42 replies to this topic

[Canuck]

If one spends any time reading over at Bob Behn's site, you quickly get the impression he has a significant distaste for billet aluminium Harley-type v-twins, especially when it comes to cylinder heads. A distilled version of his rant is that F1 and NASCAR (amongst others) don't use billet heads so they must not work, thus the cast heads he uses are better than the billet offerings.

Now, I haven't been ripped off by a billet engine manufacturer so perhaps my perspective is a bit different. Fpr a one-man, one-off type of scenario, making a billet head is far easier than making molds and finding someone to cast them, especiallly not having to deal with water jackets.

Do the NASCAR and F1 - run heads have 'perfect' ports that they run without any work, or is the casting fully worked over by hand or CNC, or are they somewhere in the middle of things? Is the sand-cast finish significantly superior to that left by a sanding roll on a die-grinder?

[shaun979]

The castings are CNCed and/or hand finished. If you have enough time (small stepover) you can CNC an almost "grooveless" finish.

[mtkawboy]

Mr Behn got ripped off bigtime both finacially and technology wise by Mr Patrick, owner of Patrick Racing that makes billet HD heads and runs Yamahas Prostar V-Twin program from what Ive heard and read. How much is true and what the other side of the story is I cant say.

[Canuck]

Originally posted by shaun979
The castings are CNCed and/or hand finished. If you have enough time (small stepover) you can CNC an almost "grooveless" finish.

True, but the argument centres around the relative texture of the finish. The sand-cast port being compared to a "killer whale's skin" to keep air-flow from sticking to the port walls. Working in a CNC shop, I'd have to agree with Behn that there's no tool that I'm aware of that could duplicate that in a fully CNCed port.

I guess my question really is will a port with an intact sand-cast texture on it's surface flow 'better' than a port with identical geometries but a CNCed or sanding-roll-on-a-die-grinder texture?


OT: I've talked to Bob & Co a number of times. Great guy to talk to but slower than molasses to deliver product. I orderd (and paid for) 2 identical exhaust systems in September, recieved one in November and the 2nd after countless phone calls, in March the next year. One time I was told they couldn't get the parts out to the chromers because it was raining to hard. Great folks, great products, lousy delivery.

[shaun979]

There are too many successful championship winning race heads across many different engine types, that are CNCed and/or hand finished for one to conclude that sand cast finish is large enough a factor to make a difference vs 60,80 or 100 grit cartridge rolling. In fact I have never seen or heard of any race port (F1, NASCAR, Champ car, IRL, all sorts of drag race, GT, DTM) that is left as-cast - unless mandated by the rules.

[desmo]

Originally posted by Canuck



I guess my question really is will a port with an intact sand-cast texture on it's surface flow 'better' than a port with identical geometries but a CNCed or sanding-roll-on-a-die-grinder texture?

Mightn't sandblasting or abrasive extrusion of the appropriate grit replicate the texture of a raw sand cast port? I'd think almost any texture wanted could be done post CNC somehow.

I didn't read the link, what is the theorized issue with billet heads? Thermal stablilty? Surface textures from machining causing stress raisers?

[Canuck]

I thought the same thing of the sandblasting. Surely if the port's texture is THAT critical, media-blasting of some sort ought to do the trick - especially if traditional ported heads have a sanding roll texture. And then we go and coat them up with a thermal barrier coating anyway...

You mean other than having a bone to pick? Apart from the cost vs. a mass-produced head, the cleaning of a giant chunk of billet and the fact that most consumers of billet motorcycle engines don't know horse **** from horse power, his argument is presented below.

If you think your billet wonders flow more than these raised port special castings then you still think Henry Ford should have stopped with the Model T. It simply does not make sense to take 40 lbs of aluminum and try to carve all the nuances that thousands of manhours of development can put into a dedicated special purpose casting. Ferrari V10 Formula 1, Illmore (Mercedes) V8 and V10, Porsche 6 Cyl, Factory Chevrolet Corvette V8 Racers, Dodge Viper V10 (world champion), Jaguar(Ford) V10 Formula 1, Honda V10 Formula 1 all use dedicated castings for their cylinder heads. You figure out who's stupid...

