How good are 'typical' cylinder heads from the factory?

[MC Bodge]

How well do typical modern petrol or diesel cylinder heads from VAG, Ford, GM etc. flow air? I know that there were often large improvements to be gained from 'porting' standard cylinder heads in the past, but is this still the case with improvements in production methods?

Nowadays it seems common to take a turbocharged engine and increase output by increasing the air flow forcibly to rather than following the (expensive) traditional route of porting, polishing and fitting bigger valves until later in the process.

ps. I'm not intending to attempt a home porting job.

[anonymous-user]

Well it depends on the manufacturer!


If you are really chasing every last horse, then some port work (in conjunction with larger valves etc) WILL release more power at higher rpm, but almost certainly significantly reduce low rpm performance. You are much more likely to be able to "improve" say a head from a Ford focus, than a fezza F430!


generally, engines are so well optimised for there intended role, that you will not be able to improve any particular facet without negatively impacting on another

[Pumaracing]

MC Bodge said:

How well do typical modern petrol or diesel cylinder heads from VAG, Ford, GM etc. flow air? I know that there were often large improvements to be gained from 'porting' standard cylinder heads in the past, but is this still the case with improvements in production methods?

There has not really been any significant change in recent production methods that affects how much air a head flows for a given valve size. That depends, as it always has done, on what the engine designer intended to achieve or was targeted to achieve by the marketing department. No doubt any competent designer with flow bench experience could if he wanted design and cast a port shape that flowed at a very high level but that might impose packaging constraints such as extra engine height that simply can't be tolerated or just not be wanted or needed for the target power output.

The Porsche 911 air cooled heads from 30 or more years ago were almost impossible to improve so clearly it could be done perfectly well back then if desired. The casting or seat cutting technology was never the issue - it was the port shape design someone specified.

So as it always has been heads still range from the very hard to improve to the very easy. There is no "typical". I suppose one might surmise that the higher the specific bhp output per litre for a normally aspirated engine the more likely it is that the head flows well as standard but I wouldn't bet my shirt on it.

The Subaru Imprezza and Mitsubishi Evo heads from about 10 or so years ago make a nice comparison. The cars are in direct competition with each other, they thus need very similar power outputs and both are designed in Japan so you might assume both heads are similarly good. In fact the Evo head has very nice ports which flow very well for the valve size whereas the Imprezza head is dreadful but has bigger valves which compensate somewhat for the awful port design. I did one of those a few years back and added 80 bhp to the engine with some simple porting mods retaining the standard valve sizes. With the Evo you really need to fit bigger valves to get a decent gain which is obviously a lot more expensive.

So why did one designer use small (relatively) valves but really nice ports to optimise what the valves could flow and another use bigger valves but restrict them down with small ports? Your guess is as good as mine. Probably so they could make the ports bigger for competition purposes without having to redesign anything else or fit bigger valves yet still be competitive with the Evo at similar boost pressures for road use. Much like what Cosworth did with the Sierra Cosworth engine. The target was only 200 bhp for road use but up to 500 for competition so they put big valves in to start with but made the ports tiny but with plenty of metal around them so they could be machined out later. They could of course have very easily made the original motor produce 300 bhp or more for road use with bigger ports to start with but that wasn't what Ford wanted.

[Pumaracing]

Max_Torque said:

If you are really chasing every last horse, then some port work (in conjunction with larger valves etc) WILL release more power at higher rpm, but almost certainly significantly reduce low rpm performance.

I'm afraid that isn't even remotely true. Properly modified ports, with or without bigger valves, don't hurt low rpm performance. That only happens when too much metal is hogged out in places it didn't want to be so gas speed goes down but flow doesn't go up. I've made my reputation for over 20 years by being able to design and produce port shapes that improve the torque curve everywhere with dramatically more high and mid rpm power and no loss low down.

Obviously if you fit longer duration cams then you can't expect the same low rpm torque but it isn't the head that determines that if it's been done properly.

In fact I'd say it's the other way round. If you want the best possible torque curve for a specific target peak bhp output then fit the largest valves you possibly can, optimise the port shapes for those valves and then use LESS cam duration than you'd otherwise need. If you can find 20% more head flow and use 20% less cam duration you'll end up with similar peak power but a damn sight more tractability.

Perhaps a read of this might elucidate.

http://www.uksaabs.co.uk/UKS/viewtopic.php?f=35&am...

