Jaguar XJ-S. Service manual — part 43

167

diodes. If you do not use Germanium then you will need to increase the bias voltage on the sensors by about 0.5 V to
make up for the additional voltage loss.

Each time the MOSFET's fire they short out the RC circuit formed by R4 and C1 (also R6 and C2). This RC circuit
takes about two seconds to charge, and when it does voltage then flows into D2, lighting it up and warning the driver
(or flowing through D5 to the NPN transistor which lights the lamp, etc...). When a pulse hits Q1 (or Q2) it drains C1
(or C2) preventing the indicator circuit from ever getting enough voltage to light. As long as there is a steady supply of
pulses the RC circuit never charges and the light never lights.”

The polarity of the pickups on the plug wires makes no difference, since the pickup will put out a voltage pulse in one
direction when the spark starts and a very similar pulse in the opposite direction when the spark stops. Corkan: “...and
thus the reason for the first diode. In fact the circuit is sensitive enough that polarity will not make a difference, even
though you can see a small difference in the signal on an oscilloscope.”

An alternative warning system idea would be to detect the overheating of the cat itself, as discussed on page 317. This
has the advantage of indicating a problem no matter what the cause.

AUTOMATIC SHUTDOWN: Once you have some sort of failure detection system that will tell you that the Marelli
has quit firing one bank, there is still the question of what to do with the indication. You could provide an idiot light,
but anyone that drives despite a sudden lack of horsepower may likewise drive despite a warning light; some XJ-S
owners allow others to drive their cars and don’t want to have to lecture them on the importance of a particular idiot
light each time. Another possibility is to have the sensor shut down the car. If Corkan’s ignition monitoring circuit can
light one of the dash indicators, it could just as well be used to operate a relay to shut off power to the ignition -- but
you’d have to install the relay so that it doesn’t cut off the ignition when the ignition switch is in the “start” position or
you’ll never get the car started. The OEM starter relay has a contact that will work perfectly, left over from the days
when a ballast resistor was used with a coil but was bypassed during starting; just use the same contact to bypass the
cutout relay.

A better idea, though, is to use the relay to shut off power to the EFI system or to the injectors themselves. That way
starting won’t be a problem; as soon as the starter turns the engine over and the system detects some ignition pulses, the
EFI will be enabled and the car will start.

LIMP-HOME MODE: Automatically shutting off the engine in the event of six-cylinder operation protects the car but
aggravates the driver, especially if it happens in traffic or in a bad neighborhood. A better idea is to shut off fuel to the
bank that has lost spark, thereby making it safe to drive the car on six cylinders. That same relay idea should work for
this, using Corkan’s circuit with two separate Q3’s (you can omit diodes D5 and D6 if each subcircuit controls its own
Q3) to control separate fuel injector shutoff relays for each bank. Starting should take care of itself; as soon as sparks
are detected in each bank, that bank’s fuelling will be enabled and the engine will start.

The idea of shutting off fuel to one bank and continuing to run does have one important benefit: it is likely to provide a
better cooldown of the overheated parts. Several owners, having already been made aware of the Marelli problems,
nevertheless reported that the right side cats were cherry red hot by the time they figured out what was happening, shut
it off, and got the hood open. When the engine is simply shut off under such conditions, the residual heat within the cat
may nevertheless fry nearby rubber parts, sensors, wires, etc. If the fuel is shut off and the engine continues to run,
cooling air continues to flow through the engine compartment, and air continues to flow through the cylinders of the
inop bank and out through the exhaust system. A running cooldown is better on the engine and ancillaries than a hot
shutdown. Perhaps even if a sensor provided only a warning light indication, a switch could be provided to manually
shut off fuel to the inop bank to allow the car to continue moving.

CATALYTIC CONVERTORS -- MELTDOWN/FIRE (SENSOR-CAUSED): Greg Maddison lamented a large loss
to his wallet: “The front two had melted into a solid mass and the debris ruined the back two.” Obviously, this

168

symptom differs from the typical Marelli six-cylinder operation described above, since the Marelli failure will toast the
cats on one side only.

“It seems there is yet another reason the Marelli system can cause a misfire that brings the cats up to kiln temperatures.
This would be problems with the TDC sensor/magnets. There are 3 equally-spaced bar magnets around the crank
damper and a sensor that "looks" at the magnets and sends a pulse directly to the Marelli computer. There is a specific
air gap between the magnets and sensor that must be maintained for the system to work properly.

The Jag dealer replaced the damper/magnet assembly on my car saying the rubber had split. This caused the air gap to
be too wide between one of the magnets and the sensor. They first tried to grind down the other two magnets and move
the sensor closer (sounds like the damper deformed and became eccentric), only lasted 600miles so I had to have
damper replaced. The dealer's mechanic explained to me that each magnet produces a timing pulse for 4 cylinders. If
one of the pulses are missing, 4 cylinders don't get spark and send unburned fuel through the cats. The mechanic said
that a different damper is fitted to cars with the 100-amp alternator (like mine) and fails more often than the older cars.”
Note: the 100-amp alternator requires a grooved belt instead of a V-belt, and the alternator belt is driven by the damper
itself rather than the bolt-on pulley, so the 100-amp alternator does require a different damper. Perhaps it fails more
often because of the increased torque load of the alternator being applied to the rubber.