ORCAS take this airflow business seriously because these are nothing but two lunged air compressors with two big inlet valves that have about as much chance of being efficient as a modern Formula 1 engine as does that snowball cruising through hell. ORCAS hate polished ports which grab the air like peanut butter sticks to the roof of your mouth. ORCAS employ roughened ports that emulate their sea dwelling cousins to keep the port speeds high. ORCAS don't use flattened or excessively radiused valves because they either kill the peak flow or compromise the low speed flow. ORCAS employ sophisticated short turn radius trench cut techniques developed in Porsches and NASCAR to increase the flow on the short or inner surface of the port radius. ORCA SA motors staighten out the ports like any modern race engine tries to do...granted the valves aren't set at 10 degrees to the bore, but what the hell, we gotta get them in the frames. Valve seats and angles are decidedly non tricky because they have to survive in the real world where razor thin seats disappear faster than you know who at closing time.

Bigger ports aren't better ports...but small valves and smaller ports will limit horsepower in these big bore motors. Velocity is king. Piston speeds are a variable in the four stroke cycle and you must flow the ports to reflect this. Testing the ports at 10 to 12" of water is a waste of time as was proven by Smokey Yunick over 40 years ago. A little bit of sandpaper can do something a canned cnc program can't...and all the ball end mills in the world can't machine a real ORCA's skin. Nature rules. Art can't always be shrink wrapped.

Orcas would be his line of big-inch, big-turbo, big power engines.

Nigel Patrick builds complete billet engines for the show crowd and folks with lots of money to throw away. At some point in time, Patrick pissed in Bob's cornflakes and the rest is history. You can read more of Bob's *ahem* enthusiastic writing on the seond link which takes you to (and I'm not kidding) the Eat **** and Die section of the company website. Could never accuse him of being subtle.

[McGuire]

Not that I would attempt to translate Bob Behn when he is in full rant, but I don't think he has a problem with billet as a material, or with port finishes specifically, so much as with the Cult Of Billet. Just because a part is "billet" it is supposed to be trick, and we all know that ain't so. Billet is just a word describing a material, but somehow the word has gained magical properties.

Anyone with the $$$ can now walk in and buy a 5-axis CNC machine. They don't ask for your ID, let alone your technical credentials. People are making stuff with no learning curve behind it. That is one unfortunate downside of all these wonderful technological resources available today.

[desmo]

I've always assumed cast and billet machined parts, all else being equal.... are equal.

I live near an Aluminum plant and the billets I've seen are castings.

I'm even skeptical of the claimed upsides in the mechanical properties of forging steel and Al vs. stock removal though

[Canuck]

I've often wondered how a set of billet titanium rods would fare vs. a forged set (again because I can make all the billet stuff I want).

But, and I could be wrong of course, I don't think it being machined billet or casting would have a significant impact on the integrity of something like a cylinder head, and to my limited knowledge, there are no forged automotive/racing engine blocks.

[NTSOS]

Actually there are:

Forged Blocks

John

[shaun979]

Also LSM will reverse engineer any block and machine it out of billet.

[Canuck]

I really need to keep prefacing my comments with "I could be wrong but" 'cause every time I do, I am wrong and learn something new. I'd heard of CNCed blocks before but not forged.

I wonder what their forgings look like before they're machined.

[McGuire]

John Rodeck (TFX) makes 'em, for Top Fuel racing. That block is the weapon of choice in NHRA anymore.

[Stian1979]

I exspect annything machined to be less durable.

Anny scratch in metal from tool's are the place crack's will start to grow.

[J. Edlund]

Originally posted by Canuck

True, but the argument centres around the relative texture of the finish. The sand-cast port being compared to a "killer whale's skin" to keep air-flow from sticking to the port walls. Working in a CNC shop, I'd have to agree with Behn that there's no tool that I'm aware of that could duplicate that in a fully CNCed port.

I guess my question really is will a port with an intact sand-cast texture on it's surface flow 'better' than a port with identical geometries but a CNCed or sanding-roll-on-a-die-grinder texture?


