When one considers the Ford IAR series of alternators,
then examines the new 4G alternator, it seems that Ford went overboard
to correct all the problems with the IAR series.
The IAR series was designed for rapid assembly
and disassembly with plenty of snap-on connectors and mounting
hardware. With this design approach came problems, particularly
with the burnt out B-S-S connectors. The stator was easily removed
by simply disconnecting spade lug connectors. Remove a couple
of screws and "presto," the rectifier is in your hands.
The same holds true with the voltage regulator.
Many believe in the adage, "the best connector
is no connector." How many connector problems have you encountered
in an automobile? Corrosion, poor spring contact, or connectors
not firmly seated or fastened probably are the major problem areas
found in troubleshooting many electrical systems.
Connectors are used for convenience, the idea
being to ease replacing a defective module. Ironically, a corroded
connector can actually be the cause of the module burning out.
If the alternator rectifier output were soldered directly to the
battery post, (omitting a sulfated battery problem), load dump
would be eliminated, as the greatest load dump voltages occur
with a poor battery connection. This will not eliminate the need
for load dump protection, however, unless someone can figure out
how to eliminate all of the heavy loads, such as the starter,
windshield wipers, power motors, etc.
The flyback diode is mounted within any solid
state voltage regulator. A poor field connection exposes the entire
electrical system to huge voltage transients. The slip ring brushes
are another form of connector relying upon adequate spring pressure,
sufficient brush length and clean rings for good contact. Without
these, the voltage regulator can experience damaging voltage transients
as the protecting diode is removed from the circuit. Outside of
the brush-ring contact and vehicle connector, the 4G eliminated
all connectors. That may have improved reliability, but it made
it more difficult to rebuild; with this unit you can’t have both.
Taking a closer look
Let’s take a closer look at this unit. The
first issue deals with pulse testing diodes at 70 amperes in a
CS-130 rectifier. Common sense should tell you that each diode
must carry, in turn, the full output current. As this current
is at least 105 amperes, testing at 70 amperes is worthless! In
reality, each diode conducts for only a fraction of its allotted
time, so the peak or pulse current in practice would be much greater
than 105 amps, more like 150 amperes! Load dump currents are even
greater! Pulse testing should be at least two to six times the
rated load of the diode, in other words, test at 300 amperes minimum
to detect if faults are present.
I once read an article which stated that only
three of the six diodes in a CS-130 rectifier should be of the
avalanche type, as "the diodes are in series, and, therefore,
the load dump clamping voltage would be twice as much." First,
consider the source of the voltage – the stator. Depending on
when the load dump takes place, each stator will have its own
unique degree of magnetic energy stored in the stator core. The
polarity of the stator could either be positive or negative and
the amplitude of the self-induced voltage is magnetic-dependent.
As the stator is the source and only source of load dump voltage,
it is how the diodes are connected in relationship to the stator
that determines their connection.
In Figure 1, the left side of the drawing shows
a standard method of drawing a diode stator circuit. Initially,
it looks like the positive and negative diodes are in series,
giving truth to the above misconception. But if you trace the
diodes in relationship to the voltage generating stator, as is
done in the right side of Figure 1, you will see that the clamping
diodes are in parallel.
If D5 and D6 were conventional diodes, either
D1 or D2, (depending upon the polarity) would have to clamp the
entire stator output; thus one diode will be doing the work of
two. D6 and D5 would always be reversed biased and would not conduct
during load dump as their breakdown voltage is much greater than
an avalanche diode.
There is another old adage which says that
noise should be treated at the source. Avalanche diodes are great
in this respect. In addition to acting as normal diodes for rectification,
they serve a dual role in clamping load dump transients at the
source (stator). Load dump occurs by disconnecting the alternator
output lead while running at full speed and drawing maximum output
current. It is measured with an oscilloscope connected directly
to the alternator output terminal and must remain there when the
load is removed.
