Leak in Brand New Fuel Tank

[highcking]

This week I had a brand new fuel tank ($500 with shipping) installed in my 1958 Roadmaster Buick. [Tank vendor name available on request - one of the "big" companies offering parts for Buick and Olds.] It was a good fit, no installation issues. The shop added a gallon of so of gas and then I drove to the gas station and filled it up to full. When I got home I parked it on a concrete pad for a few hours, and when I came back I saw a large wet spot under the left rear of the car near the exhaust pipe. The spot smelled of gasoline. Looking underneath, I saw that the new fuel tank was leaking steadily from the seam at the left rear corner of the tank. See attached pictures. You can see large drops of yellow gasoline along the seam. I actually watched the gas seep out at the corner. 

 

Whoever is making these tanks is not leak-testing them before shipping and does not employ competent metal fabricators. This tank was stored on a shelf in my shop for months in the original shipping box and then was carefully installed with new pads and straps.

 

I am going to try to seal this tank to avoid having to remove it from the car loaded with 20 gallons of gasoline and having no drain plug. The seam will dry out when the gas level falls below the seam.

 

Advice? Anyone done this kind of sealing job, and if so, with what product? Thanks in advance.

 

Bill in Luray, VA

[JFranklin]

You might serve yourself better by pumping out the tank and do a proper repair billed to the company that sold you the tank and having a drain plug installed during the repair.

[edinmass]

Pull the tank......it’s a fuel leak. Fix it right, or better yet, replace it. Send photos to them ASAP. They won’t cover the labor.........it’s actually no big deal to R & R the tank.........you just did it. Is a fire and burning the car or your garage to thr ground worth it? 

[Ben Bruce aka First Born]

Duct tape comes to mind?😁

 

  Ben

[JohnD1956]

It will be difficult to fix a gas leak while the tank is in the car. Gas leaks could come from anywhere and migrate to the temporarilly lowest spot of the tank.  This occurred to me when the rivit holding the new sending unit together leaked.  

[highcking]

Thanks for all the replies. The vendor has agreed to swap the tank, though I doubt it will be tested first. I’ll get to eat the labor cost, of course.

[60FlatTop]

If it was mine I would siphon and drain the gas. We have a radiator shop that would solder the seam. We have a shop, BJR Radiator, in Rochester that would do that. You may have one near you.

A problem with old cars is buying parts in advance of their need. It is hard to go back to a vendor a long time after the purchase. I have a different type of company but I have a warranty account set up if an unfunded issue arises. The answer is always "Yes, Sir, no problem". Ask them, I can't be the only one.

 

Those of you with a tank or other sealed vessel on the shelf can test the new part with a couple caps and a vacuum tester. A shop with an evap tester can do it for you as well.

[highcking]

Very good suggestions on testing. If the vendor provides a new tank, I can test as you suggest. I have a sniffer that could check for escaping fumes with the lines capped off. I also have a vacuum tester. I never thought to leak test a new tank - I’d assumed the vendor would do that.

[NTX5467]

Am I seeing that flange held together with spot welds???  Rather than the continuous "stitch" weld of an '80s OEM fuel tank?  If so, no wonder it is seeping fuel.  IF the new tank has the same thing I think I'm seeing, it'll probably be no better than the one you have.

 

BUT considering the issues with welding on a tank that's got fuel fumes in it, get the "warranty" tank and then take it to a radiator shop that also fixes gas tanks.  Get them to put some interior sealer in the tank and then pressure-test it.   Or see if they can braze that seam to completely seal it, which ever might be the best long-term fix that they will warranty themselves.

 

Just some thoughts,

NTX5467

 

 

[highcking]

I'm not a welder, but in good light the weld looks like a common zipper. I believe it's leaking at the front and dripping towards the rear -- can't see the front at all. It has to be the seam somewhere because it leaked no gas with 3-4 gallons in the tank and leaks steadily along the top of the flange with a full tank.

 

My local repair shop works on all kinds of big trucks and the manager said they're experienced at pressure and vacuum testing tanks for leaks. 

 

I like the brazing idea - I have an excellent local welder who could braze the seam of the replacement tank.

 

Thanks for the suggestions!

[edinmass]

Am I missing something? It’s a new tank. Don’t you want what you paid for? A repaired gas tank is a sub standard service/repair. Welding or brazing on that tank makes zero sense..............it’s new, under warranty, they already will take it back, and any repair will be half assed. Makes NO sense. 

