Received-SPF: pass
 receiver=logan.com; client-ip=64.136.30.75; envelope-from=alwick@juno.com
To: flyrotary@lancaironline.net
Date: Tue, 20 Dec 2005 09:18:48 -0800
Subject: Re: [FlyRotary] Re: Fuel filter selection
Message-ID: <20051220.091906.3260.6.alwick@juno.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
From: al p Wick <alwick@juno.com>

Wow, good questions, good response. There is great learning opportunity
here. If you guys don't mind,  let me break it down.
This really shows the value of using facts (estimates) instead of our
natural method of solving problems.

Another way of wording all this is to ask:
"What's my biggest risk item? Debris killing pump, water damaging
injector, or vapor lock caused by inlet pressure drop?"

We can look at these issues individually and arrive at conclusion. If we
jumble them all together, it's more difficult to make an assessment.

Debris killing pump:
Let's guestimate how often this happens when you have no filtering of
fuel inlet. How about one pump shut down every 500 hours of operation? So
on scale of 1 to 10, pump shut down is a 6. But hold it, we have two
pumps, two independent failure odds. So that risk is 1 in 250000 hours.
Because we only loose fuel if BOTH pumps die. So now that risk is a 1 on
scale of 1 to 10. One of our lowest risk items on the plane. 
To keep this analysis simple, I'll ignore odds of noticing before flight,
and effect on flight components. 

Water damaging injector:
This is interesting one, as I never found evidence of this ever happening
in 10 million hours of vehicle operation. So I know it's extremely low
odds under normal circumstances. Perhaps there is some unusual feature of
our application that could cause this? Like "planes sit longer than cars,
so rusting occurs". I don't know. I'd guestimate one injector rusts up
each 10 million hours. You can use a different number. Once again, this
puts risk at a 1 on scale of 1 to 10. Extremely low risk item. It's even
lower risk, as when it does rust up, we just lose partial power, not full
power like in the fuel pump scenario.

Vapor lock caused by pressure drop:
Let's get a perspective here. The solution of adding inlet water
sep/filter reduces risk of "debris in pump" and "water affecting
injector". But that solution ALWAYS includes a pressure drop on pump
inlet. In fact, the pressure drop gradually increases....and  you don't
have a clue that it increased. I personally know crashes caused by vapor
lock. Loss of life. What are the odds of this happening? That's tough, as
it so depends on the details of your fuel sys design. But I'd guestimate
1 time in 900 hours for auto conversions. If you live in hot climate, I'd
way lower that number. On scale of 1 to 10, that's risk component of
about a 6. When the vapor lock occurs, often happens during takeoff, so
the EFFECT risk component is a 6 on scale of 1 to 10. You lose power at
inopportune time. Also, you guys don't measure your inlet fuel pressure.
So you have increased risk of not seeing gradual pressure drop change
over time. That risk component is a 7 on scale of 1 to 10. You multiply
these 3 risk components to arrive at measure of just how risky this is. 6
x6x 7 =   254  . Vapor lock total flight risk is 254, which puts it in
the top 5 risk items on your aircraft.

No wonder people crash!

In our daily lives, we often make decisions based on anecdotes and
feelings. This seems to work ok. These are life and death decisions on
our aircraft. We can improve our risks substantially by using unnatural
methods.


-al wick
Artificial intelligence in cockpit, Cozy IV powered by stock Subaru 2.5
N9032U 200+ hours on engine/airframe from Portland, Oregon
Prop construct, Subaru install, Risk assessment, Glass panel design info:
http://www.maddyhome.com/canardpages/pages/alwick/index.html


On Tue, 20 Dec 2005 10:43:53 -0500 Ernest Christley
<echristley@nc.rr.com> writes:
> Buly wrote:
> 
> > Guys, I don't understand why you are installing filters AFTER the  
> 
> > pumps? I sell large SEPAR diesel fuel filters and I always 
> recommend  
> > all the filtering to be done before the pumps and out of the 
> tanks.
> >  Any water or debris if it's allowed to reach the pumps, will be  
> 
> > beaten into smaller particles and may be PUSHED thru the filter, 
> or  
> > even will damage the pumps.
> > I personally suffered on my boat by frequent engine stopages in 
> the  
> > worst of times, only to discover that the manufacturer have 
> installed  
> > a screen on the pick up tube inside of the tank, where can not be  
> 
> > seen or cleaned.
> > The same manufacturer had to fix or replace number of engines only 
>  
> > because he installed the fuel pumps BEFORE the filter. Any water 
> in  
> > the fuel was beaten in to fine emulsion and pushed thru the filter 
>  
> > into the injectors. Results were rusted injectors, rusted high  
> > pressure pumps and more.
> > Buly
> 
> Seems to be a lot of debate on this one, Buly.  Filter before the 
> pumps 
> and you'll die from vapor lock.  Filter after the pump and you'll 
> die 
> from rusted injectors.
> 
> My solution, as recommended by others, is a PRM-81794 filter/water 
> seperator before the pump.   This is a high flow unit with dual 1/2" 
> 
> inlets and outlets that filters down to a few microns.  Then a 
> Summit 
> Racing 40-micron inline filter after the pump. The first should 
> insure 
> that everything is clean and dry.  The second will catch most debris 
> 
> that a failing pump or pressure regulator might throw.
> 
> The PRM-81794 has a valve on the bottom of the filter.  I presume 
> that 
> is to drain any collected water.  Has anyone investigated 
> semi-permanently attaching a hose leading to an easily accessible 
> drain?
> 
> -- 
>          ,|"|"|,                                    |
> ----===<{{(oQo)}}>===----        Dyke Delta         |
>         o|  d  |o          www.ernest.isa-geek.org  |
> 
> --
> Homepage:  http://www.flyrotary.com/
> Archive and UnSub:   http://mail.lancaironline.net/lists/flyrotary/
> 
> 


-al wick
Artificial intelligence in cockpit, Cozy IV powered by stock Subaru 2.5
N9032U 200+ hours on engine/airframe from Portland, Oregon
Prop construct, Subaru install, Risk assessment, Glass panel design info:
http://www.maddyhome.com/canardpages/pages/alwick/index.html