This post ended up being pretty long, I did it in a few stages as time permitted. If I've missed the end of a sentence or something isn't clear, please let me know and I'll edit the post and sort it out.
First things first - I keep repeating this, but it's important - I didn't come up with this theory/technology. Plenty of extremely intelligent people who make me look like an idiot worked it all out, tested, refined it, and were also generous to share it (sometimes for a very very reasonable price of a book, other times for free!). That list of people would include (and I'm missing a few no doubt) people that would be well known and others who wouldn't, depending on the specific car marque/circles one occupies. They also might be as infamous for getting some stuff a little screwy and getting other things on the bullseye (which is probably a necessity in car modifications, if you never went down the wrong path occasionally, you've never been anywhere etc). That short list would include the likes of : Smokey Yunick, David Reher, Clive Trickey, Carroll Smith, David Vizard, Larry Widmer, John Lingenfelter, Bill Jenkins, Dave Emmaneul, Jim McFarland, Doug Roe and a bunch of others. I'd also add in Mark Ellis (rip) , who I knew extensively through posts/emails on moparmarket, but sadly never got to meet in person.
The stuff I'm posting below represents my understanding of things, and that understanding was effected by the writings of the above people (and many others). I'm not doing any of this for profit, nor am I directly duplicating anything they've written, so hopefully the post is reasonable and not any attempt to rip them off in any way. I understand if the mods have to edit/delete it.
ok lets talk about why extactors 'work' -
#1 is of course that they reduce backpressure massively (well most do, though some factory exhausts over the years are actually pretty good in this respect - two that come to mind are the cast iron dual outlet torana xu-1 manifolds (for some Aussie content) and for overseas guys, the 'rams horn' chevy small block exhausts. Not perfect but far less restrictive than other factory cast iron exhausts. Less back pressure means the piston moving up the bore doesn't have to push as hard to get the exhaust out. Simple enough, a reduction in pumping losses. On 'most' setups from the 70s backpressure is relatively high, as it's 'quiet' it's also a killer on power output.
There 'was' a myth about requiring backpressure to function. This is about as logical as saying Paris Hilton is a virgin. There's _ZERO_ benefit (for power levels) for having exhaust back pressure, plain and simple it's an old wives tale. For the record, I had an exhaust done by a place local to me around 10-12 years back and I specified 2 inch system (*it was a 1275 mini, in a fairly mild combo, probably putting out around 95-100bhp). I 'picked it up' and they informed me that they had 'done me a good turn' and put a 1 3/4" system on there as otherwise 'there would be no backpressure' - arrggh. I needed the car that day, as aside from being a fun car, I blew the auto in the other car, so needed the transportation, but I never went back there. Point being, that attitude still prevailed at some places in the late 1990s, it might still today at some places so buyer beware. If anything we've probably gone a little the other way, where people are fitting 3-4 inch exhausts on small 4 cylinder cars. Above and beyond the exhaust size/design that gets rid of back pressure (it'll never be truly zero all the time, but let's say 'gets rid of back pressure for all intents and purposes') - any bigger than that will just cost more money, for no benefit, probably drone a lot at part throttle, and present clearance issues both from the ground, and between the body and the axles etc. It also likely gets a bit more attention from the boys in blue, which I'm pretty sure nobody in their right mind wants!
So basically the first big thing of note is the reduction in backpressure. This applies to the exhaust header/manifold/extractor, and of course the rest of the exhaust from there back.
In the most simple of terms if you can't get an exhaust place to make an exhaust system to the dimensions you require, and they don't have a lot of either competition experience in general, or don't want to tell you why, just a condescending pat on the back and a 'we know best' only to give you a restrictive exhaust because 'you need backpressure' - just find somewhere else. The better places out there aren't necessarily the dearest either, ironically enough.
