What's new
Van's Air Force

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

Van's Fuel Tank Sealing video vs my reality

claycookiemonster

Well Known Member
Firstly, the video (easily available on You Tube) is wonderful! No hedging my appreciation for seeing before doing. I watched and rehearsed several times before doing it.

Secondly, just an FYI, I was finishing up project tanks which were completed up to the baffle installation. I attempted to "leak test" just the nose of both tanks before the baffle installation by resting them in the assembly cradles, surrounded by brown craft paper with water up to the vent line. The water was dyed red to make leaks more obvious, but none were found. Not much hydrostatic "head" but better than nothing.

I used 50cc veterinary syringes to hold the mixed Proseal. They are fine, and the exit of the nozzle is 3/16ths, so - perfect. They are a bit cumbersome to fill, and then initially really a two hand job to squeeze a bead out of them. I found that each tank baffle installation could be done in less than 90 cc of Proseal. So, I mixed up 1 1/2 syringes and got busy.

Be ready: have your garbage bags or bins open and ready so you can throw sticky things adhering to you there without having to hold anything else. Have boxes of rubber gloves opened and ready. Have several rolls of paper towels ready. Have all the popsicle sticks or tongue depressors ready. I'd recommend a small open jar of MEK because your squeezer dies will get sticky and dropping them in MEK will help. Set the music and put your phone away. Go to the bathroom. I found that each baffle installation from start to finish took about 3 hours. Forewarned.

The initial move of laying down the perimeter bead and dropping the baffle on it was fine. Just as the video shows. All parts were cleaned ferociously first and all points where Proseal might want to adhere to aluminum were scratched up with Scotch Brite Pads first.

The next step was challenging. The video talks about "twirling" the AD-41H closed-end rivets into the holes with a bit of Proseal on the shafts to create a seal.My frustration was that they didn't actually easily "twirl" into the holes, despite clecoes in all other holes from baffle to rib. These rivets are NOT tapered, but rather sharp on the edges, so they do not help aligning rib to baffle. What do you do? The baffle has been dropped in place. It's aligned by other clecoes, and yet the AD-41H's (with a coating of Proseal) do NOT want to go easily into the hole. I'm not sure if Van's reamed or relived these holes prior to installation. All I can say is that some FORCE was required via the rivet puller to get them down into the holes. It worked, but I'm going to lay a dome over these rivet heads as I finish up. Because.

The AN-470 AD-4-4 solid rivets on the outer ribs generally went well, though I think the previous owner of my project did not drill out these to full size. Did you remember to do this? What's one more #30 drill covered with Proseal?
At this point, with pull rivets on the baffle to rib, and solid rivets on the outer rib to baffle, I turned the tanks baffle-side down to encourage the Proseal to ooze down. It's a bit of a PITA to rivet upside down, but laying the tanks across the cradle (actually both cradles) allowed me to get the skin-to-baffle 426 AD 3-3.5 rivets while imagining the sealant flowing down perfectly to cover all sins. (Leak testing not done yet. Stay tuned)

Seriously, other than the unexpected drama of the closed end rivets not wanting to easily "twirl" into their homes, all went well. Daytime temps here in UT are mid 60's, so I'm giving the baffles a week to get to know their new neighbors.

Tomorrow will be "Z" brackets, and I'm anticipating the same problems with the longer AD 44-H rivets to attach the "Z" brackets to the baffle. I do not know what help that anticipation will be, other than to minimize the nervous breakdown I experienced yesterday.

After that will be access plates and senders. Yes, I know: "twirl" the screws through Proseal so the threads are sealed too.

Tanks. Tanks a lot.
 
I'm firmly in the camp that Van's sealant gun is well worth 17 dollars.

Scott's video is full of terrific insights and is a must-watch... but I'd also recommend these two videos: a quick one showing how the pre-packaged tube is mixed for the gun, and this one (skip to about 13:30) about how easy it is to use the gun and not make a giant mess.

The disadvantage is that you can't really use it for small jobs (unless you find a small disposable container that fits in the gun, and a nozzle for it). But for doing something like a whole fuel tank, it makes life easier.
 
Last edited:
Small batches of sealant mixed on a metal plate with a spatula, then scooped into a zip top bag (lots left from parts packaging), snip the end and squeeze away. Toss when complete.

Cost is 0, mess is small, and a smallish bag is easy to navigate in any tight spot.

Like the OP said, just like painting etc., it's super important to have your setup ready to go, Mise en place style.

Have a good stand to hold the tank, and expect to take an hour or so for each rib ... think I would do 3 a session and be done for the day.

