Introduction

I am going to refer to the K-series engine from Mazda as the KL-series as this avoids confusion with the legendary Honda K-series engine. The main purpose of this article is to capture historical knowledge from forums that were active in the the early 2000's and capture information on how the KL engine can be swapped into the MX-5 NA and NB chassis. It has always been a more DIY affair, with only a small facebook group that as of 2022 is still active. 

With the KL, wonderful engineering has been overlooked and misunderstood, and this article is a love letter to a hidden gem of internal combustion. This engine was designed during the golden age (90’s) of Japanese performance cars, and while it has been disregarded because of its lack of power, it makes up for it through it’s short stroke architecture which was designed for "acceleration and top-end feel" and it’s emphasis on creating a subjectively beautiful engine note.

The KL series engine range in displacement from 1800cc to 2500cc. They employ a short stroke, 60° V-shaped 24 valve DOHC architecture with an aluminum cylinder block. They featured a 27-degree DOHC valve-train with directly actuated hydraulic (HLA) bucket lifters, which was later changed to solid lifters on the last version of the engine the KL-G4. It has 4 & 6 bolts for each main journal, an internally balanced forged crankshaft and long, lightweight forged carbon steel connecting rods.

Identifying the differences between the DE and the ZE engines can be difficult, use the stamps on the heads to make a clear distinction:

KLZE = KL31-101 & KL31-1A1

KLDE = KL-101 & KL-1A1

The KL31 head is best for those who do not want to touch the cylinder head as they provide the greatest flow from the factory. KL-ZE heads worked best with very minimal porting and with port matching. However KL-DE heads can be made to flow better than KL-ZE heads because of the differences in port design and available material for porting work.

The best combination of parts out of the box is high compression bottom end using the KL-ZE pistons, matched to the KLDE heads with un-modified KLG4 intake manifold and KLZE cams with KL-G4 valve springs and solid lifters.

The technical article on the KL engine development by Takashi Sakono, Shinobu Takizawa, Setsuo Harada, Tatsuji Ikeda, and Hiroshi Abe goes into detail of other design innovations of the engine, and one thing I will skip over is the combustion design and variable resonance induction system that underpinned linear torque delivery and increased engine efficiency. For the purposes of my build that employs individual throttle bodies, I won’t expand more on this except to say that model codes that ended in ZE typically had freer flowing cylinder head and manifold designs and were documented to have more power from Mazda itself. 

Rod Ratio and Engine Feel

In the technical paper detailing the development of the F20C, Yutaka Otobe, Hiroshi Kawaguchi and Hideo Ueshima conducted tests on a long rod and stroke version of the H22A Honda engine that demonstrated that frictional losses increase exponentially at higher rpm. One reason is that the increased angle of the rod through the power stroke creates more lateral force where the piston gets pushed against the cylinder wall. Rod angularity is the main cause of piston side loading

In addition, in this short rod scenario, there comes a point of no return where the mechanical advantage of increased angularity near TDC is overcome by the fact the short rod jerks the piston away from TDC so quickly that the piston is actually accelerating away from the expanding flame front instead of the flame front building pressure on the piston dome, and average cylinder pressure (brake mean effective cylinder pressure) heads downward. So the torque drops off at higher RPM.

There is more torque at higher RPM with a long rod configuration because it holds the piston near TDC for more degrees of crankshaft travel. So, ultimately the piston stays near TDC for a longer absolute time at any given rpm with the longer rod. This gives the flame front more time to totally ignite the A/F mixture so that when the crank pin finally heads well down on the power stroke there is higher combustion pressure on the piston crown to move the crank. Thus, more torque at higher rpm with the long rod.

The KL-ZE 2.5L engine which I will be focusing on has a rod ratio of 1.86:1 which is higher than the Honda F20C’s 1.82:1 which was the highest revving mass produced car motor for some time. Interestingly, the engine stroke of 74.2mm is even shorter than the short 77.4mm stroke found in the Honda B16B engine in the EG6 SIR. The KL-ZE 2.5L V6 engine in the Mazda has a rod ratio of 1.86:1 which is extremely high for a production engine. The short stroke and long rod ratio of the KL engine will allow a sensation of acceleration as RPM increases, and combined with more aggressive cams, oiling upgrades and valve-train modification allows a reliable 8,000rpm rev limit.

While variable valve timing would have been nice, there are three key advantages of this engine without it. The first is that oil supply will be more consistent at higher RPMs, the second is that I can run more aggressive camshafts without worrying about interference and finally overall tuning and reliability can be improved as the engine build focuses on tighter tolerences. 

