When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
Hey I am all for modify cars but my concern would be resale value on a flashed RS.
Agreed, what we're working on is to see if the flashing process can be detected by factory tools. I have a factory PIWIS 2 to be able to run VAL reports on the ECU to see what is/was done. If it can, is it possible to clone the cars ECU with a spare ECU and flash it.
Originally Posted by 911_RS
How will you tune my ECU if ever? Do I need to send my ECU over and have it bench flash?
It would be a bench flash, yes, no remote tuning capability for the foreseeable future. Porsche changed to a different ECU from the turbo, so the work Cobb has done for the turbo's doesn't apply, unfortunately, (and too bad as I'm a Cobb tuner...)
I have a 991-3 so I'm very keen on finding a way to do this that doesn't make things over-complicated, is safe for the ECU and that Porsche won't deny service/warranty coverage over...
Agreed, what we're working on is to see if the flashing process can be detected by factory tools. I have a factory PIWIS 2 to be able to run VAL reports on the ECU to see what is/was done. If it can, is it possible to clone the cars ECU with a spare ECU and flash it.
It would be a bench flash, yes, no remote tuning capability for the foreseeable future. Porsche changed to a different ECU from the turbo, so the work Cobb has done for the turbo's doesn't apply, unfortunately, (and too bad as I'm a Cobb tuner...)
I have a 991-3 so I'm very keen on finding a way to do this that doesn't make things over-complicated, is safe for the ECU and that Porsche won't deny service/warranty coverage over...
Spare ECU
Hopefully they dont use a serial # to identify the ECU as some manufacturers are getting smart.
have a few friends that are master techs at dealers and will be able to run some of the plans by them to see what they would look for and maybe even send them the ecu, or check my car to see if they find anything... Will take a bit, but want to make sure we check everything...
Also Charles had a chance to sleep on the data we've generated and has a few more ideas he wants to try to work through the torque dip and still provide some resonance scavenging...
The biggest difference we've found so far between exhaust tuning a 996/997 GT3 and the 991 is that 991 uses a MAP sensor for load and 996/7 uses a MAF for load. When more air is coming through a MAF, car easily adapts to it as it's interpretted as more load on the engine. A wide open throttle a MAP sensor is just seeing atmospheric pressure and any additional flow has to wait until the 02 sensors pick it up and trim it. This is why a flash tune for the car will have a dramatic effect in conjunction with our exhaust.
I'll work on the detailed explanation of what's going on through the dip, tonight and tomorrow. Will be a bit technical, but I think will help everyone to understand what's going on and what we're trying to optimize!
have a few friends that are master techs at dealers and will be able to run some of the plans by them to see what they would look for and maybe even send them the ecu, or check my car to see if they find anything... Will take a bit, but want to make sure we check everything...
Also Charles had a chance to sleep on the data we've generated and has a few more ideas he wants to try to work through the torque dip and still provide some resonance scavenging...
The biggest difference we've found so far between exhaust tuning a 996/997 GT3 and the 991 is that 991 uses a MAP sensor for load and 996/7 uses a MAF for load. When more air is coming through a MAF, car easily adapts to it as it's interpretted as more load on the engine. A wide open throttle a MAP sensor is just seeing atmospheric pressure and any additional flow has to wait until the 02 sensors pick it up and trim it. This is why a flash tune for the car will have a dramatic effect in conjunction with our exhaust.
I'll work on the detailed explanation of what's going on through the dip, tonight and tomorrow. Will be a bit technical, but I think will help everyone to understand what's going on and what we're trying to optimize!
Jamie we appreciate the hard work. Nice to see tuners that care and not just slap a tune in and not care about future warranty issue etc....
Crazy that Porsche went away from a MAF. MAP tuning, in my little mind, has always been crude but I know nuffing compared to Porsche calibration engineers.
Impressive results, even more massive with a tune. Can't wait to see what that produces. I'm surprised the car is allowing substantial gains like this, nice work!
A couple things we know about the 991 that are different from the 997 GT3's...
1. 991 is Direct Fuel Injected
What does this mean. Essentially it means the fuel injector fires directly into the cylinder at the top of the piston. In a 997 or port injected engine, the fuel injector fires at the backs of the intake valves.
2. 991 uses a Manifold Absolute Pressure (MAP) sensor and Throttle Position Sensor (TPS) for load calculations. 997 uses a Mass Air Flow (MAF) sensor.
A MAP sensor measure the pressure inside the intake manifold, at part throttle this is a vacuum as the throttle blade is partially open and and the cylinders and "sucking" more air from the manifold than the throttle plate allows in, therfore creating a pressure differential or partial vacuum. At wide open throttle (WOT) the pressure inside the manifold is basically the same as atmospheric pressure.
A MAF based system is measuring the mass of incoming air into the engine at all times.
