What's new
Van's Air Force

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

Fuel tank leak testing with water before closing out?

N546RV

Well Known Member
OK, so while searching for something-or-other earlier today, I saw where someone mentioned performing a leak test on his tanks prior to installing the baffle. He did this by (presumably) placing the tank in the cradle and filling it with water. Apparently he was able to find a few leaky rivets this way and fix them prior to installing the baffle.

So now I'm trying to decide whether this is worth my time to do. Honestly, I'm leaning away from it, but I figure it could make for an interesting discussion.

Pros:
  • Possibly identify minor leaks while they're still accessible and can be fixed. Maybe avoid needing to create access holes in the baffle after the fact.

Cons:
  • Fuel is less dense than water and thus presumably more likely to find a leak than water. So while the water might not leak, this doesn't necessarily mean that fuel won't.
  • Increased timeline for completing tank closeout. For example, I expect to install the end ribs this week, which would allow me to do the baffles this weekend - unless I do this leak test, in which case I'd feel compelled to let the tanks sit for a couple weeks.

I suppose my overall take on this is that the water test seems like an OK way to find gross leaks, but by no means would it be a true indicator of whether the tanks were intact. If anything, the presence of water leaks seems like it might point to systemic problems with sealing technique.

On the other hand, I reallyreallyreallyreally don't want my tanks to leak. I went out today and touched up several locations that looked sketchy to me. So the water test appeals to me from that perspective - but again, thinking rationally, I don't think it's truly worth it.

What say you all? Are leaks a smaller problem than I think they are? Do I need to just shut up, close the things, and pressure test down the road?

(I think I'm just overthinking all of this because I'm nervous about leaks)
 
I filled mine with water prior to attaching the aft bulkhead , found no leaks, and closed 'em.

I figure it is one of those things that takes 5 minutes and can't hurt.
 
+1

I filled mine with water prior to attaching the aft bulkhead , found no leaks, and closed 'em.

I figure it is one of those things that takes 5 minutes and can't hurt.

However, I ended up with a small leak in the corner with the aft baffle when I did my final leak testing with air. Tanks have been holding fuel now for 2-1/2 months and 40 flight hours with no leaks - so far so good.
 
I wonder how many have developed leaks after an initial leak free test, and how long it took to develop them?
 
I wonder how many tanks end up leaking because too much sealant is used between the ribs/baffle and the skin which could make the structure flex and leak around the rivets. Then add in the poor technique of clecoing in the ribs wet and waiting for the sealant to cure before riveting. :rolleyes:
 
OK, good feedback, thanks! Kyle and others who've done the water leak test: How long did you let the sealant cure before doing the water test? I'd like to close the tanks this weekend if possible, but I don't want to potentially muck up sealant that's only slightly cured.

I'm thinking that if I focus, I can have the end ribs sealed and riveted Tuesday night. That could give me three full days of curing (in a hot Houston garage) before Saturday, when I could water test prior to installing the baffles. Alternatively, I could let the tanks cure for an extra week while I start working on the leading edges, maybe working on the landing light installation.
 
OK, good feedback, thanks! Kyle and others who've done the water leak test: How long did you let the sealant cure before doing the water test? I'd like to close the tanks this weekend if possible, but I don't want to potentially muck up sealant that's only slightly cured.

I'm thinking that if I focus, I can have the end ribs sealed and riveted Tuesday night. That could give me three full days of curing (in a hot Houston garage) before Saturday, when I could water test prior to installing the baffles. Alternatively, I could let the tanks cure for an extra week while I start working on the leading edges, maybe working on the landing light installation.

The "Thumbnail" test tells you when proseal is cured. Depending on ambient temperature and how "hot" you mixed the proseal, it could take a day to cure or a month.
 
When I was building I asked Van's about using water to test the tanks. They are adamant that you should not use water to test the tanks.

While I don't remember their exact words, it had to do with the possibility of residual water, possible corrosion, and that the viscosity of water isn't the same as fuel and that while the tanks might hold water, they may not hold fuel.

It was enough to convince me to use the balloon method after sealing everything.

In seven years the only leaks I have had were a number of rivets that blistered the top of the left tank. The reason for that was that just before painting I used MEK to remove excess proseal on just that area. Had I sanded the proseal off, I'm sure I would be preparing to repaint the top of that tank.
 
Why not use fuel?

I poured about a gallon and a half in and stood tanks on either end and flat etc. I had one leaker in the rear baffle. If I read correctly and you have not completely finished your tanks make sure you are generous with the pro seal on the rear baffle and the tank mating surfaces. When you squeeze these surfaces together some of the sealant will get squeezed out. Good Luck!
 