Engine and Component Weight 

The weight of the engine with no alternator, no powersteering, no inlet or exhaust manifolds, no clutch, no oil, no wiring loom, no flywheel and no intake manifold, is 105.3kg.

My plan is to go with the 1.8 stock flywheel and throttlebodies which will will bring the weight to 117.8kg dry. There is a marginal weight saving of a couple of kilograms, however the weight that remains moves the weight behind the front axle. Coupled with my plan to remove the airconditioning system, this should retain and enhance the driving characteristics of the car.

Here are some verified weights to consider:

Bare engine: 105.3kg
KL 2.5 Stock Flywheel: 12.7kg 

KL-ZE 2.5 Flywheel : 10.4 kg
K8 1.8 Stock Flywheel: 7.7kg
K8-ZE 1.8 Flywheel : 7.6 kg 

Unorthodox Racing Flywheel: 4.9 kg  

Fidanza Flywheel: 4.3 kg 
Cast Intake Manifold: 12kg
Throttle Bodies: 4.8kg

Engine Friction - The Fun Killer 

As we have seen, friction stands in the way of a free revving engine. Using the diagrams above I hope to achieve the following with the engine build itself.

- Reducing the size of the piston skirt
- Removing the airconditioning pump
- Removing the power-steering pump (switch to electric)
- WPC treatment everywhere possible

Camshaft Options 

While there have been builds chasing even higher RPMs, my personal feeling is that cam selection moves the torque curve to the point where daily drivability, emissions and fuel economy are too compromised. In addition, higher RPM applications on other motors usually employ variable valve timing or variable valve lift that allows more aggressive profiles to be employed. Since the KL series engine has neither, using cams that are slightly more aggressive than OEM will likely suit the engine better. Here are the of the shelf camshaft options for the KL-DE/ZE:

Block Architecture and Rigidity

Moving back to the engine block, the split case two-piece cylinder block is designed around rigidity. In the upper block, a 3mm-thick cast-iron cylinder liner is added for durability. The lower block was given a ladder frame construction integrating the main-bearing cap and bearing beam. Contrary to popular belief, this was not done for performance purposes (although the ladder frame construction suggesting that), but rather in pursuit of a pleasant sound by eliminating unwanted sound frequencies. 

The study above shows the resonance differences between the single piece casting vs the split case design that was chosen. After the Finite Element Method (FEM) analysis was completed, there was a significant increase in the natural frequency of the block itself that reduced unwanted low end frequency noises from the engine. Here are some pictures of 3 different castings on the engine block, all external dimensions are the same. 

Another area of sound that was addressed was unpleasant rumbling noises caused by crankshaft bending vibration due to flywheel face runout (slight distortions on the flywheel face). This vibration is propagated through the cylinder block main bearing and into the car resulting in an unpleasant rumbling interior noise. To address this, the lower deck's No.4 journal near the flywheel was made wider than the other journals. In addition, a forged steel crankshaft was used to increase crankshaft bending rigidity. While the KLG4 has the lightest crank (-4 lbs) compared to other versions, however it is not forged but instead cast. 

The rotating inertia weight of the flywheel, crankshaft, and connecting rod was drastically lowered by making full use of FEM analysis and acceleration response was improved. The die-cast aluminum short-skirt pistons were designed for lightweight and continuous high-speed operation. All of these developments reduced the second-order vibration level which caused most low-frequency noise. In the clip below, you can hear how exotic the motor sounds. I personally feel that it is the best sounding V6 in the world, even tipping the legendary Busso V6. 

Packaging

Another key advantage of the KL series engines lie in its low weight and modularity. Despite the differences in capacity, all engines weigh approximately the same, and are somewhat dimensionally identical (620mm(l)x675mm(w)x640mm(h)), which allows for mixing and matching of some components. The KL engine had the most compact packaging size in its displacement class in the time of its release. 

V6 Mazda Miata Prototypes

Mazda did toy with the idea of V6 miata, both pre-ford ownership and post ford ownership. Under the now defunct M2 Incorporated brand which was a kind of R&D and exclusive arm of Mazda, they developed the M2 1006, also known as the Cobraster. That name was probably directly related to the Shelby Cobra, which pioneered the big engine in the small roadster concept. It was powered by the 929’s 3.0L JE-ZE V6 (tuned to produce 220PS) and rode on suspension components from the RX-7. Due to cost and complexity of production the project was canceled, but the prototype still exists today. There was a second prototype, this time based on the NB chassis that was rumored to use the 3.0L Ford Duratec aluminum engine from the late 90’s. The objective of this Advanced Powertrain Development project was to provide Ford/Mazda with a working prototype that could be put into production using as many parts from existing Ford/Mazda parts bins as possible, and it used the FD RX7 transmission and differential. That car made it to testing on a race track, but went no further. 

