Close

    • The big Porsche 911 header/manifold test on a turbo upgraded Porsche 991.2 Carrera 9A2B6 3.0 - Kline vs. Vektor vs. Fabspeed

      The PorscheBoost Porsche Performance Forums were able to acquire the three major designs of aftermarket 991.2 9A2B6 3.0 turbo headers. From top to bottom what you see are Fabspeed, Vektor Performance, and Kline.


      These will be tested on the PorscheBoost Project 991.2 which is currently running the BoostAddict Stage 1 package comprised of Tial M660 turbochargers, Fabspeed bypass pipes, IPD Plenum, and Gintani tuning for water-meth injection on 91 octane.

      The car is currently at 540 whp on a Mustang dyno on stock inlets, stock intercoolers, and stock manifolds.

      A turbo upgraded 991.2 3.0 will be moving much more air than stock so it is an excellent test bed for the aftermarket manifolds. A restriction will be quickly exposed.

      Take a look at the factory manifolds courtesy of Fabspeed:




      Fabspeed believes there is room to improve especially in the top end.

      Let's compare the design elements of the three aftermarket options starting with Kline.

      Kline Pictures:












      Vektor Pictures:






















      Fabspeed Pictures:




















      So which is the best design? Well, it depends. PorscheBoost believes all three will show an improvement over stock and the choice comes down to what you want to pay and your goals.

      Vektor's headers for example are ceramic coated by Swain Tech. This adds quite a bit to the price. Vektor insisted on sending a coated set for testing so this is not apples to apples but if a ceramic coated set offers a noticeable difference over the non-coated headers we certainly will see that.

      Fabspeed is non-equal length and has the shortest runners. It looks like a very simple, efficient, and straightforward design. It also feels very solid in your hands.

      Kline is equal length but at points has some bends that look crimped as if the diameter might not be the same throughout. Like Fabspeed, Kline's material is 304 Stainless Steel.

      This brings us to the major difference in material between Kline, Fabspeed, and Vektor. 321 Stainless Steel in the Vektor manifolds is stabilized with titanium. 304ss is cheaper and not as good at high temperatures as far as longevity is concerned. For a street car application both should likely last a very, very long time though.

      304 Stainless Steel Tubing
      304 stainless steel is an ideal stainless for the construction of performance headers. It is widely used because of its ability to withstand corrosion. Because itís ease of weldability and excellent forming capacities, 304 stainless steel makes an ideal choice.

      321 Stainless Steel
      Considered the preferred material for high temp applications such as turbo headers, 321 stainless steel offer extreme resistance to high operating temperatures. Used extensively for turbo applications, 321 stainless steel can operate at above 1650 degrees without sacrificing its structural integrity.
      Now, if the material is a concern to you or you intend to track in desert conditions meaning long WOT periods in hot weather you may want to go to Inconel which Kline offers their headers in. Of course, this means a price increase for the material.

      The inconel manifolds also weigh less:

      Quote Originally Posted by Kline
      Materials
      Available in premium grade Inconel 625 superalloy

      Available in 304 grade stainless steel

      Manifolds (Inconel 625), 1.5kg each.
      As buying three sets of headers and then testing them is already a major expense it is obvious why we went with 304ss for the Kline headers.

      Vektor also appears to have the largest radius bends and runners merging around a central point rather than a flat merge. The Vektor headers also certainly are physically the largest and heaviest.

      Pricing:

      Fabspeed: $2295.00
      Vektor: $3250.00 + $325 coating = $3575.00
      Kline $2350.00

      I will say Fabspeed and Kline jumped at the chance for this comparison. No, the headers were not free but Kline did discount a bit for the test. Kline had headers to me first and it literally took a few days. How Kline shipped so fast from overseas I have no idea.

      Fabspeed was glad to participate and had a set to me quickly once they did a production run.

      Vektor took the longest and did not communicate as well as the other two. Vektor did not seem eager to participate and did not feel they needed to prove anything. PorscheBoost.com feels as Vektor being the newest of the group in the Porsche market actually has the most to prove. Vektor's coated headers were the most expensive by far.

