What's new
Van's Air Force

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

Paint Blisters/Fuel Tank Rivets

As stated in previous replies, we have popped many, and it does not stop them from coming.
Ive popped em early, squeegied em down in the hot sun when everything is flexible. Bubble gone.
Month later, bubble back.
 
i stuck a pin in blisters and mashed them down. the release of pressure made them lay flatter. the ones i removed and had treated, covered with epoxy and repainted, by my professional aircraft paint shop, came back. a waste of effort. robert
 
cawmd82 said:
I think I will sand the paint off a strip of several, maybe 10 bubbles and do a little test. The first 2 I will clean with acetone only, the next two clean and apply green loctite, next two clean and apply some epoxy. Not sure what to apply to the last 4.....suggestions?? After appropriate dry time then apply some good quality, high adherence vinyl trim tape to this row and see what happens? Opinions? Suggestions? Odds for side bets???
Click to expand...

Chuck, you can never go wrong with a good test. Worst case is simply "no conclusion"....but you usually learn something.

Remember, a good experiment is repeatable.

A word on pre-paint cleaning. I spent a 1/2 hour yesterday on a conference call with three gentlemen at Flamemaster, including a fellow identified as their chief problem solver. As of last night they have blister photos (fully developed from Woods, initial formation from Howell) and background information. Most of the conversation covered general topics, including some discussion of solvents. We will talk more.

CS3204 is paintable. Properly mixed and cured CS3204 should not be affected by pre-paint cleaning with any common solvent, including acetone and MEK. The only caution offered was for methylene chloride. The primary solvent found in uncured CS3204 is toluene. Class A has a lot of it, Class B much less. More later.
 
Update on the chemistry front.

First, part of a note in yesterday from Bob Naman, our lab guy.

Dan:

I have reviewed the info you sent. Sounds like yall are a concerned, informed and intelligent group of guys with a nice problem. I'm here to try to help...no promises tho...this is interesting!

There are my thoughts on how to proceed:

I will test for the presence of avgas in the liquid extracted from the blisters. The scan I will do will detect other paint-related compounds: like MEK (Methyl ethyl), Toluene, Xylene, Ethylbenzene, MIBK(methyl-isobutyl). Tol, Xyl,Ethbz are in avgas. The others aren't. Also I will scan for a horde of organic solvents and chlorinated compounds. I will also test the liquid for Lead. Total testing costs about $500 or so.

If just Toluene, Xylene, and Ethylbenzene are present in the liquid then the culprit is avgas. If other paint-like components are present then the paint or sealant is the culprit, i.e. a curing/mixing/surface prep conditions....cause.

Also....there may be an off-gassing taking place under the painted surface. This off-gassing may caused by incompatible surface temperature or the relative humidity during the application.

In basic, Bob will be looking for individual components as well as signatures, combinations of components indicating a particular substance.

Pete Howell will attempt to collect mystery liquid from under newly formed blisters (see his previously posted photos). I say "newly formed" only because it is possible volatile components are evaporating away through the paint on older blisters. We want them all, fresh as possible.

If we have anyone else with fresh blisters, please PM me.

I spoke with Bob yesterday regarding dry blisters. He said he can wash the inside of a supplied blister with nitric acid and do a TE lead check, but as above, most of the volatile stuff will be gone. So, let's hold up on dry blister collection for now. If we get fuel indications from the liquid tests we can use lead checks on dry blisters for confirmation.
 
I don't often make posts to the site but I have been following this discussion with some interest as I am sure several people are. As I am finishing up my 7A I am beginning to think about paint also. I don't want to be disappointed in a paint job if down the road I get tank paint blisters. For reference purposes I did my own tanks and probably used more proseal than I should mainly to insure I would not have to open up a tank later for a fuel leak. I fully encapsulated all the rivet shop heads and all the shanks were covered with sealant before I inserted the rivets in the holes before riveting.

The discussion on this issue has been lively and there have been lots of good posts of how people did their tanks and what problems or not have resulted and while informative it reminds me of what I see in manufacturing quite frequently when we run into quality or scrap problems. As a manager in an Automotive manufacturing facility, I will normally get 6 or 7 different "opinions" about the root cause of a scrap problem. The operator, supervisor, maintenance technician, manufacturing engineer, quality engineer, etc. all have the answer to why we have the problem. And usually they are all different and none of them have any data to back up their speculation. Most of the time if you get them all in the same room and do real root cause analysis and gather some data you can get to the cause of the problem and fix it. Sometimes though it takes a little more effort. From what I have read I think this is one of those problems that requires the extra effort.

In those cases where we have a problem we can't seem to solve we do what is called a DOE (Design of Experiments) run by our Quality group. This approach for one problem in my plant in 2008 saved us well over $200,000 in scrap. The DOE approach in a nutshell is to vary a few inputs to measure an output. In this case tank paint blisters would be the output measurable. My suggestion would be to take the test results from Dan's lab testing and use that as a baseline for selecting some input variables (old manufacturers proseal, new manufacturers proseal, paint prep with solvent, no solvent paint prep, fully encapsulated rivets, non-encapsulated rivets, etc. - these are only examples and by no means am I suggesting these are the variable inputs) and put together a design matrix. We would need some sort of test vessel(s) and I think someone suggested that either in this thread or another to put the variable combinations together.

Surely we have a person with a quality background among our group who has a 6 Sigma Blackbelt and could easily put together the matrix so we could present it to Vans. If not, we might suggest Vans contract somebody to do the work if we can show the data from Dan's Lab results to them and show them they need to help solve the problem. I think we have a good start with the lab testing going on but until some true data gathered from known inputs and outputs which is presented and can be analyzed for real root cause we will just continue post about the problem and not solve it. Just my $.02 worth from my years in manufacturing and seeing similar situations.
 