Cylinder Heads

The KL31 head is best for those who do not want to touch the cylinder head as they provide the greatest flow from the factory. KL-DE heads are the best to work on as they have the most material to work with and better port shapes. With the right work, they will outflow KL-ZE heads because of the inefficiency in the KL-ZE port shape. The KLZE larger ports have less material around them to work with before you get into the danger zones with too thin of walls.

KL-DE Measured flow numbers (use as a guide as source not reliable):

KL-DE Stock 

176 cfm intake

159 cfm exhaust

KL-DE with porting

185 cfm intake

170 cfm exhaust

KL-ZE head flow numbers (use as a guide as no reliable source):

Head Porting Guide 

1. Raise the roof as much as possible with DE ports, beware thin at the port roof. 

2. Remove the port divider as possible until about 1” from valves. 

3. Ensure port divider is sharp and made as even as possible 

4. As the port approaches the valve it either is even or decreasing in volume 

5. Light de-shrouding and making it even 

6. Raising the port roof should be done sparingly as it is very thin. 

7. For exhaust port, only clean up the divider slightly. 

8. Matching all the port shapes is key to making power N/A. 

9. Do not make too drastic of a pressure change in the flow in one singluar area 

Valvetrain 

Going back to the KL, there are several relatively minor weaknesses that have been identified with the motor, each which can be rectified with some time and understanding. The first is a spring defect on some ZE version cylinder heads which have two likely causes. The springs being softer allows for more valve float which could allow more slapping of the camshaft against the lifter. This smacking from the top as well as the smacking from the spring on the underside during valve float periods is very stressful on the retainers. There is a possibility that Mazda wanted to keep the HLA’s quieter, but there was also a manufacturing problem that left burrs on the ends of some of the valve springs. Mazda also used hydraulic lash adjusters primarily to reduce the maintenance requirements for the engine, rather than for performance reasons. 

Mazda addressed this with the KL-G4, the final variant of the engine produced from 1997-2002. It has Solid Lash Adjusters (SLA) vs. the Hydraulic Lash Adjusters (HLA) of the standard KL motor. The part number is KLG412183. Take note that these are shim above bucket lifters and not suitable for applications beyond 7,800rpm or so.

This change was made by the engineers to reduce problems inherent to HLA's when the tiny oil ports would become clogged. HLA noise could also arise due to pressure oscillation rocking the HLAs in their bores as they are fed from an oil hole on one side of the bore.

Since the SLA's do not put pressure on the no-lift area of the bottom of the lobes then the area does not need to be as substantial as with HLAs that put pressure on the camshaft on the bottom of the lobes. There is actually a small gap at zero lift. This is called your valve lash, unlike hydraulic ones, solid lash adjusters have to be clearanced occasionally.

A solution for even higher RPM applications is to get 12 x 33mm lifter buckets for the intake and 12 30 mm lifter buckets for the exhaust and remove the inner piston and install the inner shims from Maruha motors which does the same conversion for the Miata from HLA to SLAs.

KL03 HLA size (same for intake and exhaust)

Overall Diameter- 1.180" or 29.97mm (this is the 30mm HLA)

Diameter at Oil Relief- 1.119" or 28.42mm

Oil Relief Depth- 0.030" or 0.77mm

Overall Height- 0.948" or 24.08mm

KLZE HLA size (intake, exhaust is the same as KL03 sizes above)

Overall Diameter- 1.298" or 32.96mm (this is the 33mm HLA)

Diameter at Oil Relief- 1.235" or 31.37mm

Oil Relief Depth- 0.031" or 0.80mm

Overall Height- 1.021" or 25.93mm

KLDE springs/retainers/locks can be used in the KLZE even though the spring and retainer diameter is smaller. The KLZE exhaust springs/retainers/locks are the same as KLDE. But the KLZE intake springs/retainers/locks are larger but weaker - hence the dropped valve/broken retainer problem. For an OEM solution, valves and valvelocks are the same between both ZE and KL03 heads.

There are two companies that provide aftermarket solutions to this, Ferrera and Interprep. There is a massive price difference between them, and for the purposes of my build with its 8,000rpm redline, the Interprep solution was the obvious choice. Here is some data on valvesprings to use as a guide. 

Valve Spring Data  

KLZE has 40lb seat pressure, 140lb open pressure

KLDE has 50lb seat pressure, 150lb open pressure

Interprep 65lb seat pressure, 195lb open pressure

Ferrea 75lb seat pressure, 195lb open pressure

Mark Snell, who built one of the first documented KL engines miata discovered a valve and spring set that are a swap for the KL pieces and will allow 11mm of lift from the Toyota 7MG eninge. The valves are identical in head and stem diameter, and longer; and aftermarket spring/retainer (titanium) kits are available. With a careful spec of a cam base circle the could be used.