[JFranklin]

If the leak is at a seam just solder them. Could your leak be from the sender or filler fittings? You might opt for the replacement tank and have it tested. get a drain plug added for your convenience.

[60FlatTop]

If you have the tank in the car and it fits properly it might be better just to pull it out and have the trusted shop fix it.

 

My preference would be solder on the seams. Solder on radiator tanks is good for around 20 PSIG.

 

The worst one I saw was a Bentley tank full of rust and crud. They cut the top out for access. That tank had a number of chambers inside with passageways blocked and pinholes. Cleaned, sandblasted, patched in perforated spots, it was buttoned up and down the road.

 

One thing about buying new stuff, it is manufactured, put in the package, and sent out. A while ago I figured out refurbished items had to be tested before they went out. As good as new things usually are I like that tested product. I will lean toward the refurbished or rebuilt. A new replacement won't be tested. Fix the one you have and call it refurbished with confidence.

[highcking]

I will get a replacement from the vendor (Fusick). I then plan to have my welder solder/braze the flange and have my shop pressure test it afterwards. That should do it.

 

In general I agree on rebuilding vs replacement, but a 62-year-old tank isn't the same as a generator or master cylinder. I hadn't thought of the testing issue but in future I will make sure to test before installing.

[51dyno]

Pressure test the new tank and avoid putting a torch to it to destroy the coating of the metal .

[56 Buick]

Is the leak definitely at the seam because tanks full of gas will leak from the sender unit if the sender unit seal was not fitted correctly or is of poor quality. I would see if there is wet anywhere above the seam and I mean ALL around the tank. If the sender unit seal was leaking and fuel was rolling down onto the top of the seam lip then the fuel could also be rolling around on top of the seam lip to the other side of the tank until it pours over the seam lip at a low point. Just a thought.

[old-tank]

11 hours ago, 51dyno said:

Pressure test the new tank and avoid putting a torch to it to destroy the coating of the metal .

Pressure test the current tank before condemning it.

[highcking]

To 56 Buick - the gas gauge seal is something I hadn’t even thought of, and it’s definitely a possibility. The gauge is out of sight unless the car is on the rack and it could be a leak point in a full tank! I installed the gauge in my shop before the repair shop installed the tank. It seemed pretty simple, but I was not happy about the screws penetrating the seal. I could have tested that on the bench.

 

if we find the gauge to be the culprit, are there any suggestions for installation to get a good seal the way the gauge is designed? I know it was relocated to make it easy for mechanics to swap the gauge out, but thinking about it now, it’s a bad idea.

[195354]

 I thought I would add to this subject on the safety side. 

 

Fuel tank leaks and repairs we all just do it  and safety needs to be brought up. When the tank is empty it is just a bomb looking for an ignition source. We all just do things and once and awhile read about an unfortunate incident like a fire. 

When working with a fuel tank. Check for static it is rare but happens go ahead and discharge the static in your body just to make sure. Do you have any pilot lights in your shop or garage that are on, if you have any spillage the fumes will find the ignition source. When draining the tank into a container use a bonding cable look for a metal water pipe or a good earth ground hook each item up to the cable, fuel tank and container you will pump into. We have all just drained and filled but search the web on static and fuel. 

 

Repair 

Once the tank is empty and out this is when you need to be in the know. Airing  down can work by blowing fresh air into the tank it will take time. Another option is steam after cleaning, use a meter and confirm the container is safe. If you don’t have a way to confirm the tank is safe don’t try to repair the tank. 

The way to find a leak is seal the tank and add pressure, use regulated air and do not over pressurize, start at  one pound then brush the suspected leak with your solution. Your fittings added  to  seal tank can blow off, be careful and don't over pressure the tank.  

 

 I use Detecto mist from ken tools  to locate leaks in welds,  senders or any other items on a tank Detecto is used by tire shops they might have a small amount to sell.  Use a small paint brush and cover a weld or seam with the solution after repair repeat test. 

 

Bottom line: repairing a tank is best left to the experts. I know of a few folks that have tried and had a bad experience. They are OK and much more careful now. 

Steve 

[highcking]

Excellent advice. As a lifelong fire protection engineer I cannot agree more. Gasoline vapor is one of the most explosive mixtures known, that’s why it makes such a potent fuel for our Buick’s!

[60FlatTop]

Might want to add that filler hose has special electrical conduction properties so you need to buy the right stuff when all the rubber alligators off the 60 year old one. You shouldn't just cut off a piece of radiator hose.

