Cylinder head swirl and flow data.

foxracer1 said:

Any updates?
I'm curious to how much the exhaust port can be improved. Does anyone have modified exhaust flow numbers?

Click to expand...

No updates yet. The few side projects I took on have turned into much bigger projects than we had planned on. One was a 5 cyl Volvo R head, intake and exhaust manifold and the other one I'm still working on is a 427 SB Chevy, kind of a pong story about that one.
I have done a little work on the exhaust ports. They aren't too great from the factory so there is a good amount of improvement to be made.
~Alex

I've playing with some #s and just tryin to guess how well the exhaust would flow. I took your best with stock intake valve and baseline intake flow and plugged them into a spreadsheet i made in excel. I boosted the ex flow numbers the same percent gain @ each lift as the same percent gain you got out of the intake. This is assuming the exhaust can be increased equally as well as the intake.

I had the head man at my machine shop port my head. He went at it with expirence from porting many cummins 5.9 truck heads. He focused on the valve pockets and did more work to the exhaust than he did to the intake. Different heads i know but at least he's seen many heads and is familiar with diesels.

I hope to be able to put them on their flow bench. But they don't have an 79-81mm bore adapter so flow numbers may be off.

I don't know anything about porting, but would tuliped valves with ridges or golf-ball dimples help?



Or



From: http://www.williamsmotowerx.net/valves.htm

They seem to be race valves for gas engines, but would the concept improve airflow?

foxracer1 said:

I've playing with some #s and just tryin to guess how well the exhaust would flow. I took your best with stock intake valve and baseline intake flow and plugged them into a spreadsheet i made in excel. I boosted the ex flow numbers the same percent gain @ each lift as the same percent gain you got out of the intake. This is assuming the exhaust can be increased equally as well as the intake.

I had the head man at my machine shop port my head. He went at it with expirence from porting many cummins 5.9 truck heads. He focused on the valve pockets and did more work to the exhaust than he did to the intake. Different heads i know but at least he's seen many heads and is familiar with diesels.

I hope to be able to put them on their flow bench. But they don't have an 79-81mm bore adapter so flow numbers may be off.

Click to expand...

I've done a few Cummin's heads and without removing the intake plenum there isn't a whole lot to be done with the intake ports other than the swirl bowl and the correct valve seat profiles. I can barley get my finger in an intake port of a stock 24 valve head.

Here are some exhaust numbers I came up with on an ALH exhaust port, not a whole lot of work on this, just went for it. This is without making the exhaust port opening at the gasket any bigger.

nicklockard said:

I don't know anything about porting, but would tuliped valves with ridges or golf-ball dimples help?

Image removed

Or

Image removed

From: http://www.williamsmotowerx.net/valves.htm

They seem to be race valves for gas engines, but would the concept improve airflow?

Click to expand...

The tulip shaped valve is working better than the undercut shape on the intake side. Typically with a high exhaust port an exhaust port a large tulip will perform better than an undercut. Without testing I won't know for sure, I'm more concerned with the intake at the moment though.

That golf ball surface is a patented design so it probably won't make its way onto a TDI valve. I have seen a concentric ring improve intake flow on a 4 valve per cylinder tumble port. I should add that to my list of things to test.

~Alex

Alex22 said:

I've done a few Cummin's heads and without removing the intake plenum there isn't a whole lot to be done with the intake ports other than the swirl bowl and the correct valve seat profiles. I can barley get my finger in an intake port of a stock 24 valve head.

Click to expand...

Yeah they cut off the "shelf" on any of the cummins stg 2 and higher port jobs. Can't do much with the cast in intake in the way.

Alex22 said:

Here are some exhaust numbers I came up with on an ALH exhaust port, not a whole lot of work on this, just went for it. This is without making the exhaust port opening at the gasket any bigger.

Click to expand...

Do you feel gasket matching the exhaust port increases port volume too much? I had the valve pocket cleaned up and whole exhaust port matched to gasket tapering smoothly to the valve. The intake really just got pocket work and polished.

