[E-Type] Fuel pump

E-Type
21

Hello Mike,

Having two master cylinders does not equate with
redundancy as you describe it. Each system has
to be able to perform the job individually. That is
clearly not the case as you describe it. That is
not a redundant system.

A redundant fuel pump fuel would have to have check valves
arranged in such a fashion that both pumps can run, or
only one of the two and you would not notice the difference.
Two pumps in series would probably not meet that condition.
But two in parallel with each a check valve at the end forming
an OR gate would do. You still have to have supervisory to tell
you that both are working or that one has died so you can
replace it quickly and maintain redundancy. So, Ray, the
system you describe does not meet the definition of redundancy
if both pumps must be working for the car to be able to run.
As with Mike’s master cylinders, two does not mean redundancy.
It is all in the arrangement.

The XJ-6 fuel supply was mentioned. Technically that is not a
redundant system, because both share the same switch. And
it is the switch that often fails. It also requires operator intervention,
while a true redundant system is fully autonomous.

Yes, Eric, the ultimate redundancy is the have two E-types,
just like any other Jaguar. One for driving, one under repair…

Dick Vandermeyden
San Carlos, California

Search the archives & forums - http://search.jag-lovers.org/
Subscription changes - http://www.jag-lovers.com/cgi-bin/majordomo

22

i had a dual set up for a while, till both the SU and the CArter failed
within a couple of months… I’m now running a holley, but will get a
spare. It’s a 10 minute deal to swap so not really worth having the extra
plumbing…

and Mike, yes on the v12, you can run at a moderate pace on half a pump but
some failures will competly disable it because they do share common parts in
the middle… don’t ask how i know!

Pascal----- Original Message -----
From: “Robert and Darlene Stevenson” bobanddarl@comcast.net
To: e-type@jag-lovers.org
Sent: Friday, April 04, 2003 8:58 PM
Subject: Re: [E-Type] Fuel pump.

on 4/4/03 2:20 PM, wilko pels at pels@xs4all.nl wrote:
When I had the fuel pump fail on my 64 which is in the tank, I replaced it
and plumped in an external pump in series. By changing one wire, when this
pump failed several years later, I was running again in a few minutes
instead of being stuck in backwoods Kentucky 450 miles from home with a
dead
Jag! I do alternate which pump is running once a year or so and it’s been
10
years now with no problem with either. Of course it may be the old thing
about if you carry a spare part the part in use will never fail!
Bob

Search the archives & forums - http://search.jag-lovers.org/
Subscription changes - http://www.jag-lovers.com/cgi-bin/majordomo

23

on 4/4/03 2:20 PM, wilko pels at pels@xs4all.nl wrote:
When I had the fuel pump fail on my 64 which is in the tank, I replaced it
and plumped in an external pump in series. By changing one wire, when this
pump failed several years later, I was running again in a few minutes
instead of being stuck in backwoods Kentucky 450 miles from home with a dead
Jag! I do alternate which pump is running once a year or so and it’s been 10
years now with no problem with either. Of course it may be the old thing
about if you carry a spare part the part in use will never fail!
Bob

Hey Ray,

Do you realy understand the concept of installing redundant systems for
improving the reliability ?

Indeed it requires carefull analyses of failure modes, for example, a
stikking fuel pump could have side effects, like leaking, creating a
pressure drop. So in this case we want a constant presuure, connect them
parallel, let them working constantly (a not used fuel pump tends to stop
working). In order to prevent pressure drops when one stops, add a check
valve. Indeed the reliability of a check valve should be compared to a
sigle fuel pump. If a broken check valve causes leaks or the average
lifetime of a check valve is less the a fuelpump, then it should be better
to stick at one fuelpump.
conclusion, it is to simple to talk in general about failure posibilities,
it requires carefgull analyses.

Wilko Pels
etype 71 ots under construction

At 18:38 4-4-2003, you wrote:

In reply to a message from Vand116@aol.com sent Fri 4 Apr 2003:

Dick,
I have to strenuously disagree with you here. In your example,
the chances of both pumps failing at the same time would be
100,000 hours. But, we’re not talking about simultaneous failures
here. It only takes a single failure in the right place for
both pumps to be rendered useless. So, the MTBF in this case
does not go up, it goes down, and sharply at that. If we were
talking about truly independant redundant systems, I would agree
with you that the MTBF could be increased. However, simply
hanging another pump in the boot constitutes anything but 100%
redundancy. As a failure in one pump can easily render both
useless, you now have at least twice the opportunity for failure,
so your MTBF will go sharply down. And, as I also pointed out,
simply having a pump plumbed in but non-operational may induce
additional failure modes which will reduce the MTBF of
the ‘‘backup’’ pump well below what it would be were it actually in-
service. Also, you’re adding more plumbing and wiring, which
further reduces MTBF.

