Building a Naturally Aspirated Mazda BP Engine... Right!

In this article, we explore building more power into a Mazda BP engine without using forced induction.

For those who would rather not read the entire article, the executive summary is perhaps not surprising.  Cams, compression and tuning are the ticket, like most naturally aspirated engine builds.  However, the Mazda BP presents some unique challenges and opportunities when compared to other platforms, and these will be fleshed out in greater detail as we proceed.

Whilst the Mazda BP engine is found in numerous Mazda and non-Mazda FWD, RWD and 4WD platforms, the focus of GWR is engine building for the Mazda MX5/Miata chassis.  Regardless, most of this information will be equally applicable to all BP-engined vehicles without forced induction.

This article will be updated over time as we learn more about the BP engine.

The Mazda BP Platform

Mazda first introduced the BP engine in 1989 in the Familia/Protégé chassis.  It was not until late 1993 that Mazda began selling the MX5/Miata chassis with the BP engine.

The BP engine is a cast iron block, aluminium head, DOHC 16 valve engine design.  In standard form, the engine has an under-square 83mm bore and 85mm stroke (1839cc capacity).

The engine underwent several revisions throughout its development.  The 1994-97 version found in the NA8 MX5/Miata featured hydraulic lifters and a compression ratio of 9.4:1.  In the 1999-00 NB8A, a revised head design was introduced that included solid (shim-over-bucket) lifters, modified port profiles, and different camshaft profiles.  Compression was increased to 9.5:1.

In the NB8B MX5/Miata (produced from 2001 through to 2005), further revisions were made to the engine.  These included, most notably, the introduction of variable valve timing on the intake camshaft, and new design domed pistons resulting in a further increase in compression ratio (to 10:1). Other notable changes to the 2001-05 engines included a new, uprated high-capacity oil pump, and the addition of a main bearing support plate tying the main bearing structure of the bottom end together.



The cast-iron block of the BP engine is fundamentally a very reliable and strong design.  Originally designed for forced induction, the block has thick walls that can accept a significant overbore.  Most BP engines display suitable core stability for up to a +2mm overbore, making 1929cc achievable with 85mm pistons.  The latest BP engine blocks (from the 2001-05 MX5/Miata) often display even greater cylinder thickness, allowing further overbore in naturally aspirated applications (subject to sonic testing) – up to +2.5mm, or +3mm in rare cases.

Post-2001 blocks also enjoy the added benefit of the main bearing support plate (MBSP).  The MBSP assists the block in retaining squareness and stability under heavy loads.  Whilst not critical for most builds, any maximum effort or race-focused build should include a MBSP.  Note that a MBSP may be added to any BP engine block, but will require a 2001+ sump pan and oil pickup in order to fit.


All BP engines came from the factory with very strong forged crankshafts.  In naturally aspirated applications, these crankshafts are capable of almost anything you can throw at them.  As part of a properly balanced rotating assembly, there is no realistically achievable power ceiling for the crankshaft in naturally-aspirated applications, and they have been found reliable even in 9000+ rpm racing engines.

The one qualification to this statement is that many BP crankshafts will have cracks or flaws.  In our experience, some 40% of all crankshafts tested have included some form of crack or weakness.  These crankshafts all came from running road cars that had not been subject to unreasonable stresses.  What we have discovered is that many of these are very minor cracks.

Whilst no crack is welcome, a minor crack around the #4 bearing does not appear to be a significant problem for a BP engine, and such cranks appear to be relatively reliable in service – especially if properly balanced and not run above 200whp (150rwkw).


The cast conrods found in naturally aspirated BP engines are considered one of the weakest points of the standard engine design.  In forced induction applications, they sometimes fail at power levels as low as 250whp (186rwkw), and routinely fail above 320whp (238rwkw).  These sorts of power levels are almost unheard of in naturally aspirated applications (outside of nitrous or exotic race fuel use), but hard revving will kill them just as surely.

Any naturally aspirated engine that is expected to rev over 7500rpm should be fitted with forged conrods.  An over rev to anything over 8000rpm puts the standard conrods at risk of stretching and/or failing.

Oil Pump

The standard oil pumps found in BP engines are yet another weak point, though not as critical as that presented by the conrods.  The basic design is a rotor-type oil pump consisting of a cast oil pump housing bolted to the front of the engine over the crankshaft, and containing the front crank seal and the rotor, driven directly by the rotating crankshaft.

