What's new
Van's Air Force

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

Leak Testing Fuel Tanks

rockitdoc

Well Known Member
Sponsor
I haven't found any photos of a manometer hooked up to the fuel tank to measure the 1 psi needed to check for leaks. So, here is how I set it up:

Fuel Tank Leak Test.jpg

Hooked a bicycle pump up to the Shrader valve at the yellow arrow. Pumped it until the column of red water in the Tygon tubing (thanks Larry) reached a differential of 30" (that's about 1 psi up here in Colorado. It will be different if your elevation is different, naturally).

Leaks found with bubble liquid obtained in the aircraft aisle at a local toy store. My leaks were at the ground screw holding the float mechanism to the side of the tank. A little sealant on the threads and the washer should solve issue. I'll find out tomorrow.
 
I haven't found any photos of a manometer hooked up to the fuel tank to measure the 1 psi needed to check for leaks. So, here is how I set it up:

View attachment 24330

Hooked a bicycle pump up to the Shrader valve at the yellow arrow. Pumped it until the column of red water in the Tygon tubing (thanks Larry) reached a differential of 30" (that's about 1 psi up here in Colorado. It will be different if your elevation is different, naturally).

Leaks found with bubble liquid obtained in the aircraft aisle at a local toy store. My leaks were at the ground screw holding the float mechanism to the side of the tank. A little sealant on the threads and the washer should solve issue. I'll find out tomorrow.

Just a note for clarity. Manometers are not a requirement for checking for leaks. If you actually purchase a leak test kit from Vans there will be no parts or instructions for constructing a manometer. What they will send in the kit is a handful of small balloons. Yes, the kind you can buy at the Dollar General store for, you guessed it, a dollar! Think of the balloons as a relief valve. Seal up the tank, place a balloon on one of the available ports and seal it however you can with whatever tape will do the trick. Screw in the the shrader valve and pump a small amount of air, just enough to get the balloon to expand. Thats it. Go to town with your soapy water testing. The idea being that the cheap little rubber balloon will burst long before any of the tank parts will. “Relief valve” in action!
 
True. Manometer not required unless the actual amount of pressure applied to the tank is desired. I just wanted to make sure I didnt exceed 1 psi because my bike pump gauge doesnt even read until 10 psi. Plus, my sons are engineering geeks like me and this makes a nice ideal gas science project, especially since we are a mile above sea level and our shop is rarely above 60F in winter.
 
Last edited:
True. Manometer not required unless the actual amount of pressure applied to the tank is desired. I just wanted to make sure I didnt exceed 1 psi because my bike pump gauge doesnt even read until 10 psi. Plus, my sons are engineering geeks like me and this makes a nice ideal gas science project, especially since we are a mile above sea level and our shop is rarely above 60F in winter.

While on the topic of science. Don't necessarily expect changes in readings to represent leaks. Changes in atmospheric pressure and temperature will change the readings. Manometers are sensivity to both temp and pressure and need to be adjusted out, just like getting a reliable TAS reading from your AS indicator.
 
Of course, the bubbles are the real test. And if thete are no leaks, you can always use the hoop to make your own with the grandkids.
 
Weeping rivets

Soapy bubbles are good to find leaks. This approach falls short however of finding weeping rivets.

The last step for me is to put ~five gallons of 100LL in each tank (tanks taken off the wings once the wings are done). Cap the fuel outlet and sump connection but keep a 1/4” tube on the vent line. Store the tanks for several weeks (you have other stuff to do anyway) rotating them from time to time (laying on the top skin, laying on the bottom skin, resting upright bottom baffle down). Place some white paper towels under the tank to act as witness paper.

When ready, drain the tanks via the sump fitting via a cheap, clear inline filter. If any junk comes out put the now filtered gas back in, slosh and repeat.

Carl
 
Carl
I like it. How can you tell if evaporation hasnt foiled the plan? Will fuel stain the paper?
 
I did the whole manometer thing, and at the end of the day it wasn't as useful as I thought it would be. I imagined I would pump it up, draw a line where the water level was, and then later if the water was at the line there were no leaks. As others have said, due to temperature and air pressure variation throughout the day, that doesn't work at all. I could easily watch the water level move up and down just due to the sun coming out. What I ended up doing was just using the manometer to ensure I wasn't pumping the tank more than 1PSI, and using the soapy water sprayer to find the actual leaks.
 