Your research has been very helpful. Thank you.

If working within the constraint of not destroying the swirl-inducing features in the intake, I suspect it will be easier to get larger percent improvements in exhaust-side flow because of the straightforward geometry. But increasing exhaust flow must not take place with significantly increased cross-sections and reduced gas velocities. Many have found out the hard way what was the cause of their overheating problems...

TDIMeister said:

If working within the constraint of not destroying the swirl-inducing features in the intake, I suspect it will be easier to get larger percent improvements in exhaust-side flow because of the straightforward geometry. But increasing exhaust flow must not take place with significantly increased cross-sections and reduced gas velocities. Many have found out the hard way what was the cause of their overheating problems...

Click to expand...

Basically summed it up right there. Since the exhaust port will be a straight forward port job I've been focusing on the intake port design. I've seen exhaust ports opened up to the gaskets and even grinding away at the gasket. They flow great on the bench, but I've seen stock exhaust manifolds bolted up to that port with an exhaust opening that is much smaller than the port in the head, the exact of what you want. I tend to leave the exhaust port a little on the small side when compared to the exhaust manifold/header. When it comes to porting, bigger doesn't mean better.

~Alex

http://www.mototuneusa.com/power_news_--_homework.htm
http://www.mototuneusa.com/think_fast.htm

Comments from experienced porters out there? (Yes, I now the page references non-helical swirl ports and gasoline engines).

TDIMeister said:

http://www.mototuneusa.com/power_news_--_homework.htm
http://www.mototuneusa.com/think_fast.htm

Comments from experienced porters out there? (Yes, I now the page references non-helical swirl ports and gasoline engines).

Click to expand...

I'll second the basics of that website. "Splash Zone" epoxy will last better in the intake ports than JB weld. I haven't tried to use any epoxy in a 4 valve per cylinder tumble port yet but most of the time with that type of port will respond well to unshroudig the valves and removing material from around the short turn area. By leaving most of the port untouched and increasing the flow it will also increase the port's velocity.

Its easy to get the big flow numbers out of a cylinder head by making the whole thing big and "lazy" with low velocities, but low velocities won't make power at any reasonable RPM. As those articles said high velocity ports will give the column of air in the port momentum, and that momentum will help to prevent reversion and the port can have over 100% VE for a small RPM band.

At work I have installed a few sets of aluminum tong port inserts in old BOSS 302 and Cleavland heads in order to increase both flow and port velocity in order to give the engine a better power band for the street. There are also bolt on plates that will have a similar effect to the inserts. A lot of the other large cylinder heads of that era have good flow numbers on the bench but the engines just don't make the power that smaller port with similar flow numbers

~Alex

Fascinating, thanks Alex!

Alex22 said:

At work I have installed a few sets of aluminum tong port inserts in old BOSS 302 and Cleavland heads in order to increase both flow and port velocity in order to give the engine a better power band for the street. There are also bolt on plates that will have a similar effect to the inserts. A lot of the other large cylinder heads of that era have good flow numbers on the bench but the engines just don't make the power that smaller port with similar flow numbers

~Alex

Click to expand...

SWEET! I have a mildly built 351C in my Mustang that suffers from low end grunt.

That's a good problem to have. Or do you mean the lack thereof?

TDIMeister said:

That's a good problem to have. Or do you mean the lack thereof?

Click to expand...

HAHAHAHA Good catch... yes, lack of...

So, breaking down the intake event as is described on mototune's website....

We don't have much if any overlap at the beginning of the intake stroke and can't change it by much without piston/valve clearance issues.

The suction part is pretty straightforward - valve is open, piston is going down, air is going in.

The charging part at the end of the intake cycle where the pistion is coming back up on the compression stroke, the intake valve is still open relying on the port velocity and air momentum to continue charging the cylinder...This all makes sense to me at 10k, 15k, 20k RPM where there is very little absolute time in seconds for these things to happen and having the port velocity high can make a significant difference in net cylinder charge.

