Complimentary PDF Library

Here you will find PDF Docs of the Workshop manual to compliment the sections of this page. You will also find important safety and correct jack and axle stand positions. Always ensure your safety when working beneath a veichle.

If there is more information you require on a subject contained on this page, please feel free to contact me through the 'Links' page.

Cold Start Carbon Knock

Partial Work-Around

Design fault only on earliest KL V6 engines with engine #s pre KL326296, pre 08/03/93. Mazda TSB B00995/95, issued 16/06/95, re 626/MX6/Probe-2.5.


Diesel-like sound on winter cold-start, diminishing quickly as the engine warms (sounds like frozen peas bouncing off pistons).


Carbon-build up causes "carbon-knock", tackle when heard. The noise is not HLA/Lifter ticking, nor the (aesthetic) Friction Gear noise localised around the exhaust cam of the rear V6 bank, timing belt end and dissappears at circa 2200rpm.


Simple decarbon service at ?70/70$US, common for many engines with modern fuels.

TSB solution, #96113, May 96
  • Bring engine to normal operating temperature, then shut engine off.
  • Use a 610mm (?!) length of 1/8" I.D. vacuum hose as a feed tube for carb cleaner (Carburetor Tune-up Cleaner, ford # D9AZ-19579-BA), although it is strongly advised to use an aerosol cleaner to eliminate the risk of liquid lock (some techs have done it). Do not use any of the vehicle's vacuum hoses for feeding cleaner into the engine otherwise they may deteriorate as a result (vacuum leaks on 24 possible hoses is tedious).
  • Connect one end of the feed tube to the Purge Control vacuum port on the upper intake manifold and place the other end of the feed tube in the can of carb cleaner, or to the spray nozzle of an aerosol. If a liquid bottle is used the feed rate must be at least 30 secs to avoid hydraulic lock-up damage.
  • Pinch the feed-tube with locking forceps or locking pliars to prevent suction during start-up.
  • Restart engine.
  • Allow idle to stabilise, then feed the cleaner into the engine by releasing the locking pliars/forceps while lifting engine speed to 3000rpm until half the cleaner is used. If using liquid feed, ensure feed-rate is always erring on the safe side by kinking the tube (a thumb screw clamp is ideal). If using a spray feed (preferred) keep shaking the can vigorously. At the half point allow the engine to stall, or shut-it off.
  • The vehicle should be left to sit for 6-24hrs allowing the cleaner to saturate the carbon, the initial intake of cleaner loosens carbon by both dissolving and also by a cold-shock.
  • Flush remaining carbon by repeating the above steps using the remaining half of the carb cleaner. Do not let the engine stall during the procedure.
  • Remove the feed tube & reconnect the purge control vacuum line.
  • Change oil & oil filter to remove contaminants & carbon in the oil which will otherwise be pumped through all the HLAs.
  • Take the car on a 3-4 mile test drive, using the full-rpm range to remove any remaining cleaner or carbon from the engine.
  • Let the car sit for another 12hrs, restart - if the noise is gone then the cause of the noise was carbon build-up.
Full Noise TSB re Carbon-Knock/HLA/Friction-Gear

Here is a full TSB re Noise from Carbon-Knock/HLA/Friction Gear, issued by Ford/Mazda :

Page 1 of TSB -:- Page 2 of TSB -:- Page 3 of TSB -:- Page 4 of TSB -:- Page 5 of TSB -:- Page 6 of TSB -:- Page 7 of TSB -:- Page 8 of TSB -:- Page 9 of TSB -:- Page 10 of TSB
Can I use higher-octane fuels for more performance ?

Using high-octane fuels (98/100 RON) has benefits on the V6 engines as they have a knock-sensor to take advantage of it; UK Standard unleaded is 95RON, Super-Unleaded is 97-98RON. However using higher than 95 RON is largely unnecessary (re cost/power).