Carl
I like it. How can you tell if evaporation hasnt foiled the plan? Will fuel stain the paper?

Just like a weeping rivet, there will be a blue stain left over after the weeping gas evaporates.

The corollary, use 100LL as it has a dye, not mogas.

Carl
 
A faster method than Carl's would be to include some leak UV detection dye in the fuel (AC Delco type available through Amazon). Car gas will work fine and you can then run it in your car, lawn mower, etc. I did mine before installing the backplate (see pic) and again afterward. The fuel wicked it's way up the aluminum to the unfilled rivet holes. As you can see, a leak will light up like a disco.
 

Attachments

  • tank leak test.jpg
    tank leak test.jpg
    231.4 KB · Views: 315
If the soap bubble approach fails to find a leak when you know you have one, another highly sensitive method of locating leaks is by pressurizing (with the balloon of course) the tank with freon. Does not take much. And a simple freon leak detector sniffing around will locate the tiniest of leaks quickly. Be sure to use an environment friendly refrigerant.
 
SNIP

Pumped it until the column of red water in the Tygon tubing (thanks Larry) reached a differential of 30" (that's about 1 psi up here in Colorado. It will be different if your elevation is different, naturally).

SNIP

This is great - kills two birds with one stone. Very accurate pressure reading plus relief "valve". It is exactly what I did maybe 28 years ago...

BTW - 30" water will generate the same gauge pressure anywhere that gravity is the same... (pressure differential from inside to outside of the tank)
 
Fuel Stains

Sometime fuel will find a way to seep through where air cannot.

I have seen tanks seep/wick fuel without losing air pressure in a test. The only way to be completely sure of your tank's tightness is by safely fulling the tank with 100LL and letting it sit for a couple of days, then check for the telltale stain.

Certainly, you have to take extra precautions with the fuel and fumes, when filling and afterward, when it's drained. Fuel finds a way to wick where air can't. You don't want to find this out after it's on the plane.
 
Another Manometer

Here's my leak test at 27" of water.

The tanks bulge out noticeably with roughly 1 psi in them.

The manometer is definitely sensitive to temperature. It appeared to be leaking after pressurizing it and then I realized the soapy water in my spray bottle was cold and the combination of cold water and evaporation probably cooled the tank by 15-20 degrees. No bubbles. As soon as it warmed back up, the pressure increased and the water level returned to the original mark.

Going to let it set for a few days, keep the shop roughly the same temp and monitor atmospheric pressure for up and down trends.

So far no leaks as far as I can tell. I plan to put avgas in them as a final proof as well.
 

Attachments

  • IMG_0606.jpg
    IMG_0606.jpg
    404.3 KB · Views: 187
Last edited:
Here's my leak test at 27" of water.

The tanks bulge out noticeably with roughly 1 psi in them.

The manometer is definitely sensitive to temperature. It appeared to be leaking after pressurizing it and then I realized the soapy water in my spray bottle was cold and the combination of cold water and evaporation probably cooled the tank by 15-20 degrees. No bubbles. As soon as it warmed back up, the pressure increased and the water level returned to the original mark.

Going to let it set for a few days, keep the shop roughly the same temp and monitor atmospheric pressure for up and down trends.

So far no leaks as far as I can tell. I plan to put avgas in them as a final proof as well.

This is exactly why the test instructions in the build manual don’t tell you to make a manometer. They only say to pressurize and spray with a soap/water solution and look for bubbles, because a manometer can easily give a faults indication of a leak.
 
Fuel Tank Small Leak

Hello All:

I just finished building my RV10 fuel tanks.
I have just one tiny leak in the left tank (of course this is the one I built first). :( I am a little bummed out and was hoping for total leak free left and right tanks; but this is not the end of the world! I have attached 2 photos for more clarity. It appears it is a tiny tiny leak from under the skin in between the 1005 tank attach bracket and the 1003 inboard rib???? This is on the inboard bottom side of tank. The bubble is coming right out of the countersunk screw hole. I just wanted some feedback as to how I should make the fix. I was thinking of cutting an access panel in the inboard rib and glob a **** ton of proseal on the inside--both above and below the attach brackets near the skin??? This seems to be the best way since I will not have a lot of room to maneuver proseal if I remove the fuel sender and use that small hole.

