Failure Analysis

[jjconstant]

I'm hoping to get ideas on how my left fuel indication failed. The symptom was that the indication would read correctly for a long time and then start to sweep from correct to -10 gallons, randomly and finally only reading -10.
-10 is also the indication you get when there is either no ground or no signal from the sender. This behaviour would happen both in flight and stationary on the ground with no movement at all.

I did what troubleshooting I could and concluded that the sender (float type) had to be bad so I pulled it out of the tank and indeed it was bad. The wire was broken at the big blob of solder going from the swinging arm mechanism into the base/back of the hole you screw the wire into that then goes to the instrument. I must have over tightened the screw. The failure behaviour lead me to believe that the broken wire would variably be connected and then open or very high resistance.

I thought this would solve the problem but, after thinking it was fixed and reading correctly, it started back to the old failure of variable between correct and -10. I redid all the troubleshooting including checking the resistance of the entire harness wire from the instrument all the way to the connector at the tank. Everything was O.K. Out of desperation, I stripped enough insulation from around the 470ohm resistor to put test leads on each side of it. It ready correctly. For about 5 minutes. Then it started varying wildly. O.K...bad sender AND bad resistor. I cut out the resistor and soldered in a new one, again being VERY careful not to put much heat into the resistor. It now reads correctly. For now.

I have now been away for a day or two. I'm going back to recalibrate the tank (different sender) but want to be prepared for it to start not working again. Could the wire breaking at the sender have caused the resitor to fail? Should I be looking elsewhere for something to have caused the resistor to fail? I must say that having a broken wire in the fuel tank has given me pause...is there enough current to cause a fire if the tank were almost empty?

Thanks

Jeremy

 

[apkp777]

... I must say that having a broken wire in the fuel tank has given me pause...is there enough current to cause a fire if the tank were almost empty?

Thanks

Jeremy

That and a guided missile.

 

[Bob Axsom]

Some thoughts

Broken wire - Wires breaking at solder joints is a common problem where movement is involved, less so at mechanical attachments. You fixed it apparently but I couldn't tell exactly what you did to correct the problem.

Resistor failure - The indications are that this was an open circuit failure as well. Resistors rarely fail but when they do it is usually an overloaded burn up or a mechanical failure. You said it was a 470 ohm resistor so you may have been reading a schematic, label on the resistor body or the color code. If you could see the resistor and it appeared to be intact and you measured the correct resistance across it then the problem may be an intermittent lead connection to the resistance element inside the envelope or the resistance element itself. In the aerospace industry there have been NASA alerts from time to time about lots of resistors from specific manufacturers that contain manufacturing defects in the lead to resistance element interface as well as the resistance element itself. You replaced the resistor and restored proper function so whatever the problem was it is gone.

Usage of the resistance senders for millions of hours on thousands of airplanes without widespread cases of unexplained exploding fuel tanks indicates to me that the application is safe.

Bob Axsom

 

[elippse]

Usage of the resistance senders for millions of hours on thousands of airplanes without widespread cases of unexplained exploding fuel tanks indicates to me that the application is safe.

Bob Axsom

But just wait; when it is suspected to be the cause of a car or plane explosion, the news media will play it up as if it's and everyday occurence and then the guys in Washington - Congress, FAA, NTSB, will get into the act and outlaw all of these senders and make us put in Model-T and Cub-type floats!

 

[dmaib]

I just chased down a nearly identical problem with the right sender on my RV-10. It was lack of a good ground. The screws holding the sender in had fuel lube on them and I suspect that may have been the problem. Good ground and all works well now.

 

[jjconstant]

Thanks for the input guys.

Bob-I fixed the sender by buying a new one The resistor appeared intact in every way except by direct measurement and it took a few minutes for the intermittence to appear. It is 470ohms as described by GRT schematics, their literature, by verifying the color bands and verifying two new ones with a DVM. Replaced with a new one.

dmaib-a new redundant ground was the first thing I tried with the old failure mode, and it turns out that it is still necessary on this tank but not on the other tank

My painful process today was: Tank was empty, started the calibration ala GRT, filled the tank, continued the calibration and then tried to verify everything was working by emptying the tank, a metered amount at a time. Kept waiting for the float to start moving -- never did. Empty tank, gauge read full @#$^&$%&!!!!!! At least it didn't read -10 so the old problems were solved...

Pulled the sender, too much friction in the pivot, redid the bend with a tighter radius, changed the angle to move the float to where it would exert more force with gravity and reinstalled. It now works properly. Phew!!!!

One note: I was debating whether or not to reinstall the float with Pro Seal as I had done on the first installation or use the included rubber/silicone gasket. I used the gasket. Thank goodness...the float needed to come in and out at least 4 more times to get things right and the gasket never leaked. I used a very thin film of EZ Turn/Fuel Lube on both sides of the gasket.

All Best

Jeremy