Food For Thought

David-aviator said:
......And we could rule out fuel because there is no fuel at the aft bulkhead outside flange.
Click to expand...
rvbuilder2002 said:
....Thanks Dan...This is the type of analysis that needs to be done to gain an understanding of what has caused some tanks to do this and many others not do it.
Click to expand...
DanH said:
......If just Toluene, Xylene, and Ethylbenzene are present in the liquid then the culprit is avgas. If other paint-like components are present then the paint or sealant is the culprit.....
Click to expand...
We can continue to theorize as to what is causing so many reported cases of paint blisters. Theories abound and include some form of surface contamination, sanding methods, incompatibility or solvent issues, defective proseal, modern paint formulations, even the location of vent lines have been discussed as possible causes. All food for thought. In the end, I remain convinced the problem is not all that complicated. We'll see.

While there are exceptions to every rule, based upon polling results so far, I am encouraged that perhaps soon, we may finally be able to dismiss many theories and focus on a consistent probable cause. If nothing else, it is hoped the planned laboratory tests will confirm once and for all if the blisters originate from inside or outside the fuel tank. This is HUGE. I sincerely hope all those affected QB and SB builders take advantage of this rare opportunity to actively participate by submitting a sample or two for detailed scientific analysis. What is discovered may well answer your burning question...should I repaint or not?

Dan Horton has graciously contributed his time, money and effort organizing this inquiry for the benefit of everyone. Even if fuel is discounted as the common denominator, one should continue to respect this cold hard fact: When it comes to fuel tank construction, the basic laws of physics will not be denied. The distance between the baffle web and the center of a rivet is only about 1/2". That is not very much distance so the FAY seal must be completely intact and without void. Improperly applied proseal amounts to poor craftsmanship and given the slightest chance, poor craftsmanship will be exploited by fuel in one way or another. Uncontained fuel found immediately outside the fuel tank subassembly can be observed as an outright leak, a telltale weeping rivet, or as I continue to believe...a paint blister. Of course strictly speaking, the latter symptom I mention is an unverified theory. Soon, that theory like all the other theories may finally be put to the test by scientific testing. Again, we'll see.

141nw2e.jpg
 
Last edited:
No Sample

Another Vafinista supplied me with some small syringes and tiny needles, but there is not enough liquid in the blisters to get any even in the needle. :confused: My blisters are really small and only one of the several I pierced this morning had any liquid. Unfortunately, it was tiny amount.

Hopefully some of you with bigger blisters will have some luck extracting fluid for testing. I agree with Rick, this is important.
 
Ok Pete, good try. We all appreciate it.

Have a question about the one that had a tiny bit of fluid in it. Was it a blister in the very early stage of formation, like the pair to the right of your finger in this photo?

 
No

Dan,

It was like the one by my finger, with a more developed "dome". The smaller ones did not seem to have any fluid (or any that I could get out).
 
not sure of contents of blister

haven't popped one yet. You need to be a doctor or a junkie to get a syringe around here. I'm neither. I tried to get one off my doctor's nurse when I went in for an allergy shot but no can do. I was actually waiting for instructions from DanH on how I should go about this before I made an attempt.
 
A geometry question?

Ok, so now I have a completely new thought on these things. If the blisters start out as a flat, hard paint film, but end up as a hard spherical dome that when pressed flat prefers to return to its dome shape, isn't the paint film being stretched??? That is, perhaps the real issue here is some chemical process that swells the paint film. This extra surface can no longer lie flat and must go somewhere, so we get a blister. So perhaps it's not internal gas pressure pushing them up, but rather the paint itself, much like what happens when you heat a piece of sheet metal.
 
andrew phillips said:
haven't popped one yet. You need to be a doctor or a junkie to get a syringe around here.
Click to expand...

Obviously that means Pete is either a doctor or a junkie.

Whoops, slipped into speculation. Good science requires devising an experiment to prove or disprove the theory.

I was actually waiting for instructions from DanH on how I should go about this before I made an attempt.
Click to expand...

I have no idea how a junkie acquires a needle in Canada. Probably get them from their friends in Minnesota ;)

You need a small syringe with a very fine needle, and a clean, inert, vapor tight container, preferable a glass test tube or similar with a good stopper. Two or three drops will do the trick.

We have anybody who can help with equipment? And anyone who has blisters and can round up what you need, go for it.
 
szicree said:
... perhaps the real issue here is some chemical process that swells the paint film. This extra surface can no longer lie flat and must go somewhere, so we get a blister. So perhaps it's not internal gas pressure pushing them up, but rather the paint itself, much like what happens when you heat a piece of sheet metal.
Click to expand...

Fair thought Steve. Can you devise an experiment to prove or disprove?

Frankly I think it would just wrinkle up. Lord knows, I've seen that once or twice.

Gas pressure is a lot more likely. Do a little reading about the vapor pressure and temperature relationship, then look up the numbers for volatile solvents. If I understand the physics correctly (no training, so somebody check me), raising the temperature of toluene (just to pick an example) in a closed vessel from 90F to 180F generates about 7 psi. With numbers like that it doesn't take much actual solvent to blow a bubble in a closed vessel with one diaphram surface. Just park the wing in the hot sunshine.

Note the numbers are much higher than any reasonable tank vent dynamic pressure. And yes, I have an experiment planned. Pulled one of my tanks off a wing this afternoon. Prove or disprove, I don't care. Either is progress.
 
Last edited:
Would it be possible to tape plastic over a popped blister and see if it inflates or if any condensate forms on the underside of the plastic?
 
andrew phillips said:
haven't popped one yet. You need to be a doctor or a junkie to get a syringe around here. I'm neither. I tried to get one off my doctor's nurse when I went in for an allergy shot but no can do. I was actually waiting for instructions from DanH on how I should go about this before I made an attempt.
Click to expand...