Recent developments found that much of the valve float problem has been caused by the spring harmonics, not spring pressure. Previously the solution to valve float was to throw the motor a heavier valve spring.  One theory is that the springs reaching a certain resonant frequency is the cause of much of the problems all engines have with valve float. When that frequency is reached, the spring loses its ability to hold the valve against the seat or cam lobe. Resulting in bashing of the valvetrain breaks parts like springs, retainers and locks. What many V8 builders have found is the use of ovoid(egg shaped) spring wire reduces the tendency of the spring to resonate.

More revolutionary but unusual looking is a new, old spring design. what has been found is that the use of a "bee hive" springs design can really reduce the onset of valve float. Because the diameter of "bee hive" springs changes from the spring seat(bigger) to the retainer end (smaller) the spring(which is shaped like a bee's hive) tends to not reach that critical resonant frequency.  As a side benefit, the retainer end of the spring is smaller, resulting in a smaller, stronger and lighter retainer, which increased rpm potential.

The Spun No.6 Bearing 

The second is a common big end bearing failure on cylinder no. 6, also known as a spun bearing. The root cause of this is more varied, but this can be addressed in three ways. One method could be to increase the size of the crankshaft oil journal to supply more oil to the bearing. The second would be to use higher quality bearings and also check the tolerances of the bearings and cam clearances (extreme things guide) very carefully when rebuilding the engine. The final solution would be to invest in a boundary stage 2 oil pump gear that could help flow (no data to support this though). Good maintenance practices and watching oil temperatures on a track is a must in any case, and will help prolong the life of any engine. 

Worn oil pumps and poor clearances should be the first area to address in an oiling system. KL oil starvation comes from pumps that have become sloppy because of how the impellers ride against the aluminum sealing surface. Many old KL pumps and the aluminum face that seals against the impeller -you will find gouged areas which allow oil pressure to slip by. This can be caused by contaminated oil or from over-exceeding the pump's capacity. Oil contamination can be from filtration, bearings spinning or failure to change oil/filter often enough. Most of the time -it is caused by the latter. If clearances are not right -correct them whatever it takes -because if you don't -it won't matter how much money you put into pumps, filtration or block/crank modifications. And of course -regular oil/filter changes should be a given.

Mazda removed the oil squirters all together in the late 96-97 period, and this technical update was applied to engines manufactured after that time. Alistair Oag, owner of the respected Mazda KL tuning company Interpret confirmed that oil squirters are not necessary, and that blocking off the squirters from original engine helped to reduce the probability of oil starvation on bearings #5 and #6 due to increased pressure. He also advocated boring out the oil pump, making the oil galleys non-tapered and adding a radius to the tight turns improved the flow and pressure of the oiling system and oil cooler. 

Shimming the pump does nothing for pressure or volume if you do not have correct clearances. It is a max pressure bypass valve and is set at a given release pressure to avoid too much pressure. This is not a bad thing because too much pressure with these pumps can be a bad thing. You need more volume + pressure for high performance and not just pressure. Low volume-high pressure can diminish very quickly with slightly over-tolerance clearances -whereas high volume-standard pressure is less likely to diminish.

It is interesting to note that the JGTC and BTCC cars seemed to have a redline at around 8,500rpm with a short stroke motor, but there was a guy on Proble talk who built a dry sumped KL for his dune buggy that made just under 300hp at 9,000rpm NA. The video tells the story, but there is very little information on how it was done. Here is his engine spec:

• KL-DE Bored 0.060 with 10.7:1 diamond pistons (thanks to jmmdm2 ) 

• H beam rods ( thanks to rurockn ) 

• Ferrea +1mm valves, springs and retainers 

• Combustion chambers unshrouded to larger bore 

• Inlet ports filled and ported to match itb's 

• Exhaust ported 

• Peterson 3 stage oil pump 

• Oil galleries enlarged and intersections radiused 

• S2000 Honda injectors

• MS3 extra with canbus 6x EGT probes

Friction Gear Maintenance

The third and final one is that the friction gear (see diagram below) would sometimes create a ticking noise. Between the two camshafts are drive and driven gears with 55 teeth each and a friction gear with 56 teeth. The friction gear, superimposed on the driven-gear by spring force, was designed to be free from backlash with its extra tooth. The friction resulting from this structure absorbs fluctuations in drive-gear rotation, effectively suppressing gear rattle noise. Here is a guide pulled from a forum post that helps address this issue. 

