[xj-s] Alternator rebuild update

The story so far:

My battery died (probably of long term undercharging)

I discovered the alternator was wired wrong, i.e. not as per the ROM.

Fixed that, and voila a stable 14.0V (above 800RPM that is).

3 days later the alternator died completely. Diagnosed as a shorted rectifier
diode.

So today I went to the local alternator parts distrubutor, And told them the
story. “Yeah, they give problems, those alternators, in jags. The engine bay’s
too hot, and it kills them”. they said. " Not a good design of alternator
either, the air doesn’t flow properly".

OK I believe that.

“We may be able to get a heavy duty diode pack”

But I ordered some replacement diodes instead(NZ$7.50ea=US$4). I ordered some
1/2" ones to replace the pissy small 10mm ones. Same rating though - 25A/200V. I
figure the bigger body will transmit more heat to the diode plates.

A careful drill out and the plates and 5 minutes with the 10ton press, and
they’re installed…

I’m also throwing away the black plastic back cover (keeping the centre bit to
protect the brushes though), in order to enhance airflow. Asking for trouble
with exposed 12V, but too bad.

Looking at the design(?) of the alternator, the flow is from the back to the
front, the fan on the front acting as a centrifugal pump, drawing air through
the body, and past (i.e. not through) the diode plates. I can see they fry. The
hot air after leaving the unit, gets pushed back to the back by the hot air from
the radiator, to be sucked back through the alternator, getting hotter on each
pass.

I’m looking at adding a 12V ball bearing computer fan to the back of the
alternator to provide some serious air flow, ducted up from underneath.

Overkill, you say? Perhaps, but I want to kill this problem forever. Just
replace the alternator with a GM one you say? Not in this part of the world you
don’t. A Toyota or Nissan one is a possibility though. Or perhaps I’ll just
add an extra alternator, for that extra piece of mind.

Later.

I’m also throwing away the black plastic back cover (keeping the
centre bit to protect the brushes though), in order to enhance
airflow. Asking for trouble with exposed 12V, but too bad.

It might be a better idea to figure out how to modify that cover to
direct the incoming air directly toward parts that need cooling.

Looking at the design(?) of the alternator, the flow is from the back
to the front, the fan on the front acting as a centrifugal pump,
drawing air through the body…

I believe most alternators cool this way since the impeller is at the
front (under the pulley) and is inherently a sucker. However, there
is one notable exception: the modern CS130 GM alternator. It has the
similar impeller under the pulley, but it has a SECOND impeller
inside the alternator at the back, apparently dedicated to cooling
the diodes. It is my impression that this design came from a careful
rethink of alternator design, and the realization that the way to get
more amps from an alternator was NOT to make it bigger but rather to
improve the cooling. Even with the space required for a second fan,
this is a relatively small alt (smaller than the Lucas) and can be
assembled with a rating as high as 140 amp.

Unfortunately, after all that fine design work, GM screwed up and
made the rear ball bearing (not a needle bearing, as on earlier alts)
too small. But, this being a GM fault, somebody fixed that problem
right off the bat, and kits for installing a heavier rear bearing are
readily available.

The hot air after leaving the unit, gets
pushed back to the back by the hot air from the radiator, to be sucked
back through the alternator, getting hotter on each pass.

I’m looking at adding a 12V ball bearing computer fan to the back of
the alternator to provide some serious air flow, ducted up from
underneath.

Overkill, you say? Perhaps, but I want to kill this problem forever.

IMHO, adding a computer fan would be a complete waste of time. The
CFM rates of the typical computer fan vs. the impeller on the front
of the alt probably differ by orders of magnitude. Adding the
computer fan would probably not increase the airflow through the unit
by more than a couple of percent, and it might actually reduce the
airflow by being in the way.

A better idea, I suspect, would be to find a way to duct COOL air
towards the back of that alt. There are cars that do this from the
factory. In this case, it would most definitely be beneficial, if
you can think of a way to do it.

Just replace the alternator with a GM one you say? Not in this part of
the world you don’t.

GM parts are that hard to find? Offhand, I would expect Holden to be
using the CS130 in Australia, although somebody would have to
confirm. Can’t you get Holden parts there?

A Toyota or Nissan one is a possibility though.

