A Basic head porting guide.

[4v6]

Cylinder head modification, a basic how to. Part1.


The purpose of this article is to help the enthusiast who wishes to modify their engine, achieve a better understanding of what too look for in a modified head and suggests methods to achieve it.
The heads being used in this article belong to the VW/Audi stable but the same processes apply to most all heads.

Our first item to attend to is to thoroughly clean the head to be worked on, this allows us to see any issues and also to make for an easier to handle item.

Once we have it cleaned we can then do an examination and look for any defects that might be present, such as cracks in the head or valves that would need remedial treatment or replacement.
We can then proceed to stripping the valve gear, cams, valves, springs etc to render the head free of any items that might otherwise get in the way of testing, it also allows us to further inspect each of the items for any defects that might create an expensive failure later on, better to deal with it at this stage than when its all reassembled.
Once we have the head entirely stripped of any mechanical parts we can proceed to deburr all sharp edges to prevent cuts when handling and also to remove any hardened deposits that can accumulate around the coolant transfer ports on the chamberside, we can then continue with a few flow tests.

Deburring.



Its important to do this to gain a base starting point to gauge improvements against, something you simply cant do without access to testing equipment.
The use of such equipment however basic gives you the ability to judge what you had vs what you now have, graphs, figures for flow at different lifts etc in the form of a report, something which gives provenance rather than uninformed guesswork.

Multiple runs with slight seat modifications.




The head needs a little further prepwork before we get onto the flowtesting stage as we need to block off any points that might leak air past the ports being tested.
In the case of intake ports which will be tested under vacuum conditions, we first need to tape over the exhaust ports ( duct tape also sticks well to clean, non oily surfaces hence another reason to clean it all first) and blank off the spark plug locations with a spark plug or plasticine and the open valve guides similarly.
Once we have our head so prepared we can then mount it to the flowtesting fixture which is basically a pipe of roughly bore size and about 4 inches high that we can clamp the head to, which has two flat mounting surfaces and a sealing pad of rubber fixed to it.
With the head mounted up we then have to use either a commercially made entry radius that bolts in place of the intake manifold or the poor mans equivalent- plasticine rolled into a doughnut shape about a quarter to half an inch thick and set around the port entry to act as a guide.
Its vital to have some form of entry radius as without it the air flow results will be unreliable due to a phenomenon known as edge entry loss.
Its basically the air being forced to negotiate a 90 degree bend as it travels into the port which pinches off air flow and causes turbulence which in turn skews the results.

Poor mans entry radius. Blocked runner is to allow testing on one side only.



Head mounted up with opening fixture and dti gauge setup for generating flowgraphs.



With the head now ready we can set up our flowtest equipment (varies depending on method so wont be covered here) and determine what the flow rate through the intake port is.
As we have no valves inserted we are running a "bare port flowtest", its vital to appreciate the reason for running it like this firstly.
That reason being that with no valves installed we are testing the port at its most unrestricted, ie, what the head can flow as a maximum with no valves.
Valves always restrict flow due to their nature but may cease to do so if lifted high enough or if the design of them is conducive.
Once we have recorded our bare port flow test we can proceed to do exactly the same on all the other ports.
At the end youll have a set of flow figures to look at, notice that theyre likely to be all different depending on factors such as surface finish, carbon buildup or just casting differences when the head was produced.
Conduct the same set of tests on the exhaust ports only this time the airflow direction is reversed, blowing out of them with the intake ports being taped over as earlier alluded to.

Mazda 323 turbo intake on test. Air is being blown through the throttle body mounting and out the port runners one by one.



This is also a Mazda turbo manifold, its a two piece item and was a lot easier to work on and improve as access was more open.



Before attempting to modify a head I prefer to practice and test out on a scrap item, it allows you to gauge things such as how thick the port material is and whether or not you might break through into a coolant gallery, you can also test out modifications to valve seat angles which can give excellent results if theyre non optimal.
In some instances modifications to valve seat geometries can yield equal or better results than from portwork alone, combining improvements from both can really make a head modified in such ways an excellent performer.

Other items of interest to note are the intake and exhaust manifolds being used.
In testing an intake manifold, air can be either blown or subjected to vacuum to generate a flow figure for each runner and wide differences in flow can sometimes be found.
Testing an intake manifold is very similar to testing a head, the ports not under test need to be blanked so that air is only passing through the tested runner from the throttle body mounting position.


In many cases you will find the flow to be marginally above or below that of the head, obviously modifying the head then means the intake becomes restrictive which can represent an issue due to most manifolds being almost impossible to alter internally because of the shape of the runners, 90 degree changes in direction etc meaning tools cant be utilised to open and relieve them.
In such cases it might be possible to cut and reweld the manifolds, or have them extrude honed.
Another alternative is to build a bespoke replacement that can accomplish the task which will also need testing to ensure its perfomance exceeds that of the head.
Exhaust manifolds can be similarly tested and evaluated.
There is another issue regarding intake and exhaust manifolds whereby the manifold/s can exceed the flow of the modified head and yet when bolted up and tested with the head you end up with a flow loss.
Usually thats caused by the manifold directing the air into a direction it dosent want to go so attention must be given to ensure you minimize such issues.