I measured about 30 volts peak, testing a six
avalanche diode CS-130 rectifier, and not 60 volts as the article
I had read suggested. In a three-avalanche diode circuit, if the
energy absorption rating of each diode is not twice that of a
six-diode circuit, the load dump will fry the diodes. But don’t
take my word for it, try it out for yourself; it will be your
Another article I read stated that using an
LRC (Load Response Control) voltage regulator in an non-equipped
vehicle is not recommended as it affects the driveability. However,
the author failed to mention who does not recommend making this
change. If the OEM or an OEM dealer does not recommend using an
LRC regulator, is it because they don’t have one available? Is
it because they might lose a sale on an OEM part?
A call to a Denso dealer revealed that the
cost of a replacement regulator was in the $150 range, where an
excellent aftermarket unit can be purchased for about $10. My
own interpretation is that while having a broken ball joint or
a flat tire would certainly affect vehicle driveability, switching
to an LRC unit would not.
An ideal voltage regulator would supply a constant
field current in precise proportion to load demands. Such a unit
would require a huge expensive voltage regulator, but it would
have the advantage of minimal engine loading and EMI (Electro-Magnetic-Interference),
such as found in the aviation field.
Automotive regulators switch the field "hard
on" and "hard off" using a much smaller field driver
circuit; this causes pounding of the rotor with constant jerks.
This pounding definitely shortens the life of the alternator as
the torque jolts stress the bearings and drivetrain. PWM (Pulse-Width-Modulation)
converts a single hard pound to a series of rapid light pounding
pulses, greatly smoothing out alternator operation.
LRC added to PWM applies the rapid pulses gradually
to dampen out application of a heavy load. LRC does not change
the charging output of an alternator, but adds to its life by
removing harmonic jerks. LRC is disabled at higher engine speeds
where it is not needed.
Adding an LRC regulator is like adding shock
absorbers to a car that never had them, but at practically no
increase in cost. The benefits are smoother engine idling and
longer drivetrain life. LRC is designed into the voltage regulator
and is transparent to the vehicle. A good case in point is the
Ford 4G regulator that has LRC. If the substrate of this regulator
were mounted in an IAR lead frame, the vehicle connections would
be identical, but these earlier vehicles would gain the same advantage
of having LRC, i.e., smoother engine idle and longer drivetrain
life. But again, don’t take my word for it, try it for yourself.
There have been several articles in other publications
which tried to define "soft-start," though, in my opinion,
caused more confusion than clarification. If you recall, the Delco
10SI circuit field current was supplied by the diode trio. As
the trio is stator connected, no alternator rotation meant no
stator voltage, hence no field voltage and, consequently, no field
If you accidentally left your ignition switch
on in a 10SI equipped car, no rotor damage would occur and the
load on the battery would be about 1/4 ampere. With an HEI ignition
system, you could leave the ignition switch on for hours without
discharging the battery. But the 10SI required not only the diode
trio, but a 40 ohm power resistor as well for initial field excitation,
a cost that most OEMs wanted to avoid.
Many OEMs simply full-field the alternator,
which adds starting torque to the engine and which will discharge
a good battery if the ignition switch is left on. A cheap and
dirty solution? Turn off the ignition switch and you will not
have a problem.
With the advent of high-output alternators,
the problem has been compounded. With field currents in excess
of 8 amperes, not only will the battery discharge, but the rotor
coil will burn up. Remember that the rotor fan is not turning
in a stalled engine and is not providing needed cooling air. With
100 watts of rotor power in a closed space, temperatures in excess
of 300° F or higher will occur if the stalled engine is hot,
not only burning out the rotor, but taking out the electronics
A simple solution to the stalled engine problem
is to apply a fixed duty cycle waveform to the rotor when the
stator frequency is below about 50 Hz; normally a 32% duty cycle
is used. Due to the inductance of the rotor, the actual average
field current is about 160 milliamperes, even with an 8 ampere
rotor. This low current protects both the alternator and battery,
if the ignition switch were on, limiting rotor power to a low
In an LRC-PWM voltage regulator, soft-start
or the fixed-duty waveform is easy to accomplish by locking the
PWM at a fixed level. This locking eliminates the need for the
diode trio and the series power resistor, but serves the same
The actual frequency of the soft-start is irrelevant.
