1997 Mazda MX-5 Miata: Keegan Engineering Head And Porting Tech
May 22, 2012
We dropped in at Keegan Engineering recently where the impossibly genial and humble Mike Keegan took us to school on head porting. Specifically, the porting of Project Miata's VVT BP head.
There's a lot of subtlety to this craft. It's equal parts science, skill, sweat and experience. Hit the jump to see what I mean.
Who's Mike Keegan? He's Project Miata's engine builder (above, right). Prior to founding Keegan Engineering, Mike worked for years at Cosworth building Champ Car and NASCAR engines, working in their cylinder head department developing IndyCar and NASCAR heads and then as a track support specialist for Champ Car and Formula Atlantic. Built a few of the latter's engines, too, while he was at it. He's even dabbled in building performance diesels.
Today, Keegan Engineering specializes in building numerous production-based high-performance engines. The list is long, including (deep breath) Subaru EJ series; watercooled Porsche; Mitsubishi 4G63 and 4B11; Nissan VR38, SR20, VQ and RB series; Honda H22, F20 and B series, Aston Martin Vantage V8; and, of course, Mazda BP. The company has also developed proprietary CNC ported heads for Mazda BP, Nissan VR38 and Subaru EJ (and soon, FA) series.
All of the above is a long way of saying that when it comes to engine building and head porting, Mike speaks from experience. So let's hear what he has to say about porting Mazda heads.
First, the exhaust side ("roughing" work in progress shown above) of our head. Keegan has found that for NB-era Miata BP heads like this one, there are gains to be had by porting the intake and exhaust sides, but moreso on the exhaust side.
He stresses that if there's an overriding goal of porting these Mazda heads, it's to not overdo it. Simply hogging the ports out as much as possible might show large gains on the flowbench, but it ignores a key factor -- port velocity . "A lot of guys take too much material out. You want to increase the flow, but you've got to have velocity for efficient combustion." A small port will promote velocity and in turn provide the strongest charge motion possible, which helps speed the combustion process.
A properly ported head, then, retains velocity while maximizing volumetric efficiency (VE), which is goodness whether an engine is boosted or normally aspirated. Our goal with Project Miata's new mill is, again, to make the most of 91 octane, so we'll take all the VE and charge motion we can get.
Porting-wise, Keegan emphasizes that the BP's port roof is much, much more important than the port floor. "The goal here is minimize the turning of the flow, so the roof is key. I focus on the roof area between the port exit and the midpoint of the port, blending it [with the as-cast roof] as I go further." At left, the point in the stock port at which he finishes the blend.
He touches only the areas of the port that need it. "The stock BP port sides flatten out a bit, so I reshape them for a more gradual transition to the port exit."
Surface finish on the exhaust side doesn't need to be mirror smooth. "The exhaust ports all get a layer of carbon on them anyway. 180 grit or so is all you need."
Flow is important, but not if it comes at the expense of reliability. Keegan is wary about the practice of cutting the valve guides down and making them flush with the port.
Instead he leaves the guides intact at full length and smooths out the casting boss at the base of the guide (left; stock port). This retains the full support of the valve to maximize valve guide life.
Also, he's found that the VVT heads are generally much better from the factory than the early NB or NA heads in terms of casting quality and core shift. The early heads often have a 0.040-0.050" step where the port meets the valve seat that the factory doesn't bother to clean up.
As if you needed another reason to use a VVT head on your Miata hot rod.
Speaking of reliability, there's a reason you don't see fashionable razor-sharp port dividers on Keegan's heads -- thin dividers are prone to cracking. Mike instead retains a more generous radius on the divider to maintain robustness and minimize the likelihood of cracking. And Mike hasn't found any disadvantage, flow-wise, to the blunt-er dividers.
Are you beginning to get the impression that Keegan knows a thing or two about head porting?
On the intake side of Mazda BP heads, Keegan focuses on the port roof and seat transition (right). He removes only enough material to create a constant radius and to turn the flow less abruptly. In flow bench testing of his BP heads, he's found that this change alone increases flow 10-13% across the entire valve lift range.
Same approach goes for the short turn radius - he eases this transition to the valve seat and radiuses the abrupt turn Mazda placed here.
The sides of the intake ports generally don't need much more than 0.030-0.050" taken off, just enough to clean it up and remove the casting roughness.
Though the port entry is gasket-matched, Keegan avoids gross removal of the port floor and focuses efforts on the roof in an effort to straighten the port. Mazda nerds will recall that this is the approach taken by the factory when the NB Miata debuted - the new head boasted a more upright intake port with a straighter shot to the valve, increasing flow.
Keegan shows us that the stock exhaust valves in particular are shrouded by the walls of the combustion chamber and benefit from some careful work. He again cautions against getting too rowdy - removing too much material here might help valve flow but it will remove crucial squish area in the combustion chamber. To address this, he will make very small passes in the shrouded "smile" area shown at at top left. Just enough to unshroud the valves, and no more.
The procedure goes like this. He applies Dykem (layout fluid; the red stuff shown in many photos here) to the deck surface of the head and then secures it to the bare block, which has already been bored to size. Then he flips the assembly around, reaches down into each bore and scribes the projection of the bores onto the head (above, middle). This provides a visual boundary (above, right) for said unshrouding work.
We left just as the chamber work was about to start, so while the finished head isn't shown here, here's what happens next. He'll switch to a grinding stone to iron out any small remaining bumps and ridges in the ports and smooth out the combustion chambers. Finally, he'll apply a sandpaper flap wheel to achieve the desired surface finish.
At this point the head will be ready to receive a valve job to accomodate Supertech +1mm valves (more on this later). A proper valve job also promotes flow by providing a smooth transition from chamber to port (and vice versa) and further unshrouds the valves, as there's a nasty ridge in factory heads right where the seats sit.
Keegan has developed a proprietary 4-angle valve job for the BP heads, and in his soft-spoken way stresses that the angles he chooses are not simply xeroxed across all the different heads he works with -- the valve job is BP-specific. He also targets a specific and proprietary ratio of intake to exhaust flow in his heads. It's no wonder he'll spend 10-15 hours hand-porting a head.
Once the head is complete, it'll be time to do a mockup with our long-rod short block and continue the nerdery.
--Jason Kavanagh, Engineering Editor