increase the throttle and hold it
  • RPM rises
  • I release the throttle
  • RPM continuously falls
  • At around 1000-1500 RPM often an audible knock occurs a couple of times
  • At the very same time negative ignition timing advance is observed for a brief moment
  • RPM keeps falling below normal idle of 750 RPM
  • RPM recovers to normal idle

Does anyone have any idea what's going on here? What causes the engine to knock when decelerating? Any insight is appreciated!


To make it clearer - here is a graph I obtained via OBDII of this phenomenon occurring. I pressed the throttle, held it a bit and released suddenly a few times. This was done in a standstill at normal operating temperature of 90 °C. You can see after every fall of engine speed there is a small oscillation where the speed dips below normal idle of 750 RPM, recovers a bit higher and settles. I assume the oscillation is the effect of the IAC valve working, but the dip is definitely not normal. You can actually hear the engine struggle a bit during each of those.

I noticed that it is much less likely to happen when I simply blip it, while longer throttling and higher RPM pronounces it. If I do a hard and long pull at full throttle while driving, the drop is significantly lower, almost reaching 500 RPM. It's also a bit different day to day. Some days there are many knocks and the dips happen much more often, other days it doesn't knock or knocks once or twice and the RPM dip doesn't occur (but I can still hear the fuel pump slow down and lights dim, for some reason).


You may find a high resolution version of this graph by clicking HERE.

  • Do you have the fuel trims?
    – Zaid
    Jul 18, 2016 at 15:25
  • @Zaid don't currently have the charts, but I'm getting a +5% positive LTFT. STFT usually oscillates between 5-10%, can pull out some real time charts if they'll help. No vacuum leaks, never really figured out what causes the positive trims. Might be a leaky exhaust manifold gasket, though I haven't managed to find any proof that it's leaking. Jul 18, 2016 at 15:32
  • 1
    I was thinking the lean condition would overheat the combustion chamber, and some aspect of returning to idle combined with faulty ECT readings would cause pre-detonation - possibly over-fueling being ignited on the edge of a valve relief or something. Jul 19, 2016 at 13:38
  • 1
    With STFT value between 5 - 10 % instead of the ideal 0% I would say that something is interfering with the fuel management system's ability to compensate for changing air flows. I would be surprised if the STFT has maxed out at 10%. Could the non-zero STFT value be indicative of a lazy lambda sensor?
    – Zaid
    Jul 19, 2016 at 15:09
  • 1
    @Ben The ECU does not support freeze frames (and I'm not sure it would capture any since no errors are registered), but I will post some trim and timing advance graphs this weekend, need to take it out for a longer drive. Will try capturing the knocking as well. Jul 21, 2016 at 12:13


You must log in to answer this question.

Browse other questions tagged .