First a long backstory then I'll get to my current issue. I'm doing SkyDesigns ER tanks although I doubt that's relevant other than just being more possible rivets to leak from. For my first tank I filled it with dyed auto gas before putting on the rear baffle. I had a leak at the inboard bracket. I added sealant and waited a few days. Still leaked. Then I cleaned everything off that I could and re-did the inside sealant. Still leaked. With some good advice on here I ended up covering the whole flange of that inboard bracket with sealant and it passed the leak test. But right when I was about to drain the fuel I noticed a leak out of one of the rivets at the curved nose part of the tank -- where there is only a single-rivet tab on the inside. The scary thing is that didn't leak until day 4 of the auto gas leak test. So I covered that rivet with more sealant and put on the rear baffle.
Then after drying I did the pressure test on my first tank. It leaked. Using soap water I eventually tracked it down to (1) the fuel cap and (2) the tube connecting the manometer to the vent line. I didn't use the best kind of tape so I cleaned off the area around the fuel cap and used really good 3M packing tape. I clamped the tube to the manometer. I made the mistake of spraying soap water on the packing tape to proactively check for leaks. Well that messed the tape up. So my final pressure test had clean and dry tape and I could literally see the air bubble being captured by the tape. But for 4 or 5 days no noticeable pressure loss. The pressure did go down but then back up again. Mostly tracking with ambient temperature so that makes sense. I did put together a spreadsheet with some math to try to normalize the pressure readings based on temperature and atmospheric pressure but that still never fully evened out the ups and downs. But the important thing is that after several days it still was reading similar pressures as day one.
So I wanted to do better on the second tank. I did a better job on the inboard support bracket sealant. For the rib flanges with only a single rivet hole, I just covered them in sealant figuring that it wasn't much more sealant than covering the tail of the rivet and the perimeter of the tab. I'm sure I used 10-20% more sealant on the second tank. The great news is it passed the auto-gas leak test the first time. No issues at all. So I dried it out, installed the rear baffle, then let the sealant dry. Unlike my first tank, I lubed the o-ring on the fuel cap and tightened the nut so it sit nice and tight. I also taped it really well and clamped the manometer.
And that brings me to today. It's failing the pressure test. It's losing about 0.25psi per hour pretty consistently. I've reset to 1psi three times now. I have done three passes of soap water on every edge seam, every fitting, and every rivet. I started with the rear baffle area since that's the part that wasn't tested with my autogas leak test. But I did every other rivet as well. The only thing I haven't sprayed is the taped area over the fuel cap. But, unlike the other fuel tank, I can't see any sort of air bubble under that tape. I'm getting no soap bubbles.
Are there any ideas other than just trying again and again with the soap water? I half seriously thought about submerging it in a pool but I'd worry about damaging the tank. My dad had an idea to fill the tank with propane and then use a propane leak sensor which he says is super sensitive.
Then after drying I did the pressure test on my first tank. It leaked. Using soap water I eventually tracked it down to (1) the fuel cap and (2) the tube connecting the manometer to the vent line. I didn't use the best kind of tape so I cleaned off the area around the fuel cap and used really good 3M packing tape. I clamped the tube to the manometer. I made the mistake of spraying soap water on the packing tape to proactively check for leaks. Well that messed the tape up. So my final pressure test had clean and dry tape and I could literally see the air bubble being captured by the tape. But for 4 or 5 days no noticeable pressure loss. The pressure did go down but then back up again. Mostly tracking with ambient temperature so that makes sense. I did put together a spreadsheet with some math to try to normalize the pressure readings based on temperature and atmospheric pressure but that still never fully evened out the ups and downs. But the important thing is that after several days it still was reading similar pressures as day one.
So I wanted to do better on the second tank. I did a better job on the inboard support bracket sealant. For the rib flanges with only a single rivet hole, I just covered them in sealant figuring that it wasn't much more sealant than covering the tail of the rivet and the perimeter of the tab. I'm sure I used 10-20% more sealant on the second tank. The great news is it passed the auto-gas leak test the first time. No issues at all. So I dried it out, installed the rear baffle, then let the sealant dry. Unlike my first tank, I lubed the o-ring on the fuel cap and tightened the nut so it sit nice and tight. I also taped it really well and clamped the manometer.
And that brings me to today. It's failing the pressure test. It's losing about 0.25psi per hour pretty consistently. I've reset to 1psi three times now. I have done three passes of soap water on every edge seam, every fitting, and every rivet. I started with the rear baffle area since that's the part that wasn't tested with my autogas leak test. But I did every other rivet as well. The only thing I haven't sprayed is the taped area over the fuel cap. But, unlike the other fuel tank, I can't see any sort of air bubble under that tape. I'm getting no soap bubbles.
Are there any ideas other than just trying again and again with the soap water? I half seriously thought about submerging it in a pool but I'd worry about damaging the tank. My dad had an idea to fill the tank with propane and then use a propane leak sensor which he says is super sensitive.