The weakness in the pump is created by the sintered metal rotor gears of the standard pump.  These are known to shatter under heavy use – in particular after extended racing use or in high power forced induction applications.  This generally leads to catastrophic engine failure.

There are various ways in which the risk of failure can be reduced, and possibly eliminated.  Reducing the transmission of harmonic vibration to the pump can be achieved by the use of a quality harmonic damper.  As an alternative (or in addition), billet gears may be sourced for the oil pump.  A dry sump conversion is also an option for maximum-effort builds.

Main Bearing Caps

The main bearing caps of a Mazda BP are cast, but very strong.  We are yet to see one fail.  The Mazda main bearing bolts are also very strong.

Billet main bearing caps are available for the BP, as are ARP Main Bearing studs.  Whilst the ARP studs are relatively inexpensive and may be considered insurance against the unlikely possibility of a main bearing bolt failure, billet caps are considered overkill in all but the highest power forced induction applications.


The cast pistons found in standard BP engines are relatively heavy and are not an ideal design for naturally aspirated performance applications due to their low compression design.  They are not particularly weak and should survive any real-world naturally aspirated build, but pistons are usually replaced in anything beyond a mild build in order to boost compression.

An exception to this may be engines based on a late model (2001+) bottom end, with their 10:1 pistons.  These pistons are better suited to a performance build, but any commercially-available forged high-compression piston will still see improvements – even over these late model standard pistons.

Cylinder Head Design

During the 1994-2005 run of the BP engine in the MX5/Miata chassis, three (3) different cylinder head designs were produced.  None flowed particularly well when compared to contemporary designs from other manufacturers (particularly Honda).  However, there is scope for improvement.

Certain pre-1994 FWD and 4WD BP engines were built with a head casting coded “BP26”.  Delivered with various lifter and camshaft designs (depending on application), these heads are not well suited to a naturally aspirated build, and are not particularly popular amongst serious turbo builders either.  GWR has never seriously investigated this head for a performance build.

The 1994-97 cylinder head design (cast with the code “BP05”) utilised hydraulic lifters as standard.  This head casting design is an enigma.  It is not a well-liked design by most engine builders, who berate its arguably poor port placement and profile design.  Other engine builders have taken a different view on the BP05 head, taking advantage of the extra material around the ports and undertaking extensive port work to reshape the ports into unique profiles that would not otherwise be achievable in later model heads.

The 1999-00 head design (cast with the code “BP4W”) is the most widely sought after head design by most engine builders.  With intake ports repositioned much higher in the head, the natural flow to the valve is straighter.  This is considered an advantage by these builders over the BP05 head.  The revised design also resulted in a change in intake manifold flange design, meaning a NA8 (BP05) intake manifold will not fit a later model head (or vice versa).

The 2001-05 head design (cast “BP6D”) has a port profile very similar (or identical) to the BP4W head.  It also has the same intake manifold flange pattern as that head.  However, it has the added benefit of variable valve timing on the intake camshaft.  For this reason, it presents an opportunity for top end power figures similar to the BP4W, but with greatly increased mid-range torque (depending on camshaft selection and tuning).  It is, however, more complicated to set up and tune, and as such, is not as popular with engine builders focused on a maximum horsepower number.

Both the BP4W and BP6D model heads came standard from the factory with solid lifters (shim-over-bucket) in place of the hydraulic lifters of earlier models.  These solid lifters are quite adequate for most engine builds that are not expecting revs over 7500rpm.  However, with big lift/ramp camshafts or for goals of over 7500rpm redline, shim-under-bucket lifters should be considered in order to avoid the possibility of spitting a shim.

The hydraulic lifters do not cope well with aftermarket camshafts, in particular if the base circles of those camshafts are less than 35mm.


The standard BP valvetrain is quite solid in everyday use, though there are some interesting points to note.

OEM valve springs in all models are not generally strong.  Most will, by now, be displaying reduced seat pressure and as a result, will not perform well or for long at any rpm level over 7200rpm (OEM rev limiter), or with any camshaft profile of increased lift over the OEM camshafts.

In later VVT heads, the valve springs have a tendency to crack due to harmonics in the valvetrain at any sustained rev level over 7200rpm.  In fact, harmonics have been reported in many Mazda BP valvetrains, with certain aftermarket single-spring options also displaying an occasional tendency to fail under certain conditions.  Double springs are therefore strongly recommended in any serious build.

