Lucas has a drinking problem. No matter how I drive him, or where I drive him, the fuel consumption is horrific. I needed to fix it, or I couldn't afford to keep driving him.
I track the fuel-ups for all my cars using Fuelly. Not only does this allow me to compare consumption between different cars, but it can show me just how bad it can be.
The Honda averages 7.0L/100KM on its daily commute. This was measured over about 120,000km of fuel ups, but fluctuates a lot depending on traffic or if we take the car away on a long fun.
Lucas on the other hand... has cost me $347 in fuel, to cover 446km. That is an average of 25.2L/100km (9.3MPG).
The last trip was a round trip up the coast and back, on the open road, mostly at 80-100kph. And it averaged 26.9L/100km. WTF. I can pound on it and push it hard, and it's just as bad as if I baby it and cruise at low throttle.
It was time to see if I could fix the issue. I had my suspicions that it might be the flapper Air Flow Meter. Long story short, this is a device that has a flap inside it (hence the name, Flapper), that opens and closes depending on how much air is being drawn into the intake. The ECU measures how far it's opening and closing, by a change in resistance as the flap moves, and uses that to adjust the fueling. More open, more air, more fuel.
I recorded a short video of what the flapper is doing when you rev the engine. The U shaped copper arm, and the black track it runs on, are how the ECU knows how far open the flap is.
The main tool I needed was a wideband O2 sensor. This measures the air mixture in the exhaust fumes and gives a readout of how many parts air to fuel there are in the fumes. The "ideal" is 14.7:1 (14.7 parts air, to 1 part fuel); this indicates a nice complete combustion. A bigger first number means it's lean, and smaller rich (more air, or less air to fuel). You can't run an engine at 14.7 all the time, so the ECU will vary the fuel mixture as needed (richer when accelerating, leaner for cruise, as a rule of thumb).
I have a wideband setup already, the 14point7 Spartan 2 from when I installed the Speeduino on Effie, but that doesn't have a gauge, so no instant readout. I needed a gauge.
I tried to buy a cheap secondhand setup but unfortunately, it was DOA, so I ended up opening the old wallet and throwing money at the problem. After a long trip on the slow boat from Freedom-land, I had a nice new Glowshift digital WB02 gauge in my hands.
Along with the gauge, I also bought an Innovate tailpipe clamp, so I could use the sensor without having to drill and weld a bung into the exhaust (and so I can use this setup on other cars, like the Marina).
The other night, I soldered a 12v plug to the gauge wires, so its plug and play
And then set the gauge up and free-air calibrated it (so the gauge knows what "fresh air" smells like)
Yesterday I slipped the clamp into the exhaust and tightened it, using a bungee cord around the exhaust hanger as a safety. It didn't come loose, but I wouldn't want to lose it if it did.
The reason I specifically went for the Glowshift gauge, was that they use a remote control box. Not as neat and tidy as the likes of AEM or Innovate where the O2 sensor harness plugs directly into the back of the gauge, but it gives me a good couple of meters of extra wire length, since the O2 sensor has its own wire to the harness, and then the harness goes to the control box. The control box then has another long wire to the gauge. I placed the control box in the center console, and ran all the wires to it
The O2 sensor wire is carefully shut in the tailgate seal, and then run through the boot, over the back seats and to the control box. The gauge is duct taped to the steering column shrouds.
I plugged the 12v socket in, fired up the engine and off we went. After the initial sensor warm-up, we had this on the gauge, on a cold start
This was to be expected; rich. The ECU dumps more fuel in when it's cold, to help the warm-up phase. It should steadily lean out as the engine warms up.
It didn't.
No matter how I drove the car, it was always reading 12-13:1, except under one condition; hard acceleration, where it went VERY lean. I could back off the throttle and cruise as much as I wanted, and it was still rich. No wonder we were drinking fuel.
Now, the "lean" condition. When using this much fuel, the "lean" reading surprised me. It's very unlikely it was actually lean, and what I believe happened was what's called a "false lean" condition. This happens when there is a slight misfire (such as fuel fouling a plug slightly under load) and the sensor reads the uncombusted air/fuel mixture as lean (because there is "unburnt/unused" air in the mixture).
After a couple of laps around with the engine warm, it was obvious that the engine wasn't going to lean out the mixture as it should. I had previously checked the temp sensor, and that was well within range, so I knew it wasn't that causing the richness.
This was the gauge on a hot idle
It's slightly richer!
Madness. It explains why no one would ever follow closely, and when my colleague did, he promptly got a headache and was light-headed.