Any help will greatly be appreciated. :);)
 

Attachments

  • IMG_8877.jpg
    IMG_8877.jpg
    275.5 KB · Views: 164
  • IMG_8878.jpg
    IMG_8878.jpg
    282.1 KB · Views: 190
Hello All:

I just finished building my RV10 fuel tanks.
I have just one tiny leak in the left tank (of course this is the one I built first). :( I am a little bummed out and was hoping for total leak free left and right tanks; but this is not the end of the world! I have attached 2 photos for more clarity. It appears it is a tiny tiny leak from under the skin in between the 1005 tank attach bracket and the 1003 inboard rib???? This is on the inboard bottom side of tank. The bubble is coming right out of the countersunk screw hole. I just wanted some feedback as to how I should make the fix. I was thinking of cutting an access panel in the inboard rib and glob a **** ton of proseal on the inside--both above and below the attach brackets near the skin??? This seems to be the best way since I will not have a lot of room to maneuver proseal if I remove the fuel sender and use that small hole.

Any help will greatly be appreciated. :);)

I recommend using the Vans access cover kit on the back baffle. It gives a reasonable size hole to work through and will give a better view of the work area. It shouldn't require a ton of sealant, only a small amount is needed in the correct location where you can see the void or pinhole.
 
A little trick for identifying slow fuel seepage sites (particularly when using mogas without coloration) around rivets, screws, and seams: Use a broad-tipped black Magic Marker-type pen to draw a circle around the suspected leak source. Any fuel seepage will instantly dissolve a portion of the marker ink revealing the presence of a leak. Sure it's a bit rinky-dink but it works.
 
So a little update. I cut out an access hole in rear baffle. I reached in with long piece of aluminum tubing and globbed pro seal all around inside inboard rib area….. it is still leaking from countersunk hole. The only difference is it’s making a little weeping sound as it bubbles out….?????

I had to stop earlier before I got to frustrated:-(
 
So a little update. I cut out an access hole in rear baffle. I reached in with long piece of aluminum tubing and globbed pro seal all around inside inboard rib area….. it is still leaking from countersunk hole. The only difference is it’s making a little weeping sound as it bubbles out….?????

I had to stop earlier before I got to frustrated:-(

At the risk of sounding snarky, how long are you waiting after adding the proseal before doing the pressure test? Depending on the type of sealant used, you might want to wait at least a full day or longer. Only asking because I've seen folks try to test without waiting an adequate time.
 
Generally, you can't just glob more proseal inside and expect it to seal.
Use some Poly-Gone to remove the old sealant around the leaking area. Get it off after it softens up with a plastic scraper. As gas leaks out, it can destroy the sealant and it then becomes a gummy mess, and putting more proseal over it won't cure the problem.
Scuff up the newly cleaned area with Scotchbrite, then clean thoroughly with acetone. Mix up new proseal and apply over the leaky areas. Let it set up for 2-3 days before trying to test it. Save the remaining mixed up proseal as a test sample so you have confirmation that the sealant has cured and is no longer tacky.
 
OK, thank you… I appreciate the advice. The only problem is it is a long reach down from the rear baffle hole I cut, making an access panel… I have to reach under the 1005 tank attach bracket, which is a portion that I cannot see.

Even holding a flashlight in the access panel is difficult because my arm takes up most of the space… I used a long piece of aluminum tubing scrap and I applied some sealant that way don’t know if that’s going to work but I really do not know where the leak is at… I’m just guessing at this point.
 
That is a tricky part of the tank to seal properly even when the tank is being constructed.
You are sealing the seam across the front of the 1005 bracket and front rib blind.
You need to reach under and use a mirror/camera to be sure those corners are all well sealed.
So I’m not surprised that it’s hard to fix now. I feel for you. Probably the worst spot I could think of for a leak.
If it’s not a massive leak Id probably try the loctite trick next.
 
Yes, I just came back from Lowe’s this afternoon and I bought a bora scope… I’m going to do that tomorrow and try and check and see where there is a lack in sealant… of course silly me my crazy luck. It had to be on this difficult end area, unable to be seen.Haha!