There was a technical service bulliten that a noise emerging as loud ticking heard specifically from the top-left of the engine (rear V6 bank, near cylinder #1) and oscillatory in frequency over ~3 seconds (beats). The noise is most noticeable on cold-starts and stops past 2k-rpm thus differentiating it from carbon-knock or HLA noise.

The exhaust-cam friction-gear-spring is designed to prevent backlash through the helical cut gear at low rpm. At higher rpm (>2k) backlash isn't a problem and it isn't required, the noise stops. The problem is that the friction-gear can only eliminate backlash over a certain range - excessive HLA lash (dirty oil/HLAs), hunting idle or slight miss-fire will alter the backlash through the cam sufficiently to move it outside the ability of the friction gear to compensate, thus it becomes audible. The noise is simply annoying, the device being fitted for aesthetics and not a critical wear issue. 

The only solution is a new friction-gear for the exhaust-cam, it can be done without doing the timing belt. Friction Gear noise is often HLA noise, and is not a problem requiring remedy - it's purpose is aesthetic to achieve a totally silent V6 engine. Mazda corrected this in later engines by adding two tabs which fit into slots in the retaining nut which is in the second picture below. 

Crank Dampner 

Although there has not been documentation of this happening, at high rpms, resonance in the crank has the potential to crack the cast oil pump rotors. This same issue has been shown to be an issue in high output versions of other crank driven oil pump motors, like the LSx series from GM. Crank dampners are also critical to absorbing the vibrations generated by the engine, and they are a wear item that needs to be replaced as well. For peace of mind the  Boundary Engineering oil pumps are tempered billet gears that will be able to deal with the stronger forces at higher RPMs and may have some benefit on oil flow at higher rpms due to their claimed deep porting that reduces cavitation. A water-cooled oil cooler and piston-cooling oil jet are employed in stock form to increase durability against high-temperature loads.

Probetalk forum user has found that there can be different tensioners that may contribute to valvetrain noise or other issues. One is made up on an aluminum fulcrum piece and aluminum tensioner. The second set is cast iron for both. There is a difference in protrusion length of the cylinder rod. With normal cam bounce and possible stretched belt when fulcrum was hitting tensioner this was applying a load to the rear bank of cylinders which was causing the valves to actually make the noise that was heard (this would explain the strange cyclic wear on the HLAs). This is not something conclusive, but rather something to take note of in an engine rebuild. 

KL-G4 Engine - The Final Evolution

An interesting engine to study is the KL-G4 which produced till 2002 2.5L Mazda 626, the Japanese market Capella Wagons and Mazda Melennias. There are several differences in this iteration of the engine.

- Hydraulic Lash (HLA) changed to Solid Lash Adjusters (SLA) 

(HLA tiny oil ports would get clogged)

- Cast instead of forged and nitrided crankshaft
(missing the entire center counter weight)

- Lighter crankshaft, lighter pistons and rods.
(crankshaft around 3kg lighter)

- Oil Squirters removed in the block 

(likely to improve oil flow overall)

- Cam lobes have a slightly different profile (due to SLA) 

(narrower surfaces where the lash gap is found)

- Harmonic Balancer has a 36-1 ring on the back side

(vs. the 6-lobe of the standard KL motor)

- KLG4 cars have distributor-less ignitions and Coilpacks

- Wasted spark ignition system

- Intake manifold has a better design 

Engine Build 

The engine target for this build is the Ferrari 355 engine which produces around 350hp in crank form (actual horsepower was overstated by ferrari). The reason why is that I think this engine represents the pinnacle of non VVT engine designs from around the same era.
My build is really developed for the miata applications and relies on a combination of parts from several generations of the KL-engine as well as some aftermarket parts.

I will likely get a stock motor from a working car, and swap it in to solve the registration and insurance issues, before enjoying it for a while and looking for the right partner for a full engine build. 

Base Engine: KL-ZE (from an auto car)

Mazda Millenia’s from the early 2000’s cost nothing in Japan, and paired with the automatic gearboxes, they were likely under stressed. As long as they were maintained, these would be a great platform to start with. The motor will be disassembled, throughly cleaned and inspected. The oil squirters will be blocked off as per Alistar Oag's research. 

"The original purpose of the oil squirters in the cylinders was to keep the temps in the cylinders down by squirting oil in them but after extensive reasearch he found that the squirters were not needed because the crankshaft splashed plenty of oil into the cylinders on its own. Interprep was racing these engines 12 hours at a time. He also advocated boring out the oil pump, making the oil galleys non-tapered and adding a radius to the tight turns improved the flow and pressure of the oiling system and oil cooler. "

Pistons: Weisco 12.0:1 Pistons 85mm 

Crankshaft: Stock, fully balanced 

Balancing the rotating assembly will serve long term reliability and response. Focusing on a slight increase of the oil journals in the crankshaft should help the no 6. bearing failure area. WPC treatment on the crankshaft for reduced friction. 

