2005/08/28
- (Often referred to as TFM, as in RTFM).Shutting Off a Cold Rotary - DON'T DO IT! Either warm it up, or don't start it at all. You can read here and here to find out what to do when this advice is ignored.
With the introduction of the RX-8, Mazda left-handedly admitted this is a problem by including in the owners manual instructions to allow the engine to idle before and after driving. To further diminish the likelihood of encountering short start problems, Mazda released the M-Flash Technical Service Bulletin.
During normal use, all IC engines build up deposits, made primarily of carbon, on combustion chamber surfaces. Occasionally these buildups break loose. If these breakoffs occur during normal operation, they are blown out the exhaust in short order by the violence of the combustion and exhaust process. If it happens while the engine is off, the pieces fall.
In a boinger, there are two combustion chamber seals, the rings, and the valves. Because of the orientation of the valves, a loose piece of carbon is unlikely to fall onto a valve seat or face and prevent sealing. Because a piston ring has a square edge, if a loose piece of carbon falls against it, it is unlikely to break the ring seal. Even if it did, the second ring below it provides adequate additional sealing that a minor temporary loss from the top ring will not prevent the engine from starting.
The rotary is fundamentally different. When a piece of carbon falls with the engine off, it will fall where it will come in contact with an apex seal. Furthermore, the round shape of the apex tip lends itself easily to a piece of carbon wedging itself between the housing and apex seal. By wedging thus, it lifts the seal away from the housing, reducing compression. The reduced compression, though temporary, causes hard starting. The starting failure often results in a flooded condition. The flooding washes away the oil from the other seals, compounding the compression loss, making it still more difficult to start.
If a large enough piece dislodges and falls in a cold engine, it will wedge between the rotor apex and chamber wall and prevent rotation. This is called carbon lock.
The dislodging of carbon from the rotor face while the engine is off happens far more frequently when the engine is started cold and run very briefly. Indeed, the heat cycling process itself often induces the dislodging. So, warm it up, or don't start it at all, and you minimize your risk!
Repair Manuals
2005/08/28
- (Often referred to as TFM, as in RTFM).
A shop manual is an invaluable reference. You need one. Maybe not today, maybe not tomorrow, but someday at least. This web site is intended to supplement a manual, not replace it. If you don't find the help you need here, odds are good that the reason is it's covered in the manual.
Regardless of your level of experience repairing things in general or rotaries in particular, some things simply cannot be dealt with without reading. Shop manuals are available from several sources, but only official manuals from Mazda are close to good enough for a Mazda. While Mazda's aren't perfect, others leave things out, frequently mislead, and occasionally are simply wrong. Unless you never work on your car, you need one. And, even if you never work on your car yourself, odds are good that someday whoever does your repairs for you will be unable to solve your problem without one.
Mazda changes to a different color for original shop manuals each year. Some of these follow:
In addition to Mazda dealers, factory manuals are available from Mazdatrix and Racing Beat.
Older factory manuals don't include wiring diagrams. Kevin Wright has several PDF versions of 1st gen RX-7 wiring diagrams.
Exhaust Fasteners - The high temperatures of rotary exhaust dictate more durable materials be used in exhaust systems. Alloys that inhibit rust are part of the formula. Some of the studs and nuts Mazda uses do a good job of avoiding rust and lasting long, but their peculiar alloy is highly succeptible to galling during assembly and disassembly. Galling results in stripped threads or snapped studs. Before removing any exhaust system fastener involving exposed threads, always clean the exposed threads. Then, if no prior use of anti-sieze is evident, lubricate with a penetrating lubricant. Use of anything but 6-point box wrenches or sockets can also become trouble. Before installing, always be sure the threads are coated with anti-sieze compound.
Spark Plugs Can Destroy Apex Seals - The holes leading into the combustion chambers from the plugs aren't all the same. On production Mazda rotaries, the leading is about the same diameter as the plug threads, while the trailing is about 1/8". An inappropriate length plug threaded into the trailing will bottom out on the housing, destroying the business end of the plug, if not distorting the housing itself and reducing or eliminating compression. A long enough plug threaded into the leading can contact apex seals and produce fatal seal damage.
Improper plug selection can also cause detonation & preignition, which is hard on spark plugs and apex seals. Bits of damaged spark plugs are hard on apex seals and rotor housings.
