#131
|
||||
|
||||
Quote:
I have found about 25 pin holes the vast majority of these are so small they weep out a small droplet of water only under pressure. In other words, no water leaks from the hole under simple gravity pressure. I know that fuel/gas can/will leak when water doesn't. I'm told this is a viscosity thing where fuel is less viscous than water so it can pass through small holes with little to no pressure. Fixing the holes has been relatively easy in most of the cases. A simple fusion type pass over the hole and adjacent area. I usually add a little bit of additional filler if there was a small cavity under the surface causing the leak. With the fusion pass, I make sure the weld surface is completely smooth and the edges taper down nicely to the parent metal. I have 1 more hole to fix from the last test. I think I'll add some dish soap to the water for the final water-based test. Hopefully that would result in a bubble if any more holes exist versus a small water droplet. The water droplets have been hard to find/see as they are typically very small. Once the tank passes the water based test with no apparent leaks, I do plan to test with fuel. On second thought, I think I'll do an air only test like you suggest prior to the fuel test. I do want to be sure there's no leaks in the tank before it goes into the car. Given this is my first fuel tank build, I'm learning a lot.
__________________
Joel Heinke Be original; don't be afraid of being bold! |
#132
|
|||
|
|||
Great thread Joel. I test with air as that will find the smallest of leaks. 5psi may not sound much but on a big area there is a lot of force, which can cause the tank to bulge, so don't be tempted to crank the pressure up. I go over the seams with a foamy carpet cleaner, which really sticks and the foam grows where there is a leak. A good test is to leave the pressure in overnight, if the pressure holds up, it is likely you have a sound tank.
__________________
David Hamer |
#133
|
||||
|
||||
Quote:
While testing again today for what I thought was the last leak fix, I found another leak, this one very, very small. It only leaked a small teardrop worth of water after an hour of being under pressure. In that hour, there was not a measurable drop in pressure within the tank. What's interesting is that I tried to use the air pressure inside, soapy water outside leak detection method on the last two leaks and I wasn't able to get air bubbles from either. I even used carpet cleaner on the last one and it didn't show bubbles even though the foamy liquid stayed on top of the leak for about 5 minutes and then just turned to liquid. Both of these emitted pressurized water from the inside on a prior test and that's how I found them. I'm sure the pressurized air and soapy bubble method would have detected the leaks prior to these as they were bigger leaks. Well I think I've now fixed all the leaks in the tank. I filled it with water again and it's been under 5 lbs. pressure for an hour now. I'll leave it overnight and hopefully it will still be at 5 lbs. pressure
__________________
Joel Heinke Be original; don't be afraid of being bold! |
#134
|
||||
|
||||
I like the air only inside the tank for testing. no contamination of the metal and no chance of explosion if you need to weld any leaks. A bottle of bubbles that the kids play with and a small paint brush is cheap. HVAC tech's also have a solution that they use to check natural gas joints
your tank looks very nice! Steve
__________________
Steve Hamilton Hamilton Classics Auto Restoration & Metalshaping |
#135
|
||||
|
||||
Quote:
__________________
Joel Heinke Be original; don't be afraid of being bold! |
#136
|
||||
|
||||
Fuel tank construction – Lessons Learned
I’ll start with the good news. I filled the tank with water and put just a little bit under 5 lbs. pressure in it. I came back 24 hours later, and here’s what I observed.