OT: I've talked to Bob & Co a number of times. Great guy to talk to but slower than molasses to deliver product. I orderd (and paid for) 2 identical exhaust systems in September, recieved one in November and the 2nd after countless phone calls, in March the next year. One time I was told they couldn't get the parts out to the chromers because it was raining to hard. Great folks, great products, lousy delivery.

I read some article about NASCAR heads, it stated that they just finished the heads in a CNC machine and used that surface as it was.

I doubt that a rough surface should result in any benefit. What a rough surface can do better is to make the airflow follow the surface, but the flow in the head is probably so turbulent anyway so it doesn't matter. A rough surface also increase "skin friction".

Originally posted by desmo
I've always assumed cast and billet machined parts, all else being equal.... are equal.

I live near an Aluminum plant and the billets I've seen are castings.

I'm even skeptical of the claimed upsides in the mechanical properties of forging steel and Al vs. stock removal though

It all depends on how you do it. If we take compressorwheels as used in many racing turbochargers as an example, they are almost always billet machined from an aluminium forging and compared to ordinary compressorwheels which are cast the machined ones can provide a longer life and a higher strength. The longer life is mainly to the reduction of defects in the material and a stronger part may be possible since stronger materials not suitable for castings may be used.

If we take the billets used for crankshaft manufacturing as another example, they are often double vacuum remelted to assure that they are free from defects, then we use a material that is not normally used for casting. If we make a forged crankshaft we usually used a cast microalloyed steel billet for that, cheap, strong and suitable for mass production.

Billets used for Top Fuel blocks are usually forged, but billets may also be produced from metal and/or ceramic powders. Machined AlBeMet pistons is for example made from billets which are made, usually with hot isostatic pressing from metal powder.

There are differences between castings, forgings and sintered powders. The main differences are grain structure, isotropic properties and internal defects.

Example of cast tool steel


Example of tool steel made from a powder


The powder version has better isotropic properties and less defects (no carbide segregation, and less non-metallic inclusions).

Machined parts may have surface defects, but that may also be the case for castings, forgings and sintered parts. To prevent surface defects parts can be machined, then polished or shot peened. Also, highly stressed parts should not be ceramic coated, that's like asking for trouble.

[Bob Riebe]

If you want or need a Forged Alloy block:
http://www.cnblocks.com/

[Halfwitt]

Castings are good because

a) you can make the parts near the shape you want internally.
b) less stcok removal, therefore much cheaper if you make enough parts

Billets which have been worked (rolled, forged, extruded etc) are generally better than even very good quality castings, don't suffer the same level of inclusions, porosity or other imperfections which make castings slightly risky. Very good quality castings are rare things.

In wrought aluminium, I'd have no problem making a 1mm thick wall to seal water at 3 bar, but I wouldn't risk the same part machined from a casting.

Rods: most people prefer forgings, but some people find success machining from billet. Material quality is very important, there is a lot of sh*tty Ti out there. Beware cheap material.

It is quite usual to leave CNC machined ports as machined for use on race engines, up to and including F1. They do have the capacity for producing rougher finishes by increasing the feed between tool passes, but they are generally pretty smooth to the touch. The main drawback with a casting finish is that you're never sure about the exact sizes of the port, or their exact whereabouts. Core movement of 0.5mm is commonplace, so you are likely to see mismatches between any slight machining you do near the seat and the cast port.

[Canuck]

Originally posted by Halfwitt
Castings are good because

a) you can make the parts near the shape you want internally.



The main drawback with a casting finish is that you're never sure about the exact sizes of the port, or their exact whereabouts. Core movement of 0.5mm is commonplace, so you are likely to see mismatches between any slight machining you do near the seat and the cast port.

And this is what had me pondering the idea that top level racing ran as-cast stuff. Thanks to everyone for getting me sorted. I'll now move ahead with one-off billet heads for my v-twin without worrying that I've missed the boat because my ports aren't cast.

[Fat Boy]

Originally posted by Canuck
I guess my question really is will a port with an intact sand-cast texture on it's surface flow 'better' than a port with identical geometries but a CNCed or sanding-roll-on-a-die-grinder texture?