 

And whatever you do, don't take fuel tank advice from Corvair Club members!

[highcking]

I am now almost certain that “56 Buick” has nailed the leak source. For 20 years up to 1957, Buick put the sender unit on top of the tank where it could not be a leak source. In 1958, they moved the sender to the front of the tank just above the seam. The sender is attached to the tank by five sheet metal screws that will be submerged when the tank is about 3/4 full. TERRIBLE DESIGN. It sacrifices safety for service convenience.
 

The seal is supposed to be achieved by having the screws go through holes in a round rubber flange gasket. When I installed the sender unit in my shop, I failed to test the seal by inverting the tank and filling with enough gas to cover the sender. If I had, I would have found that gas was seeping out via one or two of the screws. I checked this today by cleaning that whole area with paper towels, waiting a minute, then feeling around for fresh gas near the screws at the lower part of the sender. Gas was seeping out.

 

This leak simulates a flange leak because the flange is bent upwards into a “trough”at this point. Gas from the gauge oozes into the trough, travels around to the drivers side and drips down.

 

My repair shop will have to drop the tank, reinstall the sender unit with a fresh seal and using Permatex on the screws and the gasket itself. Then they can pressure test and also put gas in with the tank inverted and let it sit for 24 hours.

 

I have to think this was a problem in 1958 unless the factory used a sealing method that does not appear in their repair literature.

 

Thoughts welcome.

[56 Buick]

Your leaking sender unit gasket setup may be different. My sender unit is on top of the tank and merely had a plain rubber gasket. My gasket was leaking. For me I replaced with a higher grade rubberised cork gasket. No more problems. My research showed the standard rubber gasket can be prone to leaking. I ended up locating the rubberised cork gasket from one of the tri five chev parts sellers - it was an exact fit for me on the 56. This may be a possibility for you, otherwise the permatex may work.

[highcking]

The 57 unit is the same as yours. But in 58, the sender unit was moved to the front of the tank just above the flange. This was done (so says the shop manual) so that the sender unit could be serviced without lowering the tank. But this means that the seal has to withstand continuous immersion when the tank is about 3/4 full or more, not just sloshing or occasional immersion due to car motion. I think I will switch to cork - I’ll check out the Chevy option. I think it may work.

[56 Buick]

I cannot remember where exactly I purchased the gasket I used but having a quick look on the Internet it looks like Danchuk Chevy have a rubberised cork gasket. There is a photo there and that may assist in seeing if it will suit you.

[Frank DuVal]

On 1/9/2021 at 2:09 PM, 60FlatTop said:

And whatever you do, don't take fuel tank advice from Corvair Club members!

 

What did I do? I have not even responded to this post. OK, now I did.

 

Are you speaking of vacuum cleaners and gasoline? (Google it)

Are you speaking of soldering tanks while gasoline is flowing out? (personal experience)

Are you speaking of cutting tanks open to repair rust? (from a Corvaircenter.com/phorum post)

61 to 69 Corvairs have gas tanks available for less than $200 new, is that the issue?😁

 

Oh, is it that Ralph Nader said the Corvair's gas tank was in the SAFEST place of most any vehicle during the Pinto escapade with gas tank issues? That is TRUE!

[Bloo]

On 1/9/2021 at 11:34 AM, highcking said:

I am now almost certain that “56 Buick” has nailed the leak source. For 20 years up to 1957, Buick put the sender unit on top of the tank where it could not be a leak source. In 1958, they moved the sender to the front of the tank just above the seam. The sender is attached to the tank by five sheet metal screws that will be submerged when the tank is about 3/4 full. TERRIBLE DESIGN. It sacrifices safety for service convenience.

 

I am not convinced putting the sending unit on the side is a terrible design. Ford got away with it for decades (Chrysler too IIRC) using a locking ring and a square cut o ring.

 

They still leak when the sender is up on top. The gas sloshes all over when you drive, comes out the bad seal on top, and the whole car stinks like gas. It's just a lot harder to see where it is coming from. By the time you get the tank down, the gas will have evaporated. Hopefully there's a tell-tale stain or something. It is also hard to tell whether your repair really worked.

 

On 1/9/2021 at 11:34 AM, highcking said:

The sender is attached to the tank by five sheet metal screws

 

I hope you don't really mean sheet metal screws. If so, that is a terrible design. I would expect machine screws. They are screws in name only, having bolt threads. I have a hard time believing Buick sent it out the door with sheet metal screws. I can believe it with aftermarket parts though, as they are often crap. Look at your original tank and see if it has a flange with real threads. I'll bet it does.