Andrew,

When I lived in Saskatchewan, I'm pretty sure that our local TSC had them. You might give the one in Amprior a call. Either TSC or perhaps you could call a "country" veterinarian. Another source would be a family member that is diabetic, which makes me think that a local Shoppers might have them too. The only thing that these syringes are good for liquid samples. I've noticed quite a few reports of dry blisters, in which case there would be nothing to sample.

For dry blisters, what is really needed is a gas tight syringe, suitable for taking gas samples. By gas I am referring to compressed gasses like hydrogen or helium. They are typically glass barreled with tygon pistons and have a positive locking valves on the front of the syringe body. In my former life, I used to use them to collect hydrogen samples for purity checks back in our lab. I would collect a gas sample bring it to the lab and inject the sample on gas chromatograph. In my case, I knew exactly what I was looking for in the way of impurities so a standard GC was quite suitable. In this case as this is an unknown sample, the ideal instrument would be a GC-MS (gas chromatograph coupled to a mass spectrometer detector). Almost every modern mass spectrometer has a built in library function that makes identification of simple unknowns as trivial as hovering a mouse over the chromatograph peaks and letting the computer tell you with 99% certainty what the unknown peaks are. Its quite an amazing thing to see in action, and while its not quite what Hollywood and CSI would have you believe, its not far off.

My thinking is that there are a few ways to get this done.

The preferred method (IMHO) would be to take a "sample" from an unpopped paint blister with a gas tight syringe. Ideally, this would be injected "on site" (at the airport) into a portable GC-MS. I think the guys in the oil patch can do this type of on site analysis, but I personally don't know any labs or companies set up to do this.

A second option would be to collect the unknown as above and mail the entire syringe to a lab capable of this type of analysis. This would take some coordination with the lab, but my thinking is that they would send you the syringes, you collect the sample according to their instructions, lock the syringe and overnight mail the gas filled syringe back to them. They would then proceed with a GC-MS analysis. The lab chemist would have to tell you if this is a viable option or not. The syringes are gas tight, but I don't know if that means they are gas tight for a few hours or for a few weeks. In my example above, I was doing analysis within 2 minutes of collecting the sample, so time was not a factor.

The third option would be to collect the unknown as in the above examples, but immediately inject the unknown into a pre-purged and sealed glass vials. You would then mail the vial to the lab. The laboratory would then analyze the sample with a headspace analyzer coupled to the GC-MS. I would imagine a good lab could send you several prepared and purged vials, ready for unknown injection.

The other problem I see is that Andrew is in Ontario, which means he would be sending unknown samples (either liquid or gas) across the border. There may be some issues getting samples like this to a lab in the States from Canada. The pictures he provided show ideal blisters to obtain samples from however. Andrew, do you do much cross border flying? It might be a heck of a lot easier to simply fly to an airport on this side and meet an RV'er here to collect a sample.

I feel just as frustrated as most people reading this thread. If this had happened about 10 years ago, I could have easily done all of the analysis for us for free as I had access to this equipment on a daily basis. Unfortunately, the company I work for now doesn't have a lab capable of this type of analysis.

Dan, if you're reading this, you might offer up some of the options I mentioned above to the lab you are speaking with and see if they are capable of doing any headspace analysis with GC-MS. They may be able to suggest even better gas sampling ideas than what I thought of. They also need to know that we are dealing with micro liter volumes of an unknown.
 
Here is what is needed to collect unknown gas samples....again, by gas I mean non liquid samples like gaseous nitrogen or helium.

Because of the small size of the blisters, I would recommend the smallest size (25ul). From the pictures I see, there isn't a lot of material we are dealing with. The lab we are working with should be able to supply these for use.

http://www.vici.com/syr/a2.php

050023.jpg
 
Dave, I'm forwarding your entire post to the lab. Excellent information, and given your background I do hope you will participate in this effort.

Chanler, a sealed plastic over rivet lines is the fundamental tool I intend to use for some field experiments concurrent with the lab sample efforts. The plastic I'll try is 3M vinyl sign film. It is stretchy, has a good acrylic adhesive, and I'm familiar with it, as I apply and remove a lot of it from trucks in the course of business. It's also readily available anywhere in North America.

The idea with good experiments is to control variables and make one change at a time.

My own QB tanks have never been outdoors, so the highest temperature they've seen since delivery a few years ago is about 105F on the rack in my hangar. They're still coated with an unknown shipping preservative and have never seen solvents or fuel while in my hands.

I'll start by cleaning the surface with new blue paper shop towels used dry...no solvent or cleaning solution of any kind. This should leave a minute quantity of shipping preservative in the voids around rivet heads. The only other substance present will be CS-3204 sealant. I'll apply vinyl and carefully apply fingertip seal pressure around each rivet, but not necessarily on top of each rivet.

Increased temperature seems to be a constant in all reports. Using scrap aluminum (not the tank) I intend to work out a method of raising surface temperature to the 150F-175F range and maintaining it at that level 4 to 8 hours. Likely candidates are heat lamps or a simple fabricated box oven. I think the temperature range is typical for wings in the summer sunshine; comment or actual measurements are welcome.

I expect ordinary expansion of air trapped around the rivet head to raise a bubble at elevated temperature. However that bubble should disappear when temperature is allowed to return to ambient..and ambient is constant in my shop.

A substance released from entrainment or generated by chemical reaction should not return to its previous state....and the bubble should remain.

Repeated heat-cool cycles may serve as an ad-hoc distillation process. I may be able to generate a liquid.

A negative response (no sustained bubble) means I'll repeat the experiment, but this time with a solvent cleaning of the surface....one variable at a time.