A friend of mine works at Ford doing intake tract work. Basically it comes down to intakes and heads. Several years ago he was on a project to get a little more performance out of one of their engines for a special edition car (a SVT this or that, I don't remember exactly what). Anyway, he had read about the Extrude-Hone thing where they push a thick abrasive paste through your ports to polish them. He grabbed 2 heads and had them flowed. He sent one out to be Extrude-Honed and when it got back had the same two heads flowed again. They have really accurate flow bench testing available (like only an OEM can). The head that wasn't touched repeated perfectly. The one that had the polished ports lost flow across the range. He was skeptical of the validity of the test because the Extrude-Hone port _looked so nice_. So he sent them back and had them tested again. The results were the same.

The head had better performance before the Extrude-Hone process because of the surface roughness.

That led him down the path of playing with surface roughness. Apparently it isn't terribly critical, but a glass finish hurts flow. Hit is with some 120 or 80 grit sandpaper for 5 minutes, and the flow comes back. 'As cast' finish is rougher, but doesn't hurt or help flow.

The problem with sand casting heads is more an issue of consistency than anything. You will have some core shift here and there. There's just no way around it. It's another thing that our friends from Japan seem to have figured out way better than anyone else on the planet.

[Powersteer]

Honda's aluminium alloy diesel block seems tricked out enough

In order to produce a light, compact and high performance aluminium cylinder block that provides sufficient strength to cope with the high combustion pressures generated, Honda has pioneered a semi-solid casting method using a sand core cylinder.

To provide high levels of rigidity, the cylinder block water jacket requires a hollow construction known as ‘closed deck’. It is very difficult to achieve this with only a metal mould and the general method employed is to use a cylinder core of sand during casting. Afterwards the cylinder core is destroyed and removed.

However, during the high-pressure diecast process that is generally used for the manufacturing of cylinder blocks, the cylinder core can easily break and so this method was unsuitable. Honda has therefore developed a high strength sand cylinder core with a pressure-resistant coating which in conjunction with the original Honda semi-solid casting method realises a closed deck structure with high rigidity.

The new casting method takes an aluminium alloy in a semi-solid state which is then smoothly poured into a metal mould after which high pressure is applied. The good formability characteristics of the semi-solid material and a sand cylinder core that allows precise casting mean that the thickness of the aluminium between the bores is reduced to just 3 mm to give an overall length equivalent to that of a petrol engine. Furthermore, the high quality of the casting permits heat treatment after the moulding process to give even greater rigidity and surface hardness. The end result is a product that is lighter, stiffer, stronger and more accurate

High density casting??? sounds really expensive.

[Greg Locock]

I think that's also called squeeze casting.

I suspect that 'forged' aluminium suspension arms can be made in a similar fashion. Very neat.

[Canuck]

Thanks FB. The 'hot-ticket' for your E12 BMW is to swap a late® model 3.5L with Motronic injection and a big ugly manifold in. The really trick deal is to have your manifold extrude-honed. Being the open-minded skeptic I am, I was doubtful of any significant benefit but still willing to be proved wrong. Very informative story - thank you.

So in essence we're back to the time honoured practice of porting the intake (if needed) but only polishing the combustion chamber (or coating it). CNC ports can be done in an acceptable fashion. Billet parts are expensive (and far too trendy) but they can still do the job. Excellent. It's either a bunch of billet one-ofs as I test and learn and test and learn, or I get real chummy with a pattern maker and a casting / forging outfit.

[Powersteer]

Maybe this is the absolute finishing you are looking for

[DOHCPower]

the answer is so simple.

What you want is the proper CNC port geometry, with the tooling marks warmed over with a catridge roll by hand, and a rougher finish left on the intake tracts. Ive always understood the rougher finish encourages air/fuel mixture, as well as concentrating most of the flow in the center of the port.

[gbaker]

In general, isn't some degree of surface roughness advantageous to fluid flow, ala the nose of a submarine?

[desmo]

I've read it can help keep the boundary layer attached on trailing sides of bodies. Airplanes don't routinely have deliberately textured surfaces, but maybe its a scale/Reynolds Number thing which could explain the submarine noses too.

[gbaker]

Makes sense.

