Engine technology thread.

[Homemade WRX]

did anyone else notice how much more quickly the 66 track cam opens/closes after it's .050" point?

[Boil]

Originally Posted by Homemade WRX

did anyone else notice how much more quickly the 66 track cam opens/closes after it's .050" point?

LOL…

[Dimman]

Originally Posted by Homemade WRX

did anyone else notice how much more quickly the 66 track cam opens/closes after it's .050" point?

I was gawking at its 112 degrees of overlap.

[Dimman]

HomemadeWRX:

I asked a while ago (I think before you joined/were active) about stepped headers, and have noticed you mentioned them a few times.

I've seen one dyno result that shows them gaining top-end without low-end loss. I've heard two theories as to how they work.

One is that the step is an inertial thing, and the increased size should be placed where the gases in the primary have slowed down enough to start to increase pressure.

The other is wave tuning, where a short step can be used to create a negative pulse to scavenge its cylinder multiple times (neg wave gets back to the port one more time than the positive one does) while the exhaust valve is still open, and doesn't weaken itself by being spread over other cylinders.

Care to clarify/fix/elaborate on this?

[Homemade WRX]

oh boy, manifold tuning has many theories to it and it's one of those things where you can't make fact but only find what works. Way too big of a topic to simply cover off the bat on a forum but here's my quick and short on it.

I prefer to think of a step and tri-y much like stacking spring of different rates. In the manifold, each ID/length will have it's rate (wave speed) in which it's hitting the targetted pressure drop on exhaust port but when combined they can make an average area of increased VE, much like combining spring rates. I don't think you'll keep the SAME peak power (for an equally developed manifold) or tq as it will naturally shift/tilt the VE curve (not move it) but a broader elevated VE/TQ curve will be the resultant. If peak power was the goal, I think long tube 4-1 would be the starting point, however I want to spread the 'love'. I love a wide, flat torque curve for driving.

The step itself also plays a pressure game, as well, that you mentioned. It is the high velocity portion and then gas expands into the large, slower velocity, primary. Then it'll once again go into a collector or secondary (on tri-y) which gives a similar effect. It also acts as an anti-reversion tool, same reason to leave a step on the head/exhaust gasket.

You will take a bunch of variables, such a local mach speed (air temp, exhaust densities, etc), manifold inside diameters, lengths, cam duration, targeted RPM etc and the go find your choice formula/theory. I have my personal spreadsheets that I've been using since college for this...and they've grown/progress through the years(with cam phasing). I also pair it with Ricardo Wave and my own 'virtual 4-stroke calculator' on Matlab. Pretty crude compared to what is currently used in the industry but I don't have $100k to drop on software seats Plus I use only use the 'paper' to get in the ballpark and then the rest is dialed in the real world.

Guess I got a bit off topic. I'm drawing a blank on the name/author of one of the good gas dynamics books I'd read on manifold tuning...and it's pissing me off now!!

[Dimman]

Originally Posted by Homemade WRX

oh boy, manifold tuning has many theories to it and it's one of those things where you can't make fact but only find what works. Way too big of a topic to simply cover off the bat on a forum but here's my quick and short on it.

I prefer to think of a step and tri-y much like stacking spring of different rates. In the manifold, each ID/length will have it's rate (wave speed) in which it's hitting the targetted pressure drop on exhaust port but when combined they can make an average area of increased VE, much like combining spring rates. I don't think you'll keep the SAME peak power (for an equally developed manifold) or tq as it will naturally shift/tilt the VE curve (not move it) but a broader elevated VE/TQ curve will be the resultant. If peak power was the goal, I think long tube 4-1 would be the starting point, however I want to spread the 'love'. I love a wide, flat torque curve for driving.

The step itself also plays a pressure game, as well, that you mentioned. It is the high velocity portion and then gas expands into the large, slower velocity, primary. Then it'll once again go into a collector or secondary (on tri-y) which gives a similar effect. It also acts as an anti-reversion tool, same reason to leave a step on the head/exhaust gasket.

You will take a bunch of variables, such a local mach speed (air temp, exhaust densities, etc), manifold inside diameters, lengths, cam duration, targeted RPM etc and the go find your choice formula/theory. I have my personal spreadsheets that I've been using since college for this...and they've grown/progress through the years(with cam phasing). I also pair it with Ricardo Wave and my own 'virtual 4-stroke calculator' on Matlab. Pretty crude compared to what is currently used in the industry but I don't have $100k to drop on software seats Plus I use only use the 'paper' to get in the ballpark and then the rest is dialed in the real world.

Guess I got a bit off topic. I'm drawing a blank on the name/author of one of the good gas dynamics books I'd read on manifold tuning...and it's pissing me off now!!

Hell, I'm just trying to get started with MS Works' spreadsheet. It wouldn't be too bad if the damn tubes didn't all connect and cylinders fire at different times... Heh...