What happens when there is an order of magnitude more time for these events to occurr such as at 1000, 1500, 2000 rpm such as in a typical diesel engine (ignoring the aspects of forced induction for a moment)?

My intuition is that the speed of sound of the charge is relatively constant so the fill efficiency of the cylinder during the "suction" part of the intake stroke will be much better (unless you really shrink the intake ports to achieve the same velocity). The time available to charge the cylinder would be so much longer and the fill so much better during the intake stroke that there is not much to be gained by increasing port velocities? Intake valve closure would probably be the most significant variable at low RPM's

Alex22 - Have you measured the volume of the intake ports compared to their static flow capacity? Presumably dividing the flow rate by the port volume, the larger the number the better?

Interesting stuff...

I'm the one who drew Tdimeister's attention to that website because of a simulation project that I've been working on.

Neither "ram effect" tuning nor "resonance" tuning has any meaningful effect at an engine RPM which is a fraction of what the intake system is tuned for. From what I've been finding on the other (non-TDI-related) project, the intake ramming effect starts becoming significant when the Mach number in the intake runner starts getting into the 0.5 range, and at its peak, the Mach number is pretty close to 1. The tricky bit is that the density starts varying strongly with Mach number (i.e. dropping fast as Mach increases) which has the effect of the Mach number going up fast with only a small reduction in the diameter of the intake runner (or a small increase in engine RPM). If the Mach number is below 0.5 then the ram and resonant tuning isn't doing much.

In the case of the other engine that I was working on, the stock intake runner has a throat diameter of 22mm, and reducing it to 19mm in a particular region results in a rather substantial (~10%) increase in torque and power when the engine is running in its powerband. I had already narrowed and lengthened the intake runner a little bit but what I'd done so far only had the effect of extending the length of the stock minimum cross-section, I didn't actually reduce the minimum cross-section. (This is something that was not only in the simulation but also proven out in reality. By the way, thanks for the tip on the "Splash Zone" epoxy - I didn't know about that; I've done it in JB Weld as Mototune suggests. AFAICT the only issue with JB Weld is alcohol or race fuel compatibility - I'll be using Shell V-power 91 without ethanol whenever possible in this engine).

I have since opted to leave the other engine as-is:
(1) partly because the configuration that I have now is "known good"
(2) partly because the simulation result was predicting a pressure regain from the runner expanding in diameter simultaneously with the runner turning a corner and passing the valve guide (= me skeptical of what the computer was telling me)
(3) partly because of something that happens with almost ALL situations of optimizing intake runner ram and resonant tuning ... the increase in torque and power in the 8000 - 11000 rpm range (in the case of the particular project that I was working on) was matched by a nasty *loss* of torque in the 7000 rpm range. In other words, it was predicting a very peaky powerband. Even if the simulation proved out to be correct, the bike would bog off the start then take off, followed by bogging upon shifting to second gear ... Not worthwhile. This particular bike spends a lot of time in the 8000 - 10000 rpm range, which is all well and good, but it's street ridden enough that takeoff from a stop does matter some.
(4) implementing it properly would mean that I would have to take the engine out of the bike and take it apart AGAIN, and I've already had it out and at least partially disassembled TWICE this fall, and I just want to have it together in a state that I know works!

Fix_Until_Broke said:

So, breaking down the intake event as is described on mototune's website....

We don't have much if any overlap at the beginning of the intake stroke and can't change it by much without piston/valve clearance issues.

The suction part is pretty straightforward - valve is open, piston is going down, air is going in.

Click to expand...

All correct. The implication of doing this on an engine with minimal overlap, and also on an engine that doesn't have a tuned exhaust runner and/or with excessive exhaust manifold pressure, is that you can't get the benefit of scavenging during overlap and you can't get the benefit of the negative pressure wave being sent up the intake runner to get the flow going before the piston goes down. Nothing much can be done about this in the TDI application.