I heard years ago re the subs that the criteria were optimized with scale models, but had to be changed because water couldn't be scaled.

[CFD Dude]

Originally posted by Fat Boy


A friend of mine works at Ford doing intake tract work. Basically it comes down to intakes and heads. Several years ago he was on a project to get a little more performance out of one of their engines for a special edition car (a SVT this or that, I don't remember exactly what). Anyway, he had read about the Extrude-Hone thing where they push a thick abrasive paste through your ports to polish them. He grabbed 2 heads and had them flowed. He sent one out to be Extrude-Honed and when it got back had the same two heads flowed again. They have really accurate flow bench testing available (like only an OEM can). The head that wasn't touched repeated perfectly. The one that had the polished ports lost flow across the range. He was skeptical of the validity of the test because the Extrude-Hone port _looked so nice_. So he sent them back and had them tested again. The results were the same.

The head had better performance before the Extrude-Hone process because of the surface roughness.

That led him down the path of playing with surface roughness. Apparently it isn't terribly critical, but a glass finish hurts flow. Hit is with some 120 or 80 grit sandpaper for 5 minutes, and the flow comes back. 'As cast' finish is rougher, but doesn't hurt or help flow.

The problem with sand casting heads is more an issue of consistency than anything. You will have some core shift here and there. There's just no way around it. It's another thing that our friends from Japan seem to have figured out way better than anyone else on the planet.

Ford did try to use this on a program - it was the 2.5L V6 in the SVT Contour. I know several people at Roush who were doing the engine development work. They actually told me that the extrude honing (http://www.extrudehone.com/ for more info) did help the power on the dyno, not decrease it. Their story was that for cost reasons it didn't make production - instead they resin dipped the port cores instead to get the smooth finish.

I don't want to say that either of us is right or wrong in this case - I'm going from a friends story as well - but in all of the cylinder head development work that I've been involved with I've never seen anyone rough up a port after they machined it.

[CFD Dude]

http://wardsautoworl...e_shop_extrude/

turned this up. It's a few years old, the last time I was in contact with Curt Hill he was the lead Ford engineer on the Ford GT engine program.

[Powersteer]

What about fastskin swimming suite technology. It seems that the surfaces are tune to help the air around a turn or curve. Maybe rougher edges on the inner corner of a port and smooth on the outer of even vice versa.

[Fat Boy]

Originally posted by CFD Dude
I don't want to say that either of us is right or wrong in this case - I'm going from a friends story as well - but in all of the cylinder head development work that I've been involved with I've never seen anyone rough up a port after they machined it.

Interesting what a small world we live in. It must have been the same project. Possibly this is one of those situations where the flowbench didn't necesarily agree with the dyno. It's possible that both ends of the story are correct. The flowbench portion was told to me first hand by that engineer. He's a good friend (was in my wedding), so I don't have any reason to think he would tell me wrong.

As far as roughing up a port, I got the feeling that normal tooling marks had adequate surface roughness to produce the desired effect. Certainly, we've all seen a golf ball and the surface on it has a lot to do with the drag. The 'shark skin' suits that olympic swimmers use is another example of a similar principle. I have heard of engine builders dimpling the intake port on the short side radius to help the air to turn the corner. This is not first hand, but from random bull-******* with people who should know.

In the end, there are probably many, many variables involved, and I doubt if any one solution is universally correct.

[desmo]

Isn't it a little funny how far people will go to chase tiny incremental gains in VE instead of doing the more logical thing and increasing the bore and/or stroke? Displacement limits are strange if you think about it. Or at least if I do

[Canuck]

Lots of racing classes have displacement limits so it pays in that respect. For me, more displacement is a little like cheating. It's the most obvious and the easiest method - hell, anyone can stroke (or bore) a Harley. If I can bore it rounder, seal it better, reduce the friction, increase the Ve and the mileage all in one go, then my 100" or 80" or 350" powerplant is going to perform that much better than Joe down the road's stuff. Besides, haven't you ever had a smug sense of satisfaction when your small engine whomped someone else's big-incher?

