Check my thinking here...

[Kevfra]

Comments please - I have four O2 sensors, located just before and after the first set of cats. On most header systems these cats are eliminated, and on most headers there are four bungs to receive the O2 sensors. Sometimes - often I hear - you still throw a CEL because there simply isn't enough difference in readings and the computer has a hissy fit. But even if you don't throw a code here is my question:

What is the computer doing to your mixture?

Seems to me the system will be reading more hydrocarbons without the benefit of a cat, and seeing more hydrocarbons it is going to try to run the engine more lean - cut back on the amount of gas in the mix. That's not good. Comment on that thought please.

Next, I have seen the mod using a spark plug fouling adaptor to remote the sensor away from the exhaust stream but still give it enough sniff to be happy. But again - the sample is going to be different than it would be with a cat in the loop, and the computer is going to make an adjustment, good or bad.

I'm wondering if anybody has actually been able to get a read on what the computer is doing in the case of headers, and if it is a healthy response or potentially harmful response. For example there are some folks that have experienced premature cat failure on the second sets still remaining, and they attribute this to hot gas coming through the headers and burning out the cats. Seems to me a cat makes the gases even hotter than straight pipes - that's how it works - and the exhaust without primary cats should be cooler if anything. But what if the computer made an adjustment to lean the mixture - that would make the exhaust hotter and possibly cause the damage some have seen.

Comments?

Final thought. The second set of cats have a bung not being used. When installing headers why not use O2 sensors up front as normal, but extend the second set to the secondary cats, rather than to the second set of bungs on the headers? This gives you the closest to expected sampling that the computer would be looking for. Might have to extend the wires but that's not a big deal. Anybody tried this?

These things keep me awake at night...

:huh:

Edited June 8, 2008 by Kevfra

[gaby]

I will try an electronic fix (Resistor and Capasitor) and post results.

[george996]

Comments please - I have four O2 sensors, located just before and after the first set of cats. On most header systems these cats are eliminated, and on most headers there are four bungs to receive the O2 sensors. Sometimes - often I hear - you still throw a CEL because there simply isn't enough difference in readings and the computer has a hissy fit. But even if you don't throw a code here is my question:

What is the computer doing to your mixture?

Seems to me the system will be reading more hydrocarbons without the benefit of a cat, and seeing more hydrocarbons it is going to try to run the engine more lean - cut back on the amount of gas in the mix. That's not good. Comment on that thought please.

Next, I have seen the mod using a spark plug fouling adaptor to remote the sensor away from the exhaust stream but still give it enough sniff to be happy. But again - the sample is going to be different than it would be with a cat in the loop, and the computer is going to make an adjustment, good or bad.

I'm wondering if anybody has actually been able to get a read on what the computer is doing in the case of headers, and if it is a healthy response or potentially harmful response. For example there are some folks that have experienced premature cat failure on the second sets still remaining, and they attribute this to hot gas coming through the headers and burning out the cats. Seems to me a cat makes the gases even hotter than straight pipes - that's how it works - and the exhaust without primary cats should be cooler if anything. But what if the computer made an adjustment to lean the mixture - that would make the exhaust hotter and possibly cause the damage some have seen.

Comments?

Final thought. The second set of cats have a bung not being used. When installing headers why not use O2 sensors up front as normal, but extend the second set to the secondary cats, rather than to the second set of bungs on the headers? This gives you the closest to expected sampling that the computer would be looking for. Might have to extend the wires but that's not a big deal. Anybody tried this?

These things keep me awake at night...

:huh:

from what i read the sensors ahead of the cats are for contolling the mixture and the set after the cats is to confimr that the cats are working (i.e. no effect on mixture). i guess you would get a cel that the cats are not fuctioning.

[Topless]

I agree with your reasoning. Out of 100 986 owners who have added headers most have wrestled with these problems. Several have killed their motors. It seems the laundry list of downsides far outweigh the possible upsides of adding these. I have not tested them myself.