Connecting Rods: Stock connecting rods   

There have not been any documented failures of the connecting rods, so a light debur and polishing to remove stress risers. The weight of the stock connecting rod is 594 grams. Pauter Rods (MAZ-200-560-1378F) weight around 604grams and K1 Rods (028CE16138) weigh 584 grams. 

Bearings: King/ACL Bearings 

Higher quality bearings should help address the no.6 bearing failure point. Tolerences have to be managed correctly for the oil weight that is chosen for the application.

Cylinder Head

Pressure test the cylinder head to make sure there are no cracks. Check combustion chamber volume. The real gains to be made in our heads are in the bowl area. Just smoothing out the short radious on the intake and exhaust ports in my opinion will make the head flow better then lets say plus sized valves. That along with sanding down the casting marks and the transition from the valve seat to the port area yield good results. Follow porting guide detailed above. 

Camshafts: Catcams 280°/280° - 9.40mm

The original cams specs are 243°/241° - 204°/203° - 8.55mm/8.50mm. 

Camshaft upgrade to create an engine that is even more biased to top end acceleration without too much of a sacrifice at the lower part of the torque curve. WPC treatment here again to reduce friction. Here is some data that Interprep collected on their camshaft development. I am planning to run a shorter final drive to compensate for torque loss at lower RPMs.

Cam caps and camshafts should be WPC treated to try and ensure that the surfaces do not catch an foreign parts. 

Valvetrain: Ferrera valve retainers / heavier springs / stock valves

This should address the reliability issues with the valvetrain to allow reliable operation to 8,000rpm occasionally. I need to find springs that will prevent valve float with accidental overreving to 8,500rpm, but ones that are not too stiff too. Beehive springs are untested, but a promising directon to explore. From a Probe-Talk User:

"Ferrea currently offers the most stuff, but you will pay for this setup. They offer their Competition Plus Line of Valves (the solid stem valves...not the hollow stem). Their valves are available in stock size and 1mm oversize. For intake, you're looking at $239.88 for a set of 12, and for exhaust you're looking at $251.88 for a set of 12. Stock and oversize are the same price. They also offer a spring/retainer/lock package. The spring is a steel, dual spring setup. The retainer is titanium, and the valvelocks are chromoly. The springs are $402.96 for a set of 24, $306 for a set of 24 retainers, and $95.76 for a set of valvelocks. The total for the full Ferrea valvetrain is $1296.48."

Induction: Triumph Throttle-bodies 

Individual throttle bodies designed by Claire at Crapengineering in the UK, ideally with a plenum designed to feed cold air to allow for more ignition timing to be set. Fabricating a hood with a glass area to peer down the velocity stacks would be epic. Intake runner length should longer than 80mm and designed around packaging and the torque curve feeling. I hope to be able to book out a dyno day and experiment with different runner lengths to find the optimum one. 

Spark: Coil on plug 

Audi R8 coil on plug conversion seems to be the go to, and this will allow better spark control. Another solution could be to use 1999-2008 Honda CBR motorcycle coils, note that the shorter one which is found on CBR 600 cycles will not work because it is too short to reach the plug tip. However the CBR 900, 954, 1000 and 1100 motorcycles should work. It would also run a 36 tooth crank trigger wheel.  

Studs: ARP studs

High quality studs to hold the whole engine together. 

Oil Ventilation: Catch can with indicator 

Ensuring a clean valve-train and engine by better oiling quality control. 

Oil Pump: Melling Oil Pump / Boundary billet oil gears  

Paying special attention to the oil pump, I plan to try and compare a stock mazda part with a melling oil pump and billet gears to try and find the best combination of parts to ensure the long life of the engine. 

Oil Sump: Baffled with Increased Capacity  

The KL is a two-piece block design, with the upper block and a lower cradle, so the oil pan doesn't offer really any structural integrity, so an aluminimum pan would reduce the weight even further. Stock oil capacity is around 4L and I hope to increase that by at least 50% with proper baffling to try and ensure good oil pickup under sideways driving.

Oil Cooler: Air to air 

The stock engine comes with an air to water oil cooler, but I plan to switch this to a larger air to air one, and study heat management and ducting carefully. 

Flywheel:

The options for the flywheel are as follows: 

KL-ZE 2.5 : 10.4 kg
K8-ZE 1.8 : 7.6 kg 

Unorthodox Racing: 4.9 kg  

Fidanza : 4.3 kg  

Stock flywheel weight for the 2.5 engines is 10.4kg. The 1.8 engine out of the MX-3 engine has a 7.7kg flywheel. Part of the role of the flywheel is to balance harmonics in the rotating mass and prevent flywheel runout. While I would like to stay with the stock flywheel, I will probably go for the K8 flywheel from Mazda and enure the rotating masses are properly balanced and that the flywheel face is machined to a very high tolerence. 