 

Gas will just come up the threads, which are not sealed, and are a spiral path from inside the tank. If it has real threads, there are some things you can do. 1). Make your own gasket out of cork/rubber gasket material that fits tightly around the threads. This makes it a bit of a pain to install, but worth it. 2). Put purple (low strength) Loctite on the threads. Blue will do in a pinch. It seals the threads, and also allows you to decide how much you want to squish the gasket (to prevent splitting it), with no worries about the screws backing out. Tighten evenly as if you were torquing a cylinder head. 3). Use copper washers under the screw heads as a seal. This is the old fashioned method, and likely what they did in 1958. I might use more than one of these methods, maybe even all three.

 

 

On 1/9/2021 at 11:34 AM, highcking said:

My repair shop will have to drop the tank, reinstall the sender unit with a fresh seal and using Permatex on the screws and the gasket itself.

 

That sounds like a terrible idea. A dry cork-rubber gasket with undersize screw holes is the way to go. I don't know what you mean by Permatex, as they make a bunch of sealers now, but if you mean old fashioned Permatex like #1 or #2, or Indian Head, they are Shellac based. Alcohol is the solvent for them, and modern gas is laced with alcohol. They were never ideal on gasoline in the first place. If you mean a silicone-based compound, gasoline destroys those even if there is no alcohol. The only sealer I have ever seen that will stand up to gasoline long term is Seal-All, and it is a bit brittle. I would only use it as a last resort.

 

If those are really sheet metal screws (not machine screws), I guess I would put Seal-All on them and pray.

[60FlatTop]

11 hours ago, Frank DuVal said:

What did I do? I have not even responded to this post. OK, now I did.

 

Are you speaking of vacuum cleaners and gasoline?

 

I knew there had to be one out there.

[highcking]

Question on the machine screws holding the fuel gauge to the tank. I have a replacement tank and gauge. I plan to leak test the tank and gauge before installing on the car. I tried running each of the gauge screws into the threaded holes in the tank, using fingers only. What I generally found is that the screw would go in easily for a few turns and then balk. Going further would have required screwdriver torque. Is that the way it’s supposed to be? I’m certain that damaging the threads would guarantee a leak for sure.

 

Bill in Luray

[JFranklin]

Did the screws come with the tank?

[highcking]

No, the Taiwan tank from Fusick does not come with screws. They seem to be 10-32 but I’d think they should just turn all the way with slowly increasing drag. 

[NTX5467]

How about "prevailing torque" threads?  Designed to so the same as lock washers under the screw heads, but without using them.

 

GM used that situation on exhaust manifold-to-pipe studs/nuts in the 1990s.  On the nuts, they were "squashed" a bit or had a 3-way pinch.  On the studs, the center threads were not as deep as the outer threads, for that "interference fit" action.

 

Just make sure the gasket and sending unit are in the correct place BEFORE you do the final torque on the screws.  So you don't have to remove the screws once everything is all torqued.  I suspect the screws are softer than the metal they are engaging with, so holding your mouth right and some finesse might be needed to remove the screws without breaking them.

 

Just some thoughts,

NTX5467

[EmTee]

On 1/21/2021 at 8:23 PM, NTX5467 said:

How about "prevailing torque" threads?  Designed to so the same as lock washers under the screw heads, but without using them.

 

This would make sense; similar to pipe threads which also tighten as engagement increases.

[37_Roadmaster_C]

I do not remember the exact size, but 10-32 is very close to a metric thread size.  You can force a 10-32 into the metric thread, but it is NOT right.  On every sender mounting I have ever worked with on Buick the screws would thread in all the way with finger pressure unless there was damage or contamination.  I am betting the Taiwan tank has metric screw size threads....

[highcking]

Robin - that thought has occurred to me as well. No screws were provided with the tank. Not easy to check as I have no metric screw assortments around. I'll either have to buy a set or haul the whole tank into the hardware store! It would be a terrible thing for Fusick parts to do - but I can't rule it out.

 

Bill

[56 Buick]

Can you contact the tank manufacturer and find out the thread size?

[highcking]

Hah. Made in Taiwan. I've confirmed it's 10-24. But at least one of the gauge installation kits (gasket and screws) I have is 10-32. I curse not having kept the old tank. The "manufacturer" should either supply screws or put a label on the tank!