Fix_Until_Broke said:

The charging part at the end of the intake cycle where the pistion is coming back up on the compression stroke, the intake valve is still open relying on the port velocity and air momentum to continue charging the cylinder...This all makes sense to me at 10k, 15k, 20k RPM where there is very little absolute time in seconds for these things to happen and having the port velocity high can make a significant difference in net cylinder charge.

Click to expand...

Don't think about it in terms of absolute RPM numbers. Think about it in relative RPM numbers compared to the RPM capability of the engine, with the intake runner sizing such that the Mach number of the intake flow is in a certain range when the engine is running near its peak-power RPM.

Fix_Until_Broke said:

What happens when there is an order of magnitude more time for these events to occurr such as at 1000, 1500, 2000 rpm such as in a typical diesel engine (ignoring the aspects of forced induction for a moment)?

Click to expand...

On an engine that makes peak power at 4000-ish? Ram and resonant tuning will be doing essentially NOTHING for you at less than probably two-thirds, maybe even three-quarters, of peak-power RPM. The momentum of the air column and the magnitude of the pressure waves are simply not enough to accomplish anything.

Fix_Until_Broke said:

My intuition is that the speed of sound of the charge is relatively constant so the fill efficiency of the cylinder during the "suction" part of the intake stroke will be much better (unless you really shrink the intake ports to achieve the same velocity). The time available to charge the cylinder would be so much longer and the fill so much better during the intake stroke that there is not much to be gained by increasing port velocities? Intake valve closure would probably be the most significant variable at low RPM's

Click to expand...

Yes; at low engine RPM the pressure wave effects are so small that the situation can almost be considered hydrostatic. The piston goes down and pulls in air, then the piston starts going up and pushing a bit of that air back out again, then the valve closes and traps whatever amount inside the cylinder.

Cam timing that will work for best cylinder filling at high RPM will give poorer cylinder filling at low revs. Nothing earthshattering about that.

Keep in mind that the wildest camshaft that will function in a TDI is still relatively mild by gasoline-engine comparison. Stock TDI camshaft is 189 degrees nominal duration. The (aftermarket) camshaft that's in the other project that I was working on has 280 degrees duration! And it's completely streetable. Peaks and valleys in the torque curve are a lot more pronounced when you have something like that.

I should also add this. You can calculate a nominal flow speed (without correcting for Mach number effects) in an intake runner without too much trouble.

For the other engine that I was working on

58.0mm bore 47.2mm stroke, peak power target RPM 10000, intake runner 22mm diameter

Peak (not average!) piston speed halfway down the intake stroke, disregarding the swing of the con-rod, is:
(10000/60)*0.0472*3.14159 = 24.7 m/s
In the intake runner it will be in proportion to the ratio of the areas
24.7 x (58/22)^2 = 171 m/s
With the runner narrowed
24.7 x (58/19)^2 = 230 m/s
In the simulation, the actual peak flow velocity exceeds this because of pressure wave effects. But for purposes of "getting in the ballpark" it will probably give an OK starting point if one plugs in a different set of numbers for a different case.

GoFaster - that more/less confirms what I was thinking regarding applying this "high velocity" porting to a low RPM application - There is not much to be gained.

Stock ALH cam intake valve "closes" (1mm lift) 25 degrees after BDC and is more/less fully closed (0.1mm) lift 40 degrees after BDC - that's 0.00666 seconds at 1000 RPM where the intake valve is open and the piston is going up. With the ~speed of sound in air of 1125 ft/sec the pressure wave can travel ~90 inches in 0.0066 seconds (well outside the ~3.75 inches from the bottom of the piston to the intake valve.

Back on topic - I suspect there's probably not too much that can be gained in the 1000-2000 RPM range with head porting for CFM. The cam is probably the much larger influence here. There might be some advantage to increasing the port velocity without lowering the CFM to promote a cleaner burn and more turbulance in the cylinder at lower engine RPM's

Resonance tuning is more a function of tract lengths than of diameters and Mach numbers. Of course, pressure waves travel on a relative frame to the airflow, so that if the flow is locally at a very high Mach number or almost sonic, the pressure waves will propagate correspondingly slower in the absolute frame.

