Tuning for emissions
Discussion
I have a Ford 4.6 V8 with supercharger and aftermarket ECU.
We are trying to map it specifically for the emissions requirement of the IVA test.
The test requirement is:-
Fast Idle Test(2500 to 3000rpm)
CO <= 0.2% HC <= 200ppm
Lambda between0.97 and 1.03
Idle Test(450 to 1500rpm)
CO <=0.3%
We have mapped to spot on Lambda 1 in both test conditions but the CO is way out at ~1.5% and the HC about 300-400.
We have tried advancing and retarding the ignition to no avail.
We considered exhaust air leaks but we have the same condition on both banks. There is a cross over pipe (joining the two silencers) but if one bank was leaking we would at least see a difference in the two banks.
The ECU is using a pre-cat narrow band sensor in each bank and we are measuring AFR with a wideband sensor post-cat. The ECU is applying about equal short term trim to both banks.
Any ideas how to correct this or thoughts on what is going on?
Many thanks in advance.
Steve
We are trying to map it specifically for the emissions requirement of the IVA test.
The test requirement is:-
Fast Idle Test(2500 to 3000rpm)
CO <= 0.2% HC <= 200ppm
Lambda between0.97 and 1.03
Idle Test(450 to 1500rpm)
CO <=0.3%
We have mapped to spot on Lambda 1 in both test conditions but the CO is way out at ~1.5% and the HC about 300-400.
We have tried advancing and retarding the ignition to no avail.
We considered exhaust air leaks but we have the same condition on both banks. There is a cross over pipe (joining the two silencers) but if one bank was leaking we would at least see a difference in the two banks.
The ECU is using a pre-cat narrow band sensor in each bank and we are measuring AFR with a wideband sensor post-cat. The ECU is applying about equal short term trim to both banks.
Any ideas how to correct this or thoughts on what is going on?
Many thanks in advance.
Steve
Post wideband sensor will read leaner. How much.....unsure, but they do read leaner.
What camshaft are you using ?
Hydrocarbons and CO could indicate a misfire, or overlap. Is it a non standard cam ? And do you have control over injection timing, or are you just batch fire ?
And what gas analyser are you using ? O2 readings would tell you if there is an exhaust leak, as they would be very high.
What camshaft are you using ?
Hydrocarbons and CO could indicate a misfire, or overlap. Is it a non standard cam ? And do you have control over injection timing, or are you just batch fire ?
And what gas analyser are you using ? O2 readings would tell you if there is an exhaust leak, as they would be very high.
Some answers
There is a cat in each bank.
The AFR reading after the cat is what I need as that is in effect where the IVA tester will be probing.
I have no idea what cam is in the engine but may be able to find out. I suspect it will be non standard as the engine started out as a 4.6 Mustang engine then had GT heads and the charger fitted.
I only have a two gas analyser made by Probike. I may get it tested again in a local MOT station to make sure my figures tie up.
ECU provides sequential injection.
Thanks
Steve
There is a cat in each bank.
The AFR reading after the cat is what I need as that is in effect where the IVA tester will be probing.
I have no idea what cam is in the engine but may be able to find out. I suspect it will be non standard as the engine started out as a 4.6 Mustang engine then had GT heads and the charger fitted.
I only have a two gas analyser made by Probike. I may get it tested again in a local MOT station to make sure my figures tie up.
ECU provides sequential injection.
Thanks
Steve
Steve_D said:
Some answers
There is a cat in each bank.
The AFR reading after the cat is what I need as that is in effect where the IVA tester will be probing.
I have no idea what cam is in the engine but may be able to find out. I suspect it will be non standard as the engine started out as a 4.6 Mustang engine then had GT heads and the charger fitted.
I only have a two gas analyser made by Probike. I may get it tested again in a local MOT station to make sure my figures tie up.
ECU provides sequential injection.
Thanks
Steve
IVA tester does not use a lambda sensor though. They use a gas analyser. So you cannot simply take a wideband AFR readings and assume they will translate into the same as a proper gas analyser will.There is a cat in each bank.
The AFR reading after the cat is what I need as that is in effect where the IVA tester will be probing.
I have no idea what cam is in the engine but may be able to find out. I suspect it will be non standard as the engine started out as a 4.6 Mustang engine then had GT heads and the charger fitted.