Over the years, Mazda has worked extensively with spark plug manufacturers NGK and Nippondenso to develop the plugs best suited to the unique needs of the rotary environment. For stock or near-stock Mazda rotaries, you certainly can't go wrong following Mazda's initial recommendations, which can be found in your owner's manual or the shop manual. Improved NGK and Nippondenso varieties may be available for older engine models. Beginning with the 2nd generation RX-7 models, Mazda specified different heat range plugs for leading and trailing. The NGK and Nippondenso application catalogs list the types most suitable for your Mazda rotary. Mazdatrix mirrors Mazda's recommendations and lists other plugs suited to such special situations as racing.
Before you decide to risk your engine with plugs recommended by some parts counter order taker instead of Mazda, visit Century Performance's spark plug info page and NGK's spark plug FAQ to learn more about how plugs work.
A Spark Plug Can Destroy a Rotor Housing - Rotor housings are mainly aluminum, while spark plug threads are steel. As a result, the same galling problem found with exhaust fasteners can occur with spark plugs. If one gets stuck, the only solution is overhaul, usually including a new rotor housing. Again, the solution is prevention, anti-sieze compound and correct installation torque, both of which are explicitly stated in Mazda shop manuals. Due to the anti-seize and low recommended installation torque, a torque wrench is not essential, just recommended. If you don't have torque wrench access, use a 3/8" drive "stub" (short-handled) ratchet and simply snug each plug with your thumb and two finger tips.
Dual Ignition - There are two independent ignition systems (except for the siamese distrubutor) on Mazda rotaries, one for the top plugs and one for the bottom. Mazda labels the spark plugs leading and trailing according to combustion chamber position, but the leading plugs are the lower and the trailing are the upper. Under nearly all circumstances, the leading plugs are supposed to fire first. Strange things happen if wires get crossed up from their correct positions, none of which are any good for the engine. OEM spark plug wires have blue coloring on the trailing. Usually the boots are blue instead of black. The hoods for the primary wires for the trailing are blue as well. Replacement plug wires rarely have the blue coding. On first generation RX-7s, the forward coil is for the trailing plugs, and the rearward coil is for the leading plugs, backwards from what a reasonable person would expect, given the way Mazda has labeled the spark plugs.
F Is A Good Mark - According to the shop manual, there is an "F" on the distributor drive gear to indicate the front. Putting it on backwards makes it difficult or impossible to align the distributor so that the timing can be correctly set with the cinch screw for the distributor somewhere near the middle of the slot, or even in the slot at all.
There are two ways to tell which way to install it if yours didn't get an F:
Replacing Fuel Hoses - Carbureted Mazda rotaries have two fuel hoses attached to the carb. One is fuel supply. The other is a return line. The return line has a restrictor in it to prevent excess pressure drop. If you replace the line without retaining the restrictor, fuel pressure and HP will drop considerably.
Wrong Threads - This commonly arises in swapping brakes or rear ends from later cars into earlier cars. Hydraulic Threads provides those details. Another gotcha pops up with engine removal and replacement. The 8 mm studs (14 mm nuts) used on the exhaust are both a different alloy and a coarser thread pitch than used elsewhere on the car, like for the engine mounts. The short starter bolt is one of very few other instances of 8 mm coarse threads.
Clutch Misinstallation -
If the flywheel isn't stock, the flywheel may allow the pressure plate to be incorrectly installed.
There are three ears on the pressure plate friction surface. Some replacement flywheels only have three recesses around the edge of the rim, unlike others and stock, which have six. Misinstalled 180 degrees out of correct on a three recess flywheel, the ears won't be able to fall into the flywheel recesses as the disk wears, instead landing on the flywheel rim, causing greatly reduced clamping pressure and slipping that gets bad quickly once it starts.
Clutch Adjustment - As a clutch disk wears and becomes thinner, pedal engagement occurs higher and higher. When engagement height is left adjusted as it comes from the factory, clutch engagement will become very high and clutch slippage will begin before the clutch rivets begin making contact with the pressure plate and flywheel. This is supposed to induce the driver to replace a worn clutch before it becomes a more expensive job from rivet contact. The "safety margin" is lost once you disturb the linkage adjustment. Adjusting pedal height is generally best reserved for when a clutch is brand new.
Clutch Disk Replacement - A rotary redline is higher than a typical boinger redline, which means shift RPM is typically higher. The centripital force trying to rip the face off the clutch disk during a shift is considerably higher for a rotary - the force at 8,000 RPM is twice that at 5,657!
Remanufactured clutches come in basically one flavor, stock replacement. In the remanufacturing business, this means a limited selection of product, making as narrow a selection of product fit the maximum range of applications. This is pretty much a one facing type fits all sizes, with the same standard attachment method. So, a remanufactured clutch is best suited for a typical 5,000 or 6,000 redline application. As a consequence, a much higher than normal failure rate for remanufactured clutches in rotary applications is a fact.