Nothing earth shattering about this process. Its been suggested in one form or another in past posts. It simply needs done. If you have a set of virgin tanks and wish to conduct parallel, identical testing, it would greatly improve the statictical likelyhood of valid information. Anybody game?
 
Dave, today's response from the lab (below) re your excellent comments about collection. Sounds like gas sampling (gas, not gasoline) isn't practical, as you suspected.

We need liquid, which I'm still guessing will be found in newly forming blisters. Everyone check those airplanes closely please. If you think you have a candidate, the lab is offering to supply the syringe and vial.
__________________________________________
Dan

Your builder sounds like he's a smart guy....my comments on his options:

Option 1: A portable GC/MS to be used on site. Not feasible unless you know someone with a portable GC/MS. I don't have one. Not feasible.

Option 2: This will only be feasible with a gas syringe that can take a gas sample. Gas syringes are not stable for a long time period. (the liquid in the blisters is the real unknown...not a gas). Not feasible.

Option 3: This is the most feasible. I can send glass syringes for liquid collection and transfer into vials that I will send to hold the liquid. Our GC/MS has a built-in Wiley 138 and NIST 98 library search program. This will scan for some 200,000 compounds. This seems like the best option. It was what I was talking about earlier. Getting me vials of liquids should be easy. They can send it in the container they receive from me. Also I am well aware of the microliter volumes we are dealing with. SAIA sends to Canada all the time. Also if the liquid contains dissolved Lead, our analysis will pick it up.

Advise me
Bob
 
Thanks Dan,

I was afraid of that. Sounds like the lab confirmed my suspicion that gas syringes hold gas only long enough for sample injection into a GC. Typically, this is in the minute range and not the several day range like I hoped. I figured it would be worth asking regardless.

I do take a minor point of issue with the lab in that yes, the unknown is the liquid, but that leaves out a lot of commenters that noted dry blistering.

Obviously, avgas will volatilize into its primary constituents in vapor form depending on the temp. My hope was to sample the headspace of the paint blisters, looking for the the signature components of avgas. This signature would be easily confirmed with a sample of fuel taken at the same time from the wing tanks. My thinking is that this would have confirmed either outgassing of proseal (indicating an issue there), residual solvents left over from wing prep or avgas signature (confirming either a liquid or vapor path to the outside of the tank).

Someone with liquid blisters needs to be all over option three from the lab. Sounds like they are quite willing to work with us.
 
N5XL said:
...the unknown is the liquid, but that leaves out a lot of commenters that noted dry blistering.
Click to expand...

Oh yeah, we've not forgotten dry blisters. Right now it's being treated as a backup plan. Better if we can get liquid for a full spectrum of components, but if we fail at that, the plan is to try cutting off whole blisters, washing the inside with nitric acid, and doing a lead test.

Give me your thoughts on this point; if the blister has ever contained liquid fuel, surely subsequent evaporation will leave a lead trace. Could we expect the same lead trace if the blister only ever contained fuel vapor?

Obviously, avgas will volatilize into its primary constituents in vapor form depending on the temp. My hope was to sample the headspace of the paint blisters, looking for the the signature components of avgas. This signature would be easily confirmed with a sample of fuel taken at the same time from the wing tanks. My thinking is that this would have confirmed either outgassing of proseal (indicating an issue there), residual solvents left over from wing prep or avgas signature (confirming either a liquid or vapor path to the outside of the tank).
Click to expand...

Agree. All we seem to need is a portable GC ;)

Someone with liquid blisters needs to be all over option three from the lab.
Click to expand...

Yep. Let's inspect those airplanes.
 
DanH said:
Give me your thoughts on this point; if the blister has ever contained liquid fuel, surely subsequent evaporation will leave a lead trace. Could we expect the same lead trace if the blister only ever contained fuel vapor?
Click to expand...

If they had liquid avgas in them, then yes, there should be trace quantities of TEL in the residue. Avgas vapor exposure is another matter entirely.

I took a quick look at the MSDS for TEL.

http://cameochemicals.noaa.gov/chris/TEL.pdf

The best I can tell is that it would be very unlikely that TEL would be present inside a dry blister (dry blister meaning no contact ever with liquid avgas ... avgas vapor only) as the vapor pressure pressure for TEL is very low, even at elevated temps. The hint it would be unlikely is that some MSDS sheets list TEL with an 85 degree C boiling point at 15mm Hg, meaning they have to put it under partial vacuum to determine the boiling point. At atmospheric pressures, TEL decomposes before it boils.

9.1 Physical State at 15° C and 1 atm: Liquid
9.2 Molecular Weight: 323.44
9.3 Boiling Point at 1 atm: Decomposes

Look at the second page of the MSDS under saturated vapor pressure. At 155 F, its not generating much pressure at all... 0.129 PSI.

If this data is correct, you should be able to do a very easy separation between avgas and TEL using nothing more that a heat source and an open beaker (not that I would suggest you do :D )
 
N5XL said:
If they had liquid avgas in them, then yes, there should be trace quantities of TEL in the residue. Avgas vapor exposure is another matter entirely..... The best I can tell is that it would be very unlikely that TEL would be present inside a dry blister (dry blister meaning no contact ever with liquid avgas ... avgas vapor only) as the vapor pressure pressure for TEL is very low, even at elevated temps.>>

I follow your thinking. Darn.

Still, the lead check should yes-no liquid fuel involvement, so let's keep it in reserve as a test which will nail at least one fact.

Ok, back to sampling the gas inside a blister. Can't ship samples to the lab, not stable enough for the shipping time. Don't have a portable GC (at least not yet). We do have portable blisters; they're on fast airplanes.

Dave, you've done gas sampling. Do you think those samples are good for 15 to 30 minutes in the sample syringe? The Mobile Downtown airport is less than 3 miles from the lab.
Click to expand...
 
