Can somebody help me with this scavenging dilemma? 4

[Gary_321]

I've made a short 48 second video to show the problem.

Youtube Link

I can see how scavenging would work in a system with very little back pressure, but how does it work in a system with a catalytic converter and mufflers?

 

[SwinnyGG]

Scavenging works because of pressure waves traveling through the system. You can't simulate this with a test that uses constant flow. You basically can't simulate this without an actual running engine- in the real world it happens much too quickly. Optimizing header designs (if that's what you are trying to do) is about doing some calculation to get close, and then testing a LOT.

 

[Gary_321]

I agree. As I understand scavenging, it is the same as the effect experienced at the rear of a large truck, i.e. you can be "sucked" along if you get close enough. It is the tail of pulse A passing the opening of tube B that scavenges pulse B, not the Bernoulli effect of A's movement past the opening to tube B.

I think I am showing the Bernoulli effect.

Is this correct?

 

[SwinnyGG]

You're sort of correct and sort of not. A truck is not a great analogy. What you're describing is the result of several aerodynamic phenomena happening at the same time. Vortex shedding, boundary layer adhesion, bernoulli's principle, etc.

I'm at work right now so I can't view youtube. Search 'engineering explained exhaust backpressure'. There's a video called 'stop saying exhaust systems need backpressure'. Watch that video. It's imperfect but it's a pretty decent introduction to how exhaust scavenging works.

 

[Gary_321]

It looks like you want the best compromise from
1) lowest back pressure [to increase differential pressure between cylinder and valve throat (the space between the valve and the header) when valve opens]
2) velocity [to increase inertial scavenging]
3) header tube diameter and length [to increase wave scavenging]

Each improvement in one will probably have a negative effect on the other two.

I'm going to try and find an experiment that will show wave scavenging.

 

[RodRico]

Maybe using a blower to test the effect isn't representative of what really happens. Exhaust is a stream of pulses, not a continuous flow.

 

[Gary_321]

Rod, I agree with that, but the effect is the same when I pulse the blower.

I've done a test using air and water and I've got the same result. I'll post the video later.

 

[BrianPetersen]

You cannot make pulses with your test rig that are even remotely close in frequency or pressure to what will be seen in a running engine.

Just because there is "average" back-pressure in the exhaust system as a whole doesn't mean there isn't momentary fluctuation widely above and below that average back-pressure when connected to a running engine.

 

[SwinnyGG]

but the effect is the same when I pulse the blower.

You aren't going to be able to simulate this with a handheld blower of any kind.

You're talking about pressures that could be 50+ Bar and pulses that are a few milliseconds long.

 

[BrianPetersen]

A positive pulse happens from an opening exhaust valve at (say) cylinder 1.

This travels down that exhaust header until it reaches an expansion or a junction (which, to that pressure pulse, is an expansion).

The positive pressure pulse splits at the junction and travels BACK up the adjacent header pipes connected at that junction at the speed of sound and also down the collector and because it was an expansion, it gets reflected as a negative pressure wave back to the cylinder that created the positive pulse a few milliseconds earlier.

If that negative pressure pulse happens at a favorable time in that cylinder's exhaust stroke ... THAT is where the scavenging comes from.

Of course, the positive pressure wave that went back up the adjacent header pipes has to be accounted for, too. Hopefully when that gets to the exhaust valve for that cylinder, that exhaust valve will be closed. But it might not be, because you can't tune these things to work at all possible engine speeds. It could very well happen during THAT cylinder's valve overlap period, thus causing exhaust reversion in that cylinder.

Engines that have overlapping exhaust strokes (counting the whole exhaust valve opening time, not just the "nominal" exhaust stroke!) are prone to reversion caused by adverse pressure pulses. The traditional bent-crankshaft V8 is a disaster in terms of exhaust tuning. Inline-fours with aggressive cam timing can be trouble at low revs (I deal with motorcycle engines ... it's not uncommon to have a trouble spot somewhere near half of rated RPM / two-thirds of peak-power RPM). Inline-threes are pretty good, and sixes with split headers (as on your video) such that it acts like a pair of inline-threes are pretty good - the exhaust strokes are for all practical purposes non-overlapping.