As mentioned above, my first port of call was the AFM. The one in the car wasn't the one the car came with, it was an unknown one from my spares. I swapped it when I got the car (before the engine swap) because the one with the car had been tampered with and I wanted to eliminate that.
All that turned out to be a non-issue compared to needing to replace the engine, so I just left the unknown AFM in the car.
I grabbed the AFM that came with the car and set about reinstalling it
It's very easy to do. Removed the plug, remove the hoses on both sides, remove the two front bolts and then push it towards the back of the car to disengage the rear fixing, and then remove it.
I swapped the original AFM in and fired the car up. Instantly I noticed just how smooth the idle was. It wasn't rough before, but always had a slight stumble or unevenness about it. That was gone.
The wideband was also showing an immediate difference. Sure, this was a warm start, but see the photo above for a comparison.
It's significantly leaned out, and far closer to that ideal 14.7:1 ratio.
I took the car for the same short drive I did on the previous AFM, and the AFR was drastically different. Instead of a rich 12-13:1 under all conditions, I now had a slightly lean 15:1 at cruise, high 13s accelerating, no more "false lean" when under load, and the engine felt smoother and responded better.
It was a great start! I had expected I would need to tweak the AFM settings to adjust the mixture, but it was damn near perfect as it was.
Next, I had to take it for a bigger drive, including some hills, to see how it reacted and make sure it didn't lean out under load, so today I did a 60km round trip (or the same route as my daily commute).
Lucas worked flawlessly. The engine seems smoother, but as expected it lost a little "punch" that it had when it was rich. The engine still pulls like a train, it just feels more refined.
I confirmed that the engine was not running lean under load, and although I'm still missing my extra enrichment, the engine was happy enough. Hopefully once I can fix the enrichment I might gain some of that punch back.
We stopped to check everything was OK with the car, and grabbed some photos
It's a great looking car and I do enjoy driving it.
One of the biggest differences I noticed on the drive, people now follow too close... Maybe I should have kept gassing people, just to keep them off my tail.
Once home I removed all the wideband gear, as I'm comfortable that it's running as well as it needs to. The next test will be to drive it for a full tank and see what economy the next fill records.
Interestingly the tailpipe clamp is discoloured, but isn't sooty, which is a good indication to me
I also hit a milestone when I got home. This is how many KM I have put on the car since the 3.9 engine went in
So, I have an excuse to drive the car more now, just so I can see how many miles I get out of this tank. Better get onto it then.
I track the fuel-ups for all my cars using Fuelly. Not only does this allow me to compare consumption between different cars, but it can show me just how bad it can be.
The Honda averages 7.0L/100KM on its daily commute. This was measured over about 120,000km of fuel ups, but fluctuates a lot depending on traffic or if we take the car away on a long fun.
Lucas on the other hand... has cost me $347 in fuel, to cover 446km. That is an average of 25.2L/100km (9.3MPG).
The last trip was a round trip up the coast and back, on the open road, mostly at 80-100kph. And it averaged 26.9L/100km. WTF. I can pound on it and push it hard, and it's just as bad as if I baby it and cruise at low throttle.
It was time to see if I could fix the issue. I had my suspicions that it might be the flapper Air Flow Meter. Long story short, this is a device that has a flap inside it (hence the name, Flapper), that opens and closes depending on how much air is being drawn into the intake. The ECU measures how far it's opening and closing, by a change in resistance as the flap moves, and uses that to adjust the fueling. More open, more air, more fuel.
I recorded a short video of what the flapper is doing when you rev the engine. The U shaped copper arm, and the black track it runs on, are how the ECU knows how far open the flap is.
The main tool I needed was a wideband O2 sensor. This measures the air mixture in the exhaust fumes and gives a readout of how many parts air to fuel there are in the fumes. The "ideal" is 14.7:1 (14.7 parts air, to 1 part fuel); this indicates a nice complete combustion. A bigger first number means it's lean, and smaller rich (more air, or less air to fuel). You can't run an engine at 14.7 all the time, so the ECU will vary the fuel mixture as needed (richer when accelerating, leaner for cruise, as a rule of thumb).
I have a wideband setup already, the 14point7 Spartan 2 from when I installed the Speeduino on Effie, but that doesn't have a gauge, so no instant readout. I needed a gauge.
I tried to buy a cheap secondhand setup but unfortunately, it was DOA, so I ended up opening the old wallet and throwing money at the problem. After a long trip on the slow boat from Freedom-land, I had a nice new Glowshift digital WB02 gauge in my hands.
Along with the gauge, I also bought an Innovate tailpipe clamp, so I could use the sensor without having to drill and weld a bung into the exhaust (and so I can use this setup on other cars, like the Marina).