The ECU automatically adjusts timing based on the knock sensor from 6-18 degrees, a very considerable range. Base timing is 10 degrees +/- 1 degree.

Cars with 2.0 engines, like Miata/MX5s, do not have knock sensors and thus can't take advantage of higher-octane fuels without adjusting their timing. The miata/MX5 groups often do this, but there is no safety back-track if pinging does occur due to carbon-buildup (on any engine) or simply poor quality gasoline a risk of engine damage as a result.


TSB = Technical Service Bulletin : The specific Carbon-Knock TSB refers to 'customer satisfaction'; some mechanics say if untreated long-term then valves could be damaged (burnt). Decarboning is an easy & relatively cheap service to performand benefits any engine, not just those early V6s whose head design allows carbon to build-up in an area where it will create noise.



None, all 93/94 models may suffer (Mitsubishi) ignitor module failure (located in the distributor). Althogh since this document was first published, you can pre-empt the condition by replacing the known faulty transitor with a better version and bigger heatsink.


Engine cuts-out with tachometer dropping to zero. Attempting to restart the car immediately is futile, with no spark. After a wait of 5 seconds to 20 minutes the car can be restarted. However, once the failure has occurred it will repeat unpredictably & increasingly repeatedly. No ECU codes are set and the prime indicator of the condition is no spark is generated when an ignition lead is placed on the spark plug cover with a spark plug in it.


Failing ignitor module (transistor within it) within the distributor.


US owners - rebuilt distributor discounted from Trussville Mazda, or the actual components themselves listed in Ford USA TSB #97137.

Ignitor - F32Z-12A112-BA (Ford)

Coil - F32Z-12029-AA (Ford)

Ignitor - KLY1-18-V20 (Mazda)

Coil - KLY1-18-V80 (Mazda)

UK owners - individual parts are available from Ford UK : Coil - 3974598, Ignitor - 3974599

The TSB is known to all Ford main dealers. A new entire distributor is not required - just the parts. Causation is thermal overload causing permanent damage to the transistor in the ignitor module as the semiconductor substrate breaks down. Typical causes are heat-soak or hot conditions.

An alternative failure can be the Crank Angle Sensor, at the Crank Pulley end which can have 2 failure modes - temperature (cold mornings) and complete failure. 1993/94 cars have two crank sensors therefore - distributor and crank pulley end, this is not true of later vehicles. Failures of the crank angle sensor, a hall-effect-switch with no moving parts is rare.

TSB #97137
Page 1 of TSB -:- Page 2 of TSB -:- Page 3 of TSB -:- Page 4 of TSB

Friction Gear Noise


None, aesthetic noise & not detrimental.


TSBs detail noise 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, but if it is done at a Main Dealer it may be beneficial to do the timing belt simultaneously regarding cost. The task is over 7hrs labour, with the friction gear itself just US$12 or ?12.

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
Full TSB re Noise from Carbon-Knock/HLA/Friction-Gear :
Page 1 of TSB -:- Page 2 of TSB -:- Page 3 of TSB -:- Page 4 of TSB -:- Page 5 of TSB -:- Page 6 of TSB -:- Page 7 of TSB -:- Page 8 of TSB -:- Page 9 of TSB -:- Page 10 of TSB

Hydraulic Lash Adjusters - HLA's


Regular 3k-mile oil/filter changes, preferably with synthetic.


HLA noise can affect most Mazda engines (Miata/MX5 I4, V6) at any time. The noise is a tap-tap due to non zero-lash between cam lobe & HLA.


If HLAs are subject to numerous cold stop-start driving or garage-to-drive short trips they may not pump-up with oil properly. This gives the tell-tale tick-tick/tap-tap from an otherwise super-silent V6. More regular oil changes help. If the sound stays after 20mins for several consecutive days then the cause is one or more collapsed HLAs/lifters due to dirt/varnish.


Change oil regularly, every 3k-miles and use a synthetic oil if possible along with a Mazda oil filter regarding anti-drainback valve quality.