Wipe your rivet gun tool often to prevent slipping on the ooze out.

Very doable, best of luck on your pressure tests!
 

Attachments

  • 20190906_135546.jpg
    20190906_135546.jpg
    362.5 KB · Views: 370
Great tips and wonderfully visual story telling. I’m still a few months away from this work but will bookmark for future anxiety reducing tips before starting. Good luck and I look forward to hearing about the end results!
 
Tanks

Excellent tips and nice writing style.

I too used the syringes. I put it in my caulk gun for some extra torque. Mostly I just mixed in a cup and applied with tongue depressors. Amounts per rib were almost exactly the same.
I noticed the baffle rivet issue early and used a bunch of rivets to hold position. Apparently the clekos were allowing some alignment drift.
For those coming behind, there are sealed nut plates. Very handy. Install them, seal them and screws can't leak past threads from inside. No problem removing screws.
 
Excellent tips and nice writing style.

I too used the syringes. I put it in my caulk gun for some extra torque. Mostly I just mixed in a cup and applied with tongue depressors. Amounts per rib were almost exactly the same.
I noticed the baffle rivet issue early and used a bunch of rivets to hold position. Apparently the clekos were allowing some alignment drift.
For those coming behind, there are sealed nut plates. Very handy. Install them, seal them and screws can't leak past threads from inside. No problem removing screws.

Where do you recommend that we install the nut plates? Are you suggesting using screws instead of the pop rivets on the baffle? I’m right at this point.
 
Nut plates

Where do you recommend that we install the nut plates? Are you suggesting using screws instead of the pop rivets on the baffle? I’m right at this point.

Sorry, no. NAS1473A08
They take the place of the standard nutplates for the Access Panel. Since they're sealed, the screws don't need proseal. They have an integral rubber seal but I prosealed them and added proseal around the outer edge. My Mentor gave me the tip.
20211001_170625.jpg
 
Day 2 of Tank sealing.

All "Z" brackets are now sealed and riveted in place. One T-708 access plate is in place as well. Ran out of time and sealant to do more.

Be EXTRA careful that your "Z" brackets are oriented correctly! Between the "relative to the spar, or relative to the tank?" question, there's "top down, or bottom up?" too. Just take your time.

The closed in rivets did much better today, maybe they heard me crying in my sleep last night. The video shows placing two rivets, swirled in sealant, into the end holes of the "Z" brackets to act as "pins" to locate things, and this worked like a charm! It's tedious, but not overly challenging. Most rivets did indeed twirl into their holes. A few resisted, but resistance was futile, and eventually they submitted. I used the wedge tool I made from the trailing edge piece, and it was fine.

I had enough sealant to place one access plate. Twirled sealant into the threads of the socket head cap screws I substituted for the supplied screws. (When was the last time you heard the word "twirled" this many times on this site?)

Finding the receiving holes on the end rib is harder than you'd think once everything is gooped up, so go easy on the sealant or mark them or something. You will generate a big build up on sealant that does not go down into the threads when you screw the cover down. So, the screws will be sealed no problem, but be careful to remove sealant from the hex head or screw head so they're not caked with cured sealant the next time you need to unscrew them. Q-tips with MEK did the trick.

After I was done, and just before I walked away, I remembered all the trouble I went to orient the tanks so that the baffle sealant would sink back into the joints. Which way to orient the tank now? I decided that sealant inside the tank is my best friend, so I placed the tank on end with the access plate down. Hopefully whatever sealant oozed out on the inside will settle down against the T-708 and fall asleep there. Forever.

One more day to finish up and then I'm done with Proseal FOREVER. (jinx)
 
Tank tip

Here's another tip.
Sometimes theres a bit of excess sealant that can interfere with the tanks fitting against the spar cap. Let the sealant cure a couple days. Place some strips of wax paper along the spar cap where the baffle sealant fillet gets close. Mount the tank with 1/2 the screws. Leave it overnight. Next day pull the tank and allow it to finish curing. Fillets should be shaped to fit perfect against the spar cap.
 
Tomorrow I'll finish the tanks. One question. The fuel level sender units MUST be grounded to the airframe. The instructions say to ensure at least one screw holding the sender unit in MUST be in contact with the rest of the tank. Can I assume that between all the screws there will be enough conductivity, or should I make a special arrangement? Perhaps put one screw in without sealant, and then seal it over afterwards?
 