I let out a big whoop as this test passed giving me a lot of assurance all the leaks had been fixed. Just to be sure, I drained all the water and repeated the test with air only. After sitting overnight, the pressure gauge is sitting exactly where it did at the test start. So I’m now real confident this tank will not leak when filled with gasoline which I’m fairly sure is more viscous than air. So what would I do differently if I were just starting to weld this tank up? The base sheet metal is Al 5052 alloy in .063 thickness, the fuel sender and pump flanges are 6061 and the weld filler metal is 5356. This alloy of filler seemed to work well, it flowed and adhered to both base alloys fine, and from all the research I did seems to be the best filler for welding a mix of 5xxx and 6xxx series aluminum. So I wouldn’t change the filler metal alloy. I would however change how the filler was metal delivered. All of the welding on the tank was TIG process, but in some places I used 1/8” TIG rods, others 1/16” TIG rods, and for most of the sheet metal seams 3/64” MIG wire. I used the MIG wire because I hate wasting the last 3 inches of every TIG rod and I’ve gotten used to using MIG wire when I O/A weld aluminum. When I reviewed all the places I had to fix a pin hole, none of them were in places where the 1/8” TIG filler was used. I believe there are 2 main factors behind this: 1) the bead length between start/stop was much longer with the larger filler rod and the resulting beads are a bit wider and more proud than the others. The MIG wire sections had a bead length of about 1 to 1.5 inch which is what I could get from the 6 – 8 inches of wire I could safety hold and feed into the puddle for each welding pass. I can feed a whole TIG rod through my gloves in a welding pass but have difficulty feeding MIG wire, especially when a “perfect” weld is needed. I believe most of the pin holes I had to fix were located at or near a start/stop location. I’m guessing the molten aluminum does unexpected stuff when it’s changing to/from a solid state and thus it’s more likely to get pin holes there than when continuously moving a molten puddle down a seam. Just a guess but I think a somewhat educated guess. There’s a lot of filler in a 1/8” TIG rod, so you can run a very long bead between start/stop and the amount of waste is minimal per inch of bead. You just need to make sure you’re not cooling the puddle too much as you introduce filler into the puddle as this is an “over sized” filler rod for .063 thickness base metal. I’ll continue to use the MIG wire in other welding situations but not for fuel tank welding. The other thing I’d do different is to glue on tank mounting tabs versus welding them on. Here’s an example. This tank has 4 of these tabs. There’s a tray under the tank which supports its weight and these tabs secure the tank from moving around under acceleration, braking and cornering. The tabs started as integral extensions in the .063 thick sides of the tank. After welding the top to the sides, I thought the tabs looked “weakened” due to the annealing that’s a side effect of welding. So I decided to lap weld some .090 aluminum on the outside to bring the tabs to just over 1/8” thickness. I thought this to be safe from introducing leaks because the lap welds are on the outside of the tank. WRONG! All 4 tabs had leaks, these leaks were the hardest to fix and some required multiple “fixes” before they stopped leaking. In addition, the location of the actual leak could not be easily detected since the leaking water only appeared once it got above the welded part. This is an example where I knew there was a leak from water testing but couldn’t detected it with air/bubble solution testing. To fix the leaks behind the added metal, I essentially had to remove the lap weld bead with a carbide burr, re-clean the weld area, and very carefully re-weld with a very fat/wide lap weld. Next time, I’d just glue on the extra metal with the same structural adhesive as used in the chassis. No leaks and I bet it would result in stronger tabs as the new metal wouldn’t also be annealed as well. Well there you go. Long write up but hopefully it helps someone else do a better job on their first fuel tank build.
__________________
Joel Heinke Be original; don't be afraid of being bold! |
#137
|
|||
|
|||
I remember reading in Bengt Blad's book that he paints a chalk/water mix over tank seams, allowing it to dry. When the tank is filled and pressurised, the dry chalk makes any leaks easy to spot.
Not tried it myself, though! |
#138
|
|||
|
|||
Thanks for the feedback Joel. It is very interesting that you found pressurised water better than air at finding the leaks. That's another theory blown!
The start and stop of a weld is often the weak link. I know Kent says in one of his OA welding DVDs that you should always restart the weld back from where the last one ended, so you are overlapping the area. I'm sure you know that with TIG you have to ramp the current off gently at the end. Fluids like paraffin are more searching than water so in theory should be even better, but that is just theory again. I have used diesel in the past, but it stinks. Glad you got it sorted.
__________________
David Hamer |
#139
|
||||
|
||||
Quote:
In the one case where I needed water, the leak was so small no bubbles formed before the bubble solution ran off/dried up. In the other case, the leak was behind a tab, it was a very small leak and the bubble solution didn't do well in spanning the open crack between the two pieces forming the tab. So no bubbles ever formed but water did leak out to show the leak was behind the tab. I couldn't tell where behind the tab but I knew it was there and not somewhere else. So from my experience with leak testing this tank, I would start with air testing and not water testing. I would only use water testing if I knew there was a leak from the air testing (i.e. tank could not hold air pressure over night), but I could not find the location of the leak. It is much easier to fill a tank with air than it is with water. On a related note, I did add some dish soap to the water that I put in the tank. I would not do that again. It didn't help any and made it hard to completely fill the tank due to sudsing. There were suds coming out the fill spout for the last few gallons of water.
__________________
Joel Heinke Be original; don't be afraid of being bold! |
#140
|
|||
|
|||
Thanks for clarifying that Joel. I understand better now.
I'm going to state the obvious here, but a large volume tank with a small leak will take a lot longer to lose pressure than a small volume one. If you can reduce the volume by filling with clean pebbles or similar you increase the sensitivity and so reduce the waiting time. You could reduce the volume with water but if the water is covering the small leak then a few drops per hour could still take a long time for you to see a pressure drop.
__________________
David Hamer |
Thread Tools | Search this Thread |
Display Modes | |
|
|