----
Before I go on to #2 I have to add in some more theory. An engines output is of course a product of the whole combination of parts - head/port work, intake manifolding and carburettion or efi configuration, the bore and stroke (and hence capacity) and the exhaust system, and of course the cam profile. And that is a big generalisation
The longer an intake valve is open, the more time the engine can breath in. Believe it or not, they actually close the intake valve _well_ after the piston has reached the bottom and has started to rise. At high enough rpm (and it's all relative, if the cam is short enough, this can happen at lower rpm, for a little while at least) the air coming in has momentum, and even though the piston has started to move back up, it will continue packing in the intake charge. The higher the rpm, the later you can close that intake valve and still have beneficial cylinder filling. But with later and later intake valve closure, at lower rpm, the air isn't moving fast enough to have enough inertia, so the piston can actually push it back up the intake tract - losing some of the cylinder filling, a phenomenon known as flow reversion (*in this case intake flow reversion)
Similarly, at the end of the exhaust stroke, the exhaust valve might not close till after top dead centre, for similar reasons, the exhaust might get enough momentum up to keep plucking the last remnants out of the exhaust. Of course at the end of the exhaust phase, we begin a new intake phase, and the intake valve itself will be starting to open just as the exhaust is beginning to close. This is known as valve overlap.
Imagine we have a very restrictive exhaust system, at lower rpms, when there is valve overlap, it might actually be less difficult to push that last bit of exhaust up and out the intake valve!. This is dreadful, as it slightly heats up the intake charge, as well as meaning that it gets pulled back in the cylinder for the next cycle, taking up valuable space that could be occupied by air and unburnt fuel for more power.
So on a stock engine, with stock exhaust, it's not uncommon for the intake/exhaust overlap to be relatively small.
Now let's imagine you have a free flowing exhaust, you can safely run more overlap and not have intake port contamination with exhaust gas reversion. That's fine when the throttle is open. No big deal. BUT when the throttle is closed, like at idle, it's hard to suck the intake air/fuel in there, but there's this open exhaust, with no major resistance to gases being sucked bakc in the cylinder (which is different to them being pushed into the intake port at wide open throttle). At idle, it can suck some exhaust back in there, and hence have less air/fuel in there at idle, and produce (relatively) less intake vacuum at idle, and a rougher idle. In the past, to alleviate this problem (which as mentioned is only occuring with a closed throttle/idle scenario) they used a cam (i.e. when talking about performance/non stock cams with more duration) with wider lobe separation angle, less valve overlap, and a smoother idle was to be had. The 'problem' is that this generally resulted in a cam with too wide a lobe separation angle for optimal mid-range output (which is more important than peak power for just about anything except land speed salt cars and tractor pulling). Adding an anti-reversion step at the exhaust port exit, at the extractor flange helps resist this reverse flow, so allows a cam with tighter (or in this case more optimal) lobe separation to be used without as much of a detriment to the idle quality.
Which moves us on to
#2 - exhaust scavenging.
Imagine a near perfect 'primary' header pipe. With the right length and diameter even if it just went to open air, didn't meet with any other pipe' - if it was combined with a cam with the right overlap (and duration etc in general) to work over the same rpm range, as the tail end of the exhaust rushed out the exhaust valve, it'd create a negative pressure wave that would create a momentary low pressure in the cylinder, and not only 'pull' out hte remnants of the exhaust, but it'd pull in some intake through the intake valve and with the right timing, that would be left in the cylinder, so it 'kickstarts' the intake flow without the piston moving down much to initiate it, which means you can get the cylinder filled beyond 100% with the right combo.
David Vizard has written that this outgoing exhaust pulse can have a 300% stronger 'pull' (and I know that's the wrong terminology, but it'll do for the sake of explanation) than the piston moving down. That's absolutely the truth. It's also a 'fact' claimed by the people trying to sell tuned length extractors as 'negative supercharging' as if it's a new technology or something, and they WAY exaggerate the potential increases. What those chaps DON'T seem to realise is that SURE it's 300% the pulling power, but it's at a time when the intake and exhaust valves are only open about 50-70 thou or less (and closing in the case of the exhaust) - and we all know that ports probably only flow 10% or so (I pulled that figure out of the air, whatever the actual one is in any given circumstance, the main things is it is WAY smaller than peak or mid range flow) of their potential - so it's NO WAY KNOWN going to provide a doubling of torque or some other such nonsense claim. So if you ever see an article by Vizard on this issue, keep that in mind. He isn't in error in any way, it's just some people have chosen to mis-interpret it. All I am saying is the gains are real, they are worthwhile, but they are icing on the cake, not 'triple the power' or anything.