A synthetic is known to sludge far less and has better detergents. Mobil1 is favourite, but very expensive in the UK - Comma 5wNN is a good alternative, Castrol Syntec/Magnatec is not a true synthetic.

Should a flush be used, several oil changes before & after it are necessary. The flush is a misnoma - it loosens desposits which are then circulated around until that oil & filter are changed.

  • Run synthetic for two oil changes before a flush is performed.
  • Run a flush 10 mins at idle before a change, and then change oil & filter.
  • After 250 miles change the oil & filter again (filter will be full).
  • After 500 miles, put another can of flush through 10 mins at idle, then change the oil & filter again.
  • After 1500 miles, change oil & filter again. Repeat 1-5 if noise remains.

Bilstein-2000 flush machines exist in the USA which can clean an engine without it being running, are factory approved and extremely efficient.

HLA Test Procedures

In Vehicle Test :

  • Remove all spark-plugs.
  • Remove the cam covers front & rear (including VRIS casting etc).
  • Slowly rotate the cam whilst trying to insert a 0.15mm feeler gauge between HLA & cam lobe. Slowly rotating to avoid introducing air into the HLA (which will falsely increase clearance). If the gauge can be inserted then the HLA is collapsed and nearing replacement specification of 0.18mm.

Out of Vehicle Test :

  • With a HLA in your hand, squeeze it. If it compresses the HLA is to be replaced.
  • Any scuffing of the HLA surface or the cam lobe suggests both must be replaced.
  • In removing any HLA, it must be put back in the place from which it was removed (otherwise the cam-lobe to HLA interface will be different and point-loading will occur increasing temperatures and wear rapidly.

In any assessment of HLAs, there is the remote possibility of aerated oil being the cause of HLA noise - from a faulty sump-pick-up, sump-to-block gasket in the oilpan, to a faulty oilpump. These are discussed in the Ford TSB covering the detailed & comprehensive assessment (detailed below, p1-10) of noise problems in notably silent engines. The 2.0s and other I4s can suffer the same HLA problems, it is easier on these to check HLA-to-cam clearances.

In the vast majority of cases, HLA noise is dirty oil & filter. Change them both more regularly - every 3k miles if necessary, and use a good quality 5wNN oil.

Full TSB re Noise from Carbon-Knock/HLA/Friction-Gear :
Page 1 of TSB -:- Page 2 of TSB -:- Page 3 of TSB -:- Page 4 of TSB -:- Page 5 of TSB -:- Page 6 of TSB -:- Page 7 of TSB -:- Page 8 of TSB -:- Page 9 of TSB -:- Page 10 of TSB

Oil in Spark Plug Well

Tools Required

Spark plug wrench (in the toolkit of Mazda cars).

Oil on Ignition Leads or on Spark Plug Part Outside Cylinder

Oil on the ignition leads, or on the area of the spark-plug outside the cylinder is from the cam/valve-cover gasket (lifespan is 55-70k miles for the rear bank and 70-90k miles for the front bank).

The gasket is not designed to be overtightened, it operates by forming a gap filling function between the head & cam cover which accommodates both flex of head/cam-cover & thermal expansion changes.

Over repeated thermal cycling the elasticity of this gasket deteriorates until it can no longer perform a gap filling function and oil will seep past the gasket into the non-draining spark-plug-wells or down the front/rear of the engine. Ignition leads left in oil-wet wells will suffer degradation in their performance until a no spark condition results in that cylinder.

Solutions :

Every 6 months mop-out the small amount of oil that accumulates (usually in centre wells) by rolling up kitchen tissue into a tube and feeding this down the well & rotating it. Ignition leads must also be wiped clean and allowed to dry out - in the longer-term if neglected a miss-fire at that cylinder will result.