Ancillaries:

The power steering and water pump will be replaced by electric ones to reduce parasitic loss on the engine, and alternator will be swapped out for one with higher capacity and efficiency. 

Friction Reduction:

All parts will be subjected to WPC treatment. This process reduces friction significantly, reducing heat, improving engine response and durability. Here is a short video on this incredible process:

Cars that Used the Mazda K Platform 

The smallest member of the K-series V6 family is the K8-ZE which was introduced to the Japanese market in 1991. In these early years, the K8-ZE was installed in the base Efini MS-6 and its Ford badged Telstar equivalent, base Cronos and Eunos Presso (aka 30X). The K8-ZE goes into the history books as one of the smallest mass production V6s at 1.8-litres of displacement. A high output version of the K8-ZE was also released in the 1993 Autozam AZ-3 (based on the Eunos 30X). 

Slightly larger than the K8-ZE engine is the KF-ZE 2-litre V6 which was released in Japan during 1991. Using a 78mm bore, the KF V6 sweeps 1995cc and runs a 10:1 compression ratio in DOHC, four-valve-per-cylinder guise. This engine was tied to automatic or manual gearboxes and was sold in the Efini MS-6/Ford Telstar models and mid-spec Cronos models.

In ’92, the KF-ZE 2-litre V6 was spread into the mid-spec Mazda Autozam Clef, Efini MS-8, top-line Eunos 500 and MX-6 coupe (all available with a standard auto transmission, except the Eunos 500 and MX-6 were offered with an optinal five-speed manual). 

In 1993, the Mazda Lantis appeared on the scene also with KF-ZE 2-litre power. It was also used in the Mazda Lantis Type R which also had some racing applications in that Japanese Touring car era. It’s likely that a factory high-flow exhaust is responsible for the additional top-end power.Finally, in 1998, the KF-ZE was slotted into the auto-only 1998 Millenia 20M where it made the run-of-the-mill 118kW/180Nm. Production for the Millenia ended in Aug of 2003. 

The biggest engine in the K-series family is the KL-ZE, which debuted in the Japanese Efini MS-8, Autozam Clef and Ford-badged Telstar 25Vi of early 1992. The KL-ZE uses an 84.5mm bore and 74.2mmm stroke for a total displacement of 2.5-litres and breathing is through a pair of DOHC, four-valve-per-cylinder heads. The compression ratio is relatively high at 10:1. This was the most powerful factory version of the engine at around 200hp and 165 lb/ft of torque. A five-speed manual version of the KL-ZE was offered as an option in a special version ’92 MX-6. The following year, in late 1993, the high powered Mazda Cronos Gran Turismo X and Eunos 800 25F/25G models were also released with KL-ZE power. Output remained the same and a four-speed auto came standard.

In 1997, the KL-ZE was spread into a couple of new vehicles - the Capella VRX AWD wagon and its Ford Telstar wagon equivalent. The mid-spec Mazda Millenia also adopted the services of the revised KL-ZE. The KL-ZE powered AWD wagon and Millenia come auto-only and production continued into the early ‘90s.

Here is a table of all the cars I have found that use the KL-ZE variant:

MAZDA KL-01/03 DATA

GENERAL ENGINE SPECIFICATIONS

Firing order

1 - 2 - 3 - 4 - 5 - 6

Oil Pressure

28psi @ 1000 rpm, 40-71psi @ 3000 rpm

Compression Ratio

9.2:1

Valve Bore Diameter

6.01-6.03mm (0.2367-0.2374in)

Valve Seats Width - intake/ exhaust

0.8-1.4mm (0.0315-0.0551in)

Valve Stem-to-Guide Clearance

Intake 0.025-0.060mm (0.0010-0.0023in)

Exhaust 0.030-0.065mm (0.0012-0.0025in)

Valve Head Diameter

Intake 31.85-32.15mm (1.254-1.266in)

Exhaust 27.45-27.75mm (1.081-1.093in)

Valve Face Angle

45 degrees

Valve Stem Diameter

Intake 5.970-5.985mm (0.2351-0.2356in)

Exhaust 5.965-5.980mm (0.2349-0.2354in)

Valve Spring Free Length (Approx.)