I only have a two gas analyser made by Probike. I may get it tested again in a local MOT station to make sure my figures tie up.
ECU provides sequential injection.
Thanks
Steve
For tuning the engine with a wideband, you must have it before the cats. The cat will skew the readings.
To add.
Lambda sensor only measures excess oxygen. So if there are any misfires, any air leaks prior to the sensor, it may display a leaner mixture than the engine is actually seeing. Which may go part way to explaining the higher CO and HCC readings, yet your wideband appears to display lambda 1
Edited by stevieturbo on Thursday 14th February 22:04
In no particular order:
1) AFR: you're using narrow band sensors, and they tend to be more robust than wideband versions. To get the best catalyst efficiency you will need to "jump" the fuelling either side of Lambda 1, which is what should be happening. If you have any ability to map these jumps, then small changes to the 'bias' make big effects in conversion efficieny (effectively moving the mean lambda tiny amounts towards or away from stoichiometric
2) Are you sure the catalyst had lit off for the test. You need over 400degC mid brick for any catalyst to start catalysing. it's not unheard of for cat to "go out" at idle (especially on large diameter low loss, high volume exhaust systems with downstream cats (lots of heat gets lost before it gets to the cats!)
3) Post catalyst lambda will home in on pre-catalyst lambda if speed and load are kept constant. The catalyst has significant oxygen storage capability, and during rich lean transients the post catalyst AFR will lag pre catalyst AFR significantly. Down stream lambda sensors work to measure the mean exhaust AFR, so you can only map to this value if the engine operating state is stabilised (more than 30sec at any particular load point)
4) High HC/CO suggests incomplete burning. Even at lambda 1 the installed catalyst volume may be unable to convert these large engine out emissions into CO2. If you have an aggressive cam specification, the overlap at low rpm will make life difficult. Deffinately try moving the End of Injection angle at low rpm / idle, as this can have a massive effect on combustion stability. Also, a more advanced ignition angle might allow longer for combustion to occur and help too.
I would try and get an emissions sample at a fast idle (say 3000rpm) and see what results you get, and also try 2-3 min of elevated rpm before the idle test to make sure the cats are lit off!
If you have too little catalyst volume, or too open a catalyst matrix density (CPSI), the system may not be able to meet the emissions requirements.
Obviously you need to check that there are no obvious misfires occuring too! Plug condition and coil dwell can make a significant difference at light load ignition states (like at idle)
1) AFR: you're using narrow band sensors, and they tend to be more robust than wideband versions. To get the best catalyst efficiency you will need to "jump" the fuelling either side of Lambda 1, which is what should be happening. If you have any ability to map these jumps, then small changes to the 'bias' make big effects in conversion efficieny (effectively moving the mean lambda tiny amounts towards or away from stoichiometric
2) Are you sure the catalyst had lit off for the test. You need over 400degC mid brick for any catalyst to start catalysing. it's not unheard of for cat to "go out" at idle (especially on large diameter low loss, high volume exhaust systems with downstream cats (lots of heat gets lost before it gets to the cats!)
3) Post catalyst lambda will home in on pre-catalyst lambda if speed and load are kept constant. The catalyst has significant oxygen storage capability, and during rich lean transients the post catalyst AFR will lag pre catalyst AFR significantly. Down stream lambda sensors work to measure the mean exhaust AFR, so you can only map to this value if the engine operating state is stabilised (more than 30sec at any particular load point)
4) High HC/CO suggests incomplete burning. Even at lambda 1 the installed catalyst volume may be unable to convert these large engine out emissions into CO2. If you have an aggressive cam specification, the overlap at low rpm will make life difficult. Deffinately try moving the End of Injection angle at low rpm / idle, as this can have a massive effect on combustion stability. Also, a more advanced ignition angle might allow longer for combustion to occur and help too.
I would try and get an emissions sample at a fast idle (say 3000rpm) and see what results you get, and also try 2-3 min of elevated rpm before the idle test to make sure the cats are lit off!
If you have too little catalyst volume, or too open a catalyst matrix density (CPSI), the system may not be able to meet the emissions requirements.