–
Ray Livingston
Santa Cruz, CA, United States
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–

Search the archives & forums - http://search.jag-lovers.org/
Subscription changes - http://www.jag-lovers.com/cgi-bin/majordomo

Wilko Pels
Amsterdam; The Netherlands
Alfa 156 2.0 TS 98: daily use
E-type 70 OTS just passed the dutch government car inspection

Search the archives & forums - http://search.jag-lovers.org/
Subscription changes - http://www.jag-lovers.com/cgi-bin/majordomo

Search the archives & forums - http://search.jag-lovers.org/
Subscription changes - http://www.jag-lovers.com/cgi-bin/majordomo

24

In reply to a message from wilko pels sent Sat 5 Apr 2003:

Wilko,
Yes, I do understand how to design reliable systems. I think
you and I are agreeing here. My point is, as I believe you are
suggesting: it’s impossible to know what the effect of adding a
second pump is, as we have no information on the MTBF of any of the
components, nor do we know what the most common failure modes are,
or their probabilities. In general, simply adding more hardware
will make any system less reliable, unless the failure modes oif
all the components and their probabilities are known and taken into
account in the design.
Bottom line for me is still: A single high-quality fuel pump
should be a very reliable piece of hardware. If you don’t trust
your current pump, install a single good pump, and be done with
it! I believe that will always be the most reliable solution. If
it makes you feel better, carry a spare, but don’t connect it until
it’s actually needed.–
The original message included these comments:

Hey Ray,
Do you realy understand the concept of installing redundant systems for
improving the reliability ?
Indeed it requires carefull analyses of failure modes, for example, a
stikking fuel pump could have side effects, like leaking, creating a
pressure drop. So in this case we want a constant presuure, connect them
parallel, let them working constantly (a not used fuel pump tends to stop
working). In order to prevent pressure drops when one stops, add a check
valve. Indeed the reliability of a check valve should be compared to a
sigle fuel pump. If a broken check valve causes leaks or the average
lifetime of a check valve is less the a fuelpump, then it should be better

–
Ray Livingston
Santa Cruz, CA, United States
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–

Search the archives & forums - http://search.jag-lovers.org/
Subscription changes - http://www.jag-lovers.com/cgi-bin/majordomo

25

In reply to a message from Vand116@AOL.COM sent Sat 5 Apr 2003:

Dick,
No, it’s not all in the arrangement. Besides, we seem to have
totally lost touch with the original topic. Someone was wanting to
put a second pump in to improve reliability. There was no mention
of additional check valves or other hardware, simply plumbing in a
second pump, either in series of parallel with the existing one.
Certainly a proper redundant system could be built, but simply
slapping a second pump in there somewhere just isn’t going to cut
it, and none of us have the statistical data to enable us to do the
job in such a way that we can have any real confidence that we will
be improving reliability. Without knowing the failure modes and
probabilities on every single component, you have no way of knowing
whether the hardware you’re adding is increasing or decreasing
overall reliability. For any configuration you come up with, I
promise you I can find at least one failure mode that will render
the ‘‘redundancy’’ useless. The only thing that really matters is
how likely that particular failure is to occur. In your example
with the two parallel pumps with ORed check valves, in addition to
knowing the probability of a pump failure, you’d have to know what
the reliability of the check valves was. What is the probability
of one failing open vs closed, or even springing a leak? A closed
failure may not cause a problem. However, any single check valve
connected in an OR configuration, if it fails open or springs a
leak, would render even 10 parallel backup pumps totally useless.
If the pump has an MTBF of 10000 hours, and the check valve an MTBF
for open failures or leaks of only 1000 hours, you’ve spent a lot
of money to reduce the reliability by a factor of 10. That’s not
even taking into account the increased chances of a leaking fitting
or hose, an electrical short or open, etc., etc. now that there are
more hoses, fittings, wires, connectors, etc.–
The original message included these comments:

A redundant fuel pump fuel would have to have check valves
arranged in such a fashion that both pumps can run, or
only one of the two and you would not notice the difference.
Two pumps in series would probably not meet that condition.
But two in parallel with each a check valve at the end forming
an OR gate would do. You still have to have supervisory to tell
you that both are working or that one has died so you can
replace it quickly and maintain redundancy. So, Ray, the
system you describe does not meet the definition of redundancy
if both pumps must be working for the car to be able to run.
As with Mike’s master cylinders, two does not mean redundancy.

–
Ray Livingston
Santa Cruz, CA, United States
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–

Search the archives & forums - http://search.jag-lovers.org/
Subscription changes - http://www.jag-lovers.com/cgi-bin/majordomo

26

works for me, one to drive while working on the other one :-}
how can you put sensible & practical in the same sentence with english cars
:-}}----- Original Message -----
From: “Eric MaLossi” emalossi@austin.rr.com
To: e-type@jag-lovers.org
Sent: Friday, April 04, 2003 10:34 AM
Subject: RE: [E-Type] Fuel pump.