So what does this mean, essentially if you do something to change the volumetric efficiecy (VE) of the engine (add headers, change to a high flow air filter etc...) a MAF based system can easily see that more air is entering the engine and calculates it as a higher load and adds more gas. A Map based system speed density) only measures pressure and sees that atmospheric pressure is reached at WOT and tries to trim the fueling with the oxygen sensors, after the gas and air are combusted. The 991 system probably has a RPM v TPS map controlling the main fuelling and the rate of change via TPS sensor is a lot faster than any manifold pressure change. The MAP sensor is probably one of many corrections that the main fuelling algorithm uses to calculate final injection values. The car can't sense that more or less air is in the cylinder and adapt accordingly
3. The 991 has variable valve timing and an aggressive amount of valve overlap at WOT operation. The intake valve can advance up to 50 degrees crank angle and the exhaust cam can retard up to 30 degrees crank angle. This allows for a large amount of valve overlap. And the 991 uses full overlap (exhaust cam at full retard and intake at full advance) at WOT to help over come the stock cars exhaust restriction.
What is Overlap?
This is when the two valves (exhaust and intake) are open at the same time. It’s when the Exhaust system can have an influence on the Intake system. The most ideal situation is the exiting exhaust gases pull the intake air into the cylinder without any of it passing straight through and back out into the exhaust.
What we want is for the Exhaust gases to help pull more intake air into the Cylinder than just that of the cylinders displacement or swept volume. This is why engines exceed a certain VE number.
At WOT, the air is less restricted entering the cylinder (no throttle plate obstruction) and so the air in the manifold is closer to atmospheric pressure. As the Piston descends on the intake stroke from Top Dead Center (TDC) to Bottom Dead Center (BDC), the pressure in the cylinder across the piston drops, which creates a differential with the pressure within the intake manifold (atmospheric) and air rushes into the cylinder. This pressure differential is not constant and changes with RPM and valve event timing. Variable valve timing allows us to maximize the differential pressure by moving the cams and opening and closing the overlap period.
Why the Dundon/RK headers are a bit different.
We use the reflected resonance pulses of the exhaust gasses expanding into the primary tubes to help scavenge (suck) the air out of the cylinder. This reflected pulse needs to arrive at combustion chamber when the intake valve and exhaust valve are both open during the overlap period to take advantage of the scavenging effect of the exhaust system. Ideally you tune the resonance pulse to help the engine where it's VE is starting to decrease so we can improve peak horsepower. That is what our headers are tuned for and why you see a big increase in power from 6000-9000rpm over stock.
Why there is a torque dip
OK, now that we've gone through the background here's what we believe is going on. Our headers are increasing the volumetric efficiency (VE) of the engine and allowing the spent combustion gasses to get out of the cylinder faster and more completely. This in turn allows more fresh air and oxygen to get into the cylinders.
At 5000-6000rpm at WOT the engine's valves are at full overlap, the exhaust valve is open and due to the lower gas flow restriction and the resonance tuning of our exhaust, the burnt combustion gasses leave with a higher velocity than with the stock exhaust. The intake valve is open at this time as well, and new fresh intake air enters, but because the exhaust valve is held open some of the intake air also leaves the cylinder out through the exhaust valve.
Typically in a port injected engine (like the 997) the gas is already mixed with the intake air as it enters and is therefore more dense and heavier than just air so will sink into the cylinder and is less likely to just flow through. Since the 991 is DFI the intake air is lighter and more likely to just flow back out. Although this is a common trait with overlap, the overlap period in a DI engine has a greater effect on cylinder filling.
So now we have a cylinder with less than a full amount intake air. Now the ECU tells the fuel injector to add the normal amount of fuel as if all the air was in the cylinder. This creates a rich combustion and lower torque than normal.
Now wouldn't the oxygen sensors see this rich condition and start to trim the fuel back? The O2 sensor readings we see at 5000-6000 rpm indicate about 0.95 - 0.96 lambda, this means a little more fuel than air. The thing is the O2 sensors are averaging from the entire bank and are relative slow to react. So we have this fresh air exiting the cylinder mixing with the rich combustion from one of the other cylinders. So in essence the O2 sensors are getting fooled a little bit.
How else we can tell this is the likely scenario is if the car was running lean (.96 lambda at WOT is a bit lean) it should be pulling a bit of ignition advance to prevent knock or have a bit of a knock signature. We see none of that, car isn't showing any knock and isn't pulling any ignition timing.
Another indication that we know the cam overlap is too much at WOT is the fact that at 50-70% throttle where the cam overlap is not as much as at WOT, the car has prodigious power, more so than at WOT at the same 5000-6000 rpm. We plan to get a dyno run of this to illustrate but ran out of time
So where do we go from here.