Do you want to risk having to redo your clutch job after only a few break-in miles? Don't risk it. Buy your new clutch either from Mazda or a high performance clutch manufacturer.
Clutch Pilot Bearing - Most professional mechanics won't do a clutch job without replacing the pilot bearing. The part is cheap, they've spread their cost of the removal tool over many jobs, inspecting and cleaning the old bearing takes more time than replacement, the new part is comeback insurance, and a new bearing means additional profit. There really is no good reason for them not to.
That said, if you are doing the job yourself, you don't have ready access to a proper removal tool, and the bearing isn't damaged, don't disturb it. Clean it, lube it & install new a seal.
Here's why:
The main caveat is determining whether the bearing in place is suitable for reuse. To make that determination, you must be able to get it thoroughly clean, you must be able to visually inspect it carefully, and you must be able to feel that there is no damage that you can't see.
Another caveat is in parts supply. Independent parts houses commonly sell clutch kits that include the disk, a pressure plate, a throwout bearing, an alignment tool, and a pilot bearing. Usually a pilot bearing seal is omitted from kits sold by non-Mazda specialists. This happens because most boingers use a bronze pilot bearing that requires no seal, instead of a needle bearing. The parts catalogers as a group are simply not up to speed on this exception.
Last caveat is the clean & lube process. This is difficult unless you have compressed air. With compressed air, you can preclean with the air jet, then use a solvent, then remove the solvent and any debris with the air jet. Without the air jet, you risk leaving behind both contaminants and solvent, both of which will accelerate failure of the new grease and/or the bearing.
Flywheel Resurfacing - A stock Mazda rotary is built with an imbalance to counter the eccentric shaft's offset journals. Under the front cover is another counterweight. These counterweights must match to keep your engine running smooth as it should. In order to maintain adequate imbalance precision, extremely high accuracy in the face machining process is required. If your machine shop doesn't have this capability, or the ability to restore the exact same imbalance through a balancing facility, you are far better off using the flywheel as is instead of machining the clutch face. Most of the time, a new clutch will work and last just fine even though the flywheel looks and feels really bad. If yours really need a reface, take the opportunity to switch to a light flywheel. Aftermarket flywheels use an automatic transmission's counterweight instead of having the imbalance built in. If the replacement ever needs resurfacing, rebalancing is less of a challenge to a machine shop.
Changing the Pulley on the Front of the Eccentric Shaft (or the seal) - When you pull the pulley off after removing the big bolt, there's close to a 50/50 chance the key will stay put. That means there's roughly a 50/50 chance it will be stuck to the pulley in your hand. Another possibility is that it will slide out just a little, then stay there. If this happens, you won't even know there is a problem until you find the oil leak that results from excessive end play after the thrust bearing invisibly falls out of position behind the key.
While performing the swap with transmissions and clutches through 1991 models, the clutch pedal should be held to the floor. This holds the eccentric shaft in its most forward position, minimizing the chance of parts dislocation during the procedure. A piece of wood wedged between the pedal and the front edge of the seat will do the trick. With clutches and transmissions newer than 1991, the clutch can't hold forward pressure on the flywheel, so you must use other means to do so.
If you don't first, before removing it, turn the pulley so that the shaft keyway points straight down, and the key does come out, the counterweight inside will probably fall out of alignment, turning your "simple" swap into a major ordeal. With the key pointing straight down, the weight will also be pointing down and so won't try to move if you are unsuccessful in keeping the key in place removing the pulley.
You must prevent the key from moving while you pull the pulley off. Failure to do so means removing the front cover to reinstall everything correctly. If the key does come out just a little, the bearing can fall out of place even if the weight doesn't. A successful pulley change must be confirmed by checking for proper end play after the pulley bolt is correctly torqued.
The bolt that holds the pulley to the eccentric shaft is supposed to be torqued to about 90 lb-ft. A common 1/2" breaker bar should be sufficient to break it loose, but a number of reasons might be responsible for one being particularly stubborn to remove. For example, if some fool has used that bolt for turning the engine over, it's possible that the torque value has been exceeded by a factor of three or more, particularly if the engine was ever carbon locked. You may find it impossible to remove without a 3/4" drive impact gun, which may require radiator removal to use.
Note - Mazdatrix has more detail on this entire subject.
Sloppy Steering - One of the worst features of the 1st generation RX-7 is recirculating ball steering. Not the most sensitive feel when new, it gets worse with wear. On the early models, a slotted bolt secured by a 14 mm nut could be used to remove excess lash and reduce steering wheel free play to an acceptable level.