But if that does not work, I was told to use green Loctite is that correct?
Just apply the Loctite on the outside, where that countersunk screw hole is and let it absorb into the seam, and hopefully that will solve my problem.
On the flipside and positive note, the right fuel tank is perfect
 
I would place a flashlight inside the tank, then reach in.

A thing you could do would be try to seal it with the thinner composition of proseal - the flowable version, rather than the goo you're gonna have to force in place.

Regarding locktite, I think you'll eventually regret that solution. Plenty of people have had success with it, but for how long? A properly sealed tank is going to go 20+ years. Locktite? Dunno. Think about how hard it'll be to fix *next* time (flying, painted, in your hangar, not your shop)...
 
That is a tricky part of the tank to seal properly even when the tank is being constructed.
You are sealing the seam across the front of the 1005 bracket and front rib blind.
You need to reach under and use a mirror/camera to be sure those corners are all well sealed.
So I’m not surprised that it’s hard to fix now. I feel for you. Probably the worst spot I could think of for a leak.
If it’s not a massive leak Id probably try the loctite trick next.

So yea it’s just a tiny leak…. However I don’t know how long this green loctite will hold up versus the pro seal. I am thinking all the pro seal I added to the outside near and around that cs hole on the nut plate side quite possibly rerouted the leak forcing it through another path.

I just used the borescope I bought today to look down under the bracket and didn’t see much. I am going to try and lay a small flashlight inside the tank and try to lower my go pro in for a 4K video. I hope that will show any missing proseal spots. Looking down in tank from the baffle access hole I see the one side of tank attach bracket with decent squeeze out of sealant between the bracket and skin. I looked at my pictures as I was building and the side I can no longer see also has sufficient squeeze out…..

So that leaves me back to the beginning.
 
So yea it’s just a tiny leak…. However I don’t know how long this green loctite will hold up versus the pro seal. I am thinking all the pro seal I added to the outside near and around that cs hole on the nut plate side quite possibly rerouted the leak forcing it through another path.

I just used the borescope I bought today to look down under the bracket and didn’t see much. I am going to try and lay a small flashlight inside the tank and try to lower my go pro in for a 4K video. I hope that will show any missing proseal spots. Looking down in tank from the baffle access hole I see the one side of tank attach bracket with decent squeeze out of sealant between the bracket and skin. I looked at my pictures as I was building and the side I can no longer see also has sufficient squeeze out…..

So that leaves me back to the beginning.

Hmm. Not much fun. I think Kyle is right in that the green loctite is probably not a guaranteed long term solution. Maybe the thinned proseal and gentle suction.
Im not sure you’ll see a smoking gun bare patch with a camera tbh. Hopefully you do and you can then get to it with your fingers or other effective tool to put more sealant on the offending spot.
Good luck
 
Thx I will keep everyone posted. I am making a video as a learning experience and I will def post and edit my er tank build videos within the week on YouTube.
Something tells me the spot I am not seeing with leak is right there staring at me in plain sight…. I’m pretty tired but tomorrow is another day:)
 
My experience with doing a re-seal is that the Proseal needs to be "worked in" to the area and if you cannot readily see what your are doing then try donning some gloves (they need to be thick and fit tight) and smear the sealant in by feel with your finger...
 
Special Rivet

Paul Dye has an excellent article in the July 2023 edition of Kitplanes on the subject of fuel tank leaks. If you have a weeping rivet, he shows how to modify an AD48H pulled rivet to replace the leaky one. All done from outside with no damage to paint. I wish I had thought of this!
 
Nice article. I will keep in mind once plane is flying. Thank you for sending article. I am proud to report I fixed the leak coming from the countersunk screw hole. I cut access panel hole in rear baffle and used my gloved finger to smear a line of proseal on the areas inside tank. I patched and all is great:)
 
You can save some effort here, as they make a countersink version of this rivet. The AD42H (good luck finding one - AS out of stock for months) is a dome rivet. The AK42H is the same, but with a countersink head.
 