Retorque the cover slightly: do not overtighten, the torque is 5.8-8.0Nm which is finger tight. Torquing is to be peformed in a specific pattern as per "Cam/Valve-Cover Gasket" in the Maintenance Page.

Replace the affected cam/valve-cover gasket replaced. The front bank is DIYable in 30mins, the rear bank requires 4hrs labour. The gaskets are 20ukp/20$US from Mazda Dealers. Details of the procedure can be found in the Maintenance Page.

Oil on Spark Plug Base & Electrode inside Cylinder

Oil on spark-plug threads alone should not be confused with oil on electrodes. When removing spark-plugs even from seemingly dry wells, any oil hidden under the spark-plug-to-head washer will rapidly wick down threads.

Oil on the spark-plug electrodes is due to one of two things :

If all the spark-plugs are similarly wet, there is a problem and a compression test needs to be done to verify valve-guides are not worn. An extremely small number of 94 V6s suffered premature valve-guide wear, typified by compression around 120psi - all such engines would have had the failure several years ago.

More usually the electrode on only one spark-plug is wet, due to a blocked fuel-injector causing oil to be drawn through under vacuum. Injector cleaner (Techron/Redline-SI1/SI2/STP-Black-Bottle) in the petrol tank will resolve the problem.

Probe Rear Hatch


Over time rear hatch gas-lift struts become weak, such that the hatch will no longer stay raised. The cause is from loss of pressure in the gas charged struts, themselves subject to temperature variation between winter/summer.


NAPA, USA offer replacement hatch struts :

93-ProbeGT Without Spoiler - BK.819-5555 & BK.819-5556

93-ProbeGT With Spoiler - BK819-5608 & BK819-5609

Note that Left & Right struts differ.

Installation Notes

Two people are required for safe fitting, as the 100kg hatch must be held up during the procedure. It may be possible to fit one at a time, but care is required whilst working to avoid damage to oneself & rear window glass.

Torque figures for the fasteners: Body Fitment end - 8.9-12.7Nm - Hatch Fitment end - 16-22Nm

Probe Widow Weathershield


Simple adjustment


Front doors - glass closure at top of drip rail, leaks, wind noise, adjustment.

1993-1996 Probe service manuals do not provide a clear procedure for adjusting the front door glass for windnoise, water leaks, window stutter/jumping and door closure. This document is based upon a Ford TSB.

There is a very simple solution to frame-less window noise problems, which stem from the window/door not sealing properly on closure. The simple solution is to automatically drop the window 1/2" on the door handle-trigger on opening and raise the window that 1/2" after closing the door. Technology is useless without initiative.


The specifications vary because door fitment varies - it may be necessary to change the door glass adjustments to the high or low end of the specifications to correct the problem on a particular vehicle.

Front door glass adjustment procedure

Check the door & striker for correct adjustment (door height must be level with quarter panel at the top of the door, and the door must be flush with the quarter panel within 1mm at points above the door handle & middle of the door (figure below). Only remove the door panel if adjustments need to be made. Do not remove the door glass, and especially do not remove the door glass guide brackets from the glass which are not adjustable and are very difficult to reinstall properly.

Door Panel Removal & Installation

Remove side view mirror inside cover. Remove the door grasp screw, fig below, (1993-94 years).

Contrary to the TSB, do not remove the door handle as it may break.

If fitted, remove the window regulator handle. If fitted and you need to later, disconnect the switch panel electrical connectors. Note the switch panel plastic tabs are very fragile and the unit is expensive. Remove the inner door handle trim screw, fig below, (1993-94 years).

Remove the 3 remaining door trim panel screws and 1 retaining clip, fig below, (1993-94 years).

Carefully pull the door trim panel away from the door to disengage the clips. Guide the inner door handle trim through the door trim panel. Remove the door trim panel.

Door Trim Panel Installation :

This is the reverse of the above, with the the door clips replaced to prevent door panel squeaks/rattles.