At first glance, the arrangement of the later 1st gen models appears similar. Under a small, nearly round cap lies a slotted bolt and a 14 mm nut to lock it in place. Though the shop manual calls it that, it is not a free play adjustment in the same sense as earlier boxes. Adjusting it won't materially affect steering wheel slop unless the amount you need to eliminate is small. The free play adjustment you probably need is the cap with four holes in it locked in place by a 40 mm nut that early boxes don't have. Mazda calls this a "preload" adjustment. As a practical matter, to remove excess steering wheel free play on the later boxes, the big nut must be loosened and the spanner turned clockwise sufficiently to remove excess slop. Once you do this, the more obvious adjustmnet probably won't need to be touched.
Overtightening, besides stiffening the steering, will accelerate the wear that creates the slop.
Two Barrel Carburetors - If you haven't already digested 'Why does a rotary engine make as much HP as a piston engine twice its size?', do that now, as the discussion of power cycles is pertinent to understanding what follows.
No two barrel carburetor has ever been designed for use on a production model Mazda rotary engine. No seriously high-performance (Weber, Mikuni, SK, Dellorto) two barrel carburetor has ever (to my knowledge) been designed originally for use on more than two cylinders of a boinger. High-performance two barrel carburetors are designed and intended for use of one carburetor for each two cylinders, one throat for each combustion chamber.
That said, why are single two barrel intake kits for rotaries available from reputable rotary performance shops? Under racing circumstances, low speed performance and fuel economy matter none, while torque and HP production from the big twos is typically quite impressive, if the selected carb is big enough. Because such applications are available, customers who aren't familiar with their drawbacks want them for street use. Adaptations of those kits for street use represent compromises. One carb is cheaper than two. In many cases, they are extreme compromises. If you want detail on how this is so, you will have to read books on the subect of carburetor and intake manifold theory.
To itemize the components of the problem:
Two barrel carburetors designed for multi carb installations are expected to be used with short intake manifolds, necessitating only a small accelerator pump shot. Also, the multi arrangement provides multiple accelerator pumps, so each individual pump doesn't need to be very large. Consequently, such carbs universally have accelerator pumps of smallish capacity, with the sole exception to my knowledge of the no longer available Dellorto 48 as modified by Racing Beat specifically to deal with the pump capacity problem.
Many of these problems are precisely the reasons why four barrel carbs were invented. High velocity is key to good atomization of fuel being dispensed in the carb throat. The better the atomization, the better the performance. High velocity is accomplished by keeping the carb throats small. Capacity is maintained, or increased, by having additional throats that only open when additional flow is required. The same overall effect as a four barrel is achieved by using a pair of staged two barrels, such as in the no longer available Rotary Engineering Weber DCD system.
Various bandaids and operating behavior can be utilized to minimize the problem, but only to a degree. Unless the pump shot is adequate, poor drivability is inevitable. While the definition of poor is subjective, in virtually every case, it will be worse than when the car was new, particularly if it was originally equipped with fuel injection. With nearly any large two barrel carburetor installed on a Mazda rotary, you can kiss goodbye the excellent drivability offered by the stock induction system. Whether you can live with the drawbacks, only you can decide.
Holley Carburetors - Since Racing Beat makes kits that include Holley 4 barrel carbs, it's easy to assume that money can be saved by buying just the manifold or an adapter and using a Holley obtained cheaper (or free) elsewhere. While you may save money, the cost of that saving can be much grief. Holley carbs must be extensively modified in a fashion that conventional Holley tuning wisdom dictates is not to be done, particularly by an ordinary end user.
Out of the factory box, a Holley will exhibit stumble as the secondaries open and/or severe excess richness as RPM increases, and usually cause terrible performance above about 5000 RPM. Changing only main jets and/or fuel level are always insufficient to provide best performance.
Side hung Holley float bowls work best when used in drag racing applications, with the float bowl pivots oriented fore-aft, primaries to the front of car and secondaries to the rear. Center hung Holley float bowls are the opposite, working best in road racing applications, with the float bowl pivots oriented side-to-side, still with the primaries to front of car and secondaries to the rear.
On a rotary, Holleys are mounted 90 degrees from the postition they were designed to operate in, so side-hung bowl float pivots are oriented better for cornering, while center-hung bowl pivots are oriented better for drag racing. While the pivot orientation preferences are reversed with the carb oriented 90 degress from normal, the bowls were not designed for the altered orientation. Even when a Holley is properly modified otherwise, the altered orientation is usually of little significance for drag racing or typical acceleration, but it is exceedingly difficult to get a Holley to behave acceptably during extreme cornering, which makes them awful for autocross compared to other carb choices or EFI.
