I learned something today

[Bill Boyd]

It might even be a useful something - not sure.

Due to some surprisingly high CHT's on a hot-day climb to 9500 a few days back, I decided I'd better do some sealing up of neglected air leaks in the baffling and check the GAMI spread again to see if ol' #6 had gotten leaner than the rest, making it more susceptible to running hot. As part of that plan, I thought I should calibrate this latest (my 3rd) Red Cube's K-factor, rather than assuming it was "close enough" out of the box.

Since my plane is not based at an airport with a fuel pump to simplify the fill-tanks, fly, refill tanks and see if the fuel pump agrees with the EFIS totalizer routine, I thought I'd try the calibration on the ground. My red cube is between the servo and the spider, so I disconnected the line to the spider and routed it via a clear PVC tube into a graduated pitcher. Plan was to run a gallon at a time into the pitcher with the boost pump and then compare 3 gallons actual with the reported fuel used and calculate the correction needed for the K-factor.

That's where the surprises came: it was immediately apparent that the Cube was way off (3 gallons pumped showed as 2.3 on the totalizer) and the fuel stream in the clear plastic line was *full* of air bubbles. This was true for either tank, leading me to think the air leak must be on the suction side, after the Andair valve but before the pump. But I couldn't figure for the life of me why a leak that allowed so much air to be entrained was not causing at least a sniffable fuel leak in the tunnel from the static pressure in the tanks.

I tried disconnecting the line from the mechanical pump to the servo and emptying the fuel into the jug, and to my surprise/relief, there were no bubbles at all! I could only deduce that the air was somehow getting entrained into the fuel inside the servo, while under positive pressure. This seemed very counter-intuitive but I could see no other explanation.

I ran all this past my mechanic, who being on a first-name basis with Don Rivera at Airflow Performance offered to call him and get his opinion. Don realized where this was going before Mike could even get it all out, and assured him that the regulator in the servo under these conditions could indeed introduce a stream of bubbles in the fuel stream when the line was not terminated by the spider and nozzles. So I abandoned the effort and will do my flow calibration the conventional way - at the airport fuel pump after my next X/C.

Since the results were unreliable (no doubt a stream of bubbles is not what the red cube wants to see), I'm shelving my worry that my full-rich climb burn could actually be 32 gph and my LOP economy cruise 16 gph.

Inspection of the nozzles showed #5 was visibly clogged, and #6 had a tiny speck of debris at the edge of the orifice. I found a bunch of baffling light leaks in the dark hangar last night and slathered them with RTV. Hopefully I won't see #6 hit 473F in climb anymore - that has never happened before. If it does, I will bump the timing advance back on the CPi2 in addition to dropping the nose for a less aggressive cruise climb.

I feel like I know what else to pursue for the CHT issue (front dams, etc) but just thought I'd report the unusual finding of air bubbles in the hose between servo and spider and that The Don himself says this was expected. Random factoid now filed away in case I ever need it

 

[wcalvert]

Bill, good after action report. Like you said, file that one away.

As an aside, I run my 7 with the front dams removed in the summer. I'm sure many will chime in and say something similar. You're on the right track sealing things up.

Cheers

 

[Bill Boyd]

Photo added

Here's the TL/DR:

If your cube is between the mechanical injection servo and the distribution spider, as many now are, don't bother testing your fuel flow calibration on the ground using the boost pump - it won't be right. Attached is a pic of the bubbles I encountered in the tubing referred to in the original post.

My own testing and the engineer at Airflow Precision confirmed that these bubbles enter the fuel stream within the servo, and will be there unless there is resistance at the outlet of the tube - which would have blown loose the cobbled-up taped connection I used to splice the PVC onto the AN flare fitting.

Picture taken while flowing fuel at about 4 gph during the attempt at calibration.

 

[PilotjohnS]

Red Cube cal

Speaking of Red Cube Calibration, me K number is quite a bit off from the 68K standard. But today I realize both my drains are leaking, so maybe I am loosing fuel due to the fuel drain leaks?
I can think of a scenario whereas I fill up the week before a flight, go fly, and re fill right when i return, and the red cube is showing lots of fuel used. It would show the fuel used during the flight, plus all the fuel that leaked out during the week of sitting. Tomorrow I will fill up right before my flight, and then right after and calibrate the red cube , again.

Amazingly, after 60 hours, i am still fiddling with the settings.
Once this is resolved, then it will time to start IFR lessons....

 

[Bill Boyd]

I think if it was a significant leak you would see puddling on the floor - or at least a stain. My right tank drips very slowly when over half full, from the vent line. No doubt it's a B-nut issue inside the tank (tanks closed and allegedly leak-tested when I brought the kit), but the conventional S-handle wrench trick isn't going to work, as a boroscope through the drain fitting shows it slathered in a thick coating of Pro-Seal and the wrench would never fit on the flats. My fuel leak is enough to leave a small wet puddle on the hangar floor, perhaps 2" across.

Enough fuel leak to affect the totals in a week's time is going to leave a big mess and a fire hazard, IMO.

 

[DanH]

Don realized where this was going before Mike could even get it all out, and assured him that the regulator in the servo under these conditions could indeed introduce a stream of bubbles in the fuel stream when the line was not terminated by the spider and nozzles.

What was the physical explanation?

 

[Bill Boyd]

Behavior of the step-down regulator in the servo when not having any resistance to work against on the far end of the hose.

I agree very counterintuitive, Dan, but it was quite reassuring that the line going into the servo did not contain such bubbles when I investigated there. Both said lines of course are plumbed in series and are on the pressure side of the boost pump. At first I was sure I had some serious leak-chasing to do on the suction side, and wondered how my engine could even idle, or cruise w/o vapor lock, with so much entrained air. Seeing a pure stream of fuel with no bubbles exiting the line from engine pump to servo body was thus quite reassuring.

 

[DanH]

Behavior of the step-down regulator in the servo when not having any resistance to work against on the far end of the hose.

I was hoping to learn the underlying physical cause of the behavior.

Probably the bubbles are vapor, not air, formed at a point of pressure drop. There are several, the most likely suspect being the fuel diaphragm's ball valve...but I'm guessing.

 

[Bill Boyd]

Not sure, Dan. My mechanic talked with Don about this, so he has it first hand. Vapor makes sense, going from 30 psig to atmospheric in the regulator with no outflow restriction to work against. Not sure how I would safely and practically tested the bubbles to see if they are air vs gasoline vapor - if air, there would certainly be some fuel in the vapor as well...

 

[MechaSteve]

plus all the fuel that leaked out during the week of sitting.

This would only be possible if your totalizer is powered during this time.

A flow meter like this simply puts out pulses as the internal vanes spin. The totalizer has to count all of these pulses. (that is what the k-factor is, pulses per gallon)


Source: worked on much larger flow meters

 

[bjdecker]

This would only be possible if your totalizer is powered during this time.

A flow meter like this simply puts out pulses as the internal vanes spin. The totalizer has to count all of these pulses. (that is what the k-factor is, pulses per gallon)


Source: worked on much larger flow meters

Actually, the totalizer only counts the fuel flowing through it...when it's on. The fuel leaking out of the vents, the torn bellows in the fuel pump (if upstream of the totalizer), and out of the missing fuel caps, does not get counted...