[anonymous-user]

My response was in relation to std OEM ports and intakes, which are primarily designed for a good spread of torque (typically making over 85% vol eff right down at around 1400rpm these days) This requires a "small" high velocity port, and a long intake runner (so the helmholtz resonator is effective at the low speeds). (making high torque at low rpm isn't just about the port flow either, it's about getting sufficient charge motion to maximise the burn rate (detonation is prevalent at low rpm where the charge has a low dynamic component)

So, just about any mod you make to increae power (which is a high rpm factor by necessity) WILL have an impact on the systems optimisation at low rpm.

Increased "Power" is increased intake mass flow at high rpm, which means the time availible for each intake event decreases, however you need to move at least the same mass of air per cylinder filling event as at low rpm, so the "mean velocity" of the port will be higher. As velocity increases a point is reached where the balance between dynamic pressure recovery and skin friction (and inertial interuption) become limiting factors, in effect the port "chokes" (usually at a little bit above ~100m/s)

Now, with this new "large" port, you can make more power than the std, engine, but at low rpm, no matter how good you cylinder filling is, your air charge dynamic component will be lower than the std engine, decreasing you burn rate, and resulting in a higher heat loss to the chamber and less BMEP (and if you're really unlucky, tipping you over the edge into detonation, requiring a reduction in ignition advance and really hurting BMEP)

So this just leaves you with port mods that increase man vol eff, without a change to the bulk port velocityand very, very, few modern ports will have features that you can "remove" that significantly incease mass flow without changing the mean port velocity. They will have been designed (with an enormous amount of CFD) to optimise port velocity and dynamic pressure distribution along the port to just deliver the power target, but to also meet the "torque spread" targets (usually these days HARDER to meet than the power number!)

That being said, I suspect that allowing for the "mass produced" nature of the OEM design, changes to intake valve seat widths and valve seat throat profiles WOULD increase mass flow without reducing low rpm torque, but on a modern engine (with modern CNC & high accuracy engine build lines) these gains will be quite low (in the order of a few %, which on a cost per bhp basis makes them fairly pointless as aftermarket tuning). Also, you may be impacting on the engines durability capability with things like narrower seat widths etc


(if you are chasing large power gains, with work to more than just the ports, any change to cam or manifold /intake system will have an even bigger effect on the spread of torque, and of course on a modern engine with VVT and varriable intake runners, then you will probably now have a control system calibration that is no longer optimum for the changed system dynamics, so will possibly loose even more torque without a remap......... and of course thats before we even get onto "spray" or "air" guided GDi, which require precise charge motion control to optimise the AFR around the spark plug at the right moment for ignition. Bugger around with the ports on those, and you might even stop the engine actually firing properly at all !!)

So, as always it's a compromise, yes you can "improve" a modern engines intake tract, but you will have to accept some change in the operating range ;-)


(regarding the link to the saab thread, sorry, but how is that a "modern" engine?? I'd also be very suprised if you couldn't improve the output across the whole speed range for a 25yo 8v 2.0 with about 60/bhp litre std......

Have a go at "improving" this:



and get back to me with the new torque curve........ ;-)


Here's the torque curve from the saab thread:



and heres an audi FSi one:




now the audi is only 1.6 litres, but matches your 2.0 at approx 3000rpm, and beats it below 2500 (sorry, couldn't find audi curve in lb/ft, so you need to correct from Nm (div by 1.34)) (although to be fair, the saab is a chassis dyno curve, so you obviously haven't been able to load it fully below 2500rpm, and there's huge amount of drivetrain torque oscilation on the data, so you'll just have to extrapolate the likely shape of the curve below that rpm)

Look also, how at even 1000rpm the Fsi is making 110Nm (70Nm/l !!)



but aside from all that, as the OP mentioned though, it's really a pure cost thing that any forced induction engine is easier to "tune" by "showing it the boost" rather than mechanical changes. You just plug in a laptop, hit enter and bingo +30bhp and +80Nm. Compared to 12hrs labour removing and refiting a cylinder head and 20hrs labour porting it, all for maybe +5bhp, + 5Nm (and possibly -20Nm..... ;-)

Edited by anonymous-user on Saturday 19th March 12:32

[Pumaracing]

Don't forget the Saab engine has a pretty serious cam in it. About 30 degrees more duration than stock so any comparison with a standard road car engine isn't really appropriate. The point I was trying to make is that even with a rally cam which the owner wasn't expecting to do much below 3k, a big valve ported head and well over 80 bhp per litre this is still a completely road driveable engine. The aim was to produce a serious power rally engine which he hoped would still be just about useable on the road but he actually got the top end power he wanted plus something he can still trundle about in right down to 1500 rpm on part throttle.

