When I first took delivery of my R56 MCS, the first thing I noticed was how quiet the exhaust was; I could barely hear the engine with the windows down going through a tunnel! This had to change. With the hype of ALTA and Milltek's exhaust success, I decided to take on a turbo-back exhaust challenge of my own.

I used 100% 304 stainless material for this project; from the turbine exit flange, the 3" 16 and 18 gauge tubing, hanger rod, v-band clamp, catalytic converter housing, muffler construction, and flex bellows are all 304 stainless. I used my Millermatic 210 MIG welder with 308 welding wire to tie it all together.

The downpipe consists of plate stainless cut to the same dimensions as the stock downpipe flange with a 2.6" outlet hole. I used 16 gauge 3" tubing for the downpipe, containing two 3" radius bends [very tight], and a Dynatech 3" in/out catalytic converter. The 3" tubing was compressed to match the hole size of the flange to ensure excellent flow with minimal turbulance. The shape of the downpipe closely resembles the stock piece, to entertain the flow conversion from radial flow coming out of the turbine to laminar flow. I decided to push the v-band clamp and flex bellows aft of the front subframe to give myself maximum ground clearance with the 3" tubing and the subframe.

This is the stock downpipe. Actually very nice flow characteristics for a downpipe....except that restrictive cat!


My downpipe with high-flow cat and higher diameter tubing


From the downpipe, the rear section is very simple, being essentially straight back. I used a Magnaflow 3" in with dual 2.5" out polished 304 stainless perforated core muffler. You can see lots of daylight looking in one end and out the other of this muffler; it mainly smooths out the acoustic tones for a deep mellow sound. I used polished stainless tips that flare from 2.5" to 3.5", and angle cut them to match the angle of the rear valence panel for maximum aesthetic integration.

New exhaust tips, top view


New exhaust, bottom view


The sound quality is fantastic! The system is quiet when you're cruising, and deep and loud when you're on the throttle. The sound is quite similar to a Subaru, with lots of fun warbles and throaty character without ever being ricey or shrill. On the dyno the exhaust got rave reviews with its sound quality.

On the dyno, performance wise, is a different story. While the exhaust freed up power, it also changed the way the ECU controls the boost controller. Below 3000 RPM, the ECU runs the turbo with 1.4-1.5 PSIg less than stock. Peak boost of 206 KPa is the same as stock that I've seen on the dyno, however at higher RPM the ECU now allows slightly more boost than with the stock exhaust, the caveat being it clamps down on the boost hard near redline. The overall is; above 3k RPM, the turbo-back averages 9wHP gain and 10wLb-Ft gain. Including the change in boost level across the board, the system averages a 6.0wHP gain and a 5.9wLb-Ft gain over stock. Of note, average boost level is about 0.5 PSIg over stock. Here is my dyno plot of an average run my R56 did stock, and an average run my R56 did after the turbo-back. These numbers were done on the same DynoJet with the same operator and are SAE corrected.



The peak numbers are like this:
Totally Stock: 172wHP and 202wLb-Ft
Turbo-Back: 183wHP and 212wLb-Ft

As you can see, the turbo back doesn't have as big a torque swell at lower RPM. Boost threshold RPM to boost controller interaction is identical to stock. Good torque gains with the turbo-back exhaust from 3000 all the way to 6400 RPM.

Why the ECU decides to change the boost controller curve is beyond me. To add to the confusion, the ECU also changes the ignition timing curve slightly, consistently, compared to totally stock. The way the ECU changes it's interaction with the boost controller now makes the car much easier to putz around town and not have the turbo spike into action when it's not desired, yet when you want to get a move on, the torque gain is very nice.

My results are totally different than ALTA and Mach V Dan's results. I'm not sure if the flow characteristics of my exhaust system are wildly different than theirs, or the dynos are that different. What I want to know however, is if ALTA and Mach V Dan did any datalogging to observe their MAP, MAF, Ignition, and IAT's while doing the comparisons to stock. Why does my R56 seem to know there's a change, and alter it's boost and ignition curves to change the way the car drives versus stock. It's all very strange to me. Gasoline direct injection turbo strikes again! Let me know what you think MINIacs!