Tomorrow I'll finish the tanks. One question. The fuel level sender units MUST be grounded to the airframe. The instructions say to ensure at least one screw holding the sender unit in MUST be in contact with the rest of the tank. Can I assume that between all the screws there will be enough conductivity, or should I make a special arrangement? Perhaps put one screw in without sealant, and then seal it over afterwards?

I think Vans recommended a star washer between the fuel sender and the access plate, and also between the access plate and the end rib to ensure a path to ground.

I have put a star washer and then a connector between the fuel sender and the access plate, and am running a wire to a ground on one of the Z brackets.

https://vansairforce.net/community/showthread.php?t=160141

https://vansairforce.net/community/showthread.php?t=106317
 
If you are using capacitive fuel senders, make sure you goop all wire ends. If not you may have a leak down the wire and out the connector! Ask me how I know. Still have not figured out how I’m going to fix this leak without cutting new access holes in the tanks
 
Finished up the sealing of my tank yesterday. Added a star washer on one screw between the float mounting plate and the access panel, and then added a wire with star washer under a different cap screw and it will be grounded to one of the "Z" bracket bolts. Checked the electrical conductivity and it matches the conductivity I got before mounting the floats. It's delicate to put the float and arm and unit through the hold in the access plate once you've spread sealant! Not much room to spare, but when I turn the tanks over, I can hear the float flipping inside, so it's still free.

Leak checking soon.
 
I think Vans recommended a star washer between the fuel sender and the access plate, and also between the access plate and the end rib to ensure a path to ground.

I have put a star washer and then a connector between the fuel sender and the access plate, and am running a wire to a ground on one of the Z brackets.

https://vansairforce.net/community/showthread.php?t=160141

https://vansairforce.net/community/showthread.php?t=106317

Actually we recommend them under screw heads so that there is good conduction between the screw head and the plate when sealant is under the screw head.

On the RV-12 we now use a screw mounted spade connector tab under one of the sensor flange mounting screws and then connect a dedicated ground wire using a female spade connector.
 
Ok, I guess now, when I introduce myself in a meeting, I should begin like this, "Hi, I'm Clay, and my tanks leaked." As Ozzy Man might say, "Destination F#@%&d!"

Ok, ok. i, like all those who have gone before me, swore I did it right. I watched the video several times and when I dropped the baffle, I almost popped the champagne right then and there, because it was going to be SO GOOD!

Now both tanks have multiple leaks right along the skin/baffle seam. The balloon never even had a chance to blow up. I'm intrigued by people drilling out the skin-to-baffle rivets (presumably keeping the baffle-to-rib pop rivets?) and attempting to clean out and re-introduce proseal there. I know there's a big bead of sealant there because I put it there, so how hard is it to get back to a clean surface when the space is clearly limited to about a popsicle stick's thickness?

Also, because there are 4-5 leaks along the skin to baffle seam, I'm worried that there may be other leaks that aren't being revealed because the pressure is being bled off so quickly. I'd hate to do the hole-cutting repair only to find other leaks.

Destination F'd here.
 
That sucks!

I’m yet to attempt to put the baffle on, but have been having a look at what you can do if it doesn’t seal short of taking the whole lot off again.

There’s a guy on YouTube who uses a vacuum to suck sealant into the gap that isn’t sealed correctly. No one I can see tries to remove the old sealant without removing the baffle.
 
"The baffle" sealing

My tanks aren't leaking, but I have fixed several that are. Putting the rear baffle, which isn't really a baffle in yields the biggest challenge of sealant failure for many, and the leaks along the skin/baffle flange are most common. I have had success spreading/sealing these, but not always. I use alot of PR1422 "A" sealant which is very thin compared to the "B" consistency. It is a must have to seal thin bond lines, or you have to thin down the "B" using MEK. I am currently repairing one for a -6 from another VAF member that was so undersealed it was breached in about 6 places and reseal attempt were futile..so I have had to do the access panels cut outs in all 5 bays and go inside to fillet seal the upper and lower baffle flanges. I prefer to "butter" both surfaces of components going together and make sure there is really good squeeze-out ,cleco in every hole to prevent pillowing and then squeeze the rivets slow. After cure with tank sitting L/E up, I topcoat the external rivet tails and seams with 1422"A" brushcoat to seal all possible voids. At your point of the build, you may want to think about pulling the baffle back out and re-sealing or adding the cut outs now and seal from inside. Its messy (shavings) but not that bad. I have attached a picture of a re-seal of an RV-8 tank from the outside using 1422"A"..2 years and no leaks (fingers crossed) but the cut outs are the best way to go.
 