Of course the Audi engine has more torque per litre low down with a stock duration cam but it also has nowhere near the bhp per litre of the Saab even with 4 valves per cylinder instead of just 2. 110 bhp from 1.6 litres is equivalent to 137 bhp from the 1.985 litre Saab. If we took that rally cam out and put a stock one back in there's no way it would lose 30 bhp of its 167 bhp, more like 15 to 17 bhp, so it would still cream the Audi on bhp per litre and I'm pretty damn sure it would have more torque per litre right down to tickover. Can't see anyone doing that just to prove a point though.

However the main thing is back to where we started. Ported heads or even big valve heads don't hurt tractability if done right.

[anonymous-user]

Pumaracing said:

Don't forget the Saab engine has a pretty serious cam in it. About 30 degrees more duration than stock so any comparison with a standard road car engine isn't really appropriate. The point I was trying to make is that even with a rally cam which the owner wasn't expecting to do much below 3k, a big valve ported head and well over 80 bhp per litre this is still a completely road driveable engine. The aim was to produce a serious power rally engine which he hoped would still be just about useable on the road but he actually got the top end power he wanted plus something he can still trundle about in right down to 1500 rpm on part throttle.

Of course the Audi engine has more torque per litre low down with a stock duration cam but it also has nowhere near the bhp per litre of the Saab even with 4 valves per cylinder instead of just 2. 110 bhp from 1.6 litres is equivalent to 137 bhp from the 1.985 litre Saab. If we took that rally cam out and put a stock one back in there's no way it would lose 30 bhp of its 167 bhp, more like 15 to 17 bhp, so it would still cream the Audi on bhp per litre and I'm pretty damn sure it would have more torque per litre right down to tickover. Can't see anyone doing that just to prove a point though.

However the main thing is back to where we started. Ported heads or even big valve heads don't hurt tractability if done right.

And that was i hope the point i was trying to make with regard to "porting" modern (or even old tech) engines, yes, you can get more power, but you will loose some low end torque. And the more modern the engine is, the more its been optimised for low rpm performance etc, so the more you are likely to loose at the bottom end. I guess the point i was really trying to make (probably not very well ;-) is that on a modern engine, the "port" really is just one of the many supporting systems that are now so interelated and optimised, that just considering "port work" is not a terrbly cost effective idea etc

I think the very valid point you are making is that you don't just have to cut a massive area port to make gains, just make the shape optimum (effectively manage the geometry so that the velocity change is smooth (dv/dl "area rule"). My point is that this is already done to a high degree on a modern engine, so to make decent gains, you are to either forced increase the port area (to limit the mean intake velocity at rated power rpm) or look to other changes like cam profile etc.

Fundamentally, power is revs x torque, and torque is all about cylinder filling, so on an NA engine, once you have max'd the cylinder filling (say peak ~100% vol eff for normal road car, ~110% for fruity road cars and ~130% vol eff for F1) then you can only increase the rpm at which that peak cylinder filling occurs to make "significantly" more power. And pushing the rpm of peak torque up the rev range (which is what you have done very successfully with that saab engine!) has the side effect of reducing the torque at low rpm. Which is i think, where i came in

Of course, on that Audi 1.6 FSI, the excellent low rpm torque has much much more to do with the variable cam timimg and detonation suppressing effects of Direct Injection that the port and intake really, but that's a discussion for another day ;-)

Edited by anonymous-user on Saturday 19th March 15:21

[Pumaracing]

Max_Torque said:

I think the very valid point you are making is that you don't just have to cut a massive area port to make gains, just make the shape optimum (effectively manage the geometry so that the velocity change is smooth (dv/dl "area rule").

That's exactly right. It's pretty hard to get the discharge coefficient of a poppet inlet valve above 0.6 on a production engine designed cylinder head, even a fairly downdraft one. That tells you immediately that the straight part of the port which flows at close to 100% efficiency doesn't need to be greater than 60% of the valve area to properly supply that valve. In practice 55% suffices most of the time, especially on 2v heads with 90 degree port bends in them. If you can maximise valve flow "without" making the port any bigger than necessary by cutting the optimum seat profiles and getting the shape of the shortside bend right then you can actually increase port velocity as well as overall CFM and the torque curve improves everywhere.

Max_Torque said:

My point is that this is already done to a high degree on a modern engine, so to make decent gains, you are to either forced increase the port area (to limit the mean intake velocity at rated power rpm) or look to other changes like cam profile etc.

That's the bit I would say there is no great consensus on. The port area as a percentage of valve area varies wildly on production engines. Some are already too big and some too small. You just have to treat each head on its merits.

