Reversion and turbocharging

[brysonc]

Hey all -- I've been a long time lurker, but just now am getting around to my first post. Great info on this site! Anyway, I know it's common practice to keep the step between the exhaust port and the header, but how important is it on a turbocharged vehicle? Should there be any steps to prevent reversion, or would the pressure in the header before the turbo make the step useless except to slow flow by an increase in volume?

On one hand, I kind of feel like a small step around the outside of a port will do very little to prevent exhaust gas from entering the cylinder again, especially at the pressures in a turbo header. Also, it seems like a sudden increase in volume would decrease exhaust velocity, as well as creating a turbulent area limiting "laminar" flow.

On the other hand, maybe the step will help even more than on a naturally aspirated engine because the pressure differential between the cylinder and exhaust port are higher. So far, I'm leaning more towards the idea of eliminating the step and having nothing but smooth transitions from the valve all the way to the turbo, but I don't have the experience many of you do. Has anyone tried matching the port and header primary on a forced induction car? Did you see any significant results? Thanks!

 

[Warpspeed]

The exhaust gas pressures in a turbo engine are vastly higher than an n/a engine, typically twice boost pressure in a road engine, and perhaps equal to boost pressure in a sports/racing engine.

With such high exhaust pressures, valve overlap is going to be far more important than any small step in the pipework in limiting reversion.

There are two different approaches to turbo exhaust manifold design. The first is to keep the pipe volumes absolutely minimal to improve response. This works particularly well where there are few cylinders and very little valve overlap. It is important to avoid the situation where one exhaust valve is just opening blasting hot exhaust straight down another exhaust port which is in the final stages of the exhaust cycle. The ideal situation is to only have one exhaust valve open at any one instant. All cylinders can then be combined in a fairly crude log style manifold without any real disadvantage.

Where there are many cylinders, and very long duration exhaust timing, that will not work so well. Individual runners all the way to the turbo can help considerably in preventing one cylinder emptying straight back into an adjacent cylinder. Multiple turbos, or split pulse exhaust housings help too. I believe you will find that reversion steps, and tuned pipe lengths are far less important than with normally aspirated engines.

The best way to avoid the reversion demon is with sensible valve timing and a logical well thought out pipework layout.

 

[patprimmer]

If you cannot maintain perfect alignment over time and through the extreme thermal cycling you will get a step forming. If the step is up on the downwind side it does little damage, but if it steps down on the downwind side, it seriously hurts flow, so it is better to make the manifold flange slightly bigger than the port. The amount bigger depends on how accurately you can maintain alignment.

Regards

eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[brysonc]

The engine is a 2.3L Quad4 with custom ground Engle cams. I don't remember the exact cam specs (the spec sheet is in Charleston and I'm at school in Clemson), but there's about 4* of overlap at .050" if I remember correctly. I can't change the cams, but I might be able to change lobe centers by welding and re-drilling the cam gears, but that would be a one-time deal. The cam design was one made by Jim Feuling for a turbocharged Quad4. I'm just having a problem trying to decide if a step is really worth having. I understand the argument for maintaining alignment, but it's a little tough for me to think that a step would have a significant impact on stopping exhaust gas from coming back into the cylinder.

 

[Warpspeed]

As Pat says, a step is not going to hurt, but don't expect any really significant gains from deliberately adding a step.

That Quad 4 of Jim Feuling's was an amazing bit of development engineering.

How much valve overlap can be tolerated is strongly related to the size of the exhaust turbine. Generally, big turbo, big cam, small turbo, small cam. Being able to adjust lobe centres on the dyno is going to be well worth the effort.

The effect of having too much overlap usually ends up with an engine that feels eager and responsive at part throttle, as though it really wants to go. When you floor it, the power is just not there. Opening up the lobe centres will usually fix it, as might a larger turbine housing.

 

[GregLocock]

I agree with your doubt as to the effectiveness of a step. I can see from Pat's argument that it is a necessary evil.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[brysonc]

So, I guess the consensus is that on a turbocharged car, a small step, maybe 1mm all the way around, is needed to account for alignment.

On a N/A car, would a larger step be more effective? What about 1mm on the top, and larger on the floor?

 

[Warpspeed]

The cylinder head being water cooled is not going to get that hot, but the exhaust manifold sure will. If the exhaust flange plate is continuous, then there definitely needs to be allowance for expansion. That is usually done by making the centre manifold holes on size, and making the end manifold holes oversize or slotted. The manifold flange can then move around on the studs.

Another way is to use heavy individual flange plates over every port. These can be rigidly bolted to the head and accurately aligned, allowing no movement. This will work, provided the pipework itself is designed to absorb all the considerable expansion movement.

Long individual runner one piece manifolds should be capable of that, provided there are enough compound curves in different planes, and enough length to absorb the movement. Obviously long straight lengths of pipe should be avoided, but that is hardly likely in a typical manifold design.