“After removing the crank pulley they found the woodruff key sheared and cone split, pulley had rotated 6 degrees on
the shaft. Mechanic says the Marelli computer can vary the timing up to +/- 28 degrees. So, at this point I don't know
if they misdiagnosed the sheared key problem as a damper problem (they are claiming both are toast) all along. I can
certainly understand how having the pulley rotate on the shaft can introduce timing errors on a Marelli car.”

There are a few problems in the explanation offered by Maddison’s mechanic. First off, the tabs on the pulley are
attached to the center portion, not to the outer portion held in place by the rubber -- but they might still be damaged by
the rubber shearing. Chris Carley explains: “The timing tabs are a ring with 3 tabs sticking out every 120 degrees
bolted to the back of the pulley attached to the solid middle section and then the tabs are bent over the edge of the
pulley. There is a 1/2 inch of pulley behind the alternator belt (ribbed), where these are bent over. If the rubber
sandwich breaks down then the outside ring of the pulley will slop around and bash the tabs out, which could then hit
the sensor (.028 to .042 gap). Whether that problem sheared the key - who knows?” Whatever, make sure your pulley
is in good shape and securely torqued down (see page 89); somehow, problems here can evidently toast your cats.

CATALYTIC CONVERTORS -- MELTDOWN (SECONDARY DAMAGE): Greg Maddison says, “Also check the
engine speed sensor located on the flywheel housing. It is a $40 part located very near one of the catalysts. Mine
melted when the catalyst overheated. After that the speed sensor became temperature sensitive, car ran well when it
was cold. After it warmed up the car wouldn't develop any power and couldn't go over 40 MPH.”

Tim Jones had another secondary problem. After having the cats replaced and the car put right, it started issuing white
smoke from the right exhaust pipe. This is an indication of a blown diaphram in the vacuum modulator on the GM400
transmission; the vacuum line sucks transmission fluid out of the transmission and into the right manifold, causing the
smoke in the right bank only. Apparently the catalytic convertor meltdown generated enough heat to roast the vacuum
modulator.

INTERMITTENT FAILURES: Greg Maddison suffered from electrical connections: “Occasionally when warm/hot
the engine just stops. It will usually restart after a few tries or if left for a couple of minutes. It turned out to be the
crankshaft sensor connector. I could jiggle the crank sensor and car would start again. If you have the Marelli digital
ignition system you will have a crankshaft sensor and a flywheel sensor; engine will not run if the signal is interrupted
from either. Look for a two-wire connector on the left front top of the engine; wires should be in a hard plastic wire
protector running down the front of the engine to the back side of the crankshaft pulley. I found the contacts in this
connector to be green and nasty. Cleaning with spray-on relay and contact cleaner cured the problem. Run the engine
and wiggle this connector and see if it stops. Also check the flywheel sensor connector, it’s on the rear top left of the
engine, looks like the crank sensor connector and hard to get to.”

169

Steve Sarmanian had the same problem, only different. “In the 2-wire upper connector of the front crank sensor (1
white wire, 1 blue), about 8 inches back into the top harness the white wire connects to the wire braid that surrounds the
sheathing around the blue wire by a solder connection. That connection is protected by a clear plastic sheath (rubbery
plastic tubing about an inch long). The connection is bad, and bad from the factory. It fooled me the first time around
because underneath the clear tubing the wiring looked good. But this time I split open the tubing with a razor blade and
the white wire fell off the braid. There's no good mechanical connection there; when the loom is built they apparently
just strip the end of the the white wire, lay it on top of the braid and melt on the solder. The tubing holds the joint
together. I confirmed that that was the cause of my problem: touch the wires, the car will start; separate them and the
engine dies.

“The fix is to twist the sheath into a secure mechanical connection with the white wire, crimp it with a slip connector
and then resolder. Problem disappeared.

“I unwrapped the rear sensor to check the braid connection on that one (near the firewall) and saw that someone had
been in there before me with a mechanical fix, but no solder. I fixed it.”

John J. Lynch had similar symptoms, but the cause turned out to be the crank sensor itself: “The problem went
something like this:

1) Car started fine

2) Ran normally for 15 minutes

3) Died completely due to spark failure

4) Engine cooled down for 1-4 hours

5) Repeat step 1

“Turns out the Engine Speed Sensor was the real culprit. The specs say that the sensor should read about 700 ohms but
when I measured it after the engine died I noticed it was only reading about 260 ohms. Since I had previously tested
the sensor with an O-Scope and saw that it was producing the square wave I was expecting, I thought everything was
fine. I was wrong. For some reason when the engine warms up the sensors resistance is dropping below 700 ohms
causing the voltages that are delivered to the ignition computer to be way out of spec. Consequently the computer was
shutting down the ignition amps. For kicks I decided to put a 500 ohm resistor in series with the sensor to validate my
new conclusion and sure enough the car ran fine. As soon as I removed the resistor, the car died again. Problem
solved!!!”