I connected a pulse generator to a field driver transistor, kept
the duty cycle constant at 32%, but varied the frequency from
40 to 700 Hz. I observed no appreciable change in field current.
I can offer no logical reason as to why different OEMs use different
frequencies, except to say that whatever frequency they are using
is used in another part of their chip.
Ironically, the Ford 4G does not use soft-start,
but full fields the rotor when stalled. If you get a burnt 4G
returned, suspect that the user left on the ignition switch. You
can recharge a battery, but you cannot unburn an alternator.
Delco engineers (except for the rolled in front
bearing and the over-pushed design envelope) did a fantastic job
in introducing many innovations, including LRC and soft-start.
Proof of their innovations lies in the fact that many other OEMs
are following their lead. Both Denso and Ford recently introduced
LRC, but only Denso has soft-start; maybe Ford will add it next
Figure 2 shows the electric wiring diagram
of the 4G. The major difference of the 4G to the IAR is the addition
of the two center tap diodes and the deletion of the "S"
terminal connection from the regulator vehicle connector. These
are just the electrical differences. When comparing schematic-to-schematic,
mechanically, the components are substantially different. There
are no interchangeable parts between the 4G and IAR series.
The center tap diodes, D4 and D8, add about
10% to the output current. In other words, expect about 135 ampere
output without the center tap diodes and about 150 with them.
The center tap diodes only conduct when the output load has a
battery. The battery is required to "series aid’ the 1/2
stator voltage to overcome the full stator voltage for conduction.
The "S" terminal has been removed
from the vehicle connector socket and replaced by an embedded
ring terminal that contacts the rectifier assembly directly. This
is a pressure contact dependent on the tightness of one of the
three regulator mounting screws. Another screw provides ground
contact, while the third connects an unused embedded terminal
to the positive rectifier plate.
I can only speculate why the battery contact
screw is not used. The "A" vehicle terminal provides
all the power to the regulator, including the 5 ampere field draw.
It would make more sense to connect this terminal directly to
the rectifier plate via the embedded terminal as any conductor
carrying field current would generate plenty of EMI. The shorter
this conductor is, the less EMI would be generated.
My guess is that Ford is using the relatively
tiny vehicle "A" terminal so it can fuse it. The CS-Series
uses an internal fuse to prevent full fielding the alternator,
where Ford must be using an external fuse. In any event, the vehicle
"A" terminal is carrying a lot of current and must be
kept clean. The female spade lug must be tight, with at least
a five pound pull for removal.
Figure 3 on page 44 shows a representative
drawing of what’s inside the 4G regulator. Again, this circuit
is very similar to the IAR. The major difference is the addition
of the LRC block to the IAR circuit. This circuit has a unilateral
integrating function in that when the battery voltage lowers,
it delays this change to the comparator. If battery voltage increases,
field voltage instantly decreases.
The 4G uses a voltage controlled PWM circuit
that may operate at a repetition rate anywhere between 100 to
230 Hz. Like the soft-start frequency, the exact frequency is
not critical. Delco uses about 400 Hz. What is critical is the
duty cycle of the PWM. Each cycle or period is defined as 100%
long. The amount of time the field voltage is on during this period
is the percent of duty cycle.
At 25° C, 14.5 volts or below, the field
is on 100% of the time. By increasing the voltage ("A"
terminal to ground) from 14.5 to about 14.8 volts the duty cycle
proportionally decreases from 100% to 10%. The 10% duty cycle
remains until the voltage is increased to 16.5 volts where the
field is completely turned off.
The range between 14.8 and 16.5 volts is called
the MPW (Minimum Pulse Width) range and causes the stator to generate
just enough voltage to keep the stator pin activated. Loss of
the stator signal would cause the warning lamp to ignite. Above
16.5 volts is considered the over-voltage condition where a chip
internal circuit reignites the warning lamp. The 14.5 volt "setpoint"
voltage drops 10 millivolts per each 1° C.
