Why it pays to clear those CELs

[ChrisGB]

For some time now I have had a CEL on for post cat sensor. Initially it was the "cat below efficiency" code. I recently fitted a PPE decat pipe (to go with the PPE manifold) and an 02 sensor conditioner. All fitted, still get a CEL. Code reading (thanks Liz) showed we now had a post cat heater circuit problem.

I started looking into the resistor mods that people have applied and read over on SC that some people are using around 15 Ohms (the resistance of a cold O2 heater circuit) and some around 35 Ohms (the resistance of a hot o2 heater circuit). I elected to go with the higher resistance as people running the 15 said they still got the occasional CEL where those running 35 Ohms seemed to have it permanently fixed. As a bonus, the higher resistance means that the resistors draw less current and thus run cooler.

I also read somewhere on SC that if there is a CEL the engine pulls timing a little to run in safe mode. Not sure if this were true or not, I thought I would test this. I ran a stretch of local test track today to measure 60 - 80 in 3rd gear before and after clearing the CEL (well having the battery off for an hour) and fitting the resistor mod.

The result, 0.4s quicker on the 60 - 80 increment on the stopwatch. In the top 500rpm the result is even more noticeable through the butt dyno. So if like me you were tolerating a post cat CEL and not worrying about it, fixing it is free HP.

For reference, I used two wirewound ceramic housed 10W 68R J resistors in parallel giving 34Ohms and running them at 2.1W each. They get fairly hot, but not much hotter than everything else in the engine bay. I may go to 3 x 100 Ohm in parallel or a single heatsunk 33 Ohm to tidy it up. I just did a lash up with bits from Maplin for Tuesdays rolling road tune.

Chris

[darkday]

Hmmm... The post cat sensor is only there to show if the cat is working or not. It has nothing to do with fuel trims, that's why we have two pre-cat sensors.

If you think about what's going on throughout the whole deal it makes sense. You've got a bunch of unburnt fuel in the exhaust gases which is going through a cat and being broken down into CO2. Depending on how warm the cat is and how old it is the effeciency will vary. You'll never get a good read on how lean/rich the car is by taking measurements past the cat.

[ChrisGB]

Hmmm... The post cat sensor is only there to show if the cat is working or not. It has nothing to do with fuel trims, that's why we have two pre-cat sensors.

If you think about what's going on throughout the whole deal it makes sense. You've got a bunch of unburnt fuel in the exhaust gases which is going through a cat and being broken down into CO2. Depending on how warm the cat is and how old it is the effeciency will vary. You'll never get a good read on how lean/rich the car is by taking measurements past the cat.

I think the issue may be that the ECU runs a safe trim if it detects any codes to do with emission gear. I had a dyno run with the CEL on a while back and the AFR was running up down to nearly 11:1 at the top end. I wonder if it will be leaner now the CEL is fixed. I did four runs before and after cancelling the CEL and fitting the resistors and the difference is definitely there, the figure above being pretty repeatable. This may explain why some people reckon a periodic ECU reset sharpens the car up.

As for figuring out AFRs, not sure about the measurement technique used, but have seen many setups done measuring post cat with a wideband.

Chris

[Tem]

As for figuring out AFRs, not sure about the measurement technique used, but have seen many setups done measuring post cat with a wideband.

You CAN measure after the cat, but like dd said, it's not accurate.

It's better to measure after the cat than not at all though. That's why dynos rather use a sensor in the tailpipe than no sensor at all. But if there's a choice to measure before the cat, it's always the best method.

[SteveJ]

Just one minor snag with your theory - you didnt just clear the CEL, you have cleared all of the long-term fueling stats by disconnecting the battery, so it goes back to a richer, 'safer' tune as opposed to the state the ECU has been programmed to achieve which is fuel efficiency.

I would have been more inclined to believe your results were due to the CEL being cleared if you had used a code reader to clear it.

[ChrisGB]

As for figuring out AFRs, not sure about the measurement technique used, but have seen many setups done measuring post cat with a wideband.

You CAN measure after the cat, but like dd said, it's not accurate.

It's better to measure after the cat than not at all though. That's why dynos rather use a sensor in the tailpipe than no sensor at all. But if there's a choice to measure before the cat, it's always the best method.

If I understand correctly, the sensor uses the surface oxygen deficit to provide an output. An example 12:1 mixture would provide an output that was showing oxygen deficit (positive charge seeing as the surface is giving up electrons). Passing the same exhaust gas through a cat, the oxygen deficit will still be the same, but there will be less CO and less O2 in the gas. Fortunately for me, to measure after the cat, the guy doing the dyno is going to have to stretch the probe lead all the way to my garage  s:) s:)  

Just one minor snag with your theory - you didnt just clear the CEL, you have cleared all of the long-term fueling stats by disconnecting the battery, so it goes back to a richer, 'safer' tune as opposed to the state the ECU has been programmed to achieve which is fuel efficiency.

I would have been more inclined to believe your results were due to the CEL being cleared if you had used a code reader to clear it.