Yeah, I looked at Toyota and Nissan alts when I was considering a
replacement. Offhand, I couldn’t figure out how to mount one without
a lot of fiddling, but I didn’t get that far into it. Note that the
problem wasn’t mounting it – that was easy – it was getting the
pulley to line up. The Japanese alts all seemed to position the
pulley quite a bit farther forward than the Lucas alt did, so to get
the pulleys to line up you have to figure out how to move the mount
about a half inch rearwards.

– Kirbert | Palm’s Postulate:
| If anything is to be accomplished,
| some rules must be broken.
| – Kirby Palm, 1979From: Tony Bryant zot@paradise.net.nz

I believe most alternators cool this way since the impeller is at the
front (under the pulley) and is inherently a sucker. However, there
is one notable exception: the modern CS130 GM alternator. It has the
similar impeller under the pulley, but it has a SECOND impeller
inside the alternator at the back, apparently dedicated to cooling
the diodes. It is my impression that this design came from a careful
rethink of alternator design, and the realization that the way to get
more amps from an alternator was NOT to make it bigger but rather to
improve the cooling. Even with the space required for a second fan,
this is a relatively small alt (smaller than the Lucas) and can be
assembled with a rating as high as 140 amp.

I started off thinking that there couldn’t possibly be a problem with
the diodes burning out. However…

Let’s say that the alternator pushes out 80A. The diodes will drop a
volt or so at those currents, so a hefty 80W will be dissipated. Now
there may be the odd root-three, or root-2 in there with it being a
3-phase rectifier stack, but lets ignore that for now.

Maximum junction temperature for a silicon diode is 150C, or say 130C
above nominal ambient. That means that the thermal resistance has to be
less than 1.6 C/W, which is a pretty tall order (stud mounted rectifiers
have a thermal resistance from junction to case of about 1 C/W before
you even get as far as the heatsink).

The conclusion is: thermal management of the diodes in an alternator is
a big deal: efficient forced air cooling is absolutely essential. Given
the cramped space, and the hot environment, it seems all to easy for a
burn-out to occur.

Craig

Craig Sawyers wrote:

… I started off thinking that there couldn’t possibly be a problem with
the diodes burning out. However…

Let’s say that the alternator pushes out 80A. The diodes will drop a
volt or so at those currents, so a hefty 80W will be dissipated. Now
there may be the odd root-three, or root-2 in there with it being a
3-phase rectifier stack, but lets ignore that for now.

You might rethink this paragraph and numbers offered. A good start would be
to get the specifications for the diode in question and get a schematic of
the rectifier bridge circuit. The junction voltage drop would leave the
battery sourcing current and not sinking it. End results a is dead battery.

Ned

had two diode failures out of six due to stress on leads, one was
intermittent after being reassemabled. new assy had S shaped leads for
stress relief, this has work ok for now. John 86 xj-s

It might be a better idea to figure out how to modify that cover to
direct the incoming air directly toward parts that need cooling.

It probably would, but its a little hard to tell if it helps or not.
Its probably not a helper. They problem is the diodes.

The diodes are not schottky, therefore they tend to drop about 1V @ 75A. Two
diodes conducting at a time, = 2 X 1 x 75 = 150W. The diode plates are the heat
sink. They are certainly not better than 1 degC/W heatsink, giving a 150 degree
rise @ 150W, on top of an worst case ambient of around 70deg in the engine bay.
= 230deg C. Diodes are usually only rated to 150-200deg C. I can see why they go
fut.

I’m tempted to buy 6 50A schottky stud diodes, maunfacture my own diode plates
(from beefy heat sinks), arranged to duct air properly thorugh them. Schottky
diodes typically drop 0.5V at high currents, thus taking the power down to a
managable 75W.

I believe most alternators cool this way since the impeller is at the
front (under the pulley) and is inherently a sucker. However, there
is one notable exception: the modern CS130 GM alternator. It has the
similar impeller under the pulley, but it has a SECOND impeller
inside the alternator at the back, apparently dedicated to cooling
the diodes. It is my impression that this design came from a careful
rethink of alternator design, and the realization that the way to get
more amps from an alternator was NOT to make it bigger but rather to
improve the cooling. Even with the space required for a second fan,
this is a relatively small alt (smaller than the Lucas) and can be
assembled with a rating as high as 140 amp.