Regarding headspace sampling of blisters with appropriate syringes...its likely that 15 minutes would be fine Dan. If I were the analyst doing this, I would certainly give it a shot.

My thinking is that within 15 to 30 minutes, there would be vapor in a gas tight syringe to get some telltale peaks on a chromatogram. The thing in our favor is that we aren't looking to quantify amounts, just see if the peaks match known avgas signatures (or not). With any sort of peak being detected, the mass spec should have no trouble identifying it.

Its going to take the lab to buy into the plan you're suggesting..they may still balk at the idea of blister headspace sampling, but I'd certainly give it a shot if I were running the lab. If they go for it, I would think someone from the lab would want to do the headspace sampling. Its not hard to do, just a bit of technique involved. They will also want to take a sample of avgas at the same time.


Good stuff going on here Dan...
 
Just heard back from the lab in Mobile. Bob's machine isn't set up to inject a gas sample; his usual line of work (environmental remediation testing) involves liquid. Says he would need to buy and install a "gas receiving aperture" (whatever that is), but is apparently willing to do so. I'll forward your comments. More later.

Dave, really appreciate your participation. Us poor 'ole dumb rednecks need all the expert guidance we can get.
 
The gas in the syringes could be concentrated. The way to do so is to trap the contents with a "cold finger" trap. This is a u shaped tube that dips down into a cryo liquid (probably liquid nitrogen in this case) and back out to a vacuum pump. The gas is admitted to the other end of the trap and then the trap is sealed and pumped down. The hydrocarbons freeze to the wall of the cold finger. By doing this, the contents of several syringes could be gathered, and a more concentrated sample given to the GC. This is fairly routine stuff when trace samples are all that's available.

I got paid $2.75 an hour to gather samples and test them in the GC and MS during my University days, as a lab aide to a research chemist. Eons ago. :eek:

Come to think of it, if anyone knows a chemist (or chemists aide) in a university doing airborne pollution research, this is exactly the sort of thing they are set up to do. They might just do it for fun.
 
Last edited:
Ted Johns said:
Come to think of it, if anyone knows a chemist (or chemists aide) in a university doing airborne pollution research, this is exactly the sort of thing they are set up to do. They might just do it for fun.
Click to expand...

Ted, thats a great idea. Most university analytical labs usually have some old 5890's set up somewhere...the Chemistry prof would be the person to talk to.

Dan, I guess I'm a bit surprised by the labs comments that they are not set up to do "gas injections." I think what they probably mean is that they are not set up to do "headspace analysis", which is very understandable as it takes special equipment. However, I have to explain why I am surprised by their comments, so forgive me while I get off of track for a moment as it will explain itself when I'm done. For the real chemists among us, forgive me for what I am about to say.

Headspace analysis is interpreted (at least by me) to be a special device that is coupled to the front end of a GC. Its specifically designed to sample and analyze gaseous headspaces of samples. Contrast that with the typical GC analysis where a liquid sample is put in a small vial and liquid is injected, either with an autosampler (where the GC injects the liquid on its own) or manual injection (where the chemist does the liquid injections by hand). The liquid sample injected into the "inlet" of the GC, which is most of the time is heated. In the inlet, the liquid sample gets volatilized by extreme heat where it turns into a gas. The GC (which is basically a highly accurately temperature controlled oven) then uses whats called a carrier gas to push the now volatilized sample through a long thin capillary column of special design. In the capillary column is where the component separation actually takes place as unknown compounds in the sample react with special coatings on inner diameter of the column. Unknown compounds from the original sample that "like" the coatings tend to adhere to the coating and slow down in linear velocity as they are swept through the column. Those compounds that dont "like" the coating proceed at the same linear velocity as the carrier gas. At the far end of the column is a detector that looks for the different components in the carrier gas and plots them out over time for the analyst to interpret. Low boiling point compound and compounds that dont like the capillary coating come out first. Those compounds that do like the coating tend to hang around longer inside the column and only leave due to the constant pressure of the carrier gas. The entire process is done in a gas vapor form which is why its called, you guessed it, gas chromatography. Again, for the chemists in the group, I'm sorry but I think it paints an accurate picture for everyone.

The reason why I am surprised by the labs comments is that GC doesn't care if the sample is in a liquid state or a gas state during the injection sequence. When I did my analysis, I was doing manual gas injections into the heated inlet of a 5890 GC. The same GC was also set up for automatic liquid injections, so before I could do my thing, I would have to remove the autosampler tower so I could access the inlet. Again, the GC didn't care as it was all a gas by the time it got to the detector anyway.

In my thinking, the lab should be easily able to set up a quick and dirty test method if they have a small vial of avgas. They could then play with manual gasseous injections of avgas vapor to see how it behaves. Again, we don't care for peak quantification, just simple peak comparison is all that matters.

Dan, I'd have to bow to what the lab is telling you. I'm sure they know the abilities of their equipment.
 
DanH said:
Dave, really appreciate your participation. Us poor 'ole dumb rednecks need all the expert guidance we can get.
Click to expand...

Dan, I appreciate it, but I'm just like everyone else here...nothing more. I enjoy solving problems as its what I do for a living, just in a different area.
 
N5XL said:
Dan, I appreciate it, but I'm just like everyone else here...nothing more. I enjoy solving problems as its what I do for a living, just in a different area.
Click to expand...

Yeah, well, I've long wished for more participation by our members in professional fields. We have so many with talent, training, and brains, but too often they hide in the tall grass.

You're a pretty good example. I see 17 posts since 2006, yet you can contibute a very understandable tutorial on gas chromatography at the drop of a hat. It's guys like you who will help solve this problem.

Keep it up.

.....and Ted, that goes for you too!
 
In the capillary column is where the component separation actually takes place as unknown compounds in the sample react with special coatings on inner diameter of the column. Unknown compounds from the original sample that "like" the coatings tend to adhere to the coating and slow down in linear velocity as they are swept through the column.
Click to expand...