Attachments

  • Tank.jpg
    Tank.jpg
    455.4 KB · Views: 267
I have Van's repair kits on the way, so the cutout repair version can be done. I'm going to leak check again to try and get a clearer picture of EXACTLY where the leaks are. At first glance, it was all from between baffle and skin, but I want to nail it down.

The circular cutouts seem like a simpler solution when compared to drilling out all the baffle to skin and rib rivets and redoing them.

Like Tin Cup said, "Perfection is not attainable."
 
Consider pulling the baffle

In May and June I went through addressing a number of fuel tank related service bullitens for my RV-4 I recently purchased. Pulling the rear baffle wasnt all that bad. If you can find all your leaks before cutting the big holes then maybe cutting holes is the way to go. If there is any chance you miss a spot, it is going to suck to drill your repair hole out again for another repair then you might as well pull he baffle. The toughest part is cleaning up the old proseal off the tank side. Baffle side is easy to clean up.

In my case specifically, I needed to clean out slosh and everyone suggested cutting holes. I'm really glad I pulled the baffle. Would have been difficult or impossible to clean it all out. Now my tanks look good as new and I watched the tank seal videos as well and succeed first shot fortunately.
 
Problem with the Vans Video

The Vans Video is great except for one thing. When Scott is laying beads for the baffles he is putting it too far away from the rivet holes. If he would have laid the bead down touching the bottom side of the rivet holes then there would a better chance of a good seal there. Doing it like he shows in the video will most likely cause leaks along the back baffle.
 
As to the Van's Video, while I'm now in the position of trying to find a scapegoat other than me, I'd agree about the placement of the bead of proseal relative to the rivet holes. The bead should just about touch the edge of the holes. Or, maybe, just barely NOT cover the holes. Look at the geometry of baffle and skin carefully. There's not much skin beyond the holes, so take advantage of every millimeter you can.
 
The repair begins

I bit the bullet and, with the help of an A&P friend, we managed to remove both baffles today. It's not impossible, but it's not easy and I can't imagine trying to do it alone. Drilling out rivets isn't hard, but some of the pop blind rivets became "spinners" that would not drill out. The real challenge is unsticking the proseal from the rear of the tank ribs. We fashioned a long flat blade from a spare piece, and sharpened it on a sander and used it to literally cut the baffle away from the ribs. Anyway, it's done, but now I have a post-mortem on the sealant placement. I did my best to lay that long bead just as described in the video, just inside of the rivet holes, and I think I did it fairly well. However, when the parts were pulled apart, it's clear that the sealant DID NOT get wiped into the joint between baffle and skin. It just gathered in the corner joint, but was not really present in force where the two pieces lay flat against each other. My A&P friend says he's used what amounts to a paint roller to lay a thin coat of sealant right along the joint, and you can be sure I'm going to do it next time. I'll try to upload some photos of what we found when we opened the tank. Couldn't find the "How to upload photos" sticky.
 
Photos of a disassembled tank joint

Here are two representative shots of the tank. You can clearly see that the sealant did not "wipe" into the joint. There is a healthy bead along the corner, but neither side got sealant as far as the rivet holes. I know having too much sealant between baffle and skin can cause "pillowing" and other problems, but ZERO sealant there is worse.IMG_8933.jpg
 

Attachments

  • IMG_8937.jpg
    IMG_8937.jpg
    323.4 KB · Views: 262
How long had the ProSeal been mixed during the work cycle? From the time the harder went into the cup to the placement of the back baffle plate to the tank and all clecoed together? Possibly the pot life/ working time may have been exceeded.
 
Fair question. The Proseal shipped on 9/7/21 and was applied on 9/29/21. I only had a pint on hand. Everything else on the tanks was done previously, so this operation began with the baffles. I mixed up two 50 cc syringes and used them immediately. FYI, I had 25 cc left over after doing both tank baffles.
From initial mixing to usage? In total, it was a 6 hour day to do both tanks. Since both of them seemed to leak equally, I'm not sure the time delay between the first one, which would have begun within 30 minutes of mixing, and the second which might have begun perhaps 3 hours after mixing made a difference.
I can tell you that cured proseal is a BITCH. I have no doubt it will last for decades. Removing the baffles was not significantly easier on one tank versus the other. I'm happy to assume 50% responsibility for extending the time from mixing to application, but seeing the tanks and baffles after disassembly showed me clearly that there was almost no proseal at the rivet holes. All the sealant was in that nice bead along the corner, right where I put it.
I am absolutely not pointing fingers or blaming anyone else. On balance, the Van's video was spectacular, but of course, it was a guy who has built 20 sets of tanks. I can tell you that i found one of the blind rivets had not gone through the hole on the rib at all, but my efforts to put it in place had actually bent the rib flange down because the holes were not lined up. Next time all the holes get a cleco until right before the blind rivet is installed, but there is no way to line up those holes at that moment if they don't want to.
Most of us will only do one set of tanks in our building experience. So, this was my shakedown tank sealing. For better or worse.