[serialk11r]

Originally Posted by Dimman

Hell, I'm just trying to get started with MS Works' spreadsheet. It wouldn't be too bad if the damn tubes didn't all connect and cylinders fire at different times... Heh...

hehehehe
If I'm not mistaken, a twin scroll turbo effectively separates the tubes...this is part of the reason I'm excited about my turbo-compound idea. Total exhaust tube length can be cut like crazy, and you don't need to consider effects on cylinders that don't share the same exhaust tract.

Okay so I'm fully set on doing this someday. I figure a turbo that is good for about 2 times the intended mass flow would work pretty well. Normal turbos are set up to produce maximum boost and good power. For maximum efficiency, the turbo should be creating basically no backpressure, only absorbing the energy of the exhaust pulses. So I figure, on a 2L engine, a GT35xx turbine with twin scroll housing. 4:1 gear reduction, fabricate a gearbox to bolt onto it, figure out a 2.5:1 or so belt drive, and then have a flexible shaft of sorts (I wonder if reinforced silicone tubing would work...it has to transmit about 1Nm of torque, and withstand something on the order of 50psi of "centrifugal" force) connect the turbine to the gearbox in order to not load the turbine shaft.

If I don't do something horrendously wrong, should be good for at least 5% more power at the top end. Complements a long duration camshaft perfectly, as lower rpm has lower VE so less exhaust pressure.

[Dimman]

F1 is bypassing the reliability issue of the gearbox by just using the turbine to power generator and stores/releases energy through the KERS system.

Could be simpler to turbo-compound a CRZ.

[arghx7]

Originally Posted by Dimman

F1 is bypassing the reliability issue of the gearbox by just using the turbine to power generator and stores/releases energy through the KERS system.

Yes, here is a system diagram of a Honda Kinetic Energy Recovery System:

[serialk11r]

Originally Posted by Dimman

Try to find specs that have the duration @ .050" lift. That gives a better picture of 'effective' duration. Prius probably has low, gradual lift so a lot of those 270 degrees are probably at tiny (ie not really significant) lift numbers.

In the BC .pdf catalog they have some more detail about when events take place (only certain cams though, and not the stock figures).

http://www.briancrower.com/dl/06cat.pdf

Didn't notice this post.

Sorry, that doesn't make sense to me. If the lift is "not significant" then all it does is create pumping loss. The peak lift is probably not very high, sure, but that doesn't mean they don't want the valve to open quickly. I think the whole point of late intake valve closure is to incur less pumping loss than early intake valve closure because you can have the valve opened up more, along with the benefit that a late closing intake valve becomes a high volumetric efficiency valve at higher engine speed or when it's advanced (this isn't optimal though), so you can sorta control output with cam phasing alone.

[Dimman]

Originally Posted by serialk11r

Didn't notice this post.

Sorry, that doesn't make sense to me. If the lift is "not significant" then all it does is create pumping loss. The peak lift is probably not very high, sure, but that doesn't mean they don't want the valve to open quickly. I think the whole point of late intake valve closure is to incur less pumping loss than early intake valve closure because you can have the valve opened up more, along with the benefit that a late closing intake valve becomes a high volumetric efficiency valve at higher engine speed or when it's advanced (this isn't optimal though), so you can sorta control output with cam phasing alone.

The point about the duration at .050"+ is partly that flow under that is usually insignificant. Also when you compare, for example, cams advertised as 264 degree cams between different manufacturers, you can end up with huge differences in the duration @.050" and @.100", etc... Just comparing cams of the same 'duration' is usually not accurate. Partly too is different manufacturers start measuring at different lift points. 264 degrees starting as soon as the valve moves .001" off the seat is not the same if another company starts at .010".

This is why the duration @ .050" lift figure started. And it isn't perfect either.

[serialk11r]

Yea I understand that, but it doesn't make sense that it would spend a long time at low lift, that would just be more pumping loss. It only makes sense to lift the valve as quickly as possible to the desired position, and close it as quickly as possible.

If we can assume Toyota measures their cam durations the same way across different models, then it should be easy to extrapolate. The 2GR cams look like about 250 degrees based on the PDFs, and the Prius cams appear to be a bit over 270.

[Dimman]

Originally Posted by serialk11r

Yea I understand that, but it doesn't make sense that it would spend a long time at low lift, that would just be more pumping loss.

If we can assume Toyota measures their cam durations the same way across different models, then it should be easy to extrapolate. The 2GR cams look like about 250 degrees based on the PDFs, and the Prius cams appear to be a bit over 270.

I think that comes back to the ramp rates, and wear/durability issues. I can see the Prius maybe trying to reduce friction with lighter/weaker springs, maybe? So they need more gradual valve opening/closing? Cost too? Just guessing...

[serialk11r]

The problem with that in my mind at least is that the valves affect performance and efficiency much more than anything else does, and I have a hard time seeing them trying to cut costs there. I remember reading that the cost to the manufacturer to make camshafts was something ridiculously low, and the cheapo components in say the standard 1NZ-FE can take much more of a beating than they would ever encounter in a rev limited 1NZ-FXE, so there is no reason to redesign an extra cheaped out valvetrain when they can just drop the old one in with a new camshaft.