Let me briefly describe the problem and what I’ve already done — maybe someone will spot something or offer a suggestion?! I’m out of ideas at this point.

I’m the happy owner of a 2005 MERCEDES W203 1.8 Kompressor with LPG, currently at 317,000 km. The car starts instantly, accelerates well, doesn’t misfire or hesitate — generally everything seems fine... but it’s not!

Compression is even across all cylinders at 10 bar. I added 5ml of oil per cylinder, and compression rose to 11 bar, so piston rings are probably stuck. I also performed a leak test with compressed air — no significant pressure loss, so valves appear to be sealing well.

Camshaft phasers were replaced proactively, and an oil leak didn’t reach the wiring — ECU and lambda connectors checked. The entire timing set, including gears, was done 30,000 km ago.

I’ve been struggling for a long time with what I believe is excessive fuel consumption (Petrol ~14L/100km, LPG ~16L/100km — mostly city driving), lean mixture fault codes, and high LTFT at 25%.

The car starts fine cold or hot, even after long downtime. Acceleration is smooth, no jerking or hesitation.

Here's what I’ve done so far:

1. LPG off, drove on petrol only (~800km, mostly city) – didn’t help. LTFT only dropped slightly to 20%. I went to Star Diagnosis to reset trims. Immediately after starting, STFT jumped to +25%, and the O2 regulation was marked as “out of range 24%.”
   
   After driving for 3–4 km, LTFT went to 11%, and eventually climbed back to 20–25%. Still like that today.
   
2. Intake leak test – sprayed starter fluid around intake, hoses, and connections while engine was running – no RPM changes. I also used compressed air in various points – couldn’t detect any leaks. PCV (entire breather system) was replaced ~3 months ago.
   
   Even tried disconnecting and plugging hoses – no change in trims, so I assume the intake is sealed tight.
   
3. MAF sensor – at idle: 90°C coolant, 750 RPM, engine load 11%, throttle 1.5°, reading 2.4 g/s.
   
   Tried driving with MAF disconnected — no improvement. MAF cleaned with proper cleaner, OEM Siemens unit installed.
   
4. MAP sensor – at idle: 32 kPa. Ignition on: 99 kPa (atmospheric). Under ~20% load with A/C: 47 kPa.
   BOOST at idle: -0.68 to -0.70 bar. Under acceleration: 0.6–0.7 bar. LPG MAP sensor shows identical readings.
   
5. IAT sensor – compared with cold engine ambient and coolant temps — deviation of ~1.5°C.
   Also installed my own independent sensor — readings within 2°C of each other. On a hot day (30°C ambient), IAT can reach 56°C in traffic.
   
6. Throttle body – cleaned, moves freely with no resistance. Electrical connectors cleaned. Re-adapted after disconnection.
   
7. Air filter – new Filtron. Tried running without filter — no change in trims.
   
8. Fuel pressure (petrol) – stable at 3.5 bar at idle and under load, measured at fuel rail.
   
9. Petrol injectors – not tested directly, but injection times go from 1.66 ms (after trim reset) to 2.3–2.4 ms now.
   Assuming they’re not leaking or clogged — hard to believe all would fail at once.
   
10. Coolant temp – matches other sensors, reading stable at 93°C.
    
11. EGR – disconnected, confirmed sealed and working. Trims don’t change when disconnected.
    
12. Exhaust system – checked for leaks before the O2 sensors — none found.
    
13. Upstream wideband O2 sensor (5-pin) – responds well to changes in mixture, STFT adjusts accordingly.
    AFR sits at 14.6–14.7 at idle and under driving. Removing vacuum hose causes ECU to enrich mixture — sensor reacts instantly.
    

I also tested by simulating MAF signal with voltage generator:

• Set to 1.8 g/s → STFT dropped briefly, then adjusted back up.
  
• Set to 3.5 g/s → ECU enriched, but STFT quickly went to -25%.
  => Conclusion: sensor responds quickly and accurately.
  

I also removed the sensor from the exhaust pipe: ECU showed LAMBDA = 2.

1. Downstream narrowband O2 sensor (4-pin) – at idle shows 0.7–0.8V. After revving, drops to 0.1–0.2V, then returns to 0.7–0.8V.
   

As a test, I installed this sensor in the upstream position — it showed 0.850V at idle (rich), didn’t move. After revving, it dropped to 0.15V, then returned to 0.850V.

1. Emissions test – checked recently, photos attached. Looks fine to me — slightly lean lambda, but otherwise normal combustion.
   

Summary:

Everything seems fine, so why is the ECU adding so much fuel right after a fuel trim reset? It’s not learning — it just starts correcting immediately.
To me, this feels like a classic intake leak — but where? Why haven’t I found it?

Maybe someone can help identify the issue.

One last question:
Does the supercharger in this engine have any vacuum lines connected to it? Behind the main throttle body, there’s a small nipple with a check valve and a very thin hose (almost like shrink tubing) — I can’t trace where it leads.
Thanks again in advance for your help.
Best regards!

 

Hello.
I had High LTFT (Lean +25) on my petrol/lpg c200 compressor (111.955).
Problem solved completely with a new MAF (Hella)