The design of the CS130 diode pack includes a healthy looking heat sink,
go to www.transpo-usa.com, and see the pictures. The A133 relies on two
small alum. plates, not actually in the air flow path. Despite that, a little
research on the net shows a plethora of replacement HD diodes assemblies
for the CS130. Can’t be that good.

Overkill, you say? Perhaps, but I want to kill this problem forever.

IMHO, adding a computer fan would be a complete waste of time. The
CFM rates of the typical computer fan vs. the impeller on the front
of the alt probably differ by orders of magnitude. Adding the
computer fan would probably not increase the airflow through the unit
by more than a couple of percent, and it might actually reduce the
airflow by being in the way.

A better idea, I suspect, would be to find a way to duct COOL air
towards the back of that alt. There are cars that do this from the
factory. In this case, it would most definitely be beneficial, if
you can think of a way to do it.

The computer fan would be arranged to duct cold air onto the diode plates,
thus increasing the cooling by a order of magnitude. The big problem is crawling
in traffic in summer, both fans, the air con, the stereo, the brakes lights all
pulling a good solid 45A, with no other source of cold air.

GM parts are that hard to find? Offhand, I would expect Holden to be
using the CS130 in Australia, although somebody would have to
confirm. Can’t you get Holden parts there?

I wonder, but we’re still talking expensive parts. Probably significantly
cheaper to buy one from the US.

Yeah, I looked at Toyota and Nissan alts when I was considering a
replacement. Offhand, I couldn’t figure out how to mount one without
a lot of fiddling, but I didn’t get that far into it. Note that the
problem wasn’t mounting it – that was easy – it was getting the
pulley to line up. The Japanese alts all seemed to position the
pulley quite a bit farther forward than the Lucas alt did, so to get
the pulleys to line up you have to figure out how to move the mount
about a half inch rearwards.

The plan is to add a pissy & cheap ~45A toyota, as an auxillary, in the
“air-pump” location , which on my car is a mere idler pully. Whatever I
do here, a custom bracket is called for anyway. Nothing like a little
redundancy to ease your mind, and get you home.

I am interested in the exact wiring of your alternator. On my Lucas 20 ACR I
originally had the body of the regulator going to the center brush, the
yellow wire going to the outside brush, and a black wire going to the ground
(frame of the alt). A separate wire goes from the outside brush to the IND
terminal. I have seen the 4 wire kind of regulators where a 4th red wire is
supposed to sense the battery or “machine” but I have no idea where they are
intended to connect. What was yours like? What model is yours? I am told the
higher the amp rating of the diodes, the better. They allow more flow
without heating up and burning out. 25A is awfully small when I look at a
web site like http://www.transpo-usa.com/

Yes 25A is awfully small. But its stock. Been to transpo, and they offer
30A replacement rectifiers, probably the wrong form factor though.

Its a Lucas A133-75A from an '85 HE.
The regulator is a ground switching unit, i.e. it switches in and out the field
ground connection. The regulator has three connections:

Body
Field
Goes to the centre brush.

Yellow
Voltage Sense
Now goes to the main output (battery). Did go the the excitor
diodes (wrong)

Black
Ground
Goes to the alternator body.

The rear brush connects to the excitor diodes, and also to the small wire that
goes to the back of the unit, from the dash light.

For those not in the know, The excitor diodes are the set of low current diodes
that feed the power to the field windings.On Thu, 17 Feb 2000, you wrote:

From Kirbert

I believe most alternators cool this way since the impeller is at the
front (under the pulley) and is inherently a sucker. However, there
is one notable exception: the modern CS130 GM alternator. It has the
similar impeller under the pulley, but it has a SECOND impeller

For info:
V12 E types have the alternator mounted “back to front” - ie the non driven
end of the alternator faces the radiator. Impellor is at the driven end -
never noticed if it pushes or pulls - if it pulls its a neat design. Food
for thought?

Steve (jagster)

You might rethink this paragraph and numbers offered. A good
start would be
to get the specifications for the diode in question and get a
schematic of
the rectifier bridge circuit. The junction voltage drop
would leave the
battery sourcing current and not sinking it. End results a
is dead battery.

Please don’t patronise me, Ned.

Unless those diodes are power Schottky barrier diodes (they aren’t, but
ought to be) or GaAs (they definitely aren’t that!) they will drop a
nominal 1V at 20A, or 20W per diode. That is what silicon rectifiers
do, period.