Nice article. I will keep in mind once plane is flying. Thank you for sending article. I am proud to report I fixed the leak coming from the countersunk screw hole. I cut access panel hole in rear baffle and used my gloved finger to smear a line of proseal on the areas inside tank. I patched and all is great:)

Hey John

Glad you got it fixed. I had a weeping leak in the same exact spot on my -10, showed up after about 100 hours of flying!! The "weep" only showed itself when the tanks were close to full.

Took me two tries of additional "prosealing & tank-on\tank off" to get it fixed. (I tried the loctite method before opening up the tank, but it did not work).
 
I want to connect this leak testing thread with a couple others that discuss leaks found only after filling the tanks with fuel. I had this issue.

https://vansairforce.net/threads/rv-10-fuel-tank-vent-tube-leak-fixed.223992/#post-1743590
https://vansairforce.net/threads/rv-10-fuel-vent-fittings-leaking.214300/#post-1667635

In my case I had RV-14 quick build tanks (circa 2019) but I know there have been slow builders and other models that had this same issue (for 14 builders reference sect. 18-06 and note the warning in bold at the bottom of the page :)

What happens is the vent tube bulkhead union fitting is sealed to the most inboard rib but the nut INSIDE the tank that attaches the vent tube to that bulkhead union is not fully tightened or sealed. During leak testing I followed Van's procedure which included capping off the the vent tube bulkhead union outside the tank. No leaks! With the cap on the outside of the bulkhead fitting a leak through the loose connection inside is blocked and thus not detected. When it came time to calibrate our fuel gauges the tanks were filled with 100LL such that the fuel level was several inches below the open outboard end of the vent tube near the filler cap, but above the bulkhead union which was below the fuel line due to wing dihedral. I came to the hanger the next morning to find a slow drip coming out of the vent tube from fuel seeping into the loose vent tube connection, traveling through the inside of the bulkhead union and out the vent onto my floor.

I would recommend that once you have completed the standard balloon leak test you-
  1. Cap off the fuel strainer (fuel line port) so the inboard section of the tank is sealed.
  2. Fill the tank with a couple gallons of fuel.
  3. Tip the tank up so it is standing on the inboard end.
If you get no leak out the vent tube bulkhead union it will to confirm the vent connection inside the tank is properly tightened. We carefully supported the tank with 2x4's and clamped it to the side of an EAA workbench so it couldn't fall then left it for a day.

As you will see from the links above, this one extra step prior to installation could save weeks of extra work down the road. Build On!
 
Fill it with 100LL and let it sit for a couple days, accept no other method unless just doing an initial check or trying to find a leak your going to fix. I'm speaking from years of working on airliners and KC-10 tankers. We manometer pressurize to bubble check for leaks, but the final QC stamp doesn't come until after 100% re-fill and specified dwell time. I did my RV the same way..balloon and soapy water initial check, then filled with 100LL and let them sit. When I do an RV tank re-seal, usually fill to 100%, if only a rear baffle re-seal, 2-3 gallons and set it L/E up on the bench.
 
Fill it with 100LL and let it sit for a couple days, accept no other method unless just doing an initial check or trying to find a leak your going to fix. I'm speaking from years of working on airliners and KC-10 tankers. We manometer pressurize to bubble check for leaks, but the final QC stamp doesn't come until after 100% re-fill and specified dwell time. I did my RV the same way..balloon and soapy water initial check, then filled with 100LL and let them sit. When I do an RV tank re-seal, usually fill to 100%, if only a rear baffle re-seal, 2-3 gallons and set it L/E up on the bench.
I would NOT test with avgas, having had the bad experience of the proseal reverting to goo after exposure to avgas during testing. Both tanks needed to be opened up and cleaned out and re-sealed. It was a terrible experience and testing with avgas achieved nothing, other than destroying the tanks.
 
I would NOT test with avgas, having had the bad experience of the proseal reverting to goo after exposure to avgas during testing. Both tanks needed to be opened up and cleaned out and re-sealed. It was a terrible experience and testing with avgas achieved nothing, other than destroying the tanks.
Explain your reasoning here if you don't mind. I'm not sure the correlation/causation connection has been established. So the sealant should never be exposed to fuel? Are you theorizing exposure to avgas and then to air caused the reversion? As for now, I think I'm siding with someone with Bill E's decades of directly related experience.
 
