>
>
Hi Jan,
Thank you for your message and your kind words.
>I don't quite agree with your fuel pump discussion though.
>
>As I see it there are exactly 2 ways of connecting 2 pumps.
>Parallel:
>To remove a pump you replace it with an open circuit. A series check
>valve for each pump makes that possible (no backflow).
>Series:
>To remove a pump you replace it with a short circuit. A parallel check
>valve for each pump makes that possible (no blockage).
>There is no 3rd way.
>
>
A true series circuits has no parallel branch. Hence the
"series/parallel" expression.
Of course the true series mounting is not desirable in an airplane.
>A pump has a flow rate versus pressure characteristic. How steep is the
>curve?
>
>
You've got the flow/pressure curve on page 9 in the downloadable
Pierburg fuel systems
<http://contrails.free.fr/fichiers2/kraftstoffanlagen_en.pdf>
pdf document.
>At one extreme is a pump that supplies a set pressure regardless of the
>flow required to achieve that.
>Two identical such pumps in parallel would supply the same pressure and
>flow into a certain fixed orifice as one.
>Two identical such pumps in series would supply double the pressure and
>flow into a certain fixed orifice as one.
>At the other extreme is a pump that supplies a set flow rate regardless
>of the pressure required to achieve that.
>Two identical such pumps in parallel would supply double the pressure
>and flow into a certain fixed orifice as one.
>Two identical such pumps in series would supply the same pressure and
>flow into a certain fixed orifice as one.
>
>
One can clearly *hear *the pumps slowing down when backpressure builds up.
>Where in the range between one extreme and the other the Piermont pump
>falls I do not know. I suspect that Rotax found experimentally that it
>is somewhat closer to being a " flow source" (the second extreme) than
>to being a "pressure source" (the first extreme).
>
I have the notion the current or voltage source is too schematic when
applied to pumps. Remember this is an approximation to facilitate
reasoning on electronic device.
>Hence the series setup
>recommendation. But the difference is apparently not large.
>
>
I believe Rotax took the diagram direct from the Pierburg documentation,
without questioning too much...
>Practical questions:
>Is there a likelyhood that a check valve fails stuck open? That would
>weaken the case for the parallel setup.
>
>
The probability is remote indeed. And consider that such a vane pump
blocks the fuel flow when stopped, apart from some leak. Sooo, if a
check valve sticks open, the parallel setup won't be too much impaired.
>Is there a likelyhood that a check valve fails stuck closed? That would
>weaken the case for the series setup.
>In both cases the pump with the failed check valve must remain on.
>But maybe the backflow in the parallel case with the Piermont pump
>turned-off is not too bad?
>Or, maybe the blockage in the series case with the Piermont pump
>turned-off is not too bad?
>
>
>On your website you also mention the possibility that a failed pump may
>shed debris.
>That probably weakens the case for the series setup indeed.
>
>
That's the drawback of this setup. A stopped pump is not transparent to
fuel flow.
>B.t.w. The "pressure source" (first extreme case above) has the battery
>as its electrical analogue: one voltage (approximately), any current
>(almost). The "flow source" (second extreme case above) has no readily
>available electrical analogue, although current sources are quite common
>constructs in electronics. But pretending that a pump is just like a
>battery is not useful in a parallel versus series discussion.
>
>Hoping I didn't offend.
>
>
>
No problem,
Regards,
Gilles Thesee
Grenoble, France
http://contrails.free.fr
|