The easiest way to maximize the horsepower potential and remove the torque dip would be to reduce the exhaust valve overlap in this range and match the exhaust valve retard at 70% throttle with an ECU flash. We're putting all the pieces for this together. So far Porsche's stance on flash tuning hasn't changed, if a warranty issue happens and it has nothing to do with the flash or aftermarket parts, the warrany will be honored. I plan on vetting and testing the exhaust and flash tune on my own car and checking it with PIWIS2 (factory tools) for any and all issues.
Charles is working on a few other ideas with resonance tuning to dampen the effect in the midrange rpm, so that we can dampen the torque dip. This isn't the ideal solution for power as the headers we have designed are the ideal for power, but these other ideas may allow a bolt on solution with increased VE (power) and no (or at least) reduced torque dip at the 5000-6000rpm area
DFI tuning is definitely different than port injection and gives us more to look at and learn. Charles and I are excited to be working on it and extending what we know works and findng out what doesn't! That's part of the development process and as long as we let the data and knowledge direct our process, we'll get the best possible outcome!
I refuse to use 304 on a GT3 exhaust as there is no longevity in the product, only initial profit for whoever makes and sells it. 321 has always driven the price of my exhaust systems up high, but atleast this way making them 10 at a time I can keep the cost low and return the savings to the purchaser. plus with 321 I can keep wall thickness down and retain strength as well.
my 20ga 321 996/997 center exit system weighs 24 lbs
It's my understanding that the factory uses 304 for the stock manifolds...?
Were working on the tuning side. I have 2 spare Ecu's. I'm cloning one with PIWIS2 and leaving the other uncloned. We will work to clone the clone without using PIWIS and try to flash tune both. The point of this is to ensure there's no issues with the flash tuning and that PIWIS isn't necessary to perform it
We tried a few variants on the headers design, we are able to dampen the midrange effect and move it up and down the rpm band. This confirms that the issue with the midrange hp flattening out is due to exhaust cam retard causing overlap. The headers are doing their job and we don't want to sacrifice the top end. We could redesign the headers to match the stock system with a little less back pressure, but this isn't the best design and we believe the gains are there and makes the car lighter and faster especially for the track guys.
We're going to install the headers on a local unaffiliated 991-3 who is going to give us and you all an unbiased review. He may even be able to get us with and without vbox results...
The headers are available for sale now, for those that are so inclined.
It's my understanding that the factory uses 304 for the stock manifolds...?
OEM exhaust components are made from a few 300 and 400 grade stainless steel grades. 304 does have a few slight drawbacks mainly higher thermal expansion and not being a completely pure stainless grade.
I repair cracked OEM exhaust components weekly at my work. Exhaust manifolds usually crack at the flange where the pipe is welded to it. Expansion and contraction and lack of outter support usually leads to the weld cracking and the exhaust leaking drawing in air into the exhaust. I have seen 996tt boxsters and a cayman all have these crack bad enough to where the exhaust manifold pipes were barely still attached to the flange and 3 I have seen completely break off. Also these are all 2mm wall thickness material.
The 991 manifolds are thinner and lighter and the flange is only one piece one layer 1/4" thick. All of the others are thicker material and a 2 piece flange for more strength and better expansion properties. My main worry with the 991 gt3 factory headers is the catalytic converter section. Most cats have 3 layers to the outter support and heat shield. A thick section to hold and support the cat core, another stamped piece with insulation or stamping to hold the core and support it. Then another layer with an air gap and or insulation to keep the core away from the atmosphere to retain heat and lower heat transfer. All of these are important to keep the cat core from failing and melting/ coming loose. Ferrari used this same style cat on 348-355-360 models and they have a high failure rate. Very high.
I choose to use 321 as it is a pure grade of stainless that retains a high level of strength at very high temps (normal exhaust operating temps at continual use) compared to 304/409/400/mild steel. With this you can also run a thinner material and have the Same or more strength with much less weight. 321 has very minimal thermal expansion compared to 304 and also does not contaminate like 304 will. 304 is not pure and can and will most of the time begin to rust as there is still a high enough iron content in it.
I have repaired more aftermarket exhaust pieces than total setups I have built. I get many in every month that have cracked and failed due to poor design, so so fabrication and or final product had too many corners cut in the thought of saving money to have a higher profit margin at the expense of the end consumer. This is just poor quality and lack of pride in ones products that they sell. This is my opinion of course. I have probably re welded 40-50 center muffler delete pipes cracking in every way possible. A lot of it is poor welds that look great but do not have a full weld penetration and thus have no strength. In some of these cases 304 vs321 would not matter.
Really loving this thread. Im convinced with Jamie & Charles knowledge and attention to detail i would go now here else for a full system for my GT3. Great work guys.