After bare port flow testing we can then turn our attention to the testing of the ports with their valves fitted and generate useful flowtest data that shows the behaviour of the ports and valves and their effects across the entire lift curve.
Incidentally, once youve got the flow figures for the intake and exhaust ports you can then use that data to get an idea of the intake to exhaust flow ratio, known as E/I ratio.
For example intake flows 200cfm and exhaust flows 170cfm.
Plugging the numbers, exhaust / intake flow =0.85 or 85% which is more than enough exhaust flow.
Usually around 80% is enough and going higher makes no difference.
In the case above you could leave the exhaust ports alone and just modify the intakes, thatd bring that 85% number down to whatever, maybe 75% etc, in that case you would then modify the exhaust port to bring the E/I ratio back up to 80%, just remember though intakes are generally harder to modify for flow than exhausts which is why you modify the intakes first.
Its only necessary to test one intake and exhaust port with the valves in because thats our test port to which we will apply our modifications and which when completed we then recreationte as closely as possible on the remaining ports.
Using a fixture that allows the valves to be opened a precise amount at a time, (I usually go 1mm per measurement) and a dial test gauge to check the opening is accurate the valves being tested (intake or exhaust) are lifted and flow readings taken.
The action generates a graph on my flowtest equipment and tabulated data which is easily printed or compared.
Its important to note the test pressure your head is being flowed at, for example, 10, 25 or 28 inches of water, its not a lot of use to be told your head flows 225 cfm if you dont know what test pressure it was done at.
So with the intakes or exhaust flowtested we can set about the modifying task, this is where it gets interesting.

The temptation to dive in and start chopping metal out is a strong one, but we need to do other things before we even think about that.
We can get some excellent results simply from making small adjustments where possible to the valve seat architechure for minimal effort once the basic testing is out the way.

Youve likely heard the reference to a 3 angle valve seat or one with even more angles say 5 or so.
Referring to the head on test here we know it has a 45 degree seat angle, a lower 75 degree angle and a crude 30 top angle cut into the alloy, plus theres a ridge where the two materials of the seat insert and the head meet.
Examining the intake valve, we can see that its seat doesnt actually extend all the way out to outer periphery of the valve, its only a small amount, around a half mm or so but its clear that the whole diameter of the valve isnt being utilized fully, so one of the first things to try is to ensure we use that extra width to get ourselves some extra flow capability.

Showing the valve seat that has receded but alos illustrates the wasteage of diameter at the periphery.




Going on the principle that an intake seat gives best flow vs robustness at 4.5% of the valve diameter, we can see that a 38mm intake valve should have a seat width of 1.71mm which we have near as damn it from the maker at 1.70mm.
Our job so far is a simple one, expand the seat insert in the head with a carbide cutter to measure 38mm or whatever valve diameter we're working with at its outer periphery so it will match the seat on the valve.
Before we do that we can measure the throat of the port, the area just below the lower angle which is usually parallel bored.
A measurement of 86% of the intake valve diameter will usually do a good job of providing near optimal flow and velocity, going much larger will cause a slower speed port.
The makers pretty much spot on again at 33mm so we need do nothing to this area.
The lower 75 degree angle on this head can be improved upon by using a 70 degree bottom cut.
Applying a 60 degree cut improves it over standard but i feel 70 will likely be closer to optimal.
The top cut should lead out smoothly into the chamber as ridges and obstructions such as exist on this head in stock form all impact on flow.

Completed valve seat with changes tested out on the flowtester.





This flowtest shows a stock vs modified seat. The dip below stock is caused by the lip generated when modifying the valve seat, which heavily intrudes on flow.



With that out of the way the intake valve itself can be altered.
We've all heard of "backcuts" applied to valves, its essentially another angle applied to the back of the valve which removes excess material that can obstruct flow when the valves are at low lifts.
Its important to ensure whatever valve combination you use that the seats in the head and their angles are matched ( using the flowtester) to the valve's seat angle combination.

Backcut valve on the left has material removed up to the inboard side of the seat, compare with stock on right.



This particular head suffers terribly if an otherwise stock port and valve seat are used with a backcut valve as the port develops an oscillation and whistles extremely loudly which kills off flow.

This flowtest shows the damage such oscillation can do to flow.




Part 2 will deal with actual flowtests and modifications on a multivalve head as they are carried out along with modifications to the intake manifold.

[Andy10v]

Good read.

Is the one blocked runner when testing the flow purely for the valve flow bit so you can concentrate on 1 valve/seat etc?

When you do the port is it done as a pair (as that is how it operates?)?

[4v6]

Hi Andy, well i was blocking off one side to try and find what shape produces the best result, its harder to know that if the other ports bleeding air down it.
What i did find was that youd be able to generate quite good figures just down one side, say for arguments sake 100cfm , then youd modify the other side and get another 100cfm so youd think youd have 200cfm in total.
Unfortunately it seems theres perhaps some turbulence that chokes off that desirable extra flow because its always been lower when you test with both sides open, which as you say is how it operates.
Its probably the impact of the air interfering with each flow and bottling up somewhat.
I generally close off the port simply to get the best possible flow down each side and keep them balanced as closely as i can, that being the other reason for doing them that way.

Now, if i had a nice Centroid cnc machining centre........

[Andy10v]

Ok, got you, balancing makes sense, I had not considered that, just them operating as a pair and effecting each other.

Continue with your witchcraft

[20vcqdriver]

Love your idea of 'basic'
Love to know what you call 'in depth'
Look forward to seeing the progress on this

[missfire]

Still waiting for a chance to read this! Might get to it tonight, if I'm lucky..

[4v6]

Originally Posted by 20vcqdriver

Love your idea of 'basic'
Love to know what you call 'in depth'
Look forward to seeing the progress on this

Well it is basic Dave, lets face it, theres folks out there who do this kind of thing and make me look well, basic with it.
I can accept that so i just do the best i can.


Progress wise, should have a new vacuum source this thursday so i can get into flowtesting proper.
So far the heads been stripped of parts, cleaned and deburred.
Stripped it all off on christmas eve as i was bored....
Should make some fairly good progress before too long.