< chem nerd >
In the columns we used, they were actually filled with an inert adsorbent, I think they were tiny ceramic beads. I don't know if they were coated or not. I do know that Dr. O'Brien fussed around with new "magic beads" from time to time, trying to get better specificity for certain compounds. If the boiling point and molecular weight of two compounds were very close, analyzing a mixture of the two compounds with the "wrong" beads would just show a fat peak, where the right beads would show two distinct peaks as you would like. < / nerd > :)
 
So, I just did a little research on the GC topic, curious if my old alma mater was still doing air pollution research.

Lo and behold, my old professor invented something pretty cool (20 yrs after I was there :eek: ):
http://www.rdrop.com/users/voctec/
Looks to be precisely the instrument one would need for this task.

If you click on the pricing, it's only $2400 to rent one for a month. They also quote per sample analysis.

I wonder if any Chapter 105 locals have some bubbles to sample?

PS: Dan, thanks for the encouragement. :)
 
Ted Johns said:
So, I just did a little research on the GC topic, curious if my old alma mater was still doing air pollution research.

Lo and behold, my old professor invented something pretty cool (20 yrs after I was there :eek: ):
http://www.rdrop.com/users/voctec/
Looks to be precisely the instrument one would need for this task.

If you click on the pricing, it's only $2400 to rent one for a month. They also quote per sample analysis.

I wonder if any Chapter 105 locals have some bubbles to sample?

PS: Dan, thanks for the encouragement. :)
Click to expand...


For what it is worth, I had a 2 hour plus talk about blisters with a retired senior Boeing engineer on his back patio in Seattle during a visit 10 days ago while he was grilling a huge salmon on his huge grill. (He is married to one of my wife's cousins) Granted, we had a couple beers going over all this stuff, but it was covered quite extensively.

Based on his experience the cause of the blisters is fuel vapor. He knows all about pro seal, wet fuel tanks and the problems associated with the technology. I mentioned opening the aft end of my tanks to perhaps apply more pro seal and he suggested another 5 pounds or so will fix the problem. :)

Beyond all that, it would appear the overwhelming majority of builders have NOT experienced blisters. How many responded to the survey? Last time I looked it was set at 21. That's 21 out of over 6000 flying airplanes. OK, not all 6000+ airplane builders are present here and some people do not use paint, so the actual number may be more than 21. But the numbers indicate the failure rate starts at about .3%. Or looking at it another way, about 99.7% of airplanes are ok. I doubt the failure rate is over 1%. Whatever is going on, it is one rare gremlin.

But even one messed up paint job is too many. Good luck with the tests.

If I rebuild my tanks, lots of pro seal will be used. :)
 
Broken Link?

Shockwave said:
I found this link while Googling for more info. The article is from 1980! While it's a bit off topic, I did find it interesting that we are most definitely not alone on this.

http://http://74.125.47.132/search?...+wing+paint+blisters&cd=5&hl=en&ct=clnk&gl=us
Click to expand...

Link didn't work for me. Is this the one?:
http://74.125.47.132/search?q=cache:iHPikpxyFfAJ:www.pazmany.com/newsletters/PL-1_and_2/64.pdf+wet+wing+paint+blisters&=5&hl=en&ct=clnk&gl=us

Or try: http://www.pazmany.com/newsletters/PL-1_and_2/64.pdf.
 
Another Shortcut?

David-aviator said:
.....I had a 2 hour plus talk about blisters with a retired senior Boeing engineer......Based on his experience the cause of the blisters is fuel vapor. He knows all about pro seal, wet fuel tanks...... I mentioned opening the aft end of my tanks to perhaps apply more pro seal and he suggested another 5 pounds or so will fix the problem......
Click to expand...

When it comes to leakproof fuel tank construction, there are no shortcuts. I'm sure it comes as no surprise to you that I completely agree with the retired senior Boeing engineer. Did he discuss with you the vital importance of a good fay seal? If so, did you finally get past that "no fuel at the aft bulkhead rivets" thing?

2pt860j.jpg

Forget the shortcut Dave. Instead of applying what amounts to a 5 pound proseal bandaid, why not do the job right and build a pair of new fuel tanks with the parts you recently bought?
 
Last edited:
Chanler and Pete,

While everyone is looking at the links you've proved for the reasons we all expect, I found something else quite unexpectedly in them that made me pause for a long while.

I distinctly remember as a child growing up hearing stories of my distant family members and the activities they were involved in. My father and mother both talked quite fondly of our distant family members in the Kansas City area. Both distant relatives were professionals in their field and had what were described as unique jobs and hobbies. My "aunt" Mildred was an accomplished chemist and my "uncle" Ken was an avid pilot and ham radio operator. I had heard many times of the story of my uncle Ken's passing. As told to me, uncle Ken was landing a plane with a family member and had a heart attack while on the landing rollout. I've always remembered that story. As amazing as the story was on its own accord, I had a bit of awe of my distant relatives and the things they were doing.

What made me stop in my tracks this morning is on the bottom of page two. At the bottom of the newsletter on page two is a brief article that talks of the passing uncle Ken.

You see, I knew that uncle Ken was a pilot, but I had no idea he was a homebuilder, actively flying plane that he built. Finding his name in a newsletter on a forum talking about wet wing sealing was quite unexpected. Finding out something I didn't know about him from it is quite special indeed.

Sadly, I did a quick FAA N database search for his PL-2 tail number, but see its no longer a part of a PL-2, but now belongs to a Cessna 208 in Juneau, Alaska.

How strange life is to have discovered this, THIS way. What are the chances of finding this out? A few post back, I gave an oversimplified description of gas chromatography...a field that my aunt Mildred would have known well. Here I sit reading over wet wing construction difficulties with other homebuilders...something uncle Ken would have been in the middle of. I also hold an extra class amateur radio license.