Now since the current actually varies over a recified half-cycle, the
average dissipated power has to be derived by integrating the current
waveform with the voltage (via the I=I0exp(qV/kT) law). So the above
20W per diode is probably an overestimate. I can’t be bothered to do
the integration, frankly; give it a bash and let us all know the result.

As for configuration, there are six diodes mounted as a full-wave, three
phase bridge, with a further three diodes supplying current to the rotor
windings. Hence my uncertainty on the final multiplicative factor; but
if you really want a definitive answer, look up the numbers in a power
engineering book.

But at the end of the day, the diode pack will dissipate many tens of
watts when the alternator is providing its rated current, which is
difficult to dissipate within the junction temperature maximum of
silicon power diodes. And that is the point I was trying to make.

Craig============================================
Dr Craig Sawyers

Technology Enterprise

e-mail @Craig_Sawyers1
web www.tech-enterprise.co.uk

Craig Sawyers,

I am sorry that you a hurt by my response. It was just a comment on an area
that probably needed to be expanded to be understood by most “List”
members. When I write technical documents I adjust the level of for the
reader. Certainly no one expect you or me to prove or disprove Jaguar’s
design. We would like to simply know what is required to keep it going.

Unlike most of the XJ-S Jaguar owners that I respect, I will not feel a
thing when time comes to sell one or both of my Jaguars. I am not that
attached to any car, but the “List” and you helps me to keep the cars
going. For that I am thankful. I also hope that I am helping in return.
Also, I am not accustom to insulting a group or an individual that is
saving me money.

Based on the level of your initial input, I assumed that you are a BSEE or
higher. Correct?
The response I expected is: “you want design data, show me the money.”

Again,

Ned

Craig Sawyers wrote:> Please don’t patronise me, Ned.

Unless those diodes are power Schottky barrier diodes (they aren’t, but
ought to be) or GaAs (they definitely aren’t that!) they will drop a
nominal 1V at 20A, or 20W per diode. That is what silicon rectifiers
do, period.

Now since the current actually varies over a recified half-cycle, the
average dissipated power has to be derived by integrating the current
waveform with the voltage (via the I=I0exp(qV/kT) law). So the above
20W per diode is probably an overestimate. I can’t be bothered to do
the integration, frankly; give it a bash and let us all know the result.

As for configuration, there are six diodes mounted as a full-wave, three
phase bridge, with a further three diodes supplying current to the rotor
windings. Hence my uncertainty on the final multiplicative factor; but
if you really want a definitive answer, look up the numbers in a power
engineering book.

But at the end of the day, the diode pack will dissipate many tens of
watts when the alternator is providing its rated current, which is
difficult to dissipate within the junction temperature maximum of
silicon power diodes. And that is the point I was trying to make.

Craig

============================================
Dr Craig Sawyers

Technology Enterprise

e-mail c.sawyers@tech-enterprise.co.uk
web www.tech-enterprise.co.uk

Now since the current actually varies over a recified half-cycle, the
average dissipated power has to be derived by integrating the current
waveform with the voltage (via the I=I0exp(qV/kT) law). So the above
20W per diode is probably an overestimate. I can’t be bothered to do
the integration…

Pardon me for having weak electronic skills, but can’t the dissipated
power be figured more simply by noting that the alt is rated at 75
amp, the voltage across the diode pack is 1 volt, so the wattage
dissipated per diode would be (75 x 1) / 6 = 12.5 watts? Or does
that leave something out?

– Kirbert | Palm’s Postulate:
| If anything is to be accomplished,
| some rules must be broken.
| – Kirby Palm, 1979From: “Craig Sawyers” c.sawyers@tech-enterprise.co.uk

that probably needed to be expanded to be understood by most “List”
members. When I write technical documents I adjust the level
of for the
reader. Certainly no one expect you or me to prove or
disprove Jaguar’s
design. We would like to simply know what is required to
keep it going.

Well, I broadly agree with that. However, the explanation has to be
sufficiently complete; I have a strong dislike of explaining things so
simply that it misleads. And in doing so it simply invites a more
technical debate by the tecchie list members.

Based on the level of your initial input, I assumed that you
are a BSEE or
higher. Correct?