LACK OF OOOMPH: Jim Householder reports on “a problem I had with my 1991 XJS; cost me a lot of money and
stumped all the Jag mechs. At 4500 RPM the engine would just flatten out and go no more, but ran quite well up to
this speed. The story is long and sad about trying to get it fixed but decided I would try it myself. Turns out that the
number 2 coil was bad, not a big deal but seems none of the Jag pros had run into this problem.” Interestingly, running
out of juice at 4500 RPM is exactly what you would expect from a secondary coil failure on the earlier cars with Lucas
CEI ignition, but it’s not obvious how it can occur with the Marelli.

1992-ON: Richard Mansell quotes from a Jaguar Publication on the changes for the 1992 model year:

“The Marelli digital ignition ECU has been software upgraded.”

REPLACING THE MARELLI IGNITION SYSTEM: Since the Marelli has some serious problems that can cause
fires, obviously there is incentive to replace this system as a preventative measure. Those fires sometimes destroy
considerable portions of the Marelli system itself, though, and Marelli hardware is quite expensive, so owners having
already suffered such fires also seriously consider a change of system -- not only to avoid a repeat but because they
have a budget for such a replacement.

If your objective is to avoid even the slightest possibility of a fire, you’re out of luck; any car with any ignition system
can have a fire if things go too far wrong. If your objective is to eliminate even the slightest possibility of having a one-

170

bank failure of the type the Marelli system is infamous for, the only way is to replace the Marelli system entirely; as
long as there’s effectively two six-cylinder ignition systems, there will always be the possibility that one will fail while
the other keeps working. Before you latch onto that for a solution, though, you might wish to review the first fact and
realize that the Marelli, if you can fix the rotor and cap problems, may actually be relatively safe. This author suggests
your money might be better spent on an overtemp sensor for the catalytic convertors.

If you have decided that two six-cylinder ignition systems are totally unacceptable, there are two other types of system
that will work on a V12: a single 12-cylinder system or a distributorless system.

REPLACING THE MARELLI IGNITION SYSTEM PART I: The only replacement ignition systems that will
eliminate the problems with the Marelli distributor are ones that eliminate the Marelli distributor. One idea is a
distributorless system -- six double-ended coils, each connected directly to two spark plugs.

One source is Electromotive; see page 713. The installation manual for their HPV-1 includes instructions on how to
install two six-cylinder systems on a V12. You are once again back to two six-cylinder systems, one of which can fail!
But at least you won’t have the Marelli rotor to worry about any more.

Another source is...Jaguar. As discussed below, the final few months of XJ12 production came with a Nippondenso
distributorless ignition system. Yeah, that’ll be a cheap retrofit. Judging from Roger Bywater’s and Rick Wilder’s
descriptions of this system, it‘s two six-cylinder systems again!

A distributorless system can have a failure in one part of the system; for example, one of the six double-ended coils
could fail, killing two cylinders in one bank. But that would cause a lumpy-running engine, hopefully getting the
driver’s notice. It wouldn’t keep running smoothly as the Marelli one-bank failure does unless the three coils serving
one bank happened to fail together, which is highly unlikely. But if the distributorless system actually consists of two
six-cylinder systems -- as apparently many of them are -- you obviously could lose one control circuit and have a one-
bank failure, just what you feared with the Marelli.

If the system you install happens to be a “total engine management” system (ignition and EFI together), you may be in
luck. It’s still likely to be two six-cylinder systems, but the ignition pulses from each six trigger the injectors for each
six. So, if the ignition for one bank fails, it might cause the injectors for that bank to quit operating -- providing a safe
limp-home mode.

REPLACING THE MARELLI IGNITION SYSTEM PART II: Another idea is to retrofit the earlier Lucas CEI system
into the later cars. It’s kind of a shame to replace the high-tech electronically-controlled Marelli with the comparatively
archaic Lucas with its mechanical advance controls, but it will definitely eliminate the possibility of one-bank ignition
failures. If anything in the Lucas ignition system quits working, the car stops; one-bank failures are simply not
possible.

Note that the centrifugal advance mechanism in the Lucas distributor should be attended to prior to installation, as
described beginning on page 137. This mechanism then requires periodic maintenance (oiling) to make sure it doesn’t
seize, but this isn’t really a problem unless you don’t do it.

This retrofit has the additional benefit of saving money on future tune-ups. The Lucas caps and rotors aren’t cheap, but
they’re cheaper than the Marelli items.

Whit Coleman at British Parts International (page 689) reportedly has assisted several owners with installing Lucas
ignition systems in Marelli-equipped cars -- after the fire. The motive was cost, pure and simple; after the car has
suffered an engine compartment fire, it’s cheaper to convert to a Lucas system than it is to buy the Marelli parts needed
to get it going again -- possibly because the Lucas stuff is more readily available at junkyards. Coleman claims that
people usually obtain an entire donor car to do this job.

Coleman asserts that the change requires replacing the EFI ECU and the associated wiring harness with the items used
with Lucas ignition systems. He’s not entirely sure why, but if you’re using a donor car you’ve got those items on hand