Careful Dick this is a dangerous though process that could ultimately
lead one to think that owning 2 E-Types is actually a sensible and
practical way to insure non-stop availability of an English motoring
system.
Eric

Search the archives & forums - http://search.jag-lovers.org/
Subscription changes - http://www.jag-lovers.com/cgi-bin/majordomo

27

Hello Ray,

Well, the float valves can be troublesome, true enough,
but they are not check valves. A check valve permits flow
in one direction but not the reverse. The float valves do not meet
that definition. The needle valves are metering valves and their
function is different and correct operation depends on
other components like the arm and float. The XJ-6 Ser III
has a true check valve after the pump to keep the fuel line full after
the pump shuts down. Never gives any trouble. And because of that,
few people are even aware of its existence. So it can be done.

Since this all started with the fuel pump, I have always wondered
why the 2FP in-tank pump was abandoned. It has the hall mark of
good engineering: elegant simplicity. All it is, is an impeller connected
to an armature, no poppet valves, only a relief valve. The lowest
MTBF components were the commutator/brushes which can be made
to last 20,000 hrs. Compare that to the external SU pump which had low
MTBF (1000 hrs) contacts, membrane, poppet valves and I doubt if the
overall MTBF was any higher than 500 hrs. True, the solid state
contact version was a bit better, but still could not hold a candle to
the 2FP. Never had the 2FP let me down in the 8 years I had a 3.8
E-type. It also sounds that the problems with the fuel pumps are
for the most part with the external SU type. Anyone knows why
the 2FP was discontinued?

Dick Vandermeyden
San Carlos, California

Search the archives & forums - http://search.jag-lovers.org/
Subscription changes - http://www.jag-lovers.com/cgi-bin/majordomo

28

Dick,
Elimination was likely completely related to cost. Either of the part
itself or assembly time. Its a mystery just like the removal of the
cooling fan removal which happened about the same time I believe. That
was definately a bad idea but they did it anyway.
pauls 67ots

Hello Ray,
… I have always wondered
why the 2FP in-tank pump was abandoned. It has the hall mark of
good engineering: elegant simplicity. All it is, is an impeller
connected to an armature, no poppet valves, only a relief valve. The
lowest MTBF components were the commutator/brushes which can be made
to last 20,000 hrs. Compare that to the external SU pump which had low
MTBF (1000 hrs) contacts, membrane, poppet valves and I doubt if the
overall MTBF was any higher than 500 hrs. True, the solid state
contact version was a bit better, but still could not hold a candle to
the 2FP. Never had the 2FP let me down in the 8 years I had a 3.8
E-type. It also sounds that the problems with the fuel pumps are
for the most part with the external SU type. Anyone knows why
the 2FP was discontinued?

Dick Vandermeyden
<<<<<<<<<<<

Search the archives & forums - http://search.jag-lovers.org/
Subscription changes - http://www.jag-lovers.com/cgi-bin/majordomoFrom: Vand116@AOL.COM
Subject: [E-Type] Re: Fuel pump.

29

Hello Ray,

Your points are well take. The original topic was to
put a second fuel pump in to increase reliability.

But there is any number of ways of how to hook up a
second pump. Some will enhance the overall availability
some will decrease it. Thus, since the stated purpose was
to increase availability, it was assumed that a proper method
and properly matched components suitable for the intended
purpose had been utilized. I did not suggest to just slap any
old pump in.

As I said before, the design should be such that the car runs
properly with either both pumps running, or only one without
any operator intervention.

In the grand scheme of things, items that require electric power
are deemed to be “active” items, items that do not need a power
source to work are deemed “passive” items. Any passive item
is orders of magnitude more reliable than active items. Thus
a check valve of any pedigree is far more reliable than a pump.

Also redundancy is defined as that any one single failure in the
system will not lead to outage. Since the proposed arrangement
consists of an “A” subsystem of a pump and a check valve and the
“B” subsystem is identical, any one failure of either one pump or
one check valve will not render the system inoperative and is thus
permissible. We are only providing redundancy for the pump. If you
want to make the fuel system truly redundant, you need two tanks
two fuel lines, two fuel filters, two electric supply lines running off
two different fuses and two check valves. Now you can have any
failure mode, including leaks, burned wiring, clogged filter and it
still works (As long as it is a single failure). Hello NASA.

Hey! I should quit right here! I charge my customers big bucks for
this sort of dissertations and I am not getting paid here!

Dick Vandermeyden
San Carlos, California

'68 OTS, System A, up and running.
'67 FHC, System B, under repair.

Search the archives & forums - http://search.jag-lovers.org/
Subscription changes - http://www.jag-lovers.com/cgi-bin/majordomo

30

Dick,
Should have read “just like the removal of the cooling fan RElAY”
Sorry, too quick on the send key.
pauls 67ots—
Dick,
Elimination was likely completely related to cost. Either of the part
itself or assembly time. Its a mystery just like the removal of the
cooling fan removal which happened about the same time I believe. That
was definately a bad idea but they did it anyway.
pauls 67ots

Search the archives & forums - http://search.jag-lovers.org/
Subscription changes - http://www.jag-lovers.com/cgi-bin/majordomo