As GoFaster said, given the low operating speeds of the TDI engine, intake runner lengths would have to be much longer than those installed in any TDI engines from the factory in order to exploit a useful pressure wave resonance order at the desired time late in the period of intake valve opening after the piston makes its way back on the upstroke.

The plastic SDI manifold, with total lengths from valve to runner mouth into the plenum, was measured at a little over 2 feet long (~600 mm), and is of the length closest to what would actually do anything beneficial in a TDI engine. The graph below shows that maximum delivery ratio of that length corresponds to about 3800-4000 RPM, precisely where the TDI engine would usually have its peak power output.


Source: Burghardt & Arnold, 1989.

Also, what is laughable is people changing their intakes, not knowing the engineering that goes into- and the careful calculation of lengths, diameters and volumes, because the plenum and piping leading into the manifold ALSO operate as resonant systems. Of course, these systems are usually optimized in the context of stock engine specifications, tight packaging constraints and usually for minimum intake noise (whereas performance enthusiasts usually want the MOST noise), but blindly tearing out the OE intake for shiny, massive diameter aftermarket pieces is like a non-surgeon changing out the blood vessels leading to the heart that works just as it should for bigger, blingy pipes.

The diameters matter for resonant tuning, too. If the diameter is too big, the magnitude of the pressure wave will be too small to accomplish anything.

The magnitude of the pressure waves is a function of the ratio of areas at the diameter step from the pipe to the open end, not of the absolute diameter of the runner itself (2-stroke expansion chambers work on this principle). That said, you're right that it's not in the interest of scavenging performance to have diameters that are too large, as 2-stroke exhausts have tiny diameter pipes and a large expansion chambers - again, one reason for this is if you have large diameter primary pipes, the expansion chamber has to be absolutely massive according to the principle of area ratios. The effect of the diameter manifests itself when the flow is at such a velocity that pressure waves traveling against this flow is significantly slowed in the absolute frame of reference and accelerated going in the other direction.

Fix_Until_Broke said:

Alex22 - Have you measured the volume of the intake ports compared to their static flow capacity? Presumably dividing the flow rate by the port volume, the larger the number the better?

Interesting stuff...

Click to expand...

I even forgot about this one, I started to re enter the info into my spread sheet to make the table, then I noticed this. (origionaly posted in February)



The port's maximum flow was increased by 25.4%, the swirl was increased by 41.7% with the port's volume being increased by 6.6%.

At a usable lift (.400 lift) flow went up 17.1%, and the swirl decreased by 2.4%


As for the resonant frequency tuning, I don't have much to contribute on that topic so I will keep quiet on that topic. One of the guys I work with knows how to tune intake tracks (throttle body and port to the engine) and has put some serious time into motorcycle engine development.

~Alex

As was discussed earlier in this thread culminating at around post #36, the % change in swirl per industry practice is to relate it as a ratio to the overall flow. Absolute values are mostly irrelevant. One extra column in your spreadsheet should take the ratio of swirl RPM over flow CFM, or % flow increase over % swirl increase.

Resurrecting this great thread to see if any new info has developed. It would be nice to have all this fine work culminate in some concrete recommendations and protocols for "properly" porting our TDI heads. Sticky??

Loss of swirl at low valve lift can be countered using swirl chamfers, sort of like shown below, but this is highly exaggerated from what I've seen and much more than necessary to have a functional effect.



Edit: This technique is now being done by OEMs in some newer engines and only on the intake side. I think the CBEA 16V TDI does this. But the eccentricity of the swirl chamfer is very small to minimize undesirable dead-volume in the Diesel combustion chamber.

TDI Meister


I am also looking for a 1 dimensional software program, program like Fluent etc is out of my reach


Source: Burghardt & Arnold, 1989.