My lessons learned: Other than all the same cleaning and preparation as before, I'm going to make sure there is a thin coat of proseal along the facing parts of the baffle/skin. THEN I'm going to lay a bead right on top of the rivet holes. Then I'm going to do everything I can to align the blind rivet holes with clecoes before inserting those closed end rivets.

Then I'm going to pray. Hard.
 
I don't think Scott's video is fair, he make things look so easy but in reality it has not been for me :(

Today was my first day of tank sealant work and the fuel cap and stiffeners went in well but I have the hardest time to put the ribs in without touching here and there first. I have done a couple of dry run at putting the ribs and they do not go in that easy, I can't imagine what a mess it is going to be with the proseal on them.

@Clay, what type of die did you use to mix with your water for the test? I was thinking of doing this but with fuel but that is way too dangerous so I scrapped that thought.
 
I just used normal food dye in water. Now, I only did this test without the baffle in place and also before the access plate was installed, so I filled the nose of the tanks as far as I could. They sat in their cradles wrapped in brown craft paper which I hoped would show leaks. There's not much "head" pressure in this scenario, but it was the best I could do.
 
I just used normal food dye in water. Now, I only did this test without the baffle in place and also before the access plate was installed, so I filled the nose of the tanks as far as I could. They sat in their cradles wrapped in brown craft paper which I hoped would show leaks. There's not much "head" pressure in this scenario, but it was the best I could do.

The specific gravity of water is higher than it is for 100LL so it is still possible to have a fuel leak, even if no water leaked out.
 
Ok, I'm not chemistry expert, but another poster said that the water molecule was smaller than a gas molecule, so water would leak MORE.

Inquiring minds want to know.
 
Let gravity work for you

I'm still a firm believer in turning the tank upside down (standing on trailing edge of skins) as soon as you have dropped the baffle in place and placed a couple of cleckos.

More awkward working under the tank (ends sticking out a bit over the table or on sawhorses) when placing the 1/8" pull rivets and setting the 3/32" rivets, but sure beats leaks along baffle edges.

But why listen to me? I've only done one set of (non-leaking) RV-4 tanks.

Finn
 
And, back to the main point of this threat, my leaking tanks, I did the "nose-up" bit too. As soon as the clecoes were set around the perimeter and the blind rivets NOT going through the "Z" brackets were set, I turned it tail side down and did the rest of the rivets upside down. A bit awkward, but doable. Sadly, for reasons not fully known yet, it did not work. Cleaning and reloading for Baffle 2.0 is in progress. I'll report on that as soon as I test it.

But, on to the newest twist. Does anyone authoritatively know whether water or avgas will leak more easily, given a constant gap to penetrate?
 
But, on to the newest twist. Does anyone authoritatively know whether water or avgas will leak more easily, given a constant gap to penetrate?

I'm a plumbing contractor and I can tell you that yes, 100LL will leak through when water won't. Gas floats on water which means that water has a higher density than gas.

-Marc
 
I wonder if Scott has seen the attached pictures and has any thoughts or comments about placement of the proseal on the baffle. I tend to agree to put the proseal right at the bottom of the rivet holes to make sure it gets smeared with the back of the baffle otherwise it seems we are only at the mercy of gravity to get that to where it needs to go.

I have just started by tank work and honestly I am quite intimidated by it, mainly there is no room for a mistake. It seems we have to be 100% successful whereas holes need to be only 1%
 
Here are two representative shots of the tank. You can clearly see that the sealant did not "wipe" into the joint. There is a healthy bead along the corner, but neither side got sealant as far as the rivet holes. I know having too much sealant between baffle and skin can cause "pillowing" and other problems, but ZERO sealant there is worse.View attachment 17116
Clay,
Is it my eyes are playing a trick on me or do I see the blue plastic on the inside of the baffle still on, this is on the first pictures that you have taken the removed the baffle and has a number of scratches on it.
 
No, but good eyes. I didn't notice. Just a reflection of the blue skies of Utah, calling me and saying, "When are you going to finish that girl and join me?"
 