 

[Gary_321]

Thanks Brian. Good stuff to read and absorb.


I'll post this because I said that I would. It's a 4 second video that shows scavenging. I now understand this to be "inertia scavenging". If you listen closely, you can hear a popping sound when I stop the flow. Is this popping the wave that you talk about?

Link

 

[SwinnyGG]

I watched your video- I don't think I can see enough of your setup to tell what is going on.

Can you describe the arrangement you're using?

 

[Gary_321]

jgKRI - if you mean the 4-second video...

I've just noticed that you need headphones to hear the popping.

Here's a 58-second video of the setup...

Link

 

[SwinnyGG]

I can't view youtube at work, unfortunately. But I'll watch this tonight and see if I can understand.

What exactly are you trying to accomplish? Is this just an exercise in learning, or are you trying to translate this understanding into some design work?

 

[VEBill]

I recall seeing (years ago, on a video somewhere) a set - or kit - of (very) adjustable length headers so that the engine tuner could fiddle with the effective header pipe lengths on the dyno. The pipes slide into each other and were then clamped for the next dyno run. They were only suitable for a dyno of course, because they were so unwieldy (they'd never fit into an engine bay). The resultant optimal lengths from the trials and tuning would then be used as the basis for the design of the actual headers.

I don't recall if they had any options for different diameters.

Disclaimer: this isn't my area. Just mentioning it.

 

[Gary_321]

VE1BLL - That is a cracking idea if it could be made to work (it sounds like it would leak)

jgKRI - It started as a learning exercise because I like to learn about a subject before I try to modify it. Car forums are full of people who bolt on bits and say they have made an improvement. Usually all they have done is change the noise.

I am looking to improve the car's performance. I've measured (in lots of detail) the exhaust dimensions, valve lift (lift v crank angle) and the intake dimensions. I can then, hopefully, see areas for improvement.

I'm trying to prioritise what is my "biggest bang for the buck". For instance, the exhaust is "crushed" (to allow ground clearance) from 2463 sq mm to 2299 sq mm. If I can replace the crushed section with 56 mm pipe, I can reduce the back pressure in that area for almost no cost.



I've already insulated the intake and have measured improvement - that cost almost nothing.

As an idea of my thinking, I am considering dividing my plenum, so the intakes of cylinders 1, 2 & 3 are not influenced by cylinders 4, 5 & 6 - but that is a discussion for another thread.

 

[MacGyverS2000]

Pipes don't have to be perfectly round... particularly at "crushed" sections.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[Gary_321]

Yes, but the cross-sectional area is 7% less.

I couldn't find any reference to this on the internet, so I did an experiment using different shaped orifices (all the same cross-section) and it proved what you say. The only problem that I noted was that some shapes promote vortices (which significantly reduced the flowrate).

 

[SwinnyGG]

Those shapes produce vortices because they are orifices, not because of the shapes.

Round is ideal, because you want to the flow area/surface area ratio, but you're talking veeeeeeeery small changes in power required to pump air through two tubes with the same cross sectional area but different shapes.

The effect of dents on exhaust flow has been studied, and found to matter very little. A 7% reduction in flow area would be noticeable if the entire header was changed, but for one dent in one of six tubes, and with the bulk of the path having the larger flow area, that dent matters very little.

 

[RodRico]

jgKRI,

50+ bar ?! Really? I thought four-strokes were typically closer to 5 bar at the end of the expansion stroke when the exhaust valve opens.

Gary,

A lot of folks are into tuned pipes, but I'm not so much. Fixed tuned systems are peaky and will only operate ideally at a fixed engine speed (this is one reason why two strokes are so peaky). It makes a lot of sense to mechanically tune an engine that runs over a very narrow range of RPMs, but much less so to tune a typical car that operates over a wide range. Engines with close gear ratios, racing engines, etc. may all benefit noticeably, but I don't think a common road car does. Just my only slightly informed opinion. Perhaps I'm overestimating the sharpness of the mechanical resonance.

Rod