1st class honours BSc in electronics, and a PhD in laser physics.

Craig

From: “Craig Sawyers” c.sawyers@tech-enterprise.co.uk

Now since the current actually varies over a recified half-cycle, the
average dissipated power has to be derived by integrating the current
waveform with the voltage (via the I=I0exp(qV/kT) law). So the above
20W per diode is probably an overestimate. I can’t be bothered to do
the integration…

Pardon me for having weak electronic skills, but can’t the dissipated
power be figured more simply by noting that the alt is rated at 75
amp, the voltage across the diode pack is 1 volt, so the wattage
dissipated per diode would be (75 x 1) / 6 = 12.5 watts? Or does
that leave something out?

Yeah, the fact that each two diodes are conducting at once. i.e. each diode is
active for a third of the time, not a sixth.

75x1/3=25W ea, = 150W total. Yech.On Fri, 18 Feb 2000, you wrote:

Kirbert wrote:

From: “Craig Sawyers” c.sawyers@tech-enterprise.co.uk

Now since the current actually varies over a recified half-cycle, the
average dissipated power has to be derived by integrating the current
waveform with the voltage (via the I=I0exp(qV/kT) law). So the above
20W per diode is probably an overestimate. I can’t be bothered to do
the integration…

Pardon me for having weak electronic skills, …the voltage across the
diode pack is 1 volt, so the wattage
dissipated per diode would be (75 x 1) / 6 = 12.5 watts? Or does
that leave something out?

Yep! :slight_smile:

Again,

Ned

Pardon me for having weak electronic skills, but can’t the dissipated
power be figured more simply by noting that the alt is rated at 75
amp, the voltage across the diode pack is 1 volt, so the wattage
dissipated per diode would be (75 x 1) / 6 = 12.5 watts? Or does
that leave something out?

The problem is that the alternator is an AC device, giving a full-wave
rectified, three-phase output. This means that, with the battery
disconnected, it will give a “bumpy” output; ie there will be a DC
level, with an AC waveform impressed on it. The AC waveform will look
like the top bits of sine waves. In fact, the amplitude of the AC will
be sin(60) (where 60 degrees is the phase angle betweent the phases)
times the peak value ( = 0.866).

The only thing I am not certain of, is the way in which the regulator
works. It is just possible that this actually modulates the current
waveform feeding the coils to partly iron out the tendency of the ac
generator to give the above waveform. Certainly, with older, pre
silicon regulator cars (where the regulator was an electromechanical
affair) the output would have had the ac component. Even if that were
the case with the modern arrangement, the low level of AC (only
amounting to 14% or so of the peak), means that we can treat the output
voltage and current as close to DC.

With a three-phase full-wave rectifier, two diodes conduct at any one
time (where each of the three pairs conduct for one third of the time).
Since we can approximate the current as constant, the diodes will
dissipate the current drawn, times two, times the voltage drop across a
diode.

Power silicon diodes drop around 1V at 20A, and the voltage drop goes as
the natural log of the current. So for a 140A alternator, the voltage
drop will be 1V x ln(140/20) = 1.9V. The diodes will therefore
dissipate 2 x 1.9 x 140 = 530W. The Lucas 75A unit would, by comparison
dissipate about 280W.

I would argue that is an impossible power to dissipate 530 W within the
confines of the alternator body. If the silicon rectifiers are replaced
with Schottkys (which have a lower drop of 0.8V at 160A), this gives a
dissipation of 225W for a 140A alternator, which is high, but
manageable.

So the conclusion is:

for lower current alternators, ordinary silicon diodes are probably used
for high current alternators, Schottky diodes are probably used
a power dissipation in either case of 250 - 280 Watts maximum

Just as an aside, the output spade terminals are bolted directly to the
diode pack heatsink, and must get very, very hot. This is bad news for
the spade connectors, which will tend to oxidise a bit and gain some
resistance (and hence dissipate power and heat up themselves!). It is
also bad news for the insulation on the two brown wires leading from the
rear of the alternator, and could give rise to embrittling of the
insulation.

Ball park here is that 140A flowing through a resistance of a thousandth
of an ohm dissipates 20W. This seems to speak for either bolting the
leads directly to the alternator output, or even soldering on some
glass-insulated pig tails, with a connector to interface to the brown
leads say 6 to 8 inches away.

Craig