I'm not exactly sure how to describe what I'm feeling at the moment, other than a feeling of satisfaction of knowing I'm following a path in life that has been traveled before. Traveling a path covered by relatives that were thought of fondly by all in my family. Its unfortunate that I've never had the chance to meet either of them. I'm sure we would have had a few things to talk about.

To my knowledge, the information in the newsletter was never seen by anyone in my family (outside of aunt Mildred)...until today. I'll be sending an email this morning to Dad, with a copy of the newsletter, a picture or two of a Pazmany PL-2 and a few thoughts to share.

Chanler and Pete, thank you for finding this. I've had a lot to think about this morning and its not even 5am yet!

If anyone would have any old issues of Sport Aviation mentioned that would possibly have a picture of their PL-2, I would most appreciate a copy.
 
Last edited:
I worked on fuel tanks at the airlines and we used PRC 890 class B and then we painted on a second sealant known a PRC 890 class A ( This sealant is like a thick paint ) as a second chance to make sure you have a good seal. To ditto Rick, fuel leaks can travel a long way under paint.

http://buyat.ppg.com/REP_aerospace_files/Sealants/fuel_tank_proven_standards/ps_890_class_a.pdf


Bergdahl Associates, Inc. - Authorized Distributor
Home > PRC-DeSoto Products > Integral Fuel Tank & Pressurized Cabin Sealants

Pro-Seal 890 Fuel Tank Sealant
Description
P/S 890 Class A is an aircraft integral fuel tank sealant. It has a service temperature range from -65°F (-54°C) to 250°F (121°C), with intermittent excursions up to 275°F (135°C). This material is designed for brush and fay sealing of fuel tanks and other aircraft fuselage sealing applications. The cured sealant maintains excellent elastomeric properties after prolonged exposure to both jet fuel and aviation gas.

P/S 890 Class A is a two-part, manganese dioxide cured polysulfide compound. The uncured material is suitable for application by brush in thickness up to 25 mils. It cures at room temperature to form a resilient sealant having excellent adhesion to common aircraft substrates.


LET'S TELL AIRCRAFT SPRUCE WE WANT IT SO THEY WILL STOCK IT!
 
RV505 said:
I worked on fuel tanks at the airlines and we used PRC 890 class B and then we painted on a second sealant known a PRC 890 class A ( This sealant is like a thick paint ) as a second chance to make sure you have a good seal.......LET'S TELL AIRCRAFT SPRUCE WE WANT IT SO THEY WILL STOCK IT![/B]
Click to expand...

First you need a way to access the interior of each tank section. They are not all accessible through the pickup/sender opening....just the first bay.

Again, let's prove or disprove fuel involvement before we go further. I too have suspicions, but truth isn't found by playing favorites.
 
There is no easy fix to a wet wing fuel tank. I have tried alot of those quick fixes but in the end we had to go inside. On the RV You would have to open up each bay with a acess panel. You would probably be better off building a new tank. Thousands of commercial and general aviation aircraft have been constructed with wet wings. This is not a new problem. Why reinvent the wheel and call PRC DeSoto and tell them your problem.
 
Last edited:
Pete,
Thanks for fixing my link. I'm not exactly sure how you were able to untangle that mess, but I'm glad you did. :)

Dave,
That's a really neat story. I had no idea obviously and just zeroed in on the leaking fuel tank rivets. Same problem description, 30 years apart.
 
RV505 said:
Thousands of commercial and general aviation aircraft have been constructed with wet wings. This is not a new problem. Why reinvent the wheel and call PRC DeSoto and tell them your problem.
Click to expand...

Why should someone call up PPG and dump on them when many times its not even known who supplied the sealant. Do you know for sure who the supplier of "proseal" was for quickbuild tank? Someone way back in this long train of a topic even commented that they didn't know who the sealant suppler was when they build their slowbuild tanks...they just used whatever Vans supplied.

Over the years, people here have noticed that Vans has used at least two different suppliers for sealant. Has anyone else but me noticed that when you go to the Vans website that the part number for the sealant they offer is a variation of the AC Techs brand part number? The one oz container appears to be AC Techs product, however if you look at the picture of the 3.5 oz and the quart can, they are Flamemaster products. For those that dont know, the Flamemaster product code is CS3204 and not AC (or MC) 236 or 240 that Vans lists it as being. So I offer up for consideration the we have gotten all three different mfg's from Vans. So now who am I supposed to call.

Further compounding the problem, there are at least 12 different offerings from three different manufacturers that are made specifically for sealing fuel tanks...and thats only counting those that would be used in Class B applications (thixotropic pastes for fuel tank sealing like we're talking about). There are at least 4 different chemistries being used in the 12 offerings. Which one of the 12 was used? Which chemistry? They're all Mil spec so what does it matter right? If it didn't matter, why would there be 12 different offerings..why not just the one from each manufacturer. The data sheets suggest they are all different enough to have its own thread regarding discussions about which would be a better choice. ...and then we still have to apply the stuff correctly on a correctly prepped surface or else is all for nothing.

Its been said over and over in that before you can fix the problem, you have to identify what exactly is causing the problem. If it turns out that fuel vapor is in the blisters, then there is obviously a failure of creating a vapor tight barrier for any number of different reason. The root cause is not the fuel or vapor in the blister, its simply an additional symptom. Finding fuel vapor in the blisters points to problems in the application, preparation or perhaps even a mfg defect with the sealant or tank. Finding anything else in the paint blister points to an entirely different set of root causes. The reason why this is important to know is for the quickbuild tank owners that did nothing but purchase a tank ready to go. Slowbuilders were in (or should have been) in total QA control.