      As stated, all three will be dyno tested on the PorscheBoost.com Project 991.2. After which 60-130 testing will commence with one chosen option. This same option will also be tested at the drag strip.

      Ideally, we would test all three sets on the dragstrip as well but it is not feasible to rent the drag strip for a separate header test each time. BoostAddict.com rented the Famoso drag strip for May 5th and anyone is free to see how the car performs in person.

      Dyno results should come within the week.

      As I will not be able to obviously run all three designs, if you would like one of the sets left over at a discount after testing it done feel free to PM.

      #project991

      This article was originally published in forum thread: The big Porsche 911 header/manifold test on a turbo upgraded Porsche 991.2 Carrera 9A2B6 3.0 - Kline vs. Vektor v. Fabspeed started by Sticky View original post
      Comments 21 Comments
      1. T.O.'s Avatar
        T.O. -
        Cool, looking forward to seeing the test results! I have the full Fabspeed set up on my car Headers, high flow cats, x-pipe and their tips in black chrome.
      1. Aatish's Avatar
        Aatish -
        Eagerly waiting for results! Thanks for update
      1. 3DMetal's Avatar
        3DMetal -
        What are the weights of each pair ? How much does the stock manifolds weight ?
        Use a Shipping scale.
        Looking forward to the results, but also the explanation of the testing procedure.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by 3DMetal Click here to enlarge
        Looking forward to the results, but also the explanation of the testing procedure.
        Well you certainly can flow bench them if you like but I think you know the testing procedure.
      1. spdracerut's Avatar
        spdracerut -
        Can you get some closeups of the insides of the merges?

        I brightened up the pics of the runner entrances. It looks like Fabspeed now casts the whole head flange.

        Any notes on packaging and provided hardware and gaskets?

        Attachment 58583
      1. spdracerut's Avatar
        spdracerut -
        Also, is the tune going to be optimized for each header? I imagine any flow differences and pulsation differences (equal vs unequal) can allow tweaking of ignition timing.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by spdracerut Click here to enlarge
        Any notes on packaging and provided hardware and gaskets?
        I should have mentioned this, thanks.

        All were packaged very, very well. Vektor goes a bit above and beyond with their form fitted inserts.

        Vektor also provided the most hardware and best gaskets.

        Kline provided essentially what you see.

        Fabspeed also provides some hardware which is not pictured.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by spdracerut Click here to enlarge
        Also, is the tune going to be optimized for each header? I imagine any flow differences and pulsation differences (equal vs unequal) can allow tweaking of ignition timing.
        The tune will eventually be optimized but probably just for one set. We're going to see where the curve is impacted with each design.

        I want maximum top end so the tune will be optimized for that header design.
      1. spdracerut's Avatar
        spdracerut -
        Last note, I don't think the drag strip would tell you much. All it takes is 10 degrees difference or a shift in the wind to skew the results. Or surface prep. Too much variability. Hopefully with the dyno, the ambient conditions will be the same. Just have to make sure the car is at the same coolant/oil/IAT temps, etc.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by spdracerut Click here to enlarge
        Last note, I don't think the drag strip would tell you much. All it takes is 10 degrees difference or a shift in the wind to skew the results. Or surface prep. Too much variability. Hopefully with the dyno, the ambient conditions will be the same. Just have to make sure the car is at the same coolant/oil/IAT temps, etc.
        I don't disagree but we will drag test the headers nevertheless. If we pick up MPH they are doing what we hoped they would do.
      1. spdracerut's Avatar
        spdracerut -
        One last note. I've looked at datalogs from stock CS with Vektor headers, CS and GTS with Vektor headers with Cobb stage 1 and 2 tunes, and GTS with Kline headers with Cobb stage2 tune. The stock tune with headers, no base timing pulled. With Cobb stage 1 and 2 tunes with Vektor headers, there was so much individual cylinder timing pulled, it pulled timing from the base value too. With the GTS with Kline headers and Cobb stage 2 tune, there was individual cylinder timing pulled, but not enough to cause the base timing retard. The Kline header car was in cold weather, the Vektor header cars were normal temps and one hot temps as variables. But my hypothesis is the Vektor headers increased the flow so much that the fueling became insufficient causing knock and the base timing to get retarded. As your tune was optimized for the stock headers, you may run into the same issue with timing getting pulled once you install the different headers which of course leaves power on the table. It would be good to see if/how much timing gets pulled with each set of headers.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by spdracerut Click here to enlarge
        As your tune was optimized for the stock headers, you may run into the same issue with timing getting pulled once you install the different headers which of course leaves power on the table. It would be good to see if/how much timing gets pulled with each set of headers.
        Possibly but remember I'm running meth injection.