Cheers,
Ryan

 

Don't have a clue on your questions, but a pretty system none-the-less.

Matt

 

Nice job on the exhaust! Looks like many hours of work.

We did not datalog boost or any other info on our quick-and-dirty exhaust test, so I can't add much. Your stock torque curve does look very much like mine. That's strange about the dip at 2200-2900 rpm. Our mega-free-flowing dump tube simply exaggerated the stock behavior.

--Dan
Mach V
FastMINI.net

 

Thanks for the compliments guys! I'm kind of surprised to hear Matt being stumped on this! Thanks Dan for letting me know how you went about testing your dump tube. Right now I feel like I've accidentally discovered something, and I'm not sure if it's good or not.

I've looked at Dan's and ALTA's dyno charts; both are using a Dynapack which is cool, however what is strange is the characteristics of the chart, in terms of the RPM the torque peaks and dips happen are different than my MCS, especially after the exhaust. Like Dan said, their dynos show that the exhausts emphasize stock behavior, not change it. I'm wondering if somehow my exhaust design is having a negative effect. If it was however, it wouldn't be making a peak power gain of 25wHP on the top end of the RPM range! So I can't imagine there being a big restriction in flow. Is the shape of my downpipe or some strange aerodynamic restriction causing the boost reduction at the lower RPM, but then it goes away at higher flow rates? I understand how turbo exhausts are generally impervious to sloppy design because the turbo turbine messes up traditional header design with primaries and collectors and secondary length. I wonder what's going on with the ECU how it determined to change the boost curve.

 

is **** poor boost control via the wastegate. It actuall looks like an underdamped oscillator. If that's the case, the dip will occure based on time more than RPM, so maybe you guys with the dyno charts should put a time axis on it as well. Jeff and the Alta boys have been working on it a bit. But it seems that the mysteries of this ECU programming will take a bit to unwind. Thanks for being one of the first to start finding out what it does.....

Matt

 

Ryphile, what gear did you test in?

--Dan
Mach V
FastMINI.net

 

--->Dan: 4th gear, like always.

--->Matt: I've found the boost controller is very much a Boost vs. RPM situation. It doesn't matter if you're cruising on the highway, surface streets, or dyno; the boost is very predictable versus RPM. Time doesn't appear to have any relation to how it controls boost. Based on that, I think the underdamped oscillator, while an excellent theory [and I wish it was that easy!] isn't it. With the stock exhaust, if you were cruising at 2k rpm and punched it, you'd have 190KPa by 2500rpm. Now with the turbo-back, cruising at 2k rpm and punch the throttle, you'll only get 182KPa by 2500rpm. This can be repeated in any gear, regardless of time. At X RPM you'll get Y boost level. Very consistently Boost vs. RPM.

 

I guess that even educated guesses can be wrong! I have no clue why anything like this would happen. Seems really stupid to me. I guess it's a case of "good enough is good enough" or maybe "almost good enough is good enough!"

Matt

ps, the Europeans have a head start on hacking this stuff. Dimsport or anyone like Roland at GTT have any insights?

 

Yes Matt, the Europeans certainly have a head start for hacking the ECU code, however it also seems there is very little interest over across the pond. I don't think Roland has done a whole lot...at least he hasn't published anything except his test drive impressions. It seems most tuners are waiting until their warranties expire before doing anything.

dmh has announced that MTH has an "ECU flash" available for $950, but in reality he doesn't have anything in-hand to sell, let alone information on what the flash changes. That sort of ambiguity doesn't cut it! I hope dmh posts information on that flash when it actually is ready.

 

I can attest to the sound of Ryans ride, its sweet and as far as the numbers go, well, it will be figured out sooner or later and once it is, the big breathing exhaust is already there but in the meantime, an 11 hp/10 ft/lb of torque gain ON the dyno is nothing to sneeze at!