As part of a head package that does not flow well, the valves are considered undersize on both the intake and exhaust side for performance applications.  Oversize valves of up to +1mm (intake) and +2mm (exhaust) are available but headwork is required to make these options fit.


Please note that any build beyond the mildest build will require either a remap of the factory ECU, or an aftermarket ECU/tune.

At the outset, note that improving the performance of the BP is about compression, cams, and tuning.  However, the real cork in the BP engine’s bottle is flow through the head.  Head flow, and good cam choice and tuning, is the secret to big power in a BP.

Also note that unless you have already fitted quality aftermarket headers (such as Racing Beat, Maxim Works, Maruha, etc) and a suitable 2.5”+ exhaust, together with a functional cold-air high flow intake solution, you probably aren’t ready to consider further modifications to the engine.


Once the basic external mods such as intake and exhaust are addressed, aftermarket (or regrind) camshafts present the most common next step to performance improvement for a BP engine.  They can be fitted relatively easily to most engines without having to remove the cylinder head.

There are a huge variety of aftermarket camshaft profiles available.  Some more mild cams can run using the OEM ECUs, but more significant gains will always be found with tuning to suit.

The largest, most aggressive cams will require shim-under-bucket lifters, and possibly even relief of the camshaft wells in the head casting.  We have run cams with lift numbers up to 11.7mm (intake), though cams of this nature present unique tuning difficulties and generally are not streetable.

Stock valvetrains will be quickly hammered into submission by big cams, and aftermarket valve springs and retainers are also recommended.  Various offerings (notably from Eibach and Supertech) are available, in both single and dual spring options with varying seat pressure.

Cam choice, and cam tuning, is a black art and has a huge impact on the power output and delivery of the BP engine.  Adjustable cam gears should be fitted and you will do well to find a tuner who understands how to adjust cam overlap to yield the power and torque outcomes you are seeking.


As noted, earlier model BP engines are constrained by low compression.  Naturally aspirated performance is massively improved by increased compression.  With the use of 98+ octane fuels (modern premium unleaded), compression ratios of 11:1 or 11.5:1 can be safely run without fear of detonation (provided the tune is safe).

There are a huge variety of aftermarket high-compression pistons available, with ratios of 10.5:1 up to 12:1.  These range from cast offerings available on eBay, through to cheaper forged units such as Supertech and JP, to CP and Wiseco, with Mahle and Toda being the most expensive off-the-shelf pistons available.  Most of these aftermarket pistons are only offered in plus sizes, making an overbore necessary.

GWR has used Supertech, CP, Wiseco and Toda pistons.  All were excellent pistons, but each had unique characteristics that affected other aspects of the engine build (such as piston-to-bore and valve timing).

OEM NB8B (2001+) 10:1 pistons offer a cheap upgrade option for those NA8/NB8A owners on a budget.

Head Porting

The head flow in all BP engines is an inherent weakness of the design and must be improved to yield significant power gains, especially in naturally aspirated form.

Once cams/pistons have been sorted, head porting becomes necessary to yield significant additional power gains.

The OEM valves in a BP measure 33mm (intake) and 28mm (exhaust) in diameter.  Oversize valves are available, and moving to +1mm intake and exhaust valves (with accompanying headwork) can yield significant gains, and also amplify the benefits of bigger cams and higher compression.

More substantial work to the ports themselves, on both the intake and exhaust side, will yield further gains.

We have seen a number of race-prepared heads from four separate reputable engine builders across the US and Australia.  There are widely varying theories on how to yield the best flow outcomes for these heads.  Some prefer the BP05 heads, others prefer the BP4W/BP6D heads.  Some aim for greater peak flow, others compromise flow for increased velocity.  Some hand port, others use CNC machines.  All have worked to some degree, but some are better than others.

Engine Balancing

The BP engine has a tendency to vibrate.  Though tough, it is not an inherently well balanced design.

Engine balancing, done right for the entire rotating assembly, can yield significant power gains.  This balancing should include the entire rotating assembly – crank, conrods, pistons, flywheel and clutch pressure plate.


Whilst few aspirated builds will trouble the OEM conrods with their power output, the high revs generally required to get the best out of an aspirated build will require forged conrods.