The warning lamp will ignite if the stator
signal is below six volts or if the frequency of the stator is
below about 80 Hz, or if the "A" to ground voltage is
above 16.5 volts. LRC is disabled above 270 Hz to complete the
stator frequency characteristics of the 4G.
In the older IAR regulators, the PWM would
change instantly with the "A" to ground voltage change
(in the 14.5 to 14.8 volt range). With the addition of LRC, (but
note that the stator frequency must be between 80 and 270 Hz)
and increasing the voltage from 14.5 to 14.8 volts, the 4G acts
exactly like the IAR in that the duty cycle changes instantly.
But decreasing the voltage from say 15 to 14 volts, where we would
expect the duty cycle to change from 10-100%, does not occur instantly,
but takes about 10 seconds to grow from the 10% to 100% value.
Without LRC, the sudden 20-amp load caused
by the radiator fan kicking in would jerk the engine due to the
alternator load. With LRC the alternator load is applied smoothly,
giving the ECU a chance to readjust the engine speed.
Confusion exists about the "I" terminal.
In electronics, adding a pin to a circuit not only adds the cost
of the pin, but all of the wiring associated with it. The dual
or triple use of a pin is a very common practice. In Figure 2,
the "I" pin is both an input and an output. When the
ignition switch is first turned on, the engine is stalled; that
is a warning lamp "on" condition.
Q2 in the chip is off causing both Q2A and
Q2B to turn on hard; the low side of R2 is at near ground. If
12 volts were directly connected to the "I" terminal,
this voltage would appear across R2 causing it to dissipate about
15 watts! R2 is only intended to carry a maximum current of about
1/4 ampere (the warning lamp load), and R2 will burn out. Worst
yet, the high 1.2 ampere current could burn out Q2A and Q2B that
are part of the same chip.
If these transistors should short the 1.2 ampere
current into the chip, the entire regulator will burn out. When
component voltage or currents are exceeded they are going to fail,
but the nature of the failure, for the most part, is unpredictable.
A simple rule to follow is: NEVER CONNECT THE "I" TERMINAL
DIRECTLY TO THE BATTERY VOLTAGE!
The impedance of R2 and Q2B is sufficient that
even with a 1,000-ohm load connected from "I" to battery,
the "I" to ground voltage would be at least 1.3 volts.
With about 1 volt applied to the "I" terminal, this
is sufficient to turn on Gate 1 in the chip via R1. R1 limits
the current, so with Q2B off, the "I" voltage could
rise to 16.5 volts or greater without chip damage.
Regardless of whether Q2B is on or off, the
output range is from 1.5 volts to whatever the battery voltage
is; the chip would be activated as long as the "I" voltage
is greater than 1 volt. This circuit is the same as that used
in the IAR series, so never apply 12 volts directly to the "I"
terminal. Use a 560-ohm half watt resistor in series with the
"I" terminal and the ignition switch for "I"
Dave Waters of Just Parts, Chicago, IL, was
kind enough to send a sample 4G alternator on which this article
is based. Waters said that they purchased a number of these alternators
from Ford that were line rejects. In testing this alternator’s
components, they were all found good, the only problem being that
the slip rings were not turned, leaving an insulation coating
the rings that prevented brush continuity. These alternators probably
were some of the first to come off the new production line and
were likely used for line debugging.
Disassembly can be tough
Disassembling the 4G is a chore. The 4G central
component is the rectifier assembly that is riveted to the rear
case half on one side, but has the stator with through leads mounted
on the other side. The stator is bifilar wound using two parallel
conductors. Each stator winding is isolated from the other with
the two paired ends protruding through the rear case half, and
then soldered into eyelets within the rectifier assembly.
To compound this, the stator core is pressed
into the front case halve, so to separate the two halfs you either
have to desolder the six stators through paired leads, or somehow
relieve the press fit. I elected the latter not having a solder
sucker large enough to fit over the bifilar stator leads. As the
eyelets are embedded in large plated-copper conductors, it takes
a lot of heat to melt the solder.