PPG, Flamemaster, AC Tech and Van's isn't going to solve this problem for us and more than likely, they will point the finger at us. Its going to be up to us to collect sound data to prove this beyond a shadow of a doubt. Once we can prove what the root problem is, we can fix it.
 
N5XL said:
Do you know for sure who the supplier of "proseal" was for quickbuild tank?
Click to expand...

Yes......Flamemaster CS3204-B-2, shipped from the USA.

PPG, Flamemaster, AC Tech and Van's isn't going to solve this problem for us and more than likely, they will point the finger at us. Its going to be up to us to collect sound data to prove this beyond a shadow of a doubt. Once we can prove what the root problem is, we can fix it.
Click to expand...

I'm in regular contact with Flamemaster's troubleshooter. I'd describe the man as very helpful.

You're certainly right about helping ourselves. Let's face it, human nature is all too predictable. We have 27 different builders reporting blisters right here in the discussion threads, and most are NOT taking an active role. They want somebody else to solve the problem for them.

Listen up, blistered builders; you're the only people with the evidence. Your fellows can do research, set up tests, make arrangements, and provide science knowledge....but until one of the blistered bunch gathers good test samples and/or conducts controlled experiments on actual blistered tanks we won't have any concrete fact.

Personal opinion (and I hope it fires you up); if you're not willing to work to find the problem, you don't have a right to bitch about it.
 
Last edited:
DanH said:
Yes......Flamemaster CS3204-B-2, shipped from the USA.



I'm in regular contact with Flamemaster's troubleshooter. I'd describe the man as very helpful.

You're certainly right about helping ourselves. Let's face it, human nature is all too predictable. We have 27 different builders reporting blisters right here in the discussion threads, and most are NOT taking an active role. They want somebody else to solve the problem for them.

Listen up, blistered builders; you're the only people with the evidence. Your fellows can do research, set up tests, make arrangements, and provide science knowledge....but until one of the blistered bunch gathers good test samples and/or conducts controlled experiments on actual blistered tanks we won't have any concrete fact.

Personal opinion (and I hope it fires you up); if you're not willing to work to find the problem, you don't have a right to bitch about it.
Click to expand...

Actually, Dan, the response to the subject is driven by guys who have not painted and want answers before they paint - they don't want it to happen with their airplane. There are many, many more builders in that category than those who actually experienced the blisters.

The statistics (for the most part - IMHO) have identified the problem - the blisters are caused by a a gas.

At this level, identifying the nature of the gas is technically very challenging and will be expensive. Without going into a very challenging and expensive analysis that may or may not provide an answer, it would seem more practical to assume the blister gas originates with fuel fumes or surface contamination. (Now I know such an assumption will not satisfy those bent on a scientific answer - but such an answer may not be found here considering the resources available. The next best thing to do is assume what seems most obvious and move forward)

With regard to fuel fumes, there has been plenty of testimony that a diligent application of the sealant will prevent the problem. One guy has built, what did he say, 12 sets of tanks and has never had a leak or blister. Rick Galatti has been working with pro seal for years with a major aircraft manufacturer and knows the techniques that work. Unfortunately, the people who build the quick build tanks may not be doing so with adequate supervision. Of the reported failures, many are QB. (That is the #1 reason I am considering building new tanks. But I have not resolved the issue of match drilling the Z brackets to the spar sheer web, although one guy did tell me how to do it.)

With regard to surface contamination, be careful what materials are used in the process. My professional painter absolutely insists the blisters are not caused by surface contamination - and that assertion is made from a perspective of having painted many, many airplanes with and without wet wing tanks.

For my part, the matter is resolved. I am assuming without much ado that the blisters are caused by fuel fumes and will attempt to correct the issue as such - when I get around to it.

Van's will not agree with this assessment as they have a vested interest in not finding fault with the supervision of the QB tank construction process. That is a separate issue and part of doing business as economically as possible. In final analysis, the few blister events that have occurred are a great annoyance for those who have experienced them, but relative to the total value of the product and how much we like it, it is not a terribly big deal. The total failure rate would appear to be less than 1%. Above all else, it is not a safety issue. It is a nuisance issue and to a certain extent, a money issue to fix it. The airplanes fly very well with or without the problem.
 
N5XL said:
Sadly, I did a quick FAA N database search for his PL-2 tail number, but see its no longer a part of a PL-2, but now belongs to a Cesnsa 208 in Juneau, Alaska.
Click to expand...

After a brief search, I believe your uncle's plane is now registered as N709WB and is based in Florida.

It's fun to find this kind of stuff. My grandfather owned a Swift and we were happy to find it registered in Houston. Over the years my mother loved telling stories about flying with her dad in that plane.
 
Speaking of research....

The mystery "oily substance" applied to QB components is ordinary WD-40. I've reviewed the subject with Ken at Flamemaster. The MSDS says it contains the following hydrocarbons:

64742-47-8 - Deodorized kerosene
64742-88-7 - Straight run white spirit (Stoddard Solvent)
64742-88-9 - Naphtha (petroleum), hydrotreated heavy
64742-65-0 - Lube Oil

...and a proprietary surfactant (wetting agent).

Kens says the hydrocarbons would not cause a deterioration or reversion of the sealant material. The surfactant and any possible activity is unknown.

Common sense says wicking action would place WD-40 under fastener heads and inside open voids. The application of surfactant and hydrocarbons means we must clean throughly prior to paint, and even then we probably don't remove it all from the hidden spaces.

However, the above would be true for the entire airframe, not just tanks. One could argue tank rivets are the least likely to hide WD-40; they're the ones with seal material under the heads.

Probably not an issue, and no point in debating the matter right now anyway. The real purpose of identifying the oily stuff is to catalog any possible chemical traces which may show up in later analysis.
 
You must log in or register to reply here.
Back
Top