Obviously you need to check that there are no obvious misfires occuring too! Plug condition and coil dwell can make a significant difference at light load ignition states (like at idle)
Edited by anonymous-user on Thursday 14th February 23:11
Steve_D said:
We have mapped to spot on Lambda 1 in both test conditions but the CO is way out at ~1.5% and the HC about 300-400.
for comparison - I think my non cat equipped megasquirted chevy gave those sorts of figures when I asked the MOT tester to check it (I'm visual smoke only...)Should have added:
If you fit narrow band Lambda sensors downstream of the catalysts, the ratio of switching events, front / rear is an indication of the catalyst conditions. With a cold catalyst, no oxygen storage occurs, and the rear sensor will switch pretty much the same number of times as the front one. With a full working catalyst, the rear sensor will switch much much slower (10x slower for example) as the oxygen storage within the hot catalyst evens out the upstream lambda perturbations. This is an easy was of telling if your cats are hot and working.
If you fit narrow band Lambda sensors downstream of the catalysts, the ratio of switching events, front / rear is an indication of the catalyst conditions. With a cold catalyst, no oxygen storage occurs, and the rear sensor will switch pretty much the same number of times as the front one. With a full working catalyst, the rear sensor will switch much much slower (10x slower for example) as the oxygen storage within the hot catalyst evens out the upstream lambda perturbations. This is an easy was of telling if your cats are hot and working.
Max_Torque said:
If you fit narrow band Lambda sensors downstream of the catalysts, the ratio of switching events, front / rear is an indication of the catalyst conditions. ......
I will try and get the beast hot tomorrow because it certainly was not when we were doing the mapping.Next will be trying to prove the cats are working. Any ideas on a simple way of watching the Lambda switching? I don't have access to a scope nor would I know how to use one if it was sat in front of me.
Steve
Another factor to consider.
During the last year the engine has not been run much and when it was it was clear the mapping was way out. Just prior to mapping the fouled plugs were replaced to get the car to run. Short reach plugs were fitted replacing the long reach plugs fitted way back when the engine was built.
Does the panel think fitting long reach again could make a difference.
Steve
Sub note...The engine being way of map and the fouled plugs could also mean the cats are FUBAR.
During the last year the engine has not been run much and when it was it was clear the mapping was way out. Just prior to mapping the fouled plugs were replaced to get the car to run. Short reach plugs were fitted replacing the long reach plugs fitted way back when the engine was built.
Does the panel think fitting long reach again could make a difference.
Steve
Sub note...The engine being way of map and the fouled plugs could also mean the cats are FUBAR.
The rear switching will be nice and slow if the cats are working, probably a "switch" (>0.8v - <0.3v approx) every 3 to 6 seconds, so you should be able to watch it with a simple multimeter. Remember, the Lambda sensor itself must be up to temperature to work also, and you probably won't have the sensor heater connected to anything.
If the map was really miles out, and the engine was run for a long time, then yes, you could have poisoned or melted the cats (the latter damage being easily visable) but it is probably fairly unlike if the engine was only run at low speed etc.
Plug reach could have some effect on burn, depending on the chanmber shape etc, but at part loads it is unlikely to have a big effect on engine out emissions.
BTW, for future refference, a scope is really useful, i have a couple! I just got one of these:
Tiny little thing, and not great accuracy (8 bit resolution only) but really handy to have for general "emergency" diagnostics and being small and battery powered you can take it everywhere. Just over £100 on ebay etc
If the map was really miles out, and the engine was run for a long time, then yes, you could have poisoned or melted the cats (the latter damage being easily visable) but it is probably fairly unlike if the engine was only run at low speed etc.
Plug reach could have some effect on burn, depending on the chanmber shape etc, but at part loads it is unlikely to have a big effect on engine out emissions.
BTW, for future refference, a scope is really useful, i have a couple! I just got one of these:
Tiny little thing, and not great accuracy (8 bit resolution only) but really handy to have for general "emergency" diagnostics and being small and battery powered you can take it everywhere. Just over £100 on ebay etc
More than about 0.6% CO suggests its not running lambda=1.0 (extreme wild cams notwithstanding)
If it is running with a switching (upstream) lambda sensor, and it is actually switching (0.7-2.0Hz), than chances are you have an exhaust manifold leak or at least a leak upstream-ish of the sensor. Don't forget the exhaust is pulsating flow, so the leak could also be somewhat downstream of the sensor too.