First, I removed the sheave nut using an impact
wrench that spun the nut off within a second. The grooved sheave
easily dropped off the rotor shaft. Then I took out the three
housing bolts, removing the rear shield. Another three screws
removed the voltage regulator-brush assembly. I placed the alternator
in a temperature chamber and when the temperature reached 290°
F, with gloves on, I set the alternator sheave side up on a floor-type
drill press. I placed a couple of two-by-four blocks under the
ears of the front case half, put a little pressure on the rotor
shaft, and the stator core-rear case half slid out easily.
Using a 250-watt soldering iron just on the
tips of the stator leads – being careful not to burn the rectifier
plastic – I alternated between the six lead sets. Holding the
stator in my left hand while placing thumb pressure on the rear
case halves, I worked the rectifier off of the stator leads in
less than a couple of minutes. I would guess that if you did not
want to save the rectifier, you could heat it until the solder
melted, and the whole alternator would come apart.
Transpo Electronics was kind enough to send
a sample rectifier assembly. Transpo supplies the rectifier mounted
to a new rear case half and the unit is identical to OE. I removed
the three rectifier-to-case rivets by drilling them out to learn
that the diodes are pressed into a deadend hole. Ford has several
patents on its rectifier. Interestingly, the first one used a
double-spade connector for the battery terminal. Ford switched
to the stud in a later patent.
Attempting to change a diode is labor intensive
the way Ford designed this rectifier. You have to first desolder
the diode lead end from the eyelet, then very carefully drill
out the old diode. Even at that, the diode mounting is formed
from .035ý thick copper stock that expands when the first
diode was pressed in. Pressing in a second diode would not provide
a good heat conductive solid connection and may result in a return.
This rectifier was not designed to be repaired.
The rivets are also soldered, providing a ground return from the
negative heat sink. You cannot replace them with screws; remember
that these three rivets can carry as much as 150 amperes!
Regarding the oval connector Denso I wrote
about earlier this year in my feature titled "Nippondenso
Oval Connector Alternators" (see Automotive Rebuilder’s March
issue, pages 56-61), the stator lamination was epoxied together
with generous epoxy on the outer periphery. However, the 4G stator
is welded together with the uninsulated periphery making conductive
contact to the front case half. In my opinion, this is not a very
I know engineers state that welding core laminations
is effective, but these engineers are mechanical, not electrical
engineers. These weld spots are hot spots due to the high eddy
currents. Neither Denso nor the Delco CS-130HD used uninsulated
shorted out laminations in their alternators. Ford would have
been wise to follow their example. The enclosed stator cores are
required to reduce EMI emissions to avoid ECU and radio interference,
and are becoming quite common. Heating these alternators for disassembly
The rotor has special shaped fingers to produce
a modified stator waveform. Like the Delco CS-series, the sinusoidal
peaks have been flattened to reduce the ratio of peak-to-average
output diode current. Using a standard sinusoidal waveform would
overheat the diodes, leading to early failure. Refer to Figure
4 on page 44 for the OE stator waveform showing a single phase.
The rear bearing is pressed in so be sure to use a suitable adhesive
compound to prevent dropout. The front bearing uses the conventional
Finally, on the rear protective shield, with
protruding stator and diode leads, a dent on the rear shield could
result in a major short circuit; some type of fish paper insulation
would be wise.
Assembly requires heating only the front case
half. If you solder the stator to the rectifier assembly first,
and clock the rear case assembly, you will automatically get the
proper stator alignment. As only the solder is the conductor,
the lead-eyelet joint must be hot and all of the flux must burn
off before removing the iron. Working fast, you can turn the rear
case until the three attaching bolts align. Add the regulator
last. Be sure the stator diode leads are cut short. Covering these
leads with a quick setting epoxy may prevent a short to the protective
Outside of replacing the voltage regulator,
the 4G is not a "do-it-yourself" repairable alternator.
Changing just the front bearing, for example, is a major chore.
But, time will reveal how reliable this alternator performs.