When I was building I asked Van's about using water to test the tanks. They are adamant that you should not use water to test the tanks.

While I don't remember their exact words, it had to do with the possibility of residual water, possible corrosion, and that the viscosity of water isn't the same as fuel and that while the tanks might hold water, they may not hold fuel.

The molecular size of H20 is larger than that of gasoline, so a water test may not show tiny little holes which might result in a fuel leak. But a water leak can reveal gross leaks, so there is no reason not to give it a try before attaching the aft baffle.
 
Cons:
  • Fuel is less dense than water and thus presumably more likely to find a leak than water. So while the water might not leak, this doesn't necessarily mean that fuel won't.

  • While fuel is less dense than water, the molecules of gasoline are much larger than the molecules of water. The reason water is more dense than fuel is because water molecules pack together much more tightly than fuel molecules.
 
Has nothing to do with it

Surface tension. Surface tension. Surface tension. Anything else is a distant secondary influencer.
 
Water testing tanks prior to closing

I did the water test (with a fluorescent die indicator) prior to closing. I found a small leak, which I fixed and now won’t need to be fixed later. Much much easier now than later. To me that is time and effort well spent.
 
No need to reinvent leak testing.

Testing with water is simply not showing leaks, except maybe the really large holes for all the reasons mentioned.
Also, your most likely to have a leak on the rear baffle intersection as this is a point that you cannot check after the baffle is installed. You can visually verify that all possible points of leaks ahead of the rear baffle are covered in proseal before you glue the rear baffling in place. Leak testing should be a verification procedure not a corrective means to find leaks.
Vans recommended procedure of "pressure" testing with a balloon attached to one of the fittings has served me well on 4 tanks and 20 years of no leaks on any of those tanks.
 
Gas can leak through where water won’t for the reason listed above. If using water, add a surfactant to break down the surface tension. Adding dish soap is better than nothing.
 
Kerosene?

If testing with 100LL is too dangerous and water has too much surface tension, what about kerosene/Jet A fuel? Much less volatile and cheaper than 100LL!
 
The easiest and safest way to test the tanks is by inflating them to 1 to 3 psi and seeing where the leaks are with soapy water. I know the filler can present a problem but use the filler cap and some good duct tape. That way you can be certain that there are no leaks and you can see exactly where they are if there are any. Just done it so know it works
 

Couldn't tell you. I'm not a car guy; however, the components I've gotten from Summit Racing have all been of very good quality.

People state that testing with fuel is too dangerous. That's a blanket statement. More dangerous than air or water? Certainly. No way to effectively manage any related risks? No.

I tested with ethanol free car gas and UV dye; before and after tank back plate installation. This approach isn't for everybody. It only proves that on those days, there's no leaks.
 

Attachments

  • Tank UV leak check.jpg
    Tank UV leak check.jpg
    291.1 KB · Views: 184
I water tested 5-6 times because I couldn’t ever find the leak. Once I finally did and they held water, I installed the baffles and balloon tested. Balloons have been inflated to the same level for a week.

I always have the pro seal a week to cure between applications. I have plenty of other stuff to work on while I waited.
 

Not specifically in scientific terms, but they are not wetted by water but are wetted by fuel. So water will, bead and not enter the weaving. I watched them work in a demonstration at the SAE international show a decade ago and they really work. Called a water sock and typically collapse in water only.
 
One caution about testing your tanks with actual gasoline. Gas outgasses significantly, so if you pour gas in your tank, and then seal it up tight, the expansion of gasses may provoke a leak where there wasn't one before. Use gasoline if you want, but then put the balloon on it as a fail-safe.
 
.... Called a water sock and typically collapse in water only.

coalesce water only? collect?

One caution about testing your tanks with actual gasoline. Gas outgasses significantly, so if you pour gas in your tank, and then seal it up tight, the expansion of gasses may provoke a leak where there wasn't one before. Use gasoline if you want, but then put the balloon on it as a fail-safe.

Not the outgassing you have to concern yourself with. It will quickly find it's equilibrium in a confined volume. The different expansion coefficients are a different story. Neither matter here. The original topic was testing before installing the backplate. When I built these, there were a lot of long threads here about leaking tanks, and related repair methods. Seems there was fairly bad run of QB tanks that leaked. Not wanting to go there as well, I tested before BP installation. The "leaks" in the photo were where the MoGas wicked up the aluminum to the open rivet holes; a real kick in the groan at first glance until I realized what had happened. After BP install, the tank was flipped and the new joints tested with a small amount of fuel again.
 
Back
Top