Downstream lambda voltage (narrow-band sensors) should spend lengthy periods near or less than around 0.7V if all is running well. Above 0.8V and it's likely producing CO - sometimes a lot!
All of the above assumes the cat is hot and functional. If the cat is a long way (or a large surface area - think performance tubular exhausts...) from the manifold flange, it will not stay lit during extended idle periods. NB, at 1.5% CO, your cat is not likely to be the key problem
If you're trying to make it work without a closed-loop lambda function (I.e open loop base mapping), you're onto a loser
Apart from keeping the cat hot, ignition is pretty irrelevant, at least for IVA-type limits. EU-5 is a different story...
PS: just about any functional post-96 emissions system should achieve IVA limits. From a specification point of view, it is a walk in the park. Don't over think the problem
If it is running with a switching (upstream) lambda sensor, and it is actually switching (0.7-2.0Hz), than chances are you have an exhaust manifold leak or at least a leak upstream-ish of the sensor. Don't forget the exhaust is pulsating flow, so the leak could also be somewhat downstream of the sensor too.
Downstream lambda voltage (narrow-band sensors) should spend lengthy periods near or less than around 0.7V if all is running well. Above 0.8V and it's likely producing CO - sometimes a lot!
All of the above assumes the cat is hot and functional. If the cat is a long way (or a large surface area - think performance tubular exhausts...) from the manifold flange, it will not stay lit during extended idle periods. NB, at 1.5% CO, your cat is not likely to be the key problem
If you're trying to make it work without a closed-loop lambda function (I.e open loop base mapping), you're onto a loser
Apart from keeping the cat hot, ignition is pretty irrelevant, at least for IVA-type limits. EU-5 is a different story...
PS: just about any functional post-96 emissions system should achieve IVA limits. From a specification point of view, it is a walk in the park. Don't over think the problem
Edited by AER on Saturday 16th February 11:48
AER said:
More than about 0.6% CO suggests its not running lambda=1.0 (extreme wild cams notwithstanding)
If it is running with a switching (upstream) lambda sensor, and it is actually switching (0.7-2.0Hz), than chances are you have an exhaust manifold leak or at least a leak upstream-ish of the sensor. Don't forget the exhaust is pulsating flow, so the leak could also be somewhat downstream of the sensor too.
Downstream lambda voltage (narrow-band sensors) should spend lengthy periods near or less than around 0.7V if all is running well. Above 0.8V and it's likely producing CO - sometimes a lot!
All of the above assumes the cat is hot and functional. If the cat is a long way (or a large surface area - think performance tubular exhausts...) from the manifold flange, it will not stay lit during extended idle periods. NB, at 1.5% CO, your cat is not likely to be the key problem
If you're trying to make it work without a closed-loop lambda function (I.e open loop base mapping), you're onto a loser
Apart from keeping the cat hot, ignition is pretty irrelevant, at least for IVA-type limits. EU-5 is a different story...
PS: just about any functional post-96 emissions system should achieve IVA limits. From a specification point of view, it is a walk in the park. Don't over think the problem
Like riding a bike, it seems you never forget how to do an emissions cal ;-)If it is running with a switching (upstream) lambda sensor, and it is actually switching (0.7-2.0Hz), than chances are you have an exhaust manifold leak or at least a leak upstream-ish of the sensor. Don't forget the exhaust is pulsating flow, so the leak could also be somewhat downstream of the sensor too.
Downstream lambda voltage (narrow-band sensors) should spend lengthy periods near or less than around 0.7V if all is running well. Above 0.8V and it's likely producing CO - sometimes a lot!
All of the above assumes the cat is hot and functional. If the cat is a long way (or a large surface area - think performance tubular exhausts...) from the manifold flange, it will not stay lit during extended idle periods. NB, at 1.5% CO, your cat is not likely to be the key problem
If you're trying to make it work without a closed-loop lambda function (I.e open loop base mapping), you're onto a loser
Apart from keeping the cat hot, ignition is pretty irrelevant, at least for IVA-type limits. EU-5 is a different story...
PS: just about any functional post-96 emissions system should achieve IVA limits. From a specification point of view, it is a walk in the park. Don't over think the problem
Gassing Station | Engines & Drivetrain | Top of Page | What's New | My Stuff