I wonder if Scott has seen the attached pictures and has any thoughts or comments about placement of the proseal on the baffle. I tend to agree to put the proseal right at the bottom of the rivet holes to make sure it gets smeared with the back of the baffle otherwise it seems we are only at the mercy of gravity to get that to where it needs to go.

Yes, I have some thoughts and comments but I have not posted any of them to avoid joining into a blame game....

As was mentioned at the beginning of the video, it is a representation of how I have been assembling RV tanks for 26+ years. Not everyone will agree with everything I do but I have built well north of a dozen sets of tanks (I don't really know how many at this point) using exactly the method shown in the video, and as mentioned in the video, I have had one leak on one tank. it was on the left tank of the RV-8A prototype (N58VA). The leak was at the outboard bottom corner at the baffle, tank skin, outboard rib intersection. That instance is why I emphasized in the video a heavier buildup of sealant in the four corners just before putting the baffle in place.

I wont disagree that applying the sealant bead as close to the hole as possible would provide for the greatest likelihood of a smear of the sealant that some are talking about, but as I think I mentioned in the video numerous times, excess buildup of sealant at locations that can induce a dimensional change (just about everywhere on the tank that sealant is used) should be avoided and the baffle flange to skin interface is one of those places.

I think the photos provided in this thread actually prove that it is not necessary in order to get a proper seal.
Lets assume the sealant was applied exactly how I showed in the video (I think it wasn't but I will get to then in a moment). If it was, these two photos show that sealant migrated further aft on the baffle flange and skin than it would have been at application.
In fact, in the majority of the area visible in these photos, we can see evidence that sealant filled all of the void area caused by the bend radius of the baffle flange (the tangent point of that bend radius begins only about .040" fwd. of the edge of the rivet hole, and in the majority of the area visible in the photos, the sealant looks to be about that distance from the edge of the holes).

BTW, this is why we recommend machine countersinking this row of holes. Doing so removes any variable in the fit of these surfaces to each other, that can be caused by different levels of dimple quality, the the distortion that can be cause by dimpling so close to a bend radius, etc.
If a builder chooses to deviate from our recommendation, and dimple countersink these holes, I think the amount, and the application of the sealant is even more critical because it is no long a interface point that allows for a totally net fit of the parts. This is the only place that is so critical because in other instances (like the installation of the end ribs) we have the opportunity to do additional work on the sealant to assure it seals properly.

Additionally, most, but not all (see below) of the sealant on each of the matting surfaces visible on the photos appears to have been cut in order to remove the baffle. That implies that there was good adhesion to both surfaces and that the sealant did smear and flow out across that area

So what do I see that might have been the cause of a leak?

Circled in the photo is what appears to be a void in the sealant like what would be caused if application of the sealant bead was paused and then restarted. I am pretty sure I mentioned (and demonstrated) the importance of "joining" beads of sealant to assure there was no void between them. The points of the two arrows could be a couple of other areas with voids or they could be stop/start points from when the cutting of the sealant was done. Cant tell for sure. Regardless, if you zoom in on the circled point it is clearly a void in the sealant that would have the potential to cause a leak. If there were other points like this on either tank, they could be potential points for multiple leaks.

That is the best I can do as far as analysis goes with the small amount of information available, but in my opinion, blaming the leaks on the sealant bead not having been applied right at the edge of the rivet holes is not correct.
 

Attachments

  • Screenshot 2021-10-18 072743.jpg
    Screenshot 2021-10-18 072743.jpg
    216.6 KB · Views: 205
Checked out Scott's video and boy I wish that was available when I did my tanks. I didn't know what I was doing at all, and only by some great miracle did I end up with properly sealed tanks and only a tiny leak through the fuel sender connectors. I made a gigantic mess and probably used a lot more sealant than I had to. These videos are worth a thousand pictures.
 
Yes, I have some thoughts and comments but I have not posted any of them to avoid joining into a blame game....

As was mentioned at the beginning of the video, it is a representation of how I have been assembling RV tanks for 26+ years. Not everyone will agree with everything I do but I have built well north of a dozen sets of tanks (I don't really know how many at this point) using exactly the method shown in the video, and as mentioned in the video, I have had one leak on one tank. it was on the left tank of the RV-8A prototype (N58VA). The leak was at the outboard bottom corner at the baffle, tank skin, outboard rib intersection. That instance is why I emphasized in the video a heavier buildup of sealant in the four corners just before putting the baffle in place.

I wont disagree that applying the sealant bead as close to the hole as possible would provide for the greatest likelihood of a smear of the sealant that some are talking about, but as I think I mentioned in the video numerous times, excess buildup of sealant at locations that can induce a dimensional change (just about everywhere on the tank that sealant is used) should be avoided and the baffle flange to skin interface is one of those places.

