Intake and exhaust system design, a few questions

[Freddie]

I'm currently involved in designing the intake and exhaust system on our car this year, and I have ran into some troubles trying to find reasonable runner and exhaust pipe lengths.

First of all; we're going to have a 4-2-1-exhaust system on this years car, and I have been trying to design primary pipe lengths with regards to torque peaks in the system using this formula:

L = (129540*E.T)/(RPM*6)

E.T being 180+40 degrees (2007 Honda CBR600RR engine) and RPM (500rpms before peak power) being used as a variable with a step of 500 through-out the whole expected rev range. But this only gives me primary length, and to me it seems really strange to use the same formula for secondary pipe length, when the exhaust pulse will differ greatly being from two cylinders. How do you calculate secondary pipe length? I tried to reverse-engineer lengths using known diameters, but the numbers I get are not realistic. These numbers will be taken into Ansys for some CFD evaluation afterwards, but I want to be at least ballpark before I start to use the computer.

Also, what's your experience; when I design my exhaust system, should I try to get the peaks to match up with the peaks the runners produce (we're trying to design our runners to produce a broad range of efficiency peaks through-out the rev range), or use them as "cushions" and have them peak in between the peaks from the intake? Have any of you guys tried any of these theories, and which are your thoughts about the driveability of the two concepts? Personally, I'm leaning towards the "cushion" principle, but I am a bit worried that the negative aspects will overtake and eliminate all peaks instead of creating multiple peaks (more driveability).

[penna]

The "cushion" technique is what we used and it seemed to produce relatively good results. A flat torque curve that makes the car easy to drive is a lot more important than peak power

As for you exhaust runner length calculations, with 4-2-1 you generally want your tuning peak to come from the secondary merge not the primary merge for packaging and a few other reasons. Your peak will obviously not be as strong, but this is one of the good things about 4-2-1 systems they provide power everywhere rather than all coming in a rush like 4-1. Your primary merge is more there to help with scavenging from other cylinders and reduce pumping losses.

Also I'm guessing you have a formula for this as well but don't forget about your tube diameters, these are probably more important than your lengths are

[Z]

Originally posted by Freddie:
Personally, I'm leaning towards the "cushion" principle, but I am a bit worried that the negative aspects will overtake and eliminate all peaks instead of creating multiple peaks (more driveability).
Freddie,

Your above quote suggests you think,

"eliminate all peaks" (ie. a flat torque curve) = "negative", and

"multiple peaks" = "more driveability".

Why??? http://fsae.com/groupee_common/emoticons/icon_confused.gif

Z

[Freddie]

Originally posted by Z:
Freddie,

Your above quote suggests you think,

"eliminate all peaks" (ie. a flat torque curve) = "negative", and

"multiple peaks" = "more driveability".

Why??? http://fsae.com/groupee_common/emoticons/icon_confused.gif

Z

Z, I was a bit unclear of what I meant there, I'll try to rephrase myself:

If I tune the intake and exhaust to peak at different RPMs, will I still be able to exploit the extra torque/power from the peaks, or will they effectivly work as restrictors, leaving me with a flatter but lower torque curve?

[Z]

Originally posted by Freddie:
If I tune the intake and exhaust to peak at different RPMs, will I still be able to exploit the extra torque/power from the peaks, or will they effectivly work as restrictors, leaving me with a flatter but lower torque curve?
Freddie,

Firstly, I think your goal should be to have a torque curve that is as flat as possible.

Secondly, there is the question of how to achieve that, and that is more complicated. As well as intake and exhaust pressure pulses, you must also consider valve timing (which is at least as important). They are all interdependent.

I am not sure what methods (equations) you are using, but very briefly here are some tips.

1. A low pressure pulse should arive at the exhaust valve just before it closes (ie. during E/I overlap) to empty the cylinder of exhaust gases and start drawing in air/fuel.

2. A high pressure pulse can arrive at the intake valve at any time while it is open, but preferably just before it closes, so as to ram as much air/fuel into the cylinder.

3. During valve overlap, NEVER have a higher pressure at the exhaust valve than the intake, because this blows exhaust gases up the intake pipe...

4. A restricted engine develops low pressure in the intake at high revs (because of the restrictor!), so point 3 above becomes a problem if the engine has large valve overlap (which most "sport bike" engines have). So less overlap may be a good idea.

5. In summary of above, during valve overlap ALWAYS try to keep pressure at intake valve above that at exhaust valve.

Anyone who has actually tuned an FSAE engine is welcome to comment/criticise the above... http://fsae.com/groupee_common/emoticons/icon_smile.gif

Z

[Warpspeed]

All very true.

One other point, if you can get the throttle butterfly closer to the intake valve DO IT.

This really helps with exhaust reversion when running at less than full throttle, especially with crazy valve timing.

While none of this will show up on the dyno at full throttle, it can really make a big difference in the car when you are transiting rapidly on and off the throttle.

While all decent multi cylinder bikes have individual throttle bodies, if you are designing a whole new from scratch induction system, maybe with a single, or a turbo, get the throttle bodies or carbs as close to the cylinder head as you possibly can.

There is no more power in doing this, but it will be much better behaved and far more drivable.