The key thing is knowing exactly what to do and what not to do on each individual head which most so called engine tuners have little clue about. Regardless of whether a port is already too big they seem to think that in every case "make it bigger and make it shiny" will achieve something, or at least make it look to the customer like they've done something constructive.

[Kitchski]

I was going to link to a very interesting article I read on the Peugeot Mi16 head as I know it well and am always swooning over it. But then I re-read the thread and realised I'd be preaching to the choir as it's on YOUR Pumaracing website Dave

[Pumaracing]

The Mi16 head, and the similar GTi6 one, is indeed one of the heads that's almost perfect in stock form which is why I go straight to big valves rather than waste time trying to improve flow on the stock ones.

However its smaller successor the TU5J4 1600 4v which you might think would follow the same design criteria has ports which need quite a bit of work to even get the stock valves to flow properly and considerable enlarging to make bigger valves work. I've seen professionally ported big valve heads that had only had a bit of polishing done in the ports flow little more than standard and give no real power gain because of this.

A few years ago I took one such supposedly fully ported TU5J4 BV head with 31.3mm inlets (stock is 28.75mm) up from 160 CFM to nearly 200 CFM (stock is in the 150s CFM so little gain had been made) simply by getting the port sizes right to match the bigger valves. Similarly on the exhaust side which had 26mm valves fitted (stock is 24.5mm) but had not had the necessary port enlarging done properly flow went up from 115 to over 140 CFM if I recall correctly. All of that combined unleashed nearly 40 bhp on the dyno.

So it makes the point again nicely. Every head has to be treated on its merits and there's no substitute for understanding the rules which relate optimum port area to valve size.

[MC Bodge]

Thanks for the posts, chaps.

That's interesting about the differences between the Evo and the Impreza.

Edited by MC Bodge on Sunday 20th March 14:01

[stevesingo]

Dave,

Do you have any experience of the BMW s14 heads (e30m3).

The early, pre 1989 heads had what is referred to the 26mm (height at the manifold face) inlet ports with 37mm valves. The EvoII and later (1989 on) had 28mm ports with the same valves, with the EvoIII (2.5lt opposed to 2.3lt) having the 28mm ports also but with 37.5mm valves. The ports on all heads are cnc machined from the factory.

Early 26mm head





BMW motorsport developments for touring car racing increaded the port sizes with various sizes,

29.5mm with 37mm valves GpA rally and 38.5mm valves 1990-92 DTM
30.5mm 38.5mm valves 92DTM
31.5mm with 39mm valves 93on DTM

Ultimately these engines made 370bhp@9000+rpm in 1993.

Your thought would be welcome

Steve

[stevieturbo]

look like very nice ports !

[Pumaracing]

You can see that BMW exactly followed my rules. 55% port area for the road heads and just over 60% for the DTM ones where tractability was no longer the issue. I doubt you'll improve on those port shapes or sizes much. That's about as downdraft as you can get so low 60s percent discharge coefficient would be on the cards.

If I put 39mm inlet valves x 2 per cylinder x 4 cylinders at 60% flow efficiency into my power prediction programme the number it comes straight out with is 370 bhp.

[anonymous-user]

In my experience BMW are also quite good at getting decently controlled tollerances in the critical valve seat to port throat area. Their seat widths etc are obviously a slight compromise for "road" durability" but generaly they get these things pretty right from the factory!

[Pumaracing]

Valve seat to throat is indeed a very critical area of the port and where good gas speed makes much of its impact. Poorly educated engine tuners tend to enlarge the throats too much, narrow the valve seats too much and remove most of the bottom seat cut angle which I like to leave nice and long.

My starting rule of thumb is to have an inlet throat diameter about 85% to 86% of the valve size, inlet valve seat width to be 4.5% of the valve diameter and exhaust seat width the same as the inlet which will usually make it about 5.5% of the smaller exhaust valve diameter. Good wide seats are necessary for proper valve heat dissipation and I'm fiercely opposed to the 1mm seats that "race" engine builders seem to think they need to use in race engines. I even once saw a pair of small block Chevy heads with massive 2" plus inlet valves that some idiot in the USA had cut 1mm wide valve seats on. They'd just hammered concave within the first few hours of running. Some proper porting and recutting the seats to 2mm wide put 80 bhp on that engine.

I then adjust those starting dimensions based on downdraft angle, port shape, cam lift, tractability requirement and other things which might affect what I think necessary to get best flow but they make good ground rules for 45 degree seat angles. 30 degree seats are another story and rather more complex to get right.