Check door glass pitch adjustment

Check the clearance between the quarter glass and door glass along the B-pillar, the correct gap should be 9-13mm. If adjustment is needed, mark and loosen the top regulator and guide 14mm nuts - for power windows loosen the 3 power regulator 10mm bolts and reposition the door glass (fig below).

Torque the regulator nut to 14-18lb/ft (for power windows, the power regulator bolts to 62-80lb/INCH or 7-9 N-m. Note the units). Check the door glass pitch adjustment by measuring from the front edge of the regulator channel to the front edge of the glass guide channel (fig below). The distance should be 308mm. To correct, loosen the front glass guide and adjust.

Tighten the glass guide channel and recheck the distance. Check lower regulator & guide 12mm nuts for correct torque - 14-18lb/ft. Recheck the quarter glass to door glass gap.

Check the door glass rear height adjustment. Adjust if necessary.

With the door glass fully-up and door closed, draw a line along the top of the glass with a felt market. Locate the line 55mm from the rear of the glass even with the lower edge of the roof drip moulding (fig below). The mark must be 3-5mm from the glass top edge.

If adjustment is needed, lower the glass 1" and make a reference mark on the door as per the figure below.

Loosen the glass stop bolt only enough to slide the stop the same distance the mark is out of specification - eg, if glass is 2mm too high, lower the stop 2mm. Do not allow the stopper to rotate - as it can change the glass height. Tighten the glass stop bolt. Raise the door glass and recheck.

Check the door glass rear inboard and outboard adjustment. Adjust if necessary.

It is important that the inboard/outboard adjustment screw is not adjusted. With the door glass fully-up, close the door until the top rear corner of the glass lightly touches the weatherstrip (as figure below).

Measure the distance between the B-pillar weatherstrip surface and the outside of the glass (fig as above). It must be between 14-19mm. To adjust, loosen the adjustment screw locknut (fig as below) and turn the adjustment screw counterclockwise to increase the distance.

Tighten the adjustment screw locknut. Recheck the distance between the B-pillar weatherstrip surface and the glass outside.

Check the door glass front height adjustment & adjust if needed.

With the door closed, check to see if the top edge of the mirror bracket is under the weatherstrip lip. If adjustment is needed, loosen the 3x 10mm mirror bracket bolts (previous figure) and move the mirror bracket rearward and down as necessary. With the glass fully-up place a straightedge on the glass as below.

Measure the gap between the straightedge and the top of the mirror weatherstrip. The gap must be 1-2mm. If adjustment is required, lower the window 1". Loosen the front glass stop bolt (figure above that above) and move the stop enough to get the 1-2mm gap (figure above). Tighten the glass stop bolt. Raise the window and recheck.

Check the drip edge moulding-to-door glass gap. Adjust if needed (1993 models only).

Measure the gap between the door drip edge moulding and the outside glass surface along the horizontal areas of the top of the glass (next two figures, points A through B). The gap should be 3.5-4.0mm.

Use your hands to adjust the drip edge moulding by applying pressure to bend the moulding to provide the required gap. Use caution to bend the drip edge moulding evenly otherwise it may appear wavy.

Check weatherstrip for deformation at top corner of B-pillar. If the weatherstrip is dimpled inward perform the following steps :

At the top of B-pillar, measure from the bottom edge of drip moulding down 51mm. Mark the quarter glass (figure as below).

Remove the top half of the weatherstrip from the B-pillar. Remove the top weatherstrip retaining pin from the weatherstrip and discard the pin. Seal the top weatherstrip retaining pin hole in the B-pillar.

Grab the weatherstrip in front of the corner and pull the weatherstrip rearward around the radius and down the B-pillar until the parting line on the weatherstrip lines up with the 51mm mark on the quarter glass. Reseat the remaining portion of the weatherstrip into the B-pillar. Evenly distribute the weatherstrip slack down the entire length of the B-pillar.

Road test vehicle at highway speeds and perform steps 9a and 9b to verify repair.