Stock Carburetor Overhaul - Mazda rotary carburetors are difficult to remove from the manifold while still on the engine in the car, particularly 12A carbs, and even more difficult to reinstall. "Rebuilding" of the carb is simplified by doing it while still on the car. Only if it is so bad that it needs a chemical dunk does it really need to be removed. Nearly anything besides a dunk can be done with the main carb body still on the manifold. With the right tools, which includes compressed air, the carb top can be removed, giving access to all the jets, bleeds and floats. The wells and float chambers can be filled with cleaner and soaked in place. The main thing to be careful of is when using compressed air that you don't lose the check balls and ball weight from the accelerator pump circuit. An added advantage of doing it on the car is that many hoses and wires don't need to be removed, saving time and frustration in figuring out what to reattach where.
Nikki/Hitachi jetting tip: Careful shopping for the right screwdriver will allow the main jets to be loosened without removing the top off the carb or the carb off the manifold. Use the screwdriver only for loosening and tightening. A round wooden toothpick can be used to remove and install a loosened jet simply by inserting into the jet orifice, as long as the jet threads aren't damaged or corroded.
Another jetting tip: If you ever find a jet so crudded up that carb cleaner doesn't seem to get the job done, a toothpick can come to the rescue again. The wood isn't strong enough to remove metal, but combined with carb cleaner, or, better yet, magic potion, it can do a nice job of polishing away gum and oxidation without affecting jet size.
Stock Carburetor Throttle Return Springs - Most Mazda Nikki 4 barrel carbs, including all those used on 12A RX-7's, use two different types of throttle return springs. One is the normal one that provides easy throttle modulation with a light feel. The second is a much stiffer safety spring, designed to come into use only in the event that the primary spring fails or disappears. People unfamiliar with the arrangement occasionally and unknowingly engage the safety spring whilst meddling with the carb. When the safety spring is engaged, not only is the pedal stiffer than normal, but also the secondaries open late, if at all, reducing performance.
Engine Installation - A trap the average novice usually falls into is trying to put a pre-'86 engine in with the mount bracket already attached to the front of the engine. Without that bracket in the way, it is immensely easier to line the engine up to the transmission.
Another engine installation problem is the dual lift hangers Mazda uses. If you have available to use a rotor housing lift hanger off an early '70's rotary it is much easier to rotate the engine into a correct alignment with the transmission. The hanger set Mazda installed since around 1974 makes it tough to get any but one hanging position, which only about one time in fifty is the one you need to get the correct alignment. To use the factory hanger set to lift and align an engine properly requires a yoke be connected to the hangers. The yoke pivot allows a much more flexible range of motion for aligning the engine to the transmission.
Stock Carburetor Throttle Return Springs - Most Mazda Nikki 4 barrel carbs, including all those used on 12A RX-7's, use two different types of throttle return springs. One is the normal one that provides easy throttle modulation with a light feel. The second is a much stiffer safety spring, designed to come into use only in the event that the primary spring fails or disappears. People unfamiliar with the arrangement occasionally and unknowingly engage the safety spring whilst meddling with the carb. When the safety spring is engaged, not only is the pedal stiffer than normal, but also the secondaries open late, if at all, reducing performance.
Mating Engine & Transmission 
2007/11/05 -
It should go without saying that nothing should come between engine and transmission, but it seems
the oil pressure sender wire has an irritating proclivity to slip into position to do exactly that, and
not get noticed until it shows up as an improperly operating oil pressure gauge. Before tightening those
bolts, double check this hasn't happened to you. If you don't, you might find it necessary to replace that
gauge.
Custom Exhaust - Exhaust system configuration is particularly crucial to maximizing power with a rotary. Unfortunately, peripheral port rotary exhaust is extremely loud, multiple orders of magnitude more than engines that partially block exhaust flow with slowly opening and closing poppet valves. It's also considerably hotter, which causes ordinary exhaust mufflers to deteriorate rapidly.
Don't kid yourself into thinking you can design your own exhaust system if you don't have advanced exhaust engineering experience and plenty of money and time. Without them, it's highly probable that either your power level, or your sound level, or both, will be a big disappointment with anything other than the proven complete systems available from reputable providers of things rotary. If you think you can succeed with ordinary mufflers, or even popular extraordinary mufflers like Flowmaster or SuperTrapp, you are almost certainly doomed to disappointment.