Intake 43.91mm (1.729in)

Exhaust 46.92mm (1.847in)

Minimum Length

Intake 37.0mm (1.457in) at 53.40 - 60.42 lbs

Exhaust 38.7mm (1.524in) at 52.36 - 59.26 lbs

Hydraulic Lash Adjuster (HLA) Diameter

KL01-(30mm Int/ 30mm Exh)

KLZE-(33mm Int/ 30mm Exh)

HLA-to-Bore Clearance

0.025-0.066mm (0.0010-0.0025in)

Max HLA Clearance

0.18mm (0.0070in)

Hydraulic Leakdown Rate

5-25 seconds

Collapsed HLA Gap - (Desired)

0.50-1.11mm (0.019-0.043in)

Physical Camshaft Lobe Lift

43.549mm (1.7145in)

Camshaft Duration Specs

244deg. @ 0.004", 196deg @ 0.050"

67deg @ 0.300". Max. lift is 0.338"

Camshaft Lobe Schedule

Intake opens at 8 degrees BTDC 

Intake closes at 47 degrees ABDC 

Exhaust opens at 50 degrees BBDC 

Exhaust closes at 5 degrees ATDC

Theoretical Valve Lift @ Zero Lash

9.80mm (0.388in)

Allowable Lobe Lift Loss

0.2mm (0.0078in)

KLZE Physical Camshaft Lobe Lift

246deg @ 0.004", 200deg @ 0.050"

84deg @ 0.300". Max lift is 0.350"

Cylinder Bore Diameter

84.500-84.522mm (3.3268-3.3276in)

Head Gasket Surface Flatness

0.15mm (0.0059in)

Out of Round Limit

0.022mm (0.0008in)

Taper Service Limit

0.022mm (0.0008in)

Main Bearings

Bore Diameter

61.938-61.955mm (2.4385-2.4392in)

Minimum

61.931mm (2.4382in)

Crankshaft Stroke

74.2mm (2.92 inch)

Main Journal Diameter

61.938-61.995mm (2.4385-2.4392in)

Minimum Diameter

61.931mm (2.4382in)

Out of Round Limit

0.05mm (0.002in)

Taper Limit

0.05mm (0.002in)

Journal Runout Limit

0.015mm (0.0005in)

Thrust Bearing Journal End Play

0.080-0.282mm (0.0032-0.0111in)

Maximum End Play

0.32mm (0.0125in)

Connecting Rod Journal Diameter

49.970-49.990mm (1.967-1.968in)

Out of Round Limit

0.05mm (0.002in)

Taper Limit

0.05mm (0.002in)

Main Journal Oil Clearance

0.037-0.057mm (0.0015-0.0022in)

Maximum

0.064mm (0.0025in)

Connecting Rods

Large End Bore Diameter

56.000-56.015mm (2.2047-2.2053in)

Piston Pin Bore Diameter

19.974-19.980mm (0.7864-0.7866in)

Length (Center to Center)

137.82mm (5.426 Inch)

Weight 594 grams ± 2 grams

Bore to Bore Max Twist

0.050mm per 25 (0.0019 per 0.984in)

OEM Side Clearance

0.178-0.330mm (0.0070-0.0130in)

Max Side Clearance

0.40mm Max (0.016in)

Alister Oag Interprep Clearances

Cylinder
1. 2. 3. 4. 5. 6.

Piston Clearance
0.003 0.003 0.003 0.003 0.003 0.003

ring end gap:

top ring .5mm .5mm .5mm .5mm .5mm .5mm

second ring .3mm .35mm .35mm .35mm .35mm .3mm

bearing clearance:

rod diameter 2.2053 2.2054 2.2054 2.2055 2.2054 2.2055

brg shells 0.1284 0.1285 0.1285 0.1285 0.1286 0.1285

crank pin 2.0747 2.0746 2.0747 2.0746 2.0745 2.0746

clear 0.0022 0.0023 0.0022 0.0024 0.0023 0.0024

main bore

diameter: 2.6389 2.6390 2.6387 2.6390 

brg shells 0.2068 0.2068 0.2069 0.2069 

crank 2.4290 2.4291 2.4290 2.4289 

clear 0.0031 0.0031 0.0028 0.0032 

: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :

References:

Mazda New Lightweight and Compact V6 Engines

Takashi Sakono, Shinobu Takizawa, Setsuo Harada, Tatsuji Ikeda, and Hiroshi Abe

Development of the High-Power, Low-Emission Engine for the “Honda S2000”
Otobe, Y., Kawaguchi, H., and Ueshima, H., "Development of the High-Power, Low-Emission Engine for the “Honda S2000”," SAE Technical Paper 2000-01-0670, 2000,

MX-6 Forums
Probe-talk Forums
Mazda 626 Forums
Xtremethings (Micheal Perry)
Interprep Tuning (Alistair Oag)