[Pumaracing]

There's another concept I might introduce which takes port work and flow bench theory up to another level altogether. That's the concept of MCA (minimum cross sectional port area) per bhp. What this tells us is to supply an engine with a given power output we need the ports to be a certain size regardless of what the valve itself flows or how big the port needs to be to maximise the valve flow as measured on a static pressure flow bench. Basically if the port is too small it becomes flow choked as gas speeds in the actual running engine rise beyond a certain level.

So you might test a head on a flow bench and enlarge the ports enough to maximise the valve flow but still need bigger ports to feed an engine in a really high state of tune which will not show up in terms of extra flow on the flow bench.

I've done quite a lot of work in this area although I don't profess to know all the ins and outs and probably never will. However I'm also damn sure very few if any people in the UK in general race engine building even know about this concept although it's somewhat understood in the USA in the Chevy race engine world. My current calculations indicate that engine power will max out at about 5 bhp for every 1 cm^2 of inlet port MCA and 7.3 bhp for every cm^2 of exhaust port MCA.

If we run those calcs for the BMW engine you mention then 8 x 25.8mm ports will have a total area of about 42 cm^2 and max out at about 210 bhp. 28mm ports would max out at about 245 bhp. You might want to see if this correlates in any way with what these engines actually produced as I'm not familiar with them.

For 370 bhp this indicates we need somewhere around 34mm diameter ports so there is a discrepancy between the claimed DTM power and the 31.5mm port size you state which on the face of it would only be enough for 310/320 bhp.

However most power claims are overstated so not too much of a surprise.

Maybe the actual DTM engines were further worked by hand and the ports enlarged over what the factory produced.

[stevesingo]

The 26mm port engines made 200bhp
The 28mm port engines made 220bhp as a 2.3 and 238 as a 2.5 with

I have seen a engine dyno plot from a 30.5 mm port head on a 2.5 engine, 320deg inlet cam 308 exhaust and 12.3:1 compression. It made 312bhp@8250 and 212lb/ft@7000.

I have a 29.5mm head on my 2.5, 292 inlet 284 exhaust 11.9:1 compression. Intake system and exhaust are good. I'm expecting 270bhp and about 200lb/ft. Is this reasonable?

Cheers

Steve

[Pumaracing]

stevesingo said:

The 26mm port engines made 200bhp
The 28mm port engines made 220bhp as a 2.3 and 238 as a 2.5 with

I have seen a engine dyno plot from a 30.5 mm port head on a 2.5 engine, 320deg inlet cam 308 exhaust and 12.3:1 compression. It made 312bhp@8250 and 212lb/ft@7000.

So not a million miles from my predictions then. It does however make one wonder where the "missing" 60 bhp for the claimed 370 bhp DTM engines went given that's about as much cam and compression ratio as anyone could usefully use in a 4v engine. I suspect in reality they were a good bit down from that.

stevesingo said:

I have a 29.5mm head on my 2.5, 292 inlet 284 exhaust 11.9:1 compression. Intake system and exhaust are good. I'm expecting 270bhp and about 200lb/ft. Is this reasonable?

Cheers

Steve

Very few engines want less exhaust cam duration than inlet and if the valve sizes are optimal then most want more so I'm not sure that's an ideal cam choice however your power expectations seem perfectly reasonable. I'd actually be hoping for rather more torque if it were one of my engines. High 80s ft lbs per litre shouldn't be impossible from a good 4v.

I don't know what use the car is for but I rarely use more than about 285 degree duration cams in 4v engines with valves that big even for competition use. All that happens past that point is a few more bhp and lots less tractability. I'd rather have a nice spread of power than something that won't run under 5000 rpm. Seat durations mean very little though. I always go by the duration at 1mm lift so I can't really comment on what cams you actually have there but I suspect I'd have probably gone for about 285 on both inlet and exhaust rather than use a bigger inlet cam.

Edited by Pumaracing on Monday 21st March 08:08

[stevesingo]

The s14 heads flow very well on the exhaust side to the point that the BMW Motorsport 29.5mm heads dont have any work done on the exhaust ports aside from the valve throat and seats are changed slightly. The cam combo I am using is a well used path with similar engines producing 230ish at the wheels.

Almost all performance builds use longer duration on the inlet and exhaust. All the BMW Motorsport cam combos are longer on the inlet.

FWIW the bore is 95mm, stroke 87mm and the cams are timed at 102 on the inlet and 106 exhaust.

We will see in a month or so when I tune it on the rollers.

Steve