It's too cold to be in the shed today, and I've noticed in some recent threads that a lot of people still don't "get" air flow. So I thought I'd copy and paste something I did a few years ago, it's a longer version of the airflow page in the Holden six notes. I hope there's something of interest in it.
Making power is largely about airflow. Of course there's more to it than that but without adequate flow none of the other stuff really matters. It's just not possible to give any more than a brief overview in a just a few pages, but there are some very good books on the subject if you'd like to learn more and I've included links to further reading at the end of this page. But hopefully this will whet your appetite…
In the last section we looked at ways of maximising the quantity of air in the cylinder. But specifically, what's important is the mass of the cylinder charge. And that mass is in proportion to the temperature of the charge - there's more air in a 33cu.in. charge when it's at say 20 degrees than at 50 degrees. Cooling the intake charge is probably the quickest and easiest ways of improving the VE that you'll find. With most vehicles it's possible to arrange the intake to pick up air from a point that reduces the temperature and increases the pressure when compared to a standard under-bonnet setup.
What about vapourising fuel in the intake tract as a way of decreasing charge temperatures? It's currently fashionable in some circles to shift the injection nozzles upstream for this purpose. With oxygen-carrying fuels like methanol (which incidentally also has a high latent heat of vapourisation) it's pretty easy to show that early vapourisation can show a net benefit. But with fuels like petrol it isn't so clear cut. When a liquid like petrol vapourises as well as absorbing a lot of heat it also increases its volume by a huge amount. So on one hand we have a decrease in temperature helping to increase charge density, but on the other the massive increase in volume of fuel displaces some of the working fluid (air) from the charge. This may help explain why so many engines (especially those that are flow limited) are so insensitive to petrol atomisation in the intake - and in fact flow limited engines seem to make more power when the fuel is fed as a solid stream. With petrol at least, it's probably not worth the effort to encourage fuel vapourisation in the inlet.
Let's have a look at air. Like anything else with mass it's subject to Newton's laws of motion. In simple terms, when it's stationary it wants to stay that way and it takes work to get it moving. When it's moving it takes effort to change its speed or direction. And a moving volume of air contains some amount of kinetic energy, the energy that was invested in it to get it moving in the first place. Recovery of at least some of this energy is an important part of power production. But while air has mass it is also compressible and elastic. As a result of this it's usual for all the particles of air and fuel in a port or runner to be travelling at widely varying speeds. They might even be travelling in different directions. If you could see the air it would look more like a "Slinky" with waves bouncing back and forth than a column of incompressible liquid. Remember this; it's important and something that we'll touch on again shortly.
Pretty much everything we do to maximise flow has the same aim - we want the airflow to be smooth and controlled and within certain velocity and turbulence limits. Having said that pretty much every part of an engines ducting will carry flow that is basically turbulent; we just need to limit that turbulence to a point where it doesn't disrupt the flow too severely. In practice we might use a bellmouth on the carb intake to accelerate the air smoothly into the throat. In the carb we need to accelerate the air further so that we have a localised low pressure zone into which the fuel can be pushed by atmospheric pressure. Once the fuel is added we then need to allow the air to slow down again, hopefully in a way that lets us recover energy. Now this re-expansion phase is important to understand, not just in terms of carburation but throughout the engine as a whole. So we'll take a closer look at this process now.
I think everyone reading this will be familiar with the shape of a venturi. Typically it will have a large inlet and outlet diameter with a smaller throat at some point in the middle, where the velocity is high and the pressure low. The inlet transition is generally a simple radius shape, and is relatively short - usually a fraction of the length of the outlet or downstream transition. There is a good reason for this; air can be accelerated into a constricted area relatively easily over a short distance without much loss of efficiency or energy. Expansion is a different story altogether though; if we try to make it happen too quickly the flow will separate from the walls and become turbulent and this in turn leads to a dramatic reduction in flow. This is a really important concept to understand; a very large proportion of the productive port work will be related to minimising flow separation. Novice builders tend to put too much effort into easing the entry side of restrictions and not enough effort into the downstream side. In simple terms the upstream approach to any high velocity zone isn't so important, the downstream side critically so.
These "flow expansion zones" are everywhere throughout the engine flow path. Some examples: the transition from throttle body to plenum with an injected engine, or carb to plenum with a carburated engine. The port throat to turn is another area, the runner to head port transition of a siamese port engine another, the valve seat to cylinder yet another. On the exhaust side the collector is a good example. As mentioned previously novice builders often seem to sweat over things like bellmouths and port matching when the real gains are to be made at the opposite end. Keep in mind though that anywhere there is wet flow you may be working with conflicting requirements. For example the turbulence created by the rapid expansion under a carb may reduce bulk flow but it may also help with fuel mixing, and which is the dominant effect may only be determined by experimentation. This would explain why "Super Sucker" style carb spacers work very well on some engines but not others.