Ok, so we see that the outgoing exhaust pulse not only can remove all the last traces of exhaust in the cylinder, it can also pull some new intake charge in there. A _very_ handy thing. So then if you take it a step further, it's not hard to see why having the exhaust header primary pipes meet up at a certain point further back, at a collector, then each outgoing exhaust pulse from one primary, could help pull out the exhaust pulse of another cylinder about to hit it's exhaust stage, which makes that cylinder's exhaust cycle more efficient, which will then help the next cylinder's exhaust cycle and round and round we go.
So basically 4 into 1, various collector designs, that are all designed to do that. The merge collector basically does it better than the less optimally angled collectors, but they are all better than nothing. Think of merge collectors as being collectors where the pipes are all angled slightly toward the centre, instead of having a funnel at the end to redirect them, they are aimed dead on, and produce the best negative wave. Due to this improved aim/direct path, they can actually run a slightly smaller main collector (a choke point is a term I've heard used for it as well) that then tapers back out to the main size. Done right it increases the negative wave/exhaust scavenging even further, of course too small will still be too small.
At this point, without a lot more experience on it, I'd be reluctant to go tighter than 2inch OD for a 1.5 litre full race motor, tapering back out to probably 2.25 or 2.5, and I'd always wonder whether 2.25 as the narrowest point might not be better.
------
Stepped headers are a good and relatively modern refinement. IF you have an engine where you want to have your cake and eat it too, these are it. They let you get most of the torque boost of the smaller headers at lower rpm, and the larger headers at higher rpm. Ultimately the precise length of each is a matter of dyno testing, but you'd find it hard to go wrong with primary pipes going 10-12 inches, then going up to the next size for the rest of the primary. The added benefit of a stepped header is that at the 'join' it adds yet another anti-reversion step. All win win. You'd still be running a 4 into 1 setup, just that the primary pipe diameter goes up one pipe size at around 12 inches of length. Aside from having the best of both worlds kind of thing, the exhaust is still expanding a little as it leaves the exhaust system, so this slightly larger pipe a little further along has merit. Of course you don't want to go too far and lose momentum
--------
Obviously all of this stuff only works optimally when there isn't any significant exhaust back pressure. If there was, the exhaust velocities would not be reached, and even if they were, they'd be pushed against a restriction. Bottom line, they won't work nearly as well.
The next problem is of course noise. Once apon a time race cars could be loud, real loud, and a racer witout tinnitus (ringing in their ears even when there is no sound) would be a rare thing indeed. So now we have to adhere (of course it's always been the case on the street, but recently, in Australia at any rate, even old cars have now been sort of retro-legislated against, and the 'acceptable' noise level is pretty bloody low. Aside from reflecting sound waves against one another (which cancels out that sound wave) - which some (but not all) mufflers attempt to do, some of the reduction in noise will inevitibly be at the expense of raised backpressure, and therefore reduced power.
Being realistic, we'll have to accept some backpressure. I haven't seen too many hard figures on it, but if you weld a few bungs into exhaust systems, and find anything above 3-4psi, you are losing power, and if it's like 10psi, you are losing a lot of power. Down around 1-2psi, and yes you'll lose a little, but nowhere near as much. So the goal has to be, realistically, to reduce noise as much as we can with as little increase in backpressure as we can, but inevitibly we must accept a little back pressure.
So how do we do it? Well unfortunately not all mufflers were created equal, so you might have to experiment. If you can find a modern variant (in design, not just in name, as there has to be a hundred cheap knockoffs that use the name but are terrible internally) of the 'turbo' muffler, you'll have to test it. You might also find that a 2.25 muffler works well, but you only have a 2 inch system (for argument's sake) so you weld on a flange to pipe exhaust in from your system to the slightly larger muffler, and either keep the 2.25 out, or flare/funnel it back to 2 inches on the outlet.
To find out what muffler to go for, you might need to look at some sources on google where they do comparative tests back to back and show various brands and which is the quietest for each flow rate. In terms of flow, David Vizard has discussed at length testing he has done, and came up with a rule of thumb that you need 2.2cfm muffler flow (tested at 1.5 in Hg which is about 20in H2O) per flywheel bhp in order to fall into that ideal range where you get sufficient noise reduction, but sacrifice as little power as possible. So on a full race A series - I'm guessing as much as 160bhp, or more, for 1.4 ish litres, you'd need a muffler that flowed 350cfm or more.
