Pipes galore


This is what you get when you have a lot of time and get a bit OCD about pipes.

Again only modelled in EngMod4T not dyno tested. Data has a good correlation to dyno figures for the baseline. Total runtime for these graphs is about  14 to 15 hours.

Same race spec engine (Custom race cam, good head all dancing all singing) only change is the diameter of the centre branch on two variants of manifold.

One is a 3-1 manifold from down under (which was pretty optimised) the other a LCB manifold based on the lengths Vizard has published, starting at the diameters that where iteratively found to give the best curves for smaller and less racy engines using a high end simulation package. Diameters where varied using small steps (1 or 2 mm).

Graphs are split into inner and outer cylinder pairs. Absolute values are more or less wheel KW per cylinder but not really important. RPM plots are to a heady 10k rpm which is not recommended with a 3.2 inch stroke ! The model does not take into account valve train instability or crank harmonics.

What diameters ? I’m not going to tell as this is a running project and I just wanted to illustrate the general behaviour of both the 3-1 and the LCB. Not in the graphs are the effects of length. Suffice it to say that it does not seem to do much. For an LCB it is very hard to improve on the lengths as published by Vizard 40 odd years ago. Primairy length for a 3-1 (where the centre is actually a secondary length) is not very critical.


The arrows point to the ”hinge point”. Note the outer cylinders get a smidgeon better when you change a pipe of the other cylinders. how do you like them pears?


Note that the curve now has a second node at aprox. 7100 rpm around which it rotates.

Tagged , , , , , , , , ,

11 thoughts on “Pipes galore

  1. Atte Roskam says:

    What is the enginetemperature you loaded as a parmeter?

    • mowog says:

      Coolant temperature is 65c ( i have since changed this to 95c as this works better with the fmep model) . exhaust temps are variable as a siamese engine runs different temps on the inner and outer cylinders.

  2. Keith Calver says:

    For the simulation to have any meaning the tests should be run with ‘in car’ actual parameters. That would include coolant temperature. Typically 75-80 deg C. Though many run higher just because of the ambient temperature where they are. 65 deg C is somewhat low.

    • mowog says:

      Hi Keith

      The fmep model is very (i’d say overly) sensitive to temperature. For 20w50 oil and 65c you lose 25 bhp vs thinner oil and 95c engine temp ,(individual part temperatures are included in the model as well). I’m currently helping a guy for a new engine build and luckily he has his own hub dyno at home. Currently the correlation is pretty decent. More engine dyno data would be good too .

      • Keith Calver says:

        Good grief! That is wholly inaccurate! A 25BHP loss on 20W50 oil! Not what I have seen at all on engine dyno tests. Part of what makes me very wary of simulator data accumulation. Big respect for being able to use it in any case.

  3. mowog says:

    It is wholly inacurrate indeed. For some unfathomable reason it does seem to correlate quite well with gearbox losses.

    You do really have to be wary of confusing your model with reality 🙂

    The general Fmep model is based on a bunch of tests measuring actual friction (is basically a statistical fit type of deal) and then you hope that is will sort of be all rightish for your engine (always a dangerous assumption for an oddball like an A series ).

    Now it produces friction numbers that are in line with what has been measured in literature
    That is where you really need dyno data.

    The fmep model does have a huge impact on absolute power numbers (to the point of being useless for power predictions if you get it wrong) but the biggest gains IMHO (not only my opinion btw) is that you can try out a bunch of ideas that are hard to do in real life and get an idea of what does what.

  4. Keith Calver says:

    The whole ‘drive train loss’ subjectr makes me want to scream. Way too many folk quote such losses as 20-25%. That is SO stupid… so on a 1000bhp engine you are loosing 250bhp? I think not. From some real worldtesting of various gearbox builds done in conjnction with engine dyno tests applied to a constant known rolling road, typically losses on the Mini engine are in the region of 16-20 bhp depending on the gear design used and the build. In this test I purposely built 2 gearboxes badly so they would represent approximately what the first time enthusiast might achieve. The majority of tests gave a 15-18bhp deficit.

  5. mowog says:

    The Fmep model is actually just for the engine friction not the gearbox and drivetrain. So it is even worse at a 20-25 bhp loss. Friction is Bmep based and is about 3,4 isch at 8500 rpm when you get the temp up where it needs to be (the model is based around a 90 c oil temp for pushrod old engines), It also is much more in line with typical losses as found in literature. However the combined in engine loss and drivetrain/tyre friction loss would be somewhere in the 20 to 25 bhp range ( Is based on the engine from a Can-Am series mini so it is far removed from a 90 bhp street engine but about the antithesis of your motto (the valve train is far from standard using a bigger cam bearings, core, lifters etc etc ) and revs quite high).

    250 bhp gearbox loss is would need a bloody big oil cooler on the gearbox 🙂

  6. Steve says:

    I think you need to zoom in to a realistic RPM range of the engines involved, as the effects at 7500 and up are pretty useless for these motors. 1000 to 5000 RPM is not even on your chart and is where they spend 90 % of their life, unless a full on race motor built for just top end. Even then the bottom end and mid range power is just as important if you want a car that is at all drivable. It looks like the red pipe works much better at 5000 RPM and below. but that is not a surprise really. what works for top end usually does not work so good down low.

    • mowog says:

      as I stated it is for RACE engine project rpm range between 5-9 k rpm. It has been build to run this high (it has a very beefy valve train) and has been running in competition successfully for a while like this. This has been part of a whole series of simulations to first match the current set-up then try and predict (hopefully accurately) a planned large modification (that is very custom, and there is no data available).

      so my issues are unclouded and the rpm range is realistic for this application. If you want a /entry level race set-up buy a maniflow/minispares LCB 🙂 much cheaper and will work.

      You are quite right, this would be a pretty horrible engine to sit in traffic with. But it will start to go at 3500 which is pretty ok for a race engine.

      Adding another 24 hours of computation time just to see what it does from 1500 rpm to 3000rpm was not practical

      cheers JR

  7. GTT says:

    Very interesting to read all this stuff. I have a fair bit to do with exhausts and acoustics.

    There is one exhaust variant you have never thought of and TBQH I have never seen for sale.
    I put one of these on a mini I had some yonks ago.
    This consists of putting a “DUMB” branch on the (LCB) centre branch roughly 15″ down.

    This has the effect of transforming your LCB into basically a 4-2-1.
    Exhausts work in fact by acoustic resonance, not in the way that most people describe.

    That means it HALVES the length of the long centre branch and makes it behave like a primary, rather than a long peaky bit of pipe as used on a typical 4-1.

    The “dumb” branch in effect on the centre branch behaves exactly as the outer branches do (when cyl 1 exhaust is shut, so the branch is “shut” and vice versa cyl 4 to 1).

    This explains very much how the outer branches achieve resonance in this phenomenon called “interference” and tend to work quite well making a much wider power band.

    That all causes resonance (a standing wave phenomenon) to add and subtract,- the addition happening first, then the larger subtraction happening at TDC during overlap.

    Try it, you may be very suprised at the results.

    Best regards


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: