![]() |
Between pump flow reading
Is there a cause of the change in fuel flow reading when a boost pump is on or off when the transducer is installed between the pumps? I?m adding the cube to my setup, and it going to need to go between the pumps. Is there a known way to minimize? I tend to use the electric pump 100%. Low pressure.
Setup: Tanks to valve to boost pump (facet) to collator to red cube to engine pump to Ellison. The routing goes from the tradional collator position at the lower right edge of the firewall over the top of the engine to the top left front (fwd of #2)since it?s an H2. |
Quote:
Larry |
Thanks
Thanks, that?s what I thought the concern would be. Not sure how you beat that except with some form of accumulator in the system. I may have to rethink my full time use of the boost pump. Alternately, since I could always keep it on, I could calibrate around the cause, I guess.
|
Quote:
The cube's impeller is a very light plastic part. It acts as a shutter for a light beam; no magnets. Try this. Pulsation source is the engine driven pump. Consider; when the boost pump is running, both of engine-driven pump's poppet valves are pushed open. If fuel flow demand is low, they may close momentarily, but they would be open for much of the pumps diaphragm stroke. At high flow the boost pump may hold those poppet valves open all the time. Regardless of open duration, the diaphragm is still being lifted by the pump pushrod and lever, and then released so the big spring drives it downward. A descending diaphragm and an open inlet poppet would be enough create a momentary reverse flow at the cube's rotor. The result is extra interruptions of the light beam, i.e. a higher fuel flow indication. That's my story and I'm stickin' to it...unless somebody has a better one ;) |
Quote:
Larry |
Wading in way over my depth here
I'm assuming the momentary reversals of the red cube impeller that create pulse events counted as higher-than-true fuel flow are events that happen at a lower incidence than the constant pulses from fuel flow in the forward direction through the cube. As such, they would appear in an output not unlike extra ventricular beats in an otherwise normal EKG strip. Since there are no true fuel flow surge events like this I am aware of in normal operations, it seems like it would be straightforward to digitally filter out any bursts of 2x higher frequency pulses coming from the cube and count them as only two discrete pulses and not three (2 pulses from normal forward flow plus one backward pulse of the impeller in between them) or some similar type pf algorithmic interpolation that would assume steady-state flow during times of digital "noise" on an otherwise steady pulse train.
In other words, digitally excise the junk data and fill in the gap with an assumed background rate from the previous second or so of data stream. This would have to be done at the EFIS software level. And there must be compelling reasons I haven't thought of why it can't be / hasn't been implemented to mitigate this known issue. I hope the above made sense. It's hard to put these concepts into words without pretty pictures. It's harder if your background isn't technical writing :rolleyes: |
Probably be pretty easy to get some data....
Really, this doesn?t sound like a significant data acquisition task. The filtering, I can think of lots of ways that could be a challenge.
Anybody remember those pesky accumulators Toyota uses to put at the end of the V6?s they put in Camrys and es300?s? If you know what I?m talking Is the rotor buoyantly neutral in fuel? Is there a reliable check valve option (the failure modes give me the willies though). Interesting. |
Just a thought, along the lines of Dan's post, has anyone by passed the mechanical pump and tested with just the boost pump? Most being rotary vane pumps--Otis correct me--seems more constant flow stream than poppet valve pump.
Pretty simple to test--remove the supply and discharge hoses from the pump and connect them together with a AN815-6 union, boost pump on and check the flow. Tom |
Always on?
Quote:
|
As a data point I put my cube between the two pumps on the cockpit side of the firewall and have otherwise met all other red cube criteria on a RV-7 with IO360. Similar to stock RV 10 and 12 locations. With the boost off fuel flow is accurate but on is almost twice actual flow. Boost is only on for initial climb and final descent so I don?t see any significant affect. On continuously you?d have to do a lot of fudging to the ?k? factor for the EFIS but that is easy. I?d just double the impulse value as a trial and adjust from there by flight experience.
|
| All times are GMT -6. The time now is 03:49 AM. |