A wide variety of options are available, from cheap Chinese rods (rebadged variously as Eagle, Manley, etc), to the more expensive Crower, Pauter, Maruha, and Carrillo (in A, SA and H beam) offerings.  All should work.  Some are lighter than others (and therefore offer possible increases in power).  Some Chinese rods suffer from QC issues but reports of failure are rare.  Except in cases where crankshafts have broken first, we have yet to see a forged conrod fail in a Mazda BP – though we have had big end caps badly damaged by spun bearings.



Note that the OEM throttle body on all BP engines is notoriously weak and prone to fail, spitting screws into the combustion chamber and destroying the engine.  It is recommended that any serious build consider a “Skunk2” throttle body retrofit as a minimum, to avoid this possibility.

There are various improvement options available on the intake side of a BP engine (beyond a basic cold-air solution).

Firstly, the OEM “square top” NB8B intake manifold (non-USDM models only) presents a 3-5% power gain potential over the NB8A manifold.  Together with a 64mm “Skunk2” throttle body (plug and play), these gains are possibly higher.

Aftermarket manifolds, incorporating longer intake runners, have also yielded improved outcomes.  The B16 Honda intake is a favourite, as is the Skunk2 (or replica) design for the same engine.  Use of these manifolds on a BP is not easy and requires significant fabrication skills and the adaption of a suitable throttle body and TPS, as well as tuning.

From here, Individual-runner throttle bodies (aka IRTB or ITB) are popular amongst racers for their improved throttle response.  Some claim improved top-end power from ITB-equipped engines, but our experience is that top end power is not hugely affected – on-off throttle response is where ITBs shine over single throttle bodies and in racing applications, this is critical.  42-45mm ITBs are preferred, though some have employed ITBs of up to 50mm.

Oil Pump

As noted above, oil pumps are a liability in BP engines.  Billet replacement pumps/gears are available.

However, to prevent oil surge, Accusump kits or dry sump conversions are available.


Though not specifically an engine part, strong spark is critical to a successful build.

No BP engine has particularly strong coils, and some of the designs are known to be weak or fail regularly (in particular those from later model engines).

Chev LS coils are a popular replacement option, as are Toyota COPS.  Other options (such as Honda motorcycle COPS) have also been used. All of these options require rewiring of the ignition circuits, with some also requiring external ignitors.


Various internal engine coatings are available and, applied properly to the appropriate parts, can increase power and reduce detonation potential.

In addition to the common application of ceramic coating exhaust headers, these include piston skirt and bearing coatings (to reduce friction), and piston crown/combustion chamber coatings (to reduce heat transfer and increase combustion temperature).

Coatings can also be applied to other parts such as intake manifolds, intake and exhaust ports, crankshafts, etc..


Stroker Kits

There are two commercially available stroker kits available for the BP engine (Maruha and Flyin Miata).  Each have strengths and weaknesses relative to the other.  Both are expensive ($5,000USD+), but will yield power gains.  We have not experimented with stroking a BP (primarily due to rule compliance in our classes here in Australia), but empirical evidence suggests power gains are definitely possible.  To what degree these gains exceed large overbores is debateable, and as such, the financial viability of stroking must be considered in such a build.

Cheaper options are available to stroke the engine, in the form of a welded and reground stock crank, or an offset ground crank.  Longer rods and custom pistons would be recommended in such a build.  We have plans and options prepared for such a build and would be happy to progress for an interested customer.

Long rod conversion

GWR has experimented with an aggressive, long-rod conversion for a BP engine.

Standard BP conrods measure 133mm in length (centre-to-centre) which, when divided by the standard BP stroke of 85mm, gives a rod ratio of 1.56.  This is well below the "ideal" rod ratio (around 1.75-1.8) considered necessary for a free-revving engine design - as a comparison, the Honda F20c employs a rod ratio of 1.82.

All else being equal, longer conrods reduce the piston acceleration and decelleration rates at the points near top dead centre, allowing for a longer and more gradual "push" on the piston after ignition.  They also reduce the side load on the pistons in the bore as they rock around top dead centre and bottom dead centre.

The build required custom pistons and used off-the-shelf conrods from another application. Problems were encountered on this build (primarily related to problems in piston crown design, which was stuffed up by our piston manufacturer), but we are confident these issues could be solved in a future build.