Turn blower motor on at various speeds and listen for air leaks. Cycle door glass down and up to verify the door glass reseals.

Parts required :

Front Door Panel Fastener (3 required)

Bottom/Rear Door Panel Fastener (4 required)

Rear Brake Caliper - design Fault for builds pre 95


Short-lived, fix is rebuilt/new (Sumitomo redesigned).


Most Probe/626/MX6 rear callipers stick circa 55k miles.

The Problem
Handbrake pivot-pin seal defect on rear calliper

The handbrake cable moves a spring/pivot-arm which operates the piston/brake-pads against the rotor. The pivot-pin sits in a needle bearing sealed into the calliper body by a rubber seal.

Design failure is seal deterioration allowing water ingress & corrosion. The handbrake is usually found unreleased or the caliper will not release on repeated heavy brake-application. The caliper must then be replaced with a new (redesigned) or rebuilt unit. Lubricating the mechanism can release it and keep it operating for a time, but only as a short-term solution. Both front & rear caliper's must have their caliper guide-bolts lubed regularly with a high temperature anti-seize grease or the caliper will only brake one side of the rotor.


Short term: Unhook the handbrake cable from the pivot arm, spray WD40/penetrant around the area and work the pivot pin its entire range of movement. Then apply a waterproof grease (Castrol CL) to the area and work that in. It is possible on relatively new calipers to remove the pivot-pin & spring: in such cases replace the pivot-pin and seal, and thoroughly clean & lubricate the pivot-pin needle-bearing area.

Finding that lubrication is only a short-term resolution correlates with the cause being deteriorated rubber-seal & needle-bearing inside the calliper.

Regular use of the handbrake will help prevent the area seizing. Once the pivot-pin rubber seal and needle bearing inside have deteriorated, regular use/lubrication will not work in the long term.

Longer-Term Solution: Warranty: See Dealer.
Ex Gratia: Ford-UK & Mazda-UK have contributed towards the cost of replacing failed calipers being aware of the problem. Refunds are not total, usually the cost of the caliper itself.

Ford UK statement: revised level caliper was fitted in production from 5-Jan-1995, and such parts may be fitted to earlier vehicles. This revised part is supplied as a matter of course now - since the seal/needle-bearing has been redesigned.


Very hot rear wheel, speed falls back rapidly during gear-changes, deteriorating fuel mileage. If extreme, brake burning smell & glowing orange brake disc with the car being destabilised towards oversteer which the TTL suspension can't correct.

Steering Clunk


Listen for front-end/steering clunking, verifying whether power steering fluid level has dropped.


A front-end clunking may be due to the preload on the rack being set incorrectly at the factory, causing the steering wheel assembly to wear out prematurely - indication is clunking with fluid loss.

A slight noise is normal as the rack is supposed to shift in its bushings when you turn the wheel at idle - the longitudinal load is around 1,500lbs. Of course the wheel should only be turned with the car moving where possible.

Other clunking noise can be from :

Anti-roll bar linkages loose or worn (ball joints, replace)

Anti-roll bar binding in the bushings (need lubrication)


A warranty TSB exists which involves adjusting the steering lash preload. Most cars should have had this done by now.

Steering Rack Mounting Bushes & Handling / Feel

The steering rack is located in the vehicle by two large steering rack bushings, when worn they allow excessive movement in the rack.

The two bushings when worn are visibly so, and easily checked. Trussville Mazda, are discount suppliers.

Note that tie-rods, tie-rod-ends & ball-joints are not a particular wear issue on these cars, unlike 3-series BMWs where they are a 60k mile item.

It is worth replacing power steering fluid regularly, every 2 years since it can overheat and so lose lubrication & protection properties for the steering rack. Redline D4ATF is an appropriate Full-Synthetic fluid. Replacement procedure is syphon the old fluid out of the reservoir, pour in the fresh, drive the car for a few days and repeat until the old fluid has been diluted.