[CFD Dude]

Originally posted by desmo
Isn't it a little funny how far people will go to chase tiny incremental gains in VE instead of doing the more logical thing and increasing the bore and/or stroke? Displacement limits are strange if you think about it. Or at least if I do

Well, in factory hotrod programs like SVT increasing the specific horsepower of the engine is more desirable from a marketing and cost perspective. Marketing wise you want the fun version of the engine to have a connection to the garden-variety engine that the average buyer can purchase. If they can't buy the fun car, maybe they can take some satisfaction in what the engine 'could' do.

In respect to cost - large scale production engines are cheap, they're cast and machined and assembled on assembly lines. If you increase the displacement with bore and/or stroke you're looking potentially at; new machining, new honing, new pistons, new connecting rods, new crank, new installation procedures (ie it's not assembled on the existing assembly line), etc. This dramatically increases the cost. If you can reuse everything but the pistons (for compression), a new intake manifold, a reprogrammed ecu and maybe some porting to get the same increase in power you can build it on the line, you just swap out some parts, and you can take advantage of all of the other low-cost production tooling and parts because they are the same as the garden-variety version. Instead of costing $2000-3000+ more for each special hand assembled unique parts hotrod engine, you can machine and build them on line for just the cost of the few new components.

[J. Edlund]

Originally posted by gbaker
In general, isn't some degree of surface roughness advantageous to fluid flow, ala the nose of a submarine?

It depends on the application. For surface roughness to become an advantage the major losses must be due to pressure difference, not skin friction since increased roughness increase skin friction. Pressure losses are caused when the pressure is higher in front of an object, so if this is the main loss, and flow is near turbulent (depends on reynolds number), increased roughness can make the flow turbulent, and this means the flow will follow the surfaces of the object better. This will then reduce the pressure behind the object and flow losses goes down.

When flow already is turbulent and/or the main losses are skin friction there are only disadvantages in using a rough surface finish. This means that there are no general solution, each case must be studied. For most applications this isn't cost effective.

Originally posted by CFD Dude


Well, in factory hotrod programs like SVT increasing the specific horsepower of the engine is more desirable from a marketing and cost perspective. Marketing wise you want the fun version of the engine to have a connection to the garden-variety engine that the average buyer can purchase. If they can't buy the fun car, maybe they can take some satisfaction in what the engine 'could' do.

In respect to cost - large scale production engines are cheap, they're cast and machined and assembled on assembly lines. If you increase the displacement with bore and/or stroke you're looking potentially at; new machining, new honing, new pistons, new connecting rods, new crank, new installation procedures (ie it's not assembled on the existing assembly line), etc. This dramatically increases the cost. If you can reuse everything but the pistons (for compression), a new intake manifold, a reprogrammed ecu and maybe some porting to get the same increase in power you can build it on the line, you just swap out some parts, and you can take advantage of all of the other low-cost production tooling and parts because they are the same as the garden-variety version. Instead of costing $2000-3000+ more for each special hand assembled unique parts hotrod engine, you can machine and build them on line for just the cost of the few new components.

Today it's usually quite cheap to swap bores and strokes in one engine on the same production line. Infact, most engines are designed for it.

Making different intake manifolds for engines is on the other hand something that becomes more and more costly, especially if only a small production run is going to be used. Why? Intake manifolds of for example glass fibre reinforced nylon has become more common, and in massproduction they are cheap, very cheap, they are also light and decrease the heat transfer into the intake air. The tooling required for a specific design is however very expensive.

[m9a3r5i7o2n]

When I see things similiar to this I wonder just how long it will be before someone starts making Chevy based cylinder heads with D.O.H.Cs, after all its only money!!

M.L. Anderson http://www.cnblocks.com/

[gbaker]

Originally posted by J. Edlund
It depends on the application. For surface roughness to become an advantage the major losses must be due to pressure difference, not skin friction since increased roughness increase skin friction. Pressure losses are caused when the pressure is higher in front of an object, so if this is the main loss, and flow is near turbulent (depends on reynolds number), increased roughness can make the flow turbulent, and this means the flow will follow the surfaces of the object better. This will then reduce the pressure behind the object and flow losses goes down.

When flow already is turbulent and/or the main losses are skin friction there are only disadvantages in using a rough surface finish. This means that there are no general solution, each case must be studied. For most applications this isn't cost effective.