Here’s something else to consider: the energy of air. Pressure and velocity are proportional; as the velocity goes up the pressure goes down, but the energy remains constant. As we’ve already seen, we can accelerate it across something like a valve seat pretty easily. The secret is in pressure recovery - in other words, expanding the air smoothly and without creating turbulence, in order to recover as much of its original pressure as possible.
So we've established that while air doesn't like to be sped up, it likes being slowed down even less. In practical terms any surface that diverges from the flow path at an angle sharper than 10 to 15 degrees will cause flow separation. This would include any port or runner walls as well as the faces of the valve and seat and any approach and departure angles. As far as wall finish goes I think it's pretty well recognized now that a highly polished finish is not what we want. I've found that for good flow a plain old carbide finish is as good as any, though most professional porters will finish with a coarse sanding roll for the sake of appearance. This doesn't seem to hurt flow, but if you have an old-style head with polished walls it will definitely hurt performance from both a flow and fuel suspension perspective.
We could summarise porting (whether it's head porting or manifolds etc.) in just a few lines - porting in essence is simply is an attempt to do one of the following:
1.To keep the air velocity within certain limits (in order to limit turbulence and the resulting flow losses)
2.To prevent the creation of turbulent flow
3.To keep the air attached to the surrounding surfaces and prevent them from becoming detached and turbulent.
You could further distill that summary down to just five words - we port to reduce turbulence. It’d be a whole lot easier to manage the air if we could just see it - and if we could we’d probably realise that it doesn’t behave the way most of us expect it to, and that it usually doesn’t take the path that we’d expect.
It’s time for a folksy analogy to help us visualise air - airflow as highway traffic. Let’s say we have a nice, wide freeway, ten lanes wide connecting point A to point B. Point A represents the inlet of a manifold runner, while point B is just downstream of the intake valve seat. The cars on this freeway represent the particles of air travelling down the runner. The freeway isn’t a uniform width, in places it’s twelve lanes wide, in another it’s only eight wide. There are a couple of sweeping bends and one or two sharper curves as well.
In the real manifold our objective is to get the greatest possible mass of air from one end to the other. On our freeway, we want to get the greatest possible number of cars from A to B. How do we do this? Obviously we want the cars to be travelling as quickly as possible, but not so fast that they spin out of control on the bends, or are unable to change lanes where the road widens or narrows without losing control and slowing all the traffic.
The cars are travelling swiftly but still under control. Cars in each lane might not necessarily be all travelling at the same speed but they all stay headed in the right direction so the traffic flow is pretty orderly and the volume of traffic is good. This is like laminar flow. The traffic in the two outer lanes will tend to be a bit slower (the boundary layer) while the lanes in the middle will tend to go the fastest. Laminar flow.
Let’s now upset things a bit. We’ll make a chicane halfway down a straight stretch by suddenly narrowing the road to three lanes and then widen it again just as suddenly. Then we’ll dig some humongous potholes here and there, plus we’ll convince the drivers to try to drive faster than they are really able.
The result: chaos. Cars are braking for the chicane, merging suddenly and knocking other cars into a spin. The cars are still spinning when they emerge from the chicane, into the path of other traffic. Some cars manage to maintain direction, some are barely moving, some are doing 200mph, some are spinning and some are going backwards. This is turbulent flow - there’s a hell of a lot going on but there aren’t many cars turning up at point B. Just when they’ve recovered from the chicane they hit the potholes at speed and it’s on again - cars spinning, bouncing and jumping in all directions. Their progress is badly impeded and so is that of the cars in whose path they land. Turbulent flow. It’s spectacular but not very effective.
Remove the chicane, fill the potholes and get the drivers to go back to driving as fast as they can but not faster. Now sit back and watch the traffic flow for a while. If traffic really flowed like air you’d see some surprising behaviour. For a start, whenever the cars encountered a curve they wouldn’t remain in their respective lanes lanes like normal traffic. Instead, they’d all tend to crowd over to the inside of the bend, leaving the outside couple of lanes almost empty. Basically air takes the classic racing line: hugging the apex and leaving wide. Put a couple of left and right bends in series and the air will negotiate them just like a racing driver - by clipping the apexes and straightening the path as much as possible.
This corner hugging behaviour is something that may surprise some. You might expect that with increasing velocity that the flow would tend to go wider and wider, just as a liquid like water does in an open channel. In reality, air can turn sharply even at quite high velocities, and the crowding of flow to the inside of the bend tends to make the flow velocity even higher at this point.
Another characteristic of air is its “stickiness”. A moving particle of air will tend to make the surrounding particles move along with it. To go back to our cars-on-the-freeway analogy for a minute, if we were to make the centre lane of cars travel at a certain speed, the cars in the lanes at either side would also tend to move forward, and the lanes alongside those in turn would also move at a progressively reducing speed as we approach the outer lanes. This is something that has implications for fuelling, and intake and exhaust system design, and can be used to advantage.