No big deal there.
But in the middle - after the last collector and before the first (and only, if it can be done) muffler you need to add something. You need something that will allow the exhaust pulses going out the collector to still be pulses, which means the extractors etc work, but that dilutes/mixes them so that it's more like a slow continuous flow into the muffler. This will 'fool' the extractors into working at one end, and still allow the exhaust to exit at the other and the process of stopping the 'pulsed' sound will in and of itself help quieten the exhaust too.
The 'thing' is called a pressure wave termination box. To make it work, after the exhaust goes through the collector, it has to be presented with a significant change in cross section area/and a large container volume, so the whole lot can still exit said pipe. Then you basically funnel/direct it back to the first muffler. Just how big that 'container' has to be is somewhat dependant on engine size (and whether all cylinders are dumping into it in the case of a v8 vs a 4 or inline 6). The precise size would be down to testing, but a good enough rule of thumb would be to aim for perhaps 6-8 litres. The bigger the better, but once you get past around the 6-8 litre mark (for an a series up to 1500cc) the increases will be stuff all. Basically it is 'enough' volume to allow the exhaust coming out of the collector to act as it if were venting to the atmosphere.
The design doesn't have to be outrageous, and it doesn't have to be perfectly round top to bottom and side to side (ground clearance issues might force your hand on that one). The collector 'exit' should be a straigt cut, and should be inside the 'box' a good 2-3 inches before it's cut off, so that there is even 'atmosphere' behind it a little (gases do funny things). The 'funnel' shape back to the exhaust need not be fancy, just generally make sure it's 'flatter' than a 45 degree angle (probably 30 degrees would be even better) to be the final piece to make sure it doesn't try and reflect a positive wave back up the exhaust (which is done on 2 stroke exhausts on purpose!) Make it as wide and tall as you practically can - to give the biggest change in cross section area, and beyond that extend its length or alter the tapered section to get the volume up. Is this magical? Nope. Is it a closely guarded secret? Nope. Just decent application of principles found through a combo of theory, science and relentless testing.
The pressure wave terminator/cancellation 'box' or cannister in function actually does what a resonator (hot dog shaped thing) is supposed to do, but rarely does to any significant degree. It's just a better conceived/designed tool and does the job far better.
Second to last, though it's not quite as big a deal, if you get all your ducks in a row with regard to the above stuff, about the only thing else is to try and position that muffer as far back as possible - i.e. behind the axles not in front.
Very last thing. This came up in a post by Larry Widmer on his 'oldone' forums. Unfortunately the forums are down. Larry is not without his critics, but he did work for aircraft designers for the US military among other things, so does have a decent handle on fluid dynamics and so forth. I don't have the post saved or anything, but am pretty sure 1/2" was the correct measurement)
Anyway, he did a rough calculation, and if you got a piece of thick material (he used pvc pipe, which would survive the relatively more moderate temps right at the exhaust tip, but i've never located a source in Australia that is the right combination of overall diameter and thickness, so you might have to look at making one from alloy (which would experience galvanic corrosion eventually, it press fit into the exhaust tail pipe!). Anyway, it has to be around 6mm wall thickness (and thicker wouldn't hurt), and you drill holes (alternating pattern) of 1/2" diameter all around and along it's length, then slip than into a tight fitting outer pipe (i.e. the exhaust, so on a 2 inch exhaust, you'd find 2inch pvc with around 6-8mm wall thickness, and slip into a 2.5 pipe that was welded to the outlet (or similar, so the actuall inner diameter of the pvc or alloy matches the main exhaust pipe diameter). That massive saturation of holes doesn't present a massive flow restriction, but it does cancel out (via reflection) a particularly noisy range of frequencies that he observed on the 4 cylinder engines he worked with (in this case 1.6-2.2 litre honda 4s, but it should still apply). This won't quieten it across the board, but can hit some of the worst offenders. It'd cost around $5-10 to try it, can't hurt to experiment.
Transfer
![Open in new window](javascript:CaricaFoto("http://jmac.performanceforums.com/waveterminator.jpg");)