One question. Given the VE, overlap issues you've seen with the exhaust on dyno would you hesitate to comment on wether the aftermarket 3rd muffler delete pipes commonly fitted to these cars could actually reduce some power on the curve by altering the pressure in the system?
Really loving this thread. Im convinced with Jamie & Charles knowledge and attention to detail i would go now here else for a full system for my GT3. Great work guys.
One question. Given the VE, overlap issues you've seen with the exhaust on dyno would you hesitate to comment on wether the aftermarket 3rd muffler delete pipes commonly fitted to these cars could actually reduce some power on the curve by altering the pressure in the system?
If I go this way I will give you my business.
the side muffler deletes should not have any adverse effect on power in this way as this issue deals with the primary exhaust pipe length and not secondary. There is about 10hp in deleting the side mufflers and center muffler to full 3" on a 991 gt3, real world, not a 25whp gain like some claim. side deletes will save weight and make noise but i would not install them to make any improvements in HP. If i wanted to inflate my HP numbers, then i would post the very first 991 gt3 dyno run vs my very best ,and there is a spot in there where we are making about 45whp over "stock" but the engine has not been run several times to show consistency.
The issue with the power loss in the midrange is because of excess valve overlap, and this is associated with need for emissions at WOT and not power built around the stock exhaust system. Emissions is the biggest issue to deal with in this day and age when designing/building an engine for a street car in this day and age.
Compare a 991 gt3 or even a 997.2 3.8RS to say a 04-08 3.6L GT3. The 3.6 engines DO NOT have varriable exhaust cam timing where all the newer engines do. All of the engines with variable exhaust cam timing make less power in the 2500-6500rpm range PER DISPLACEMENT and OUTPUT than those that do not. On the engines with no exhaust cam adjustment, there is slight reversion but due to intake resonance tuning there is minimal loss. There is some EGR retention, but you can still have a well filled cylinder and then have a decent amount of power made with correct tuning. on a 996 gt3 there is a dip about 6000rpm in power about 20lb ft less tq there than 5000rpm or 6500rpm. this is a loss in cylinder fill due to reversion of exhaust gas back up the intake tract slightly. at 5000 and 6500rpm those are intake resonance tuning peaks so there is positive flow into the cylinder at those RPMS. on a 997 gt3 3.6, they have 2 resonance flaps to help reduce this loss by about half or more so it is a much flatter TQ curve. more air in the cylinders to combust and burn means more cylinder pressure which is more power. more cylinder pressure means more NOx.
on a 3.8+ and 991 WITH exhaust cam timing the intent is to take advantage of the short header runner length, reduce exhaust reversion but increase EGR retention. why? because EGR retention lowers NOx emissions because there is less air and fuel in the cylinder to burn and allows it to burn slower creating less NOx. This is the main reason i can see why all manufacturers are doing this. 0 HC and 0 CO emissions is easy these days, just throw 8 Cats on the car and you are good. but NOx is not that easy, you can not clean it up easy, so yo have to prevent it. The less cylinder fill (TORQUE) on the 2010+ engines in the 3000-6500 rpm range is to help battle NOx emissions, reducing cylinder fill and pressure, but also reducing reversion, helps to have a slower more controlled burn that will be NOx free or close to it.
most people forget that an engine needs to PASS emissions a idle, part throttle, WOT everywhere, not just cruise.
on the 991 GT3, i logged EX cam timing and IN cam timing vs load. up to about 70% load the engine has minimal overlap and is not at full advance/retard like at WOT. in my exhaust setup on the stock engine tune, at say 5500rpm, if you are at 50-70% throttle, there is NO loss in power as the valve overlap is minimal. it may even be more power than stock actually. once you roll into it above 70% it begins to retard exhaust cam timing pretty quick. When driving the car at part throttle through the 4-5-6000 rpm range it is almost impossible to notice a difference as the car is actually making more power than stock, but going from part throttle to full at 5-6000 just isnt a huge increase like say at 8000 that you would feel. It just isnt noticeable.
i have had 2 other people drive another 991 gt3 against the one with the header and exhaust and through 5-6000 there is no performance loss car to car, but below 4500 and above 6500 the one with exhaust pulls away with ease.
long story made longer (as it seems)
battling emissions design on these newer cars is tough as you have to understand the design and intention of the engine setup, then from there make adjustments. Once we have the ability to fully tune the exhaust cam timing and ignition timing there will be a huge gain in power everywhere.
Just one last thing to note, i built a custom exhaust system with same primary length as stock but no cat and 3" exhaust back from there, the dip is mostly gone, and it mimics the stock engine power just has about 10-12hp more everywhere. I am not sure the cats are much of a restriction, just that the power is to be had with a correct length tuned header system with corrected cam timing. the custom tuned headers make almost 20whp over this setup at 8200rpm
11-04-2015, 02:05 PM