 

Well, now he's gotta put up a sound clip

 

LOL, now I just need someone with a video camera to put together a clip!

 

great stuff as always... full of info etc etc ... blah blah... you go boyiieeee!!!

 

Thanks Tuls; you rock!

 

Ryephile,

Congrates at being hands on as always. Not to get too far off topic, but I am welding up my new DP and 3" back exhaust and I am using the 304L tubing with 308L filler on a miller matic180 mig. Any tips on welding this stuff? I am using 100% argon...You can PM if you like, just figured I would post it here since you put some details of what you were using in the post. I just started practice welding today and its definelty different than mild steel. The only way I can get a decent bead is to run the nozzle right up against the seem, making it really hard to see your puddle. I am running the voltage at setting 2 & the wirespeed at 25 with little to no burnthrough. I am finding it hard to keep the wire flowing smoothly and wasnt sure if there was a trick to it. Any insight will be much appreciated!

Thanks,

Cole

 

That's a very nice job.

The torque problem puzzles me too... Could it be a problem of emissions?
I mean, maybe the lamba feels that the kat lets more co2 pass through and turn down the boost? Just a theory

 

Adaptive ECU-not enough cycles to readapt???

 

--->TrickleX: You have a PM! Stainless sucks to weld compared to mild steel!

--->Gardus: I would doubt the boost controller mapping change is related to emissions, as emissions are only a concern for the manufacturer when it's on the dyno following the EPA's time vs. mph chart, which is very low power stuff, never at full throttle.

--->dwjj: It's possible; though from the second I test drove the new exhaust to just pulling in the driveway, the car seems to drive the same. It's been about 400 miles with lots of ignition cycles; more than enough time to get the basic adjustments out of the way. That said, I'm more interested in [i]how[i] the ECU knew there was a change to the exhaust. There is no code for the catalytic converter, so the ECU must think the cat is working well enough. An idea I had was the aerodynamics at the turbine exit have changed so much that the vacuum-operated boost controller is being pulled open at lower RPM when it doesn't think it has to pull as hard to keep the wastegate closed. If that was the case however, the ECU would see the change in boost and adjust to pull harder...unless it can't, and the ECU is on the fringe of throwing a code for being almost off the boost controller map. No, it seems that the ECU wants the boost to be different. Hmmm....

 

Ryephile,

My guess is the ECU just sees the turbo build boost faster than with the stock exhaust system and reacts to bring it back down. Of course that doesn't answer why the other exhaust systems tested so far don't seem to show this behaviour...

Do you have any pics of the turbine from when the exhaust was off (assuming it was visible)?

Did you notice if the stock downpipe was double-wall tubing by any chance

-JL

 

For our GS autocross R56, we just ran a straight pipe from the second cat. HP remained the same but we gained torque. It's a cheap and easy solution to improve the exhaust sound and lose weight. I'm sure the aftermarket will disagree as they would rather sell an expensive "must have" can.
We have determined that most of the newer BMW products do not like being on the dyno. The ECUs are very adaptive, as the numbers were all over the place with each run. It took alot of pulls to get some median results.
Oh yeah, the car sounds awesome!

 

--->jlevy: I'm not sure if I snapped any pics of the turbine exit. The compressor and turbine wheels are tiny. The turbine exit has this chamber where the wheel exducer combines with the wastegate flapper. The chamber has some decent volume to it, then it necks down to a 2 5/8" opening for the flange to the downpipe. The stock downpipe is not double wall; the upper portion is two stamped sheetmetal pieces welded together, and after the cat is standard tubing.

--->gwalsh: For what it's worth, I found the R56 to be very predictable on the chassis dyno. It either hits standard boost or overboost, and it seems to trace those two lines depending on which mode you've put the ECU in. I'm talking a +/-1.5 wLb-Ft variance provided you start the runs with the ECT's and IAT's consistent. This is far improved over the R53 and its obnoxious ignition "program".
Thanks for letting us know your cat-back findings. You're saying you gained power in lower RPM but it fell back to stock power at higher RPM? I'd love to see if your boost levels changed like mine did!