        Either way, we will tune for the set that has the curve I'm looking for.

        Keep in mind we will be doing inlets and intercoolers afterward.

        Race gas as well. Hopefully we can get these turbos up to 30 psi.
      1. 3DMetal's Avatar
        3DMetal -
        But my hypothesis is the Vektor headers increased the flow so much that the fueling became insufficient causing knock and the base timing to get retarded. As your tune was optimized for the stock headers, you may run into the same issue with timing getting pulled once you install the different headers which of course leaves power on the table. It would be good to see if/how much timing gets pulled with each set of headers.
        You would have serious issues in the tune and logs before that where the case. For example: You would see it in short and long term fuel trims if you where flowing more air than can be adjusted for, which would be the case if you had lean knock as your suggesting.
        Your calculated load vs boost pressure would show a large difference.
        You would see lean knock as misfire counts under load.


        No matter how well flowing the headers, it still backed up by the turbo/wg issues. In a serial connection, you cant flow more through the engine is its limited by the next limit, the same exit choke.. the turbo.
      1. spdracerut's Avatar
        spdracerut -
        For reference, CS before/after Vektor headers, Cobb stage 1 tune. Datalogged using a generic logger. One run on stock tune, 2 runs with Cobb Stage 1. You can see after 4k rpms, the base ignition timing got retarded twice and Lambda went richer too.

        Attachment 58584

        GTS with Vektor headers, Cobb Stage 2 tune, 115F intake air temp. Cobb AP datalog. Two different pulls, with the second seemingly a bit more heat soaked as the base ignition timing starts off lower at 3500rpm on the -5 run. Anyway, you can see all the individual timing correction across the whole rpm range along with the base timing getting retarded.

        Attachment 58585

        GTS with Kline headers, Cobb Stage 2 tune, 68F intake air temps. What's interesting is the base ignition timing is already lower (temp correction table?), but it's still getting significant individual cylinder timing correction up until 4500rpm or so, then it's clean.

        Attachment 58586

        These are all Cobb OTS tunes. As we know, lots of room for improvement because of the additional mods on these cars compared to the car Cobb tuned. I think the car that Cobb developed the stage 2 tune on only had high flow cats.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by spdracerut Click here to enlarge
        These are all Cobb OTS tunes.
        Haven't we seen these Cobb OTS tunes retard ignition and boost on the factory manifolds?
      1. spdracerut's Avatar
        spdracerut -
        I haven't, but there's gotta be someone out there with a stock hardware car with Cobb stage 1 tune that can datalog and provide data.
      1. Alpina_B3_Lux's Avatar
        Alpina_B3_Lux -
        Great project! Really excellent to go step by step and test every component like that.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by Alpina_B3_Lux Click here to enlarge
        Great project! Really excellent to go step by step and test every component like that.
        Thank you Click here to enlarge

        Just time and money.

        Plus the 991.2 Carrera is just an absolutely awesome vehicle.
      1. Sticky's Avatar
        Sticky -
        Gintani already has the car on the dyno.

        They believe the Fabspeed design is the best. We'll see.
      1. spdracerut's Avatar
        spdracerut -
        They're in Van Nuys, CA right? Looking at the weather, hopefully only a 5 degree range during testing of all the headers.