Well done. Thank you. (My background in structures doesn't help much with this topic.)

[Canuck]

Originally posted by J. Edlund


It depends on the application. For surface roughness to become an advantage the major losses must be due to pressure difference, not skin friction since increased roughness increase skin friction. Pressure losses are caused when the pressure is higher in front of an object, so if this is the main loss, and flow is near turbulent (depends on reynolds number), increased roughness can make the flow turbulent, and this means the flow will follow the surfaces of the object better. This will then reduce the pressure behind the object and flow losses goes down.

When flow already is turbulent and/or the main losses are skin friction there are only disadvantages in using a rough surface finish. This means that there are no general solution, each case must be studied. For most applications this isn't cost effective.

Ideally then, without access to flow, velocity and with any luck some visualization (wet flow bench) tools, a 'typical' finish is likely to be just fine. Good to know. Also then, if port velocity is maintained, and flow increased, Ve and BMEP should increase over the same RPM no?

[CFD Dude]

Originally posted by J. Edlund

Making different intake manifolds for engines is on the other hand something that becomes more and more costly, especially if only a small production run is going to be used. Why? Intake manifolds of for example glass fibre reinforced nylon has become more common, and in massproduction they are cheap, very cheap, they are also light and decrease the heat transfer into the intake air. The tooling required for a specific design is however very expensive.

Sorry, but I'm going to have to disagree with you on this. I've worked on a number of large production and niche engine programs, and except for cases where one builds tens of thousands of both bore/stroke options I haven't worked on a program where we swapped bores and stokes. Not to say that they aren't out there, but I'm having trouble thinking of any engines here in North America where they sell small volumes of a stoked and bored engine. I know that there are tuners out there that do that, but thats different that production engines, and much more $$$.

In the case of building ~10 thousand 'hot rod' engines a year on the same line that 2-300+ thousand 'regular production' engines, the cost of a new crankshaft forging and machining for a new stroke, and a new piston blank for a new bore are vastly more expensive than a new low pressure sand casting for a new intake manifold and new pistons machined from the same blanks as their lower compression, standard production counterparts. That's assuming that your wall stock allows for machining a larger bore, if you have to change the waterjacket core and/or sleeves that's even more money. A sand casting for an intake manifold for low volumes can be made from prototype style molds and machined on flexible tooling. On large prototype orders I've had castings about the size of a 4 cylinder intake manifold casting in the door fully machined for $150 - 250 each, and that's only for orders of ~100 parts, not thousands. Then you grab the throttlebody and all of the sensors, egr valves, etc for the from the company parts bin, so they're relatively cheap, and you've got your high output version of the engine on the cheap. It's the tooling that kills you on small production volumes, you try to have as little new as possible.

[CFD Dude]

Oh, and as for composite intake manifolds - I understand your argument on absorbing less heat, but a surprising number of intake manifolds are still sand cast aluminum. Because the composite intake manifolds can't handle the cylinder head so they are often coupled with a cast aluminum 'lower intake manifold that works as a buffer, and then there is the sealing between the upper and lower manifolds to consider. There seem to be fads that go to the composite, then to the cast aluminum and then back again.

[J. Edlund]

Originally posted by CFD Dude


Sorry, but I'm going to have to disagree with you on this. I've worked on a number of large production and niche engine programs, and except for cases where one builds tens of thousands of both bore/stroke options I haven't worked on a program where we swapped bores and stokes. Not to say that they aren't out there, but I'm having trouble thinking of any engines here in North America where they sell small volumes of a stoked and bored engine. I know that there are tuners out there that do that, but thats different that production engines, and much more $$$.