Cheers,
Ryan

 

Another great job Ryan! Can’t wait to see you in a couple weeks and have a chance to hear your new pipes.

 

Thanks Sndwave for the compliments!

--->eagletangogreen; I quoted your post over in my thread because it seemed more on-topic than highjacking the Milltek thread.

My main reason for maintaining a catalytic converter is the smell of cars without them irritates my sinuses. I had a Mustang Cobra years ago with long-tube headers w/out cats and every time I got out of the car I smelled like emissions and gave me a headache for the rest of the day. No more test-pipes for me.

Take a look at my dyno chart above. Change the units from HP to PSIg and you have an idea what the boost curve looks like. It's a roller coaster ride. My stock R56 peaked at 15.3 psig, and after the exhaust it peaked at 15.6 psig, however the shape of the boost curve changed too.

The R56 MCS's static compression ratio is 10.5:1. The only reason it's not exploding pistons from the factory [at 15.3 psig with premium unleaded] is because of direct injection creating a virtually homogeneous mixture.

I haven't taken apart a Prince engine yet, so how sturdy the internals are [as well as the new aluminum block] are up in the air.

I saw a consistent 183 wHP and 212 wLb-Ft on the Dyno Jet I frequent, versus 172 wHP and 202 wLb-Ft of bone stock. I used the same oil and wheels/tire for both sessions, and those figures are SAE corrected for atmospheric conditions.

Cheers,
Ryan

 

Wow! 15 PSI I was not expecting it to be that high? A BOV would not be a bad idea with all that pressure. Easy and cheap to install, a little insurance.

I see you chart.. Considering you have no piggyback or reflash (IMO re flash are a joke!) the stock ECU is not doing that bad. Your torque looks a a bit cleaner than the bumpy stock one. The dip in the low end with the aftermarket exhaust.. Looks like the ECU is pullings some timing on both stock and aftermarket exhaust. You can see that boost spike at the end of the graph . I hope someone makes a harness for a piggyback system to clean that torque/power up. It would control all timing and fuel once boost is built with the ecu controlling everything underboost, a stand alone would be costly but more resolution in tuning.

Give me some time play with this car a bit, but most important a way to control the ECU's timing/A/F (i.e AEM, HKS Fcon, UTEC) is needed to safely start exploring with this engine. Reflash is not going to cut it.

Bye the way I commend you for building your own exhaust system. The best part of having a car, is working on it.

 

The R56 MCS already has a BOV, it's electronically controlled by the ECU, and it's built into the turbo compressor housing. The reason the R56 runs such high boost is because the little K03 turbo is not very efficient at higher flow rates, so they need that much boost to get the torque output desired. A larger turbo, like a GT2560R, would make mondo more top-end power at that boost level.

Regarding the change in boost curves, it's not verbatim related to the power curve, just similar, especially at lower RPM. The ECU is not changing it's timing or AFR's between exhausts for the entire RPM band, I have the datalogging saved showing that. The boost at the upper end, around 5900 RPM, is shelved, not spiked as the power plot would imply.

I'd suggest reading up on Bosch's DI-Motronic system like I'm doing right now. There's a crossover point in torque output where the system switches from stratified charge combustion to homogeneous charge combustion, and wherever that point is I'm not sure, I haven't done much part-throttle datalogging to find out. Basically in stratified charge combustion, the throttle body is opened to reduce pumping losses, and torque is determined by the AFR [which is always in lean-burn mode during this combustion state]. Only during homogeneous charge combustion does the system switch over to more traditional throttle body usage with relatively traditional AFR's.

Tapping into the ECU code and developing a new flash will be the ticket, despite your bad history with flashes this ECU would require a massive amount of piggy-back interfacing to get even minutely useful improvements in certain operational parameters. ALTA found out that the R56 expects to see a certain boost level within a certain time at certain RPM points, and if it doesn't it shuts down the party and goes into "half power" mode.

Food for thought,
Ryan

P.S. thanks for the compliments on my exhaust!