The long term fuelling stats are, I believe, only applicable to the closed loop operation mode of the ECU. Full throttle high RPM stats will be fixed values from a lookup table. Possibly the TPS only, or in some cars the TPS and MAF are providing the load data. Timing on the MR2 is fixed, but with knock protection to back it off where required, it does not learn it, just applies it as it goes along. When the CEL was on, the mixture was rich anyway, with nearly 11:1 at the top end.

Will be interesting to see how it all works out on the dyno Tuesday.

Chris

[muffdan]

I thought the long term fuel trim was still applied when running in closed loop   s:?: :?: s:?:

[SteveJ]

So why do people report similar results from an ECU reset, even though they are not showing a CEL?

I maintain there are a number of parameters that are stored in volatile memory that are NOT exposed to the outside world via OBD-II which the ECU uses to achieve a longer term optimal economic tune, and erasing them puts it back to a base level which (on UK petrol at least) provides more power.

The 3rd O2 cant be used in the fueling corrections, as that would have required a major re-write of the code for the J-spec cars that dont have the sensor, instead of simply de-activating the cat failure logic.

[ChrisGB]

I thought the long term fuel trim was still applied when running in closed loop   s:?: :?: s:?:

That is what I said?

Chris

So why do people report similar results from an ECU reset, even though they are not showing a CEL?

I maintain there are a number of parameters that are stored in volatile memory that are NOT exposed to the outside world via OBD-II which the ECU uses to achieve a longer term optimal economic tune, and erasing them puts it back to a base level which (on UK petrol at least) provides more power.

The 3rd O2 cant be used in the fueling corrections, as that would have required a major re-write of the code for the J-spec cars that dont have the sensor, instead of simply de-activating the cat failure logic.

I must say that when I have done an ECU reset and MAF clean with no CEL showing, I have never noticed any gain at all. This gain is not subtle. I believe the ECU to use set parameters in open loop mode (that is why it is called open loop mode) and cannot reliably tune from the OEM narrow band O2 sensors anyway. Evidence from my last dyno run would support this as the engine was running rich at the top end. If it had "learned" fuelling for economy, it would be trying to go closer to stoich' and would have surely been running leaner?

I don't know for sure, but this is where the evidence seems to point. I have not worked in engine management for a while, but the systems we used to work on for pretty much everyone in the world from Ford Fiesta single point to all the worlds F1 engines all used TPS vs RPM as input for fuelling in WOT condition. Some, and I mean a very few, used MAF compensation as well. None ever used feedback from exhaust in big load index situations.

Chris

[muffdan]

I thought the long term fuel trim was still applied when running in closed loop   s:?: :?: s:?:

That is what I said?

Chris

Oops, you certainly did! I meant the long term fuel trim is still appied in open loop.

As Steve says, the long term and short term trims are reset when disconnecting the battery and this is always what I believed made the car feel a little more gutsy when reset.

I don't have access to the BGB so can't check it myself, but I thought the fuel trims are always applied, open and closed loop, but they only get adjusted during closed loop operation.

[ChrisGB]

I thought the long term fuel trim was still applied when running in closed loop   s:?: :?: s:?:

That is what I said?

Chris

Oops, you certainly did! I meant the long term fuel trim is still appied in open loop.

As Steve says, the long term and short term trims are reset when disconnecting the battery and this is always what I believed made the car feel a little more gutsy when reset.

I don't have access to the BGB so can't check it myself, but I thought the fuel trims are always applied, open and closed loop, but they only get adjusted during closed loop operation.

I see where you are coming from with this. The fuel trims are generated from closed loop operation and will apply to closed loop load indexes. The information is gleaned from wandering or more specifically dithering around an ideal stoich mix. Large throttle openings apply open loop. Here the system has no way of measuring the AFR as the O2 sensors are not wideband. Dithering will not be possible at large throttle openings without compromising the running of the engine. It will be interesting to see how the AFRs are before tuning now I have the codes out of the system.

Chris

[Tem]

If I understand correctly, the sensor uses the surface oxygen deficit to provide an output. An example 12:1 mixture would provide an output that was showing oxygen deficit (positive charge seeing as the surface is giving up electrons). Passing the same exhaust gas through a cat, the oxygen deficit will still be the same, but there will be less CO and less O2 in the gas.

Yes, O2/lambda sensors measure oxygen. But the measured amount of oxygen changes over the cat, so it will not see same values.

That change of oxygen over a working cat is being used in our '2 as well. The 3rd sensor only measures cat efficiency, by comparing the output of the sensor after the cat against the sensors before the cat. If the cat dies, or you remove it, there's no oxygen difference anymore and you get a CEL.  s;) s;)


I once run two Innovate widebands (LM-1 and LC-1) before and after the cat, just to see how much it affects the value. The difference was over 1.0AFR, which could be dangerous to your engine, if you assume it's the same before and after the cat.

[ChrisGB]

I once run two Innovate widebands (LM-1 and LC-1) before and after the cat, just to see how much it affects the value. The difference was over 1.0AFR, which could be dangerous to your engine, if you assume it's the same before and after the cat.

That makes sense. Tuning catless tomorrow, so what we see is what we get.

Chris