Also, what is laughable is people changing their intakes, not knowing the engineering that goes into- and the careful calculation of lengths, diameters and volumes, because the plenum and piping leading into the manifold ALSO operate as resonant systems. Of course, these systems are usually optimized in the context of stock engine specifications, tight packaging constraints and usually for minimum intake noise (whereas performance enthusiasts usually want the MOST noise), but blindly tearing out the OE intake for shiny, massive diameter aftermarket pieces is like a non-surgeon changing out the blood vessels leading to the heart that works just as it should for bigger, blingy pipes.[/QUOTE]

TDIMeister said:

Loss of swirl at low valve lift can be countered using swirl chamfers, sort of like shown below, but this is highly exaggerated from what I've seen and much more than necessary to have a functional effect.

Edit: This technique is now being done by OEMs in some newer engines and only on the intake side. I think the CBEA 16V TDI does this. But the eccentricity of the swirl chamfer is very small to minimize undesirable dead-volume in the Diesel combustion chamber.

Click to expand...

If you look at TDIsyncro pics of CBEA head.. he choose the CBEA head because it has this on the OEM surface he said..

Turbo Z said:

Also, what is laughable is people changing their intakes, not knowing the engineering that goes into- and the careful calculation of lengths, diameters and volumes, because the plenum and piping leading into the manifold ALSO operate as resonant systems. Of course, these systems are usually optimized in the context of stock engine specifications, tight packaging constraints and usually for minimum intake noise (whereas performance enthusiasts usually want the MOST noise), but blindly tearing out the OE intake for shiny, massive diameter aftermarket pieces is like a non-surgeon changing out the blood vessels leading to the heart that works just as it should for bigger, blingy pipes.

Click to expand...

[/QUOTE]

My thoughts exactly =)

But in TDI engine runners are pretty big and really short, so resonances are almost non existing.
I also believe (but don't really know) that swirl type of runner will do really bad for the returning pulse even if it might be small.
Personally I like long SDI type of runners as they offer some mass inertia, especially for longer camshafts to support later IVC.
Also rew range in these engines is so low that it does not give help in induction tuning, or pipes would need to be really long to give benefit.

I removed all numbers, but here is a graph showing some development stage cam when I was designing new cam profiles.
This is without turbo to keep things simpler as I believe there should be ~same EMP and MAP on performance engines.. boost would not change much as there is not overlap and ideally all pressure traces would just raise.
Red is cylinder pressure, green exhaust pressure, black intake pressure and blue is mass flow trace from intake side.
traces are simulated from valve.
Its not perfect, but close enough to "see" whats happening inside of engine dynamically.


Edit: sorry, did not realze this was Dave's text.
but then again I must argue about "Resonance tuning is more a function of tract lengths than of diameters and Mach number"
Resonance tuning is really tight to mach numbers, which is related to diameter.
The more you have speed in runner, the more you have mass inertia and more energy "level" you get for the pulses.
Also pipe diameter will have direct impact on wave speed, if flow speed is 1Mach, then as pulse goes with speed of sound, it never reaches other end as flow speed against it is the same.
Or if you put taper in pipe, then pulse travels faster as there is less speed in air it needs to fight against.
So you can change tuned lengths by changing the diameter of pipe.

I was trying to ask tdimeister about if he had some idea very to find a scientific paper about ram charging system with double resonance chambers and combined with helmholtz, in that I was trying to post a figure but I can not still find out out how to do it, and in that trying the original text abow got away, sorry for that

I was perticular interested in the mathematical correlation that was done in Burghardt & Arnold, 1989, since the correlation of 4 valve looks strange

storx said:

If you look at TDIsyncro pics of CBEA head.. he choose the CBEA head because it has this on the OEM surface he said..

Click to expand...

well, yes, the CBEA head does have these chamfers built in. Originally I was interested in retaining them. After seeing dismal flow numbers with the deep recess of the valves, I decided to eliminate them and go flush with the valves, as flow numbers show vary good improvement just making this one change. Cost is swirl. Perhaps I may create more smoke than I wish. Time will tell.