Digest Number 2509
Archiver
Posts: 46,084
FYI
> In endurance racing the class we run is limited to a
max of 10
> gallons fuel per pit stop. In a 25hr race, being able
to pit with a
> 5% error margin would be excellent. This .5 gallon is
what I would
> consider accurate. In racing, fuel gauges do not read
well with the
> fuel against the side of the tank.
What your discussing is monitoring consumption within
5%, or when the tank is nearly empty, to determine when
to come in.
Provided the rules allow for gages or other data
acquisition systems, better fuel tank gages will
provide a better result.
Without characterization of the flow transducers the
point of determining the actual performance is very
difficult. The problem is that using two transducers in
an A-B scenario pits the error of one transducer
against the other. They may add to a greater number
than actual or subtract from each other.
The error your manufacturer is typically talking about
is percent error of full scale and is also under steady
state conditions. If your flowing in excess of 300
pounds of fuel an hour to supply a 500 hp motor one
transducer will measure the 300 pounds/hr supply and
the other might be measuring 50 on the return at full
throttle with at times 300 being the return when
throttle is lifted. The percent errors may add or
subtract from each other, you have to bench test what
your describing to know for sure. During acceleration
transitions if the transducers have any capability to
free wheel (or are they positive displacement?) they
will add to the errors.
This scenario of A-B is simply full of accuracy issues,
though your welcome to try.
A simple float switch mounted in the right area of the
tank will probably give you the best minimum fuel level
indication.
I have built capacitance sensors for fuel tanks with a
0.1 gallon resolution for use in aircraft.
Your scenario of A-B consumption being totalized has
one other error feature, when the tanks are "filled",
who resets the totalizer back to full status? did the
tank really get filled on the pit stop or is this
assumed?
Point is don't assume anything, test everything....
Sincerely,
Ron
> Message: 11
> Date: Sat, 12 Jun 2004 14:38:38 -0000
> From: "daytonacopue" <turbo_tex@h...>
> Subject: Re:Fuel Injection Counting
>
> Ron
>
> Consumption only is what I need to measure. If I
could get 5%
> accuracy, I would be happy. One sensor states .025%
repeatability /
> Linear to 1% on frequency output versus flow
(graphed) / upto 3%
> error between sensors.
>
> You stated "One problem is that most transducers to
not have a
> constant pulse per gallon over a variable gallons per
> minute when graphed."
>
> Are you saying that the tranduscer manufactures 1%
linearity graphing
> is incorrect. Example: With 1% linearity, .5
liters/min is 60 pulses,
> 1.5 liters/min 180 pulses.
>
> Maybe my statement of accuracy is a little off.
>
> In endurance racing the class we run is limited to a
max of 10
> gallons fuel per pit stop. In a 25hr race, being able
to pit with a
> 5% error margin would be excellent. This .5 gallon is
what I would
> consider accurate. In racing, fuel guages do not read
well with the
> fuel against the side of the tank.
>
> With the electric fuel pump and a return line, the
fuel flow in gals
> per sec / minute / hr etc will be constant for the
send transducer.
> The return transducer will only see a flow reduction
of .005 gal per
> second max at full throttle.
>
> A calibration can be performed between the two
sensors for freq
> offset each time the fuel pump has power before
starting. This should
> narrow the upto 3% error between sensors.
>
> --- In basicstamps@yahoogroups.com, "Ronald Anderson"
<ronaldsa@e...>
> wrote:
> > FYI
> >
> > The use of fuel measurement transducers is
typically
> > used on Dynos where the engine speed and load is
> > controlled. Even in these controlled environments
what
> > you are describing is very difficult to obtain
> > accuracy. Thus what is typically done is to measure
> > consumption only, by feeding the return line back
into
> > the inlet of the pressure pump. Into this loop fuel
is
> > feed at 5 psi through a "consumption only
transducer".
> > Fuel measurement of A-B fuel channels is full of
> > problems related to accuracy.
> >
> > For all other vehicle operations, the pulse width
of
> > the injector times the frequency results in a
> > reasonably good piece of data for observational
> > purposes. Still not as good as consumption only
> > measurements for accuracy.
> >
> > One problem is that most transducers to not have a
> > constant pulse per gallon over a variable gallons
per
> > minute when graphed. Thus for accuracy the
> > characteristics of the flow meter have to be
> > incorporated into a lookup table to linearize the
data
> > and obtain the accuracy desired.
> >
> > I do this type of thing for a living and one can
spend
> > thousands of dollars to do it right (meaning
accurate)!
> > Thus there is nothing simple about it, the job
requires
> > attention to details typically not thought of. If
you
> > do not have the capability to characterize the
> > transducers for output counts/gal vs. flow rate
> > gallons/hr. for accuracy then the result will be
> > suspect.
> >
> > Sincerely,
> > Ron A.
> > DDS, Inc.
> > ronaldsa@e...
> >
> >
> >
> > Message: 1
> > Date: Thu, 10 Jun 2004 18:26:27 -0000
> > From: "daytonacopue" <turbo_tex@h...>
> > Subject: Re: Stamp counting capabilities
> >
> > I want to build a simple fuel differentiator. On a
> > Electronic Fuel
> > Injection sytem. The fuel system has a send and
return
> > line. A
> > transducer has to be placed into each line. The
> > transducers will
> > pulse 3300/liter or 7000 depending on application.
The
> > fuel pump will
> > pump 110 to 140 liters per hr depending on pump.
For
> > 3300
> > pulses/Liter usage 140 Liter/hr = 2.33 Liter/min =
7689
> > pulses/min =
> > 128 pulses/sec. For 7000 pulses/Liter usage 140
> > Liter/hr = 2.33
> > Liter/min = 16310 pulses/min = 271 pulses/sec.
> >
> > The two transducer pulse counts will be subtracted
to
> > determine the
> > pulse count of fuel used. I am reading about the
Stamps
> > and do not
> > know their capabilities. I thought I might have to
> > build counters but
> > wanted to make sure.... I could use the counter to
> > interupt the Stamp
> > to read, subtract, clear the counters, and display
to
> > update. I also
> > thought about sampling the pulses for frequency and
> > calculate but it
> > would not be as accurate. I need accuracy...
>
>
>
>
>
_______________________________________________________
_________________
>
_______________________________________________________
_________________
>
>
> To UNSUBSCRIBE, just send mail to:
> basicstamps-unsubscribe@yahoogroups.com
> from the same email address that you subscribed.
Text in the Subject and Body of the message will be
ignored.
>
>
> Yahoo! Groups Links
>
>
>
>
>
>
>
> In endurance racing the class we run is limited to a
max of 10
> gallons fuel per pit stop. In a 25hr race, being able
to pit with a
> 5% error margin would be excellent. This .5 gallon is
what I would
> consider accurate. In racing, fuel gauges do not read
well with the
> fuel against the side of the tank.
What your discussing is monitoring consumption within
5%, or when the tank is nearly empty, to determine when
to come in.
Provided the rules allow for gages or other data
acquisition systems, better fuel tank gages will
provide a better result.
Without characterization of the flow transducers the
point of determining the actual performance is very
difficult. The problem is that using two transducers in
an A-B scenario pits the error of one transducer
against the other. They may add to a greater number
than actual or subtract from each other.
The error your manufacturer is typically talking about
is percent error of full scale and is also under steady
state conditions. If your flowing in excess of 300
pounds of fuel an hour to supply a 500 hp motor one
transducer will measure the 300 pounds/hr supply and
the other might be measuring 50 on the return at full
throttle with at times 300 being the return when
throttle is lifted. The percent errors may add or
subtract from each other, you have to bench test what
your describing to know for sure. During acceleration
transitions if the transducers have any capability to
free wheel (or are they positive displacement?) they
will add to the errors.
This scenario of A-B is simply full of accuracy issues,
though your welcome to try.
A simple float switch mounted in the right area of the
tank will probably give you the best minimum fuel level
indication.
I have built capacitance sensors for fuel tanks with a
0.1 gallon resolution for use in aircraft.
Your scenario of A-B consumption being totalized has
one other error feature, when the tanks are "filled",
who resets the totalizer back to full status? did the
tank really get filled on the pit stop or is this
assumed?
Point is don't assume anything, test everything....
Sincerely,
Ron
> Message: 11
> Date: Sat, 12 Jun 2004 14:38:38 -0000
> From: "daytonacopue" <turbo_tex@h...>
> Subject: Re:Fuel Injection Counting
>
> Ron
>
> Consumption only is what I need to measure. If I
could get 5%
> accuracy, I would be happy. One sensor states .025%
repeatability /
> Linear to 1% on frequency output versus flow
(graphed) / upto 3%
> error between sensors.
>
> You stated "One problem is that most transducers to
not have a
> constant pulse per gallon over a variable gallons per
> minute when graphed."
>
> Are you saying that the tranduscer manufactures 1%
linearity graphing
> is incorrect. Example: With 1% linearity, .5
liters/min is 60 pulses,
> 1.5 liters/min 180 pulses.
>
> Maybe my statement of accuracy is a little off.
>
> In endurance racing the class we run is limited to a
max of 10
> gallons fuel per pit stop. In a 25hr race, being able
to pit with a
> 5% error margin would be excellent. This .5 gallon is
what I would
> consider accurate. In racing, fuel guages do not read
well with the
> fuel against the side of the tank.
>
> With the electric fuel pump and a return line, the
fuel flow in gals
> per sec / minute / hr etc will be constant for the
send transducer.
> The return transducer will only see a flow reduction
of .005 gal per
> second max at full throttle.
>
> A calibration can be performed between the two
sensors for freq
> offset each time the fuel pump has power before
starting. This should
> narrow the upto 3% error between sensors.
>
> --- In basicstamps@yahoogroups.com, "Ronald Anderson"
<ronaldsa@e...>
> wrote:
> > FYI
> >
> > The use of fuel measurement transducers is
typically
> > used on Dynos where the engine speed and load is
> > controlled. Even in these controlled environments
what
> > you are describing is very difficult to obtain
> > accuracy. Thus what is typically done is to measure
> > consumption only, by feeding the return line back
into
> > the inlet of the pressure pump. Into this loop fuel
is
> > feed at 5 psi through a "consumption only
transducer".
> > Fuel measurement of A-B fuel channels is full of
> > problems related to accuracy.
> >
> > For all other vehicle operations, the pulse width
of
> > the injector times the frequency results in a
> > reasonably good piece of data for observational
> > purposes. Still not as good as consumption only
> > measurements for accuracy.
> >
> > One problem is that most transducers to not have a
> > constant pulse per gallon over a variable gallons
per
> > minute when graphed. Thus for accuracy the
> > characteristics of the flow meter have to be
> > incorporated into a lookup table to linearize the
data
> > and obtain the accuracy desired.
> >
> > I do this type of thing for a living and one can
spend
> > thousands of dollars to do it right (meaning
accurate)!
> > Thus there is nothing simple about it, the job
requires
> > attention to details typically not thought of. If
you
> > do not have the capability to characterize the
> > transducers for output counts/gal vs. flow rate
> > gallons/hr. for accuracy then the result will be
> > suspect.
> >
> > Sincerely,
> > Ron A.
> > DDS, Inc.
> > ronaldsa@e...
> >
> >
> >
> > Message: 1
> > Date: Thu, 10 Jun 2004 18:26:27 -0000
> > From: "daytonacopue" <turbo_tex@h...>
> > Subject: Re: Stamp counting capabilities
> >
> > I want to build a simple fuel differentiator. On a
> > Electronic Fuel
> > Injection sytem. The fuel system has a send and
return
> > line. A
> > transducer has to be placed into each line. The
> > transducers will
> > pulse 3300/liter or 7000 depending on application.
The
> > fuel pump will
> > pump 110 to 140 liters per hr depending on pump.
For
> > 3300
> > pulses/Liter usage 140 Liter/hr = 2.33 Liter/min =
7689
> > pulses/min =
> > 128 pulses/sec. For 7000 pulses/Liter usage 140
> > Liter/hr = 2.33
> > Liter/min = 16310 pulses/min = 271 pulses/sec.
> >
> > The two transducer pulse counts will be subtracted
to
> > determine the
> > pulse count of fuel used. I am reading about the
Stamps
> > and do not
> > know their capabilities. I thought I might have to
> > build counters but
> > wanted to make sure.... I could use the counter to
> > interupt the Stamp
> > to read, subtract, clear the counters, and display
to
> > update. I also
> > thought about sampling the pulses for frequency and
> > calculate but it
> > would not be as accurate. I need accuracy...
>
>
>
>
>
_______________________________________________________
_________________
>
_______________________________________________________
_________________
>
>
> To UNSUBSCRIBE, just send mail to:
> basicstamps-unsubscribe@yahoogroups.com
> from the same email address that you subscribed.
Text in the Subject and Body of the message will be
ignored.
>
>
> Yahoo! Groups Links
>
>
>
>
>
>
>