[Freddie]

First of all, thanks Penna, Z and Tony for answering and trying to help me out with this. My only problem now is that I thought I had a grasp on the subject prior to this ...

Penna; I have done some calculations on diameter, the problem is that the formulas I have are based on pipe length, so it's not really helping me at the moment.

Z; Most probably, yes, but if possible I'd like to know a bit more about my options on this. If I lose 2-3%, I won't mind but if the losses are 10-15-20% it's another decision. We can't let the drivers have it too easy, can we? http://fsae.com/groupee_common/emoticons/icon_cool.gif
Since we won't alter the valve timing (this is really not my area of expertice, but wouldn't that require new camshafts?), maybe I should start there? Also, we are looking into a fairly large plenum that hopefully will help us a bit with the backpressure issue (given that the pressure difference in a larger plenum will be smaller due to [delta]V being smaller).

Warpspeed; We really haven't thought about that at the moment, but wouldn't placing the butterfly make the fluid dynamics inside the plenum really scary, with wakes, turbulence and such?

[Warpspeed]

Originally posted by Freddie:

Warpspeed; We really haven't thought about that at the moment, but wouldn't placing the butterfly make the fluid dynamics inside the plenum really scary, with wakes, turbulence and such?

At full constant wide open throttle, it probably does not matter where you place the throttle.
Once wide open, the throttle is supposed to offer zero resistance to flow, so it virtually disappears.

What happens in the plenum at WOT is a completely separate issue to throttle placement.

[Buckingham]

First of all; we're going to have a 4-2-1-exhaust system on this years car, and I have been trying to design primary pipe lengths with regards to torque peaks in the system using this formula:

With your 4-2-1 exhaust, try the following:

1. design your primaries for as short as possible, then build two sets of primaries that have an extra 12" of primary length.
2. Put one set on the dyno and test it.
3. Cut 2" off the primary, re-test.
4. Repeat step 3 until you can't cut it shorter.
5. Cut your 2nd set of primaries to the length that tested best on the dyno.

[Freddie]

Originally posted by Warpspeed:
At full constant wide open throttle, it probably does not matter where you place the throttle.
Once wide open, the throttle is supposed to offer zero resistance to flow, so it virtually disappears.

What happens in the plenum at WOT is a completely separate issue to throttle placement.

Yes, but wouldn't the butterfly affect the fluid dynamics in the plenum at all times, WOT or not? This year we had a really bad design (mine in panic-mode, after a few problems with older designs) which resulted in an uneven distribution of air to the cylinders. I don't want that to happen again, since the fuel-air-ratio will be a nightmare to adjust in that case.


Originally posted by Buckingham:
With your 4-2-1 exhaust, try the following:

1. design your primaries for as short as possible, then build two sets of primaries that have an extra 12" of primary length.
2. Put one set on the dyno and test it.
3. Cut 2" off the primary, re-test.
4. Repeat step 3 until you can't cut it shorter.
5. Cut your 2nd set of primaries to the length that tested best on the dyno.

That is obviously an option, but would like to have all "my" parts done in CAD for various reasons (packaging, weight balance etc), and my hope to be able to design both these system for an "optimal" compromise, and not go for the "design one, adjust the other" route.

[Warpspeed]

Originally posted by Freddie:
Yes, but wouldn't the butterfly affect the fluid dynamics in the plenum at all times, WOT or not?

At less than WOT, the throttle is there to deliberately kill Ve and strangle the engine in the most brutal way possible.

At WOT the throttle effectively disappears, and that is the only time you really need to become much more interested in induction flow dynamics and resonances.

[Freddie]

Originally posted by Warpspeed:
At less than WOT, the throttle is there to deliberately kill Ve and strangle the engine in the most brutal way possible.

At WOT the throttle effectively disappears, and that is the only time you really need to become much more interested in induction flow dynamics and resonances.

Maybe so, but this is my way of thinking about it: At less than WOT, the throttle will in some way disrupt the flow from the throttle head to the plenum (regardless of it beeing a butterfly or slider, or any other valve type), resulting in a non-uniform flow into the plenum where it will continue to flow through the intake runners and towards the intake valve (I can't look at the plenum as an infinite source of air for the intake valves, can I?). If we're unlucky, at some point below WOT the flow can be non-uniform over the four-cylinders due to the non-uniform flow into the plenum due to pressure reflections, turbulence and maybe even wakes within the plenum. This will result in a lower output, but more importantly; it will change the fuel-air ratio of the combustion and may cause unneeded stress on the engine components. Isn't it an easier approach to have a tapered runner from the throttle to the plenum, allowing the air flow to stabilize and entering the plenum in a sort of uniform way, just with a lower speed? If that is possible, of course.

Or am I completely misunderstanding you, and you are talking about the size of the plenum?

[Warpspeed]

Originally posted by Freddie: allowing the air flow to stabilize and entering the plenum in a sort of uniform way, just with a lower speed?
That is exactly what does happen at very small throttle opening.
Less flow = less airspeed everywhere.
Everything evens out with very low pressure differences between runners.

WOT there is a hurricane in that plenum, high airspeeds, much dynamic impact pressure, much pulsing resonance, larger pressure drops and far more turbulence.

If you can sort that mess out at full power, small throttle air distribution takes care of itself.