ZHP Cylinder Head Optimization

[DeathTrap]

As most of you may or may not know, I bent a valve in my ZHP a few weeks ago. Ever since, I've been on a philosophical journey to see how the stock head can be improved. I've devoured several forum threads and even bought a book on cylinder head porting, which led me down an even larger rabbit hole of fluid dynamics type education. It has been interesting...

I guess we can start with this guy named Adam or "PEI330Ci" who's done a lot of footwork for the M54 engine. I really appreciate that he's openly sharing what he has learned about the platform. He has spent, who knows how much, time and money on obtaining hard data for this engine.

With that being said, the exhaust port is the weakest link in the chain. It appears that some work here produces tremendous gains in flow across the board from low lift (2.5mm) to high lift (9.5mm). Almost 40 CFM difference in some spots! What that means for power in an N/A application, with stock ZHP cams is what I want to find out. I also need to figure out how to get those gains. Sure I could hog out the exhaust port, but that doesn't learn me anything about what I did and also increases the likelihood of me getting a new cylinder head.

Next, improvements in intake flow vary from around 20-40 CFM and start at around 6mm of lift and end at the peak lift of 10mm for stock cams. Not bad gains there either!

BMW has made some compromises (not their fault, simply part of the engineering process) in the cylinder head as it is a mass produced head, that fits multiple engine displacements, that has to meet a general populations' idea of a strong engine, and also emission requirements, yada. The goal here is to attempt to optimize cylinder head breathing (note that I did not use the word 'flow' here) for our more aggressively cammed, 3 liter engine.

I'm starting from zero...no tools, little knowledge of how air works, not much time either as I want to be up and running by summer...This is going to be a long journey and I'm here to document it. Let's go...

[DeathTrap]

"A well developed high-performance engine actually has two induction phases."
One is from the exhaust scavenging pulse and the second is from the suction of the piston going down the bore. I have researched this and there appears to be data that agrees with this statement. I don't think any gearhead would disagree either. But exactly how "strong" are these events? Well, an engine with large overlap (where intake and exhaust valves are open at the same time) will have more suction on the scavenging pulse than the piston caused suction. Therefore, it is critical that our engine flows efficiently at low lifts because overlap and scavenging is a mostly low lift occurring event. Poorly flowing overlap conditions rob the engine of high RPM power. Because of dual VANOS, our engines have the ability to adjust overlap duration and lift. I'm not 100% certain on the overlap specs, and is something I may have to measure.

http://www.epi-eng.com/piston_engine_technology/exhaust_system_technology.htm

It looks like the scavenging wave or "pressure spike" as the link describes it, is what we need to focus on. "That pressure spike, or pressure wave, moves down the pipe at the sum of the local sonic velocity plus the particle velocity of the gas flow." I'm going to make an assumption here that the more efficient valve will cause a stronger "pressure spike" than an inefficient valve and a higher amplitude pressure spike will also cause a stronger suction on the reflected wave. (If the wave goes up higher, it has to come down lower right? :) ) The graph in that link is just an engine simulation, but it gives a good picture of how much stronger the scavenging suction is.

Now that we can see how important low lift flow is, how do we improve it? How do we make the exhaust valve more efficient, without killing velocity? It requires careful shaping just before and after the valve seat. Port velocity is very small compared to seat velocity. Imagine your garden hose is the port and you have one of those twist style nozzles which is the valve and seat. If you made your garden hose opening a few mm larger or smaller, you may not see a big change. If you open that nozzle a few mm, there is a large change in how the water flows out. That's how I'm imagining the valve-seat area. It sounds right in my head and we will verify later.

What about high lift flow? For the exhaust, high lift kind of just takes care of itself. You have a highly pressurized chamber venting to a lower pressure exhaust manifold. That extra pressure helps push all of the extra exhaust gas out. PLUS, and this is critical, the exhaust gas can use the ENTIRE circumference of the valves and seats to exit the chamber. Remember this for when I discuss the intake side.

Questions for myself later:
How to read a flow chart and determine whether a port is valve limited or port limited?

[ZHPizza]

This is awesome man. I'll definitely be following along. I learned a lot about the exhaust pulse when doing headers and learned that just opening up the head ports to match the header diameter would actually lose power, so what you're into right now is still witchcraft to me. This should be fun!

[JPMo]

Interesting. I’m here for the knowledge

[racer2086]

Interesting and will be following along.

Some valve seats in racing applications are rounded towards the inside to reduce turbulence in the air flow. It will sometimes increase airflow by reducing metal in the path of the air and allow for more free breathing. I'm no expert, but from what I've read and learned, that is generally done on the intake side of things. Slowing down airflow and what not on the exhaust side by opening ports and increasing volume can actually work against performance due to the loss in velocity. The main question is if an expensive head job with custom valve seats is worth the cost vs the performance gain in our cars, which I suspect will be minimal.

[johnrando]

Love this, thanks for postinf. Looking forward to more.

Sent from my SM-G955U using Tapatalk

[DeathTrap]

Great! Glad you all are joining me in this journey!


This is awesome man. I'll definitely be following along. I learned a lot about the exhaust pulse when doing headers and learned that just opening up the head ports to match the header diameter would actually lose power, so what you're into right now is still witchcraft to me. This should be fun!

That is correct. "Port matching" the exhaust to an aftermarket header on our engines increases the port volume considerably! This does increase flow, but kills velocity. Flow and velocity/port volume have inverse relationships. Our goal is to maximize flow without increasing port volume too much.


Interesting and will be following along.

Some valve seats in racing applications are rounded towards the inside to reduce turbulence in the air flow. It will sometimes increase airflow by reducing metal in the path of the air and allow for more free breathing. I'm no expert, but from what I've read and learned, that is generally done on the intake side of things. Slowing down airflow and what not on the exhaust side by opening ports and increasing volume can actually work against performance due to the loss in velocity. The main question is if an expensive head job with custom valve seats is worth the cost vs the performance gain in our cars, which I suspect will be minimal.

Yep. We really don't care too much about turbulent exhaust. Intake is a whole other story. The fast exhaust pulse helps pull out combustion products through inertia, kind of like getting passed by an X6M going 180mph and you feel that pull (low pressure wave) behind it.

I don't think expensive is necessary. The S54 and S50 uses the same valve cut, with minor differences, as our M54. That's what the next post will be about. Great lead in lol.

Sent from my Pixel 2 XL using Tapatalk

[DeathTrap]

Dropping this here for future discussion.https://uploads.tapatalk-cdn.com/20200128/256e70037fb203faf8e6eb7591f6b6bd.jpg
36185
Sent from my Pixel 2 XL using Tapatalk

[DeathTrap]

Ok, let's stick to exhaust a bit more. I posted a graph in the above post which shows an improvement in flow everywhere. Please don't read this like a dyno graph. Low lift flow is not equivalent to low end torque. The improvement should actually improve power across the entire RPM range, provided velocity didn't take a hit which I don't think it did since the stock ports flow so poorly. It can also be seen that larger exhaust valves hurt high lift flow, bigger is not always better!

There is another graph that compares the S54 exhaust flow to the ported, stock sized valve M54 flow. The S54 edges it out in mid lift (4mm to 7.5mm), but otherwise the ported M54 meets or beats it. The S54 (https://www.newtis.info/tisv2/a/en/e46-m3-cou/repair-manuals/11-engine/11-12-cylinder-head-with-lid/63uDvKE), S50B30 (https://www.newtis.info/tisv2/a/en/e36-m3-cou/repair-manuals/11-engine/11-12-cylinder-head-with-lid/6CgNGrl), and the S50B32 (https://www.newtis.info/tisv2/a/en/e36-m3-cou/repair-manuals/11-engine/11-12-cylinder-head-with-lid/6Ftugqa) have the exact same outside diameter valve seat surface and exhaust valve size as the M54B30 (https://www.newtis.info/tisv2/a/en/e46-330i-lim/repair-manuals/11-engine/11-12-cylinder-head-with-lid/79gemvG). I've hyperlinked the engine codes for convenient reference to newtis. The largest difference is in the intake size and seat width. The M54 has a 0.65mm and 0.45mm thicker seat on the Intake and Exhaust, respectively. There may be gains to made by reducing the seat width! Note that all of these engines have the exact same seat cut angles. BMW did a great job with the S54, so I won't waste time or energy to see if there is a better choice seat cut. We'll run with OEM specs here.

Decision 1: Keep OEM Valve seat angles.
Decision 2: Keep OEM exhaust valve size.

"The flow of the exhaust valve is more critical from 0.100in (2.54mm) and up." Wait...go back and look at the graph, and then reread that sentence. Look at what happens exactly at 2.5mm of lift on the stock port M54.

"Seat form...can measurably affect flow over the entire lift range for an exhaust port."
"At low lift, the flow is entirely dependent on the size of the gap between the valve seat, the seat in the head, and the efficiency with which they flow."

I'm wondering how much of a difference narrowing the seat width to S54 specs would do by itself? Maybe it'll give 70-80% of the results? Not sure but the final decision of where and how much to grind away at the exhaust port needs to be made. That could be a bunch of trial and error, but looking at the CFM chart, it looks like the throat and port could be opened up just a little with some blending.

[DeathTrap]

I did some maffs.

The stock exhaust port is roughly 35mm x 27 mm. At a Cd (coefficient of discharge) of 95% efficiency, the theoretical MAXIMUM flow rate is a whopping 191.8 CFM.

If we add the 6mm valve stem into the maffs, that number drops to 186.2 CFM.

To get the exhaust port to flow 210CFM, theoretically of course, it would need to have an area of 1062 sq mm, that's accounting for the valve stem. An increase of 117 sq mm from stock. That would equate to opening the port up 2.2 mm each side. Now being that our actual port is probably not 95% efficient, it would require a bit more area than that, but I'm sure it's not far off looking at where the flow chart levels off with the stock port.


Edit: I can do the same math for theoretical Max valve flow. These equations do not take into consideration velocity. Either way, I'll have to open up the port a bit. Oh, and even though the exhaust valve diameter is 30mm, the throat is something like 20-25mm.


Sent from my Pixel 2 XL using Tapatalk

[ZHPizza]

Damn this is cool. Love seeing how perfectly tuned the S54 head is in comparison. I'm a little shocked at how much lower the flow is on the N54 than the S54, especially given the power that those things can produce.

[DeathTrap]

Right! Fresh out of the wrapper, the S54 is a beast. It probably can get massaged too. I'd love to give it a shot after I ruin a few M54 heads.

So the N54 is a strange case, because turbo. I believe BMW kept the flow low to increase port velocity for better throttle response and to make it feel more like a large NA engine.

I have a total of 4 engine airflow/cylinder head modding books. Let me study a bit and I'll be back.

Sent from my Pixel 2 XL using Tapatalk

[ZHPizza]

So the N54 is a strange case, because turbo. I believe BMW kept the flow low to increase port velocity for better throttle response and to make it feel more like a large NA engine.

Ah yeah I think you nailed it there - their goal was to make it feel like a N/A V8. Good stuff!

[KendrickC]

Awesome project - yes Adam will has done a lot of work for it.

Side question - what technical background do you have? Will be interesting to see different perspectives/approaches to the problem (or optimisation, as you will).

[DeathTrap]

Awesome project - yes Adam will has done a lot of work for it.

Side question - what technical background do you have? Will be interesting to see different perspectives/approaches to the problem (or optimisation, as you will).Thanks!

I have my education in Electrical Engineering, but I've been too far removed from my craft for a long time. School was more rigorous than my current work, so this "problem" will give me a chance to stretch my legs out.

The funny thing is, airflow seems to behave very similarly to electrical current...which would make sense, since they are both matter that's moving per unit time.

I guess I may as well throw my perspective out there.

Airflow requires a pressure differential to induce air movement, similar to how current flow requires a voltage differential in order for electrons to flow. A wire has a ton of electrons but they aren't going move unless a voltage is applied. Easy so far.

The higher the voltage/pressure diff, the more flow happens.

Current/airflow can increase, but too much current in an electrical conductor causes heat and a voltage drop due to friction. Similarly, too much air flow through an orifice or pipe causes friction and a pressure drop.

A restriction is something that resists the flow of the air and causes a pressure drop. A resistor is the electrical equivalent. These restrictions can be precisely controlled for a plethora of uses...like flow benches. *Hint*

Funny enough, the vacuum industry uses the term "airwatts" which is pressure times flow... similar to real "watts" which is voltage times current. Both represent power or the usefulness of the energy.

I have to think about an analogy for velocity...lol

But, that's my thought process and how I will go about looking at the air movement. Technically, I could model the system as an electrical circuit (SN: some mechanical engineers will model systems as electrical circuits because they have the exact same equations describing them...and circuit simulations are very reliable and cheap)

Sent from my Pixel 2 XL using Tapatalk

[KendrickC]

Excellent :D

I'm a mech eng now doing systems engineering. I'm always itching to do more hands on engineering. You're right that some mech engs will use some electrical 'styles' too

Sent from my SM-N960F using Tapatalk

[ZHPizza]

That was great. I'm a mech eng as well and fluid flow easily clicked for me, but I never found a good way to visualize electricity. I think you just helped me understand things a lot better - wish I had seen that a dozen years ago!

[DeathTrap]

I'm doing a job rotation to SE. I feel like it has nothing to do engineering and more to do with project management. It'll be nice on the resume though.

Glad I could paint a picture!

So a minor update, I have purchased an engine simulation software called Pipemax. Apparently it's supposed to be really good. There's a few parameters I'm not too sure about, but I'll make sure to get screenshots and post them here when I get it up and running.

2. There is a complete engine in Charlotte that just needs a headgasket. I may pick that up, freshen it up and throw it in the car. Then I can really take my time and even experiment a bit more with my old engine...

[DeathTrap]

Bump. I actually did get that engine. My RX7 is eating up my time now but I hope to get back to this soon.


Sent from my iPhone using Tapatalk