In the case of building ~10 thousand 'hot rod' engines a year on the same line that 2-300+ thousand 'regular production' engines, the cost of a new crankshaft forging and machining for a new stroke, and a new piston blank for a new bore are vastly more expensive than a new low pressure sand casting for a new intake manifold and new pistons machined from the same blanks as their lower compression, standard production counterparts. That's assuming that your wall stock allows for machining a larger bore, if you have to change the waterjacket core and/or sleeves that's even more money. A sand casting for an intake manifold for low volumes can be made from prototype style molds and machined on flexible tooling. On large prototype orders I've had castings about the size of a 4 cylinder intake manifold casting in the door fully machined for $150 - 250 each, and that's only for orders of ~100 parts, not thousands. Then you grab the throttlebody and all of the sensors, egr valves, etc for the from the company parts bin, so they're relatively cheap, and you've got your high output version of the engine on the cheap. It's the tooling that kills you on small production volumes, you try to have as little new as possible.

Today many engines are designed to use different strokes and bores; GM's HFV6, GM's Ecotec, well Volvo doesn't only swap bore and strokes they also swap cylinder numbers, from 4 to 6 cylinders with the same design and main components. This means that you can combine certain sets of standard components to make many different engines from the same basic design. Usually this is used to reduce the number of basic engine designs. Since many different components already exist, it may not even to be needed to make new components. One example of this is the supercharged Saturn LSJ Ecotec which uses all main components from Saabs turbocharged Ecotec.

There are of course even more exclusive engines built with altered bores and strokes, AMG built Mercedes engines are usually made that way. But their engines are also hand built with exclusive components ata very high price.

The cheapest method to make a higher output engine is however to modify the software in the ECU. Some engines even use the same software for different outputs, the ECU gets info about VIN by CAN and then sets the correct power output.

Originally posted by CFD Dude
Oh, and as for composite intake manifolds - I understand your argument on absorbing less heat, but a surprising number of intake manifolds are still sand cast aluminum. Because the composite intake manifolds can't handle the cylinder head so they are often coupled with a cast aluminum 'lower intake manifold that works as a buffer, and then there is the sealing between the upper and lower manifolds to consider. There seem to be fads that go to the composite, then to the cast aluminum and then back again.

When GM made a composite intake for the 1993 3800 V6 the following advantages where found:

1. Unit cost savings of 45%.
2. Mass savings of 66%.
3. Simplified assemby and serive procedures.
4. Improved emissions performance due to routing of EGR into the manifold.
5. Improved engine performance due to reduced air induction temperatures.
6. Reduced shipping costs due to lighter components.
7. Increased standardization across vehicle programs.

The latest Northstar engine also have a composite intake where the basic design is similar. The intake is sealed to the heads by elastomeric gaskets.

Many cast intakes also have a separate cast lower intake.

[CFD Dude]

I think we're talking different volumes here - yes Volvo makes 4-5-6 cylinder engines that share components, GM does the same with the HFV6 and it's family of inline truck engines. The ecotech is avaliable in 2.0, 2.2 and 2.4L versions here in North America, I'm sure that there are lower displacement versions in Europe. Ford also does this with the 'Modular' V8 and V10 engines, they produce 4.6L V8, 5.4L V8, and 6.8L V10. The difference from what I was talking in my previous posts is that they make large volumes of each of those engine variations for production lifes of 4-5+ years. Economies of scale come into play on each of the different engine variations. The SVT Contour engine that Fat Boy was talking about (Extrude Honing - that's how we started on this path...), in that case Ford was trying to make a very low volume alterative engine for a short production run starting with a stardard production engine. They were also trying to make the engine on the cheap - a fancy Ford Contour can't charge the premiums that AMG can manage on a fancy Mercedes.

If I need another 20 hp for 50 thousand+ engines over the course of the production life of one or more vehicles, then increasing the displacement like Desmo suggested makes perfect sense. My argument was that for ~10 thousand engines/year with a production run of only a year or two, like SVT does, you're better to tune the engine with a few low-cost components.

As for the intake manifolds, in my job (I work for a tier 1 automotive supplier) we routinely tear down engines from every manufacturer to see what technology and designs are out there and what our competitors are developing. About half of them have composites intakes and half are cast. In a few cases I've seen two-piece die-castings. To simply say that 'composite intakes are cheaper' isn't necessarily always true even though, like in your example, it sometimes can be. In the case of the GM 3800 engine the composite was cheaper and the prefered design at the time. The engine was recently replaced with the new 3900 engine:

http://cgi.ebay.com/...1QQcmdZViewItem

Note the cast intake manifold.