I would NOT test with avgas…
You mean the same fluid the tanks will be filled with for the rest of the airplane‘s life?

I fill up the tanks with gas with black light dye added (I have used car and 100LL) and let them soak for a while. Air is a reasonable initial check, but the proof is giving fuel plenty of opportunity to find a path to the outside. I have a tank soaking right now. Not a hint of a leak after a week.
 
Explain your reasoning here if you don't mind. I'm not sure the correlation/causation connection has been established. So the sealant should never be exposed to fuel? Are you theorizing exposure to avgas and then to air caused the reversion? As for now, I think I'm siding with someone with Bill E's decades of directly related experience.
Yes I will explain the logic. I built the tanks and sealed them carefully using fresh Proseal ordered from Van's. I tested the tanks for leaks using a balloon and soapy water (as recommended in the construction manual) and there were no leaks. Next I purchased a gallon of avgas and put half in each tank and rolled them in various positions and monitored for leaks, and there were none. I left the tanks with the caps off to allow the avgas to evaporate and stored the tanks while I continued with the rest of the slow build.

Several years later I opened the tank access plates to carry out the service bulletin to secure the fuel pickup nuts and discovered that the sealant on the inside of the tanks had reverted to goo. All of it was softened, including that on the fuel pickup tubes, which were supplied by Van's with the sealant already applied and cured (a different batch by a different technician).

The sealant on the outside of the tanks that was not exposed to avgas was NOT affected and was still sound. So the only explanation is that the combination of avgas vapor and air destroyed the sealant. I'm not a Chemist so I don't know what chemical reaction occurred.

I fixed the tanks by cutting access holes and taking off the rear baffles to clean out all of the gooey Proseal. This process took one week per tank and used gallons of solvent (toluene and methylene chloride) and was a horrible messy job. I don't wish that experience on anyone. I resealed all of the joints and leak tested with air/soap and there is NO WAY that I will put any avgas in for leak testing, until the RV is ready to fly.

20221125_155812.jpg20221127_130721.jpg20221221_174651.jpg
 
Several years later…
Well, this freaked me out a bit so I just checked my first tank (that has been in dry storage for at least a year) that initially passed my fuel soak test. The sealant feels exactly like cured PR2001-B2 should. Wonder what happened to yours? Maybe not mixed properly?
 
Well, this freaked me out a bit so I just checked my first tank (that has been in dry storage for at least a year) that initially passed my fuel soak test. The sealant feels exactly like cured PR2001-B2 should. Wonder what happened to yours? Maybe not mixed properly?
Supports the case-in-point. A sample size of one isn’t enough to make such wide claims.While he doesn’t feel it, PaulVS probably got a little lucky. The reversion was most assuredly going to happen no matter what. Better to find out before the tanks leaked or decided to puke a bunch of goo into his operating fuel system.

I’ve spoken with the (original formulation) Desoto engineers a few times over the years. They have original sealant still doing its job since the 1950s. This was before they buttered the faying surfaces and only filleted the seams. They claim dry layup will not affect its life; though, there is incorrect tribal knowledge stating tanks must remain full. This helps from a “heat sink“ effect for drawing sun heat from the sealant, not keeping them wet = longer life. My 1969 Mooney had original sealant when I sold it a few years ago (Desoto brand) and it sat on the ramp in the FL sun for a long time before I bought it. To be clear, I’ve never spoken to the Flamemaster engineers.

Could be a mixing/application error but there have been bad production batches in the past. Either way, I still believe some luck was had when it was discovered early.

@fixnflyguy. Does your business make test/validation samples or coupons when they apply it? Wouldn’t be surprised.
 
Well, this freaked me out a bit so I just checked my first tank (that has been in dry storage for at least a year) that initially passed my fuel soak test. The sealant feels exactly like cured PR2001-B2 should. Wonder what happened to yours? Maybe not mixed properly?
I don't believe the failure was caused by mixing, or by a bad batch, because the sealant that was on the Van's supplied fuel pickup tubes failed also. That was a different batch mixed by a different person.

It is not uncommon for tank sealant to revert to goo and there are reports on VAF of it occurring, typically in the vicinity of a leak e.g. on the rear baffle.
 
Back
Top