What's new
Van's Air Force

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

Manometer Testing of Fuel Tanks

Pmerems

Well Known Member
Advertiser
For those who have come before me.......

I am in the final stages of my RV-7A and I just sealed up the tanks. That is installed the senders and sealed the access plates. I started leak testing this evening using the manometer method. I must be dense or something because I can't get the manometer to work right.

My setup is simple. Tank is sitting on the bench (36") off the floor. Drain hole is plugged, cap is sealed and 3/8 fitting is capped. 10 foot of clear tubing is attached to the vent fitting and draped to the floor and taped to a vertical support 3 feet above the tank. When I pour colored water into the tube the water fills the tube partially then air gets entrapped in the line then the line files a bit more. So when you look at the line you see colored water then air then some more colored water. I tried several times and I get the same result. When I do get more water in the tube it fills beyond the vent fitting and into the vent line. I quickly remove the fill tube from the vent line and vacuum out the small amount of colