What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Testing fuel tanks with water?

jeffw@sc47

Well Known Member
I read a post a few weeks ago where someone tested their tanks with water so they could check them and fix any leaks before sealing them up with the rear baffle. Can't find that specific post again.

That sounded like a good idea. I found some fluorescent water dye concentrate and have been seriously considering doing that. The tanks are complete except for the rear baffle installation and am letting them sit for two weeks before proceeding - about ten days to go.

Since than, looking for that earlier post, I have come across posts that mention testing with water - and a few of those suggesting that testing them with water should not be done, but with no explanation why.

I cannot think of why it would do any harm to the tanks and only be beneficial if a leak was found that could be fixed a lot easier without the rear baffle installed.

Are there any compelling reasons why testing with water should not be done - what would it harm?
 
Can't hurt as long as the tank is supported with something that can hold over 200 pounds without damaging the structure. I do think it is a waste of time since you would be able to see any voids large enough to pass water under atsmopheric/gravity pressure.
 
Jeff, I did it and am happy I did so. I had a few areas that were big leaks but hard to see via a visual inspection. I used the cradles and filled them before the rear baffles were on. I knew at least then future leaks were likely concentrated around that. Check my blog for details. Not I didn't do a final water test but used the pressurized air method.
 
1 psi of air was very easy to do, with none of the potential hazards of a water test. I used soapy water to isolate my leaks, which turned out to be at the fuel caps in the stem assembly. I added some fuel lube in the stem assembly and problem solved.

I expect that water, having a higher viscosity than fuel, would not detect leaks that fuel might find. But I don't know that for sure.

Dave
 
what hazards?

. . . with none of the potential hazards of a water test.

What are the hazards of a water test?

I have constructed a more substantial tank cradle to better distribute and support the weight. If the tank is dried well afterward, what can the water damage?
 
What are the hazards of a water test?

I have constructed a more substantial tank cradle to better distribute and support the weight. If the tank is dried well afterward, what can the water damage?

Nothing. The issue with water is that it even if the tank doesn't leak, you haven't proven that it won't leak when you fill it with fuel. Fuel will pass through smaller openings than water.

Still, I tested mine with water... ;-)
 
I did not do this, but I thought about it: what about taking the tank/cradle outside and filling it with mogas?
 
I did not do this, but I thought about it: what about taking the tank/cradle outside and filling it with mogas?

I did this with the tanks for my bi-plane. Besides testing for leaks, I needed to know the capacity.
 
I think a full tank of gas is the absolute test. If you can fill it outside the risk is way less. I certainly wouldn't paint a fuel tank until it has sat full of gas awhile.
 
Water has very high surface tension. Fuel does not. Water can easily bridge small voids, corners, etc....
This is why water beads up easily on a surface and fuel doesn't.
Water would be a poor choice for testing a tank compared to alcohol or other fuels.
 
High viscosity? Add some dish soap.

Anybody run a test to check compliance with FAR ?23.965?

? 23.965 Fuel tank tests.
(a) Each fuel tank must be able to withstand the following pressures without failure or leakage:
(1) For each conventional metal tank and nonmetallic tank with walls not supported by the airplane structure, a pressure of 3.5 p.s.i., or that pressure developed during maximum ultimate acceleration with a full tank, whichever is greater.
(2) For each integral tank, the pressure developed during the maximum limit acceleration of the airplane with a full tank, with simultaneous application of the critical limit structural loads.
(3) For each nonmetallic tank with walls supported by the airplane structure and constructed in an acceptable manner using acceptable basic tank material, and with actual or simulated support conditions, a pressure of 2 p.s.i. for the first tank of a specific design. The supporting structure must be designed for the critical loads occurring in the flight or landing strength conditions combined with the fuel pressure loads resulting from the corresponding accelerations.
(b) Each fuel tank with large, unsupported, or unstiffened flat surfaces,whose failure or deformation could cause fuel leakage, must be able to withstand the following test without leakage, failure, or excessive deformation of the tank walls:
(1) Each complete tank assembly and its support must be vibration tested while mounted to simulate the actual installation.
(2) Except as specified in paragraph (b)(4) of this section, the tank assembly must be vibrated for 25 hours at a total displacement of not less than 1∕32 of an inch (unless another displacement is substantiated) while 2∕3 filled with water or other suitable test fluid.
(3) The test frequency of vibration must be as follows:
(i) If no frequency of vibration resulting from any rpm within the normal operating range of engine or propeller speeds is critical, the test frequency of vibration is:
(A) The number of cycles per minute obtained by multiplying the maximum continuous propeller speed in rpm by 0.9 for propeller-driven airplanes, and
(B) For non-propeller driven airplanes the test frequency of vibration is 2,000 cycles per minute.
(ii) If only one frequency of vibration resulting from any rpm within the normal operating range of engine or propeller speeds is critical, that frequency of vibration must be the test frequency.
(iii) If more than one frequency of vibration resulting from any rpm within the normal operating range of engine or propeller speeds is critical, the most critical of these frequencies must be the test frequency.
(4) Under paragraph (b)(3) (ii) and (iii) of this section, the time of test must be adjusted to accomplish the same number of vibration cycles that would be accomplished in 25 hours at the frequency specified in paragraph (b)(3)(i) of this section.
(5) During the test, the tank assembly must be rocked at a rate of 16 to 20 complete cycles per minute, through an angle of 15? on either side of the horizontal (30? total), about an axis parallel to the axis of the fuselage, for 25 hours.
(c) Each integral tank using methods of construction and sealing not previously proven to be adequate by test data or service experience must be able to withstand the vibration test specified in paragraphs (b)(1) through (4) of this section.
(d) Each tank with a nonmetallic liner must be subjected to the sloshing test outlined in paragraph (b)(5) of this section, with the fuel at room temperature. In addition, a specimen liner of the same basic construction as that to be used in the airplane must, when installed in a suitable test tank, withstand the sloshing test with fuel at a temperature of 110 ?F.
[Doc. No. 4080, 29 FR 17955, Dec. 18, 1964, as amended by Amdt. 23-43, 58 FR 18972, Apr. 9, 1993; Amdt. 23-43, 61 FR 253, Jan. 4, 1996; Amdt. 23-51, 61 FR 5136, Feb. 9, 1996]
 
Ah yes, dish detergent

Water has very high surface tension. Fuel does not. Water can easily bridge small voids, corners, etc....
This is why water beads up easily on a surface and fuel doesn't.
Water would be a poor choice for testing a tank compared to alcohol or other fuels.

Marc Bourget's suggestion > mix in some dish washing detergent to increase viscosity. That should help water to squeeze through any pin holes.
 
Just for accuracy, adding the soap decreases surface tension.

Viscosity is a measure of a fluid's resistance to flow. Surface tension is a contractive tendency of the surface of a liquid that allows it to resist an external force. In this case, surface tension, interfering with passage through a small hole, mimics viscosity, but is not quite the same.

I comment only because this technical difference may lead others astray, on a different problem/situation. For all intents and purposes for this thread, it was OK.

mjb
 
Just for accuracy, adding the soap decreases surface tension.

Viscosity is a measure of a fluid's resistance to flow. Surface tension is a contractive tendency of the surface of a liquid that allows it to resist an external force. In this case, surface tension, interfering with passage through a small hole, mimics viscosity, but is not quite the same.

I comment only because this technical difference may lead others astray, on a different problem/situation. For all intents and purposes for this thread, it was OK.
mjb

Marc, I appreciate the follow-up and explanation; it's good to know the difference between terms so the conversation makes sense.

So I take it from your note that by adding some dish washing detergent, the water does a better job finding leaks than if it is not added - by decreasing the surface tension so the water can find its way around corners and through small holes a bit better. For 25 gallons of water - what would you recommend for the amount of dish detergent added?

You learn something new more than a few times every day, then you die and forget it all.
 
You learn something new more than a few times every day, then you die and forget it all.

It doesn't take me near that long to forget stuff.
Just keep in mind, the vast majority of leaks reported are back baffle leaks. Usually in the lower corners. I don't think your test will hurt anything. It would certainly tell you something is wrong with your technique before you tackle the trickier back baffle.
 
Jeff,

My exposure to the soap technique arose out my training for liquid cooling systems for large trucks, typically 17-22 gallons, IIRC. My notes are packed away from a move of residence, but I'll guess 3-4 tablespoons for a tank. (this was for increasing cooling efficiency, not leak testing)

I once fabricated 20 aluminum tanks that I welded using GTAW (TIG) before the Tinman sold me on OA for aluminum. I tested them simply by filling with water, no soap mixture.

But, I also don't think it would be necessary to fill the tank to run the test, and now that I've thought about it, why are you considering water, why not the "balloon" on the filler neck and soapy water on the outside? (in which case I'd use a 50-50 mixture).

mjb

mjb
 
Water first, then air balloon

Marc, I plan to do the air and balloon pressure test after I install the rear baffle.

My reason for wanting to do the test first with water before installing the rear baffle is, if there are any leaks in the rib/skin connections they will likely be easier to fix before the rear baffle is installed.
 
using air pessure beyond the balloon method

All I did was the balloon method after closing up the tanks, but that is at a very low psi. I am wondering if instead of using a balloon, you would completely seal off the tank, and add a little more pressure, say 5-10 psi. Would I be risking damage to the tank from too much air psi??? And does anyone know what one could safely pressurize to on an rv tank?
 
I first did both methods simultaneously. The balloon popped when the manometer was about halfway up to the target, so at about 0.5 psi. I then capped this off and continued the test per vans manometer instructions without the balloon, that is taking the water to 27 inches or 1psi (if I remember correctly). The water level will fluctuate a little (a few inches) depending on atmospheric pressure and temperature changes. Someone has spelt out in another post how much this is per degree or inhg. What ever you do, don't fill it to 5-10 psi above atmospheric pressure. At one psi I could see clear deflection in the skin surface. At 10psi, you'll have a IED on your hands. At 10psi, you'll have the equivalent of 5 tons of force trying to tear the skin away from the ribs. 1psi for a couple of hours is heaps, and twice the pressure of balloon method already.
Tom.
 
....... What ever you do, don't fill it to 5-10 psi above atmospheric pressure. At one psi I could see clear deflection in the skin surface. At 10psi, you'll have a IED on your hands. At 10psi, you'll have the equivalent of 5 tons of force trying to tear the skin away from the ribs. 1psi for a couple of hours is heaps, and twice the pressure of balloon method already.
Tom.

Yikes!! :eek::eek: Me thinks i'll just be happy with the successful balloon tests!! Now that you mention it, 5-10 psi is no different than an explosive decompression on a pressurized airplane! What was I thinking????
 
I considered installing access holes and covers on my rear baffles during the assembly, figuring that it would be much harder to do after sealing the baffles, and be so helpful patching leaks. I eventually decided against it, though, because that would add even more possible leak paths. Got lucky and the tanks didn't leak except a bit at the cap stems.

FYI, I used some type A tank sealant (from Skygeek.com) to dot the rivets and blend the edges of sealed joints. Type A is more like a thick syrup than a paste:

293ewsm.jpg


I used a hardware-store syringe to apply that and a Popsicle stick where needed to blend it. Every edge, every rivet.

Here's the guide I followed (a mil is .001 inch, so 150 mils, for example, is .150 inches):

2612t6g.jpg


Dave
 
Back
Top