I think the photos provided in this thread actually prove that it is not necessary in order to get a proper seal.
Lets assume the sealant was applied exactly how I showed in the video (I think it wasn't but I will get to then in a moment). If it was, these two photos show that sealant migrated further aft on the baffle flange and skin than it would have been at application.
In fact, in the majority of the area visible in these photos, we can see evidence that sealant filled all of the void area caused by the bend radius of the baffle flange (the tangent point of that bend radius begins only about .040" fwd. of the edge of the rivet hole, and in the majority of the area visible in the photos, the sealant looks to be about that distance from the edge of the holes).

BTW, this is why we recommend machine countersinking this row of holes. Doing so removes any variable in the fit of these surfaces to each other, that can be caused by different levels of dimple quality, the the distortion that can be cause by dimpling so close to a bend radius, etc.
If a builder chooses to deviate from our recommendation, and dimple countersink these holes, I think the amount, and the application of the sealant is even more critical because it is no long a interface point that allows for a totally net fit of the parts. This is the only place that is so critical because in other instances (like the installation of the end ribs) we have the opportunity to do additional work on the sealant to assure it seals properly.

Additionally, most, but not all (see below) of the sealant on each of the matting surfaces visible on the photos appears to have been cut in order to remove the baffle. That implies that there was good adhesion to both surfaces and that the sealant did smear and flow out across that area

So what do I see that might have been the cause of a leak?

Circled in the photo is what appears to be a void in the sealant like what would be caused if application of the sealant bead was paused and then restarted. I am pretty sure I mentioned (and demonstrated) the importance of "joining" beads of sealant to assure there was no void between them. The points of the two arrows could be a couple of other areas with voids or they could be stop/start points from when the cutting of the sealant was done. Cant tell for sure. Regardless, if you zoom in on the circled point it is clearly a void in the sealant that would have the potential to cause a leak. If there were other points like this on either tank, they could be potential points for multiple leaks.

That is the best I can do as far as analysis goes with the small amount of information available, but in my opinion, blaming the leaks on the sealant bead not having been applied right at the edge of the rivet holes is not correct.

Scott,
Thank you for chiming on this and sharing your thoughts. This is exactly what I was hoping to see/hear as over the years I have come to value your opinion very highly and have learned many things from your posts. Sharing the reason for countersink vs. dimple is an example of that which help very much to understand the reason and possibly apply the same understanding at other/similar situation, so again I appreciate it. I have watched your video more than once and like others have side and have founded enormously helpful.
 
The Saga Continues...

Ok. Getting back on the horse here.

My build partner and I spent several days scrupulously cleaning every bit of proseal from the disassembled tank. We managed to drill out all solid rivets without ruining any holes. The few closed end pop rivets were a bit more challenging, but they were removed too. The outside of the skin wasn't involved with any of this effort, so they will still look ok, but the baffle, particularly on the inside was a bit scratched up from the effort to cut the rib to baffle proseal adhesive. The rear of several ribs were a bit twisted too. All of that was attended to, and special efforts were made so that everything would fit perfectly the next time it was assembled. Serious vacuuming and cleaning of the tanks for drilled out rivets and old pieces of proseal happened.

Prep for Sealing Day 2: I had to order more closed end rivets, since the kit only includes enough for one go. Each tank would get it's own mix of proseal, made just before being used. In the end, we did BOTH tanks fully (again, only the baffle and Z brackets) with one pint of Proseal and there was some left over. We used big 60cc syringes for application as before, with one SIGNIFICANT addition.

My build partner is a retired A&P from a major airline, and his contribution was to use a small paint roller to apply a very thin (coat of paint thin) layer of proseal to all joining surfaces. We taped off the work area as you might when scotch-briting the surface. So, on the baffle, the places where ribs come into contact was rolled with proseal. The outer edges where the baffle will slide against the skin was prosealed. On the tank, the rear of all ribs and the places where the skin will be in contact with the baffle was rolled. Initially, we thinned the lump of fresh proseal with acetone, but later we just used it straight.

IMG_9036.jpg

Obviously the goal is to literally apply a very, VERY thin layer of proseal to both sides of the joint, but the sort of layer you'd be unsatisfied with if it were paint. The surface is not actually heavily "coated", it looks like you'd want another coat if it were paint. Using a tiny "trim" roller allowed us to do this very quickly and uniformly on all these mating surfaces, without spending precious curing time with a popsicle stick. Sadly, we did not have a third worker, and our hands were a mess, so no photographs of this stage.