Now is probably a good time to mention the Coanda effect. Simply put, it’s the tendency of a moving stream of fluid (like air) to be attracted to a nearby surface. Stand at the kitchen sink and turn on a tap. Now hold a spoon so that the back of it just touches the stream. See how the water flow curves around, following the shape of the spoon? That’s the Coanda effect in action. The faster the flow, the closer the fluid will hug the surface.
This attachment is quite a handy phenomenon for tuners. It help us to turn mixture around the short turn radius above a valve seat for example. Providing the surrounding surfaces are smooth and don’t diverge too sharply, the air will want to remain attached, and this helps to minimise turbulence. As we saw earlier, maximising flow is all about reducing turbulence, and making the flow as laminar as possible.
You’ll often see head porters fill in the “dead spots” - ie. the areas of a port where the flow velocity is very low, often on the outside of a bend. Flow and performance usually improves and the man with the die grinder may say it’s because he’s reduced the cross-sectional area at this point, increasing the flow velocity. I tend to disagree - the bulk of the flow will be in exactly the same spot as it was before, so velocity will be largely unchanged. But with the outside wall crowded in, the flow is now attached to both walls and this has a guiding and stabilising effect.
Anytime the flow becomes detached there will be turbulence and bulk flow will suffer. If you check an old-school intake port with a velocity probe and a string you’ll probably find that the flow over the short turn radius is very fast but quite stable. The long side flow on the other hand will often be very slow, very turbulent and may even be rolling or going backwards in spots. Anything you can do to keep the flow attached will help. If you’ve never used a flowbench before you’ll soon discover a couple of things: one, most of what you thought about airflow behaviour simply isn’t true. And two, air isn’t the light, thin fluid you thought it was. Instead, it’s a heavy, springy, squishy, sticky medium.
Hopefully by now you have some idea of the way air flows through a passage and also have an understanding of laminar and turbulent flow. There is a practical aspect to these two different types of flow, and it’s essential that the tuner understands this, otherwise he could end up doing a lot of work for little or no gain.
This is it: when the flow is largely laminar, there is the possibility for power gains through keeping it laminar. For example, anything you can do to keep the velocity within reasonable limits, keep the flow attached to the port walls, smooth and straighten the flow path - in other words anything to prevent turbulence - will improve flow and performance.
The flip side to this is the way we manage turbulent flow. There are some situations in an engine - flow through the exhaust seat and port during blowdown for example - where the flow is very very turbulent. This type of flow couldn’t care less about all those niceties we spoke about with laminar flow, all that matters is that we have sufficient cross sectional area. In other words, if the flow is unavoidably turbulent, about all you can do is make sure the hole is big enough. All those other things - smoothness, straightness and so on - don’t mean a thing, the flow is already turbulent anyway.
An associated side-effect of the above is this: there’s really very little point to developing an exhaust port on a flowbench. The difference between the type of flow that the bench produces and the type that occurs when the valve opens on a running engine is totally dissimilar, so much so that flow testing the exhaust port achieves very little. Of course it follows that traditional rules-of-thumb (eg. exhaust flow should be 75% of intake flow) are basically worthless. If the cross-sectional area of the exhaust port is around 80% of the intake, and isn’t too ugly, then you’ll at least be in the ballpark.
Where the turbulent flowpath is fairly long - as in the case of an exhaust port and primary pipe, it’s certainly a good idea to encourage that flow to become laminar (or at least less turbulent) as soon as possible. Smooth walls and gradual changes in cross sectional area are the general idea. If you do have to have sudden changes in direction or area - where a primary pipe dives downwards from a port face perhaps - then you might as well get it over and done with as soon as possible. There’s nothing to be gained from the old ideas of having a few inches of straight pipe after the port exit; the flow is already turbulent anyway. You’ll do just as well to swing the pipe down straight away, and then worry about reducing the turbulence later.
Up til now we’ve been talking mainly in terms of flow as being steady, like that produced by a flowbench, when obviously a running engine produces flow that is very unsteady. Does this mean that a flowbench is irrelevant? Not at all. Within limits, flow gains in the entire intake tract that are made on the flowbench will translate into gains at the flywheel. The reason I said “within limits” is this: if the engine already has available to it all the flow it can use then there’s not much to be gained by increasing it further. If however the engine is flow limited - a very common situation, especially so with older 2 valve engines - then it will respond extremely well to improvements to flow capacity. Pick up 15% more intake flow and it’s likely that the engines output will increase by a similar amount. Good flow can completely fill a cylinder; by taking advantage of some unsteady-flow phenomena we can improve that even further, and that’s what we are going to look at next.
Edited by oldjohnno, 25 June 2016 - 02:52 PM.
View Garage