A good set of 285 track tires might shave 6 seconds off a 90 second lap. A good set of headers might shave 1/10 of a second. I don't see any point in adding them except for the peacock factor of a small bump in your dyno run.

[gaby]

I will try an electronic fix (Resistor and Capasitor) and post results.

I try with the resistor and the ecu gave me couple O2 codes. I put in the sparkplug foulers and haven't got a code just yet 3 days.

[ofishbein]

I have a 2001 Boxster S. My local independent Porsche specialist installed Dansk headers and extended the second set of O2 sensors aft of the secondary cats. In conjunction with that - and an EVO cold air intake - he had Powerchip program the ECU with the modifications in mind and I've had no trouble at all. In fact, my car still passed the smog sniffer for the DMV with no problems. The sound is fantastic and it seems to have way more power than before (you can save your somments about my butt dyno being unreliable - its the only one that matters to me). Much more responsive and no CEL after about 5,000 miles.

[jmatta]

Are the factory exhaust manifolds with cats that restrictive that you are seeing material gains by going through the whole replacement process? I left my original manifolds/cats in place, but replaced the secondary cats with bypass pipes on my '02 Box S. Not only did I remove a lot of weight, but the car sounds remarkably better and certainly breathes easier. Don't know if replacing with headers is really worth the bother; anyone have reliable dyno results that show differently?

[Topless]

I have a 2001 Boxster S. My local independent Porsche specialist installed Dansk headers and extended the second set of O2 sensors aft of the secondary cats. In conjunction with that - and an EVO cold air intake - he had Powerchip program the ECU with the modifications in mind and I've had no trouble at all. In fact, my car still passed the smog sniffer for the DMV with no problems. The sound is fantastic and it seems to have way more power than before (you can save your somments about my butt dyno being unreliable - its the only one that matters to me). Much more responsive and no CEL after about 5,000 miles.

This does seem to be the most thorough and effective way to add headers and get the most out of them. Others have gone this route as well with good results (i.e. car runs well no CEL). Not cheap. Will still probably not pass smog cert. in most states. Not sure how much tangable performance gains (faster lap times) to be had vs tires/suspension mods.

[Jake Raby]

Most producers of headers are not engine developers- if they were they'd be able to provide a known arrangement with known results on the topic of CEL and power output with their products. Most of these people don't own a Boxster and certainly aren't concerned with what occurs after they sell the unit to you.

This has been a battle that we have fought through our development of the M96 engine and it can be both a hassle and it can cost performance or even an engine as some have posted. Getting samples from different portions of the system, as far apart as possible is the key. We have also noticed that differences in fuel blends from one region to the next can also be a determining factor as the change of the "sniff" is just enough to trigger the CEL on some cars, but not on an exacting car and engine with the same system on the other side of the country, or even a state away.

The condition of the engine and what oil is used also makes a difference, this is because higher mileage cars will have more oil consumption as oil passes by the rings and goes out the exhaust system, or when it is pulled past the intake valve guides on deceleration. Some oils are rich in zinc and phosphorous and these also have an impact on the "sniff" of the sensors.

We are working to create an arrangement that offers known results with each model of the M96 engine, meaning that it can be installed with a known power increase and RPM range as well as a guarantee that no CEL will illuminate from it's installation onto the car. This is necessary due to the amount of calls and emails we get on the subject from people who are having these issues and costing themselves a lot of money in repairs or fighting the CEL for months before swapping back to stock exhaust.

Most of our performance engines require headers, so it is a must for us to understand this portion of the engine well enough to provide an optimized arrangement that won't create issues down the road for the customer, or for us.

[jmatta]

Jake, thank you for your comments, as it is always a pleasure to hear from a real professional. Going back to the earlier question on a stock, street motor; do aftermarket headers cause more problems than they are worth and do they provide any reliable HP gain for the effort (and pain)?

I'm curious because I'm assuming it's not worth the effort.

Edited November 1, 2008 by jmatta