Then, thin roller coat of proseal applied as described above, we did the exact same Van's approved process. Bead of proseal along rivet holes top and bottom. Not covering holes, but right up to the edge of them.

The first time I did the tanks, I did them alone, and from start to finish, it was about 6 hours, without doing the central Z bracket blind rivet process. This time I had a partner. I mixed the first batch of proseal at 10am and the entire first tank was done before noon. Including the Z brackets. There was just a bit of squeeze out along the baffle/skin, and those rivets were effectively set "wet." The second tank went even faster. Mixed proseal at noon, and completely done 90 minutes later.

So, a partner for tank sealing was incredibly helpful. Mixing just one tank's worth of proseal (1 1/2 60cc syrings for a single tank baffle) at a time worked great too. The proof is in the leak checking, and that is a few days away, but the addition of the rollering on of a thin layer of proseal along mating surfaces was easy, and just a bit of squeeze out with no pillowing at all gives me confidence that we may have done something right.

Stay tuned...
 
1.5 to 2 hours sounds like a long time.

I assume that it was just a few minutes from mixing to installing the baffle?

Finn
 
Yup. As soon as the proseal was mixed, I filled syringes and laid out a glop for my friend to begin the roller application. He was done about the time I had 2 syringes mostly filled and we laid out beads and dropped the baffle on. Probably the baffle was dropped within 10 minutes of mixing. For some reason, we found the outer blind rivets on the baffle, the ones that do not go through the Z brackets really hard to get lined up. We had an awl that we pushed through the holes to line things up, not sure why they were so challenging. All the blind rivets that went through the Z brackets went straight in. Since the perimeter rivets were "wet" it was easy to push them into the holes. My friend loaded them and I have a pneumatic squeezer, so it went fast. We did all the riveting, despite Van's video saying to save the blind rivets that went through the Z brackets until another day. So, 120 to 90 minutes from mixing to completely done. they've been stored nose up since then. Testing soon.
 
Ok. Major improvement after the last sealing session, and after adding the roller of proseal to lightly coat the mating surfaces. One tank is rock solid and tight, so that's a good thing. The other one has ONE and only one very tiny pin hole. Amazingly, it was along the dreaded baffle-to-skin border. So, we drilled out 5 rivets and pried the skins apart by a fraction and pushed some thinned proseal (with a bit of a vacuum pulling too) into the gap. We're in the waiting period again. I can tell you that one very specific and tiny leak is a VAST improvement on our initial test, so progress is being made.
 
I will be installing my baffles tomorrow. I am hoping for a success but will not know till it is done + another week. I am planning on the normal method as described in the plans and demonstrated by Scott's video.

I will say, I was very nervous before started the tank project but it was not nearly as bad as I had thought. After installing the first rib, things got much easier with practice and my wife and I did all the ribs in one day and another day I did the J channel and rest of it to the point of baffles. Visually looking at it, it looks like it is going to hold but sadly my vision does not see any of the small holes that air/fuel could scape so I will need to hold my breath till it is tested.
 
Ok. Final soap bubble test complete. Both tanks now soap bubble tight. The second one that had a single, small leak has now sealed. As it happens, a friend of mine loaned me a fuel tank leak detection machine. It's made by a company called, "Redline Detection." It's intended for diesel trucks. This device had the ability to pressurize a tank to a very low pressure, about 1psi, and then detect small leaks by seeing if the pressure holds. It can also inject smoke into the pressurized tank, so you can see where the leaks are!
Since, as of right now, I think we're good to go, I'm going to give both tanks the final QC check on this machine, and pray all is still tight.

So, what have we learned?
1. While Van's video is great, and I take nothing away from it, the builder still has to do it! Laying a bead down like the video isn't enough. Rollering on proseal to the mating surfaces isn't enough. Yes, I know you're rushing and nervous and the pot life is ticking away, but LOOK at your work and ask yourself if you've really done everything you can to make it tight.
2. Don't discount the possibility of drilling out the baffle and doing it over again. It isn't easy, but it isn't terrible either. If your leak situation is relatively localized and clearly identified, I'd recommend drilling out the baffle. The hard part there (in my experience) is to release the back of the tank ribs from the baffle when the sealant is effectively adhesive. The other alternative is to cut BIG holes there and hopefully seal up the first leak, and then seal up the big hole you just made.

I lost my builders virginity to the tanks, but, like Doris Day, I regained my virginity later. Best of luck to you guys out there. If I can do it, so can you.
 
Back
Top