The other night, I soldered a 12v plug to the gauge wires, so its plug and play
And then set the gauge up and free-air calibrated it (so the gauge knows what "fresh air" smells like)
Yesterday I slipped the clamp into the exhaust and tightened it, using a bungee cord around the exhaust hanger as a safety. It didn't come loose, but I wouldn't want to lose it if it did.
The reason I specifically went for the Glowshift gauge, was that they use a remote control box. Not as neat and tidy as the likes of AEM or Innovate where the O2 sensor harness plugs directly into the back of the gauge, but it gives me a good couple of meters of extra wire length, since the O2 sensor has its own wire to the harness, and then the harness goes to the control box. The control box then has another long wire to the gauge. I placed the control box in the center console, and ran all the wires to it
The O2 sensor wire is carefully shut in the tailgate seal, and then run through the boot, over the back seats and to the control box. The gauge is duct taped to the steering column shrouds.
I plugged the 12v socket in, fired up the engine and off we went. After the initial sensor warm-up, we had this on the gauge, on a cold start
This was to be expected; rich. The ECU dumps more fuel in when it's cold, to help the warm-up phase. It should steadily lean out as the engine warms up.
It didn't.
No matter how I drove the car, it was always reading 12-13:1, except under one condition; hard acceleration, where it went VERY lean. I could back off the throttle and cruise as much as I wanted, and it was still rich. No wonder we were drinking fuel.
Now, the "lean" condition. When using this much fuel, the "lean" reading surprised me. It's very unlikely it was actually lean, and what I believe happened was what's called a "false lean" condition. This happens when there is a slight misfire (such as fuel fouling a plug slightly under load) and the sensor reads the uncombusted air/fuel mixture as lean (because there is "unburnt/unused" air in the mixture).
After a couple of laps around with the engine warm, it was obvious that the engine wasn't going to lean out the mixture as it should. I had previously checked the temp sensor, and that was well within range, so I knew it wasn't that causing the richness.
This was the gauge on a hot idle
It's slightly richer!
Madness. It explains why no one would ever follow closely, and when my colleague did, he promptly got a headache and was light-headed.
As mentioned above, my first port of call was the AFM. The one in the car wasn't the one the car came with, it was an unknown one from my spares. I swapped it when I got the car (before the engine swap) because the one with the car had been tampered with and I wanted to eliminate that.
All that turned out to be a non-issue compared to needing to replace the engine, so I just left the unknown AFM in the car.
I grabbed the AFM that came with the car and set about reinstalling it
It's very easy to do. Removed the plug, remove the hoses on both sides, remove the two front bolts and then push it towards the back of the car to disengage the rear fixing, and then remove it.
I swapped the original AFM in and fired the car up. Instantly I noticed just how smooth the idle was. It wasn't rough before, but always had a slight stumble or unevenness about it. That was gone.
The wideband was also showing an immediate difference. Sure, this was a warm start, but see the photo above for a comparison.
It's significantly leaned out, and far closer to that ideal 14.7:1 ratio.
I took the car for the same short drive I did on the previous AFM, and the AFR was drastically different. Instead of a rich 12-13:1 under all conditions, I now had a slightly lean 15:1 at cruise, high 13s accelerating, no more "false lean" when under load, and the engine felt smoother and responded better.
It was a great start! I had expected I would need to tweak the AFM settings to adjust the mixture, but it was damn near perfect as it was.
Next, I had to take it for a bigger drive, including some hills, to see how it reacted and make sure it didn't lean out under load, so today I did a 60km round trip (or the same route as my daily commute).
Lucas worked flawlessly. The engine seems smoother, but as expected it lost a little "punch" that it had when it was rich. The engine still pulls like a train, it just feels more refined.
I confirmed that the engine was not running lean under load, and although I'm still missing my extra enrichment, the engine was happy enough. Hopefully once I can fix the enrichment I might gain some of that punch back.
We stopped to check everything was OK with the car, and grabbed some photos
It's a great looking car and I do enjoy driving it.
One of the biggest differences I noticed on the drive, people now follow too close... Maybe I should have kept gassing people, just to keep them off my tail.
Once home I removed all the wideband gear, as I'm comfortable that it's running as well as it needs to. The next test will be to drive it for a full tank and see what economy the next fill records.
Interestingly the tailpipe clamp is discoloured, but isn't sooty, which is a good indication to me
I also hit a milestone when I got home. This is how many KM I have put on the car since the 3.9 engine went in
So, I have an excuse to drive the car more now, just so I can see how many miles I get out of this tank. Better get onto it then.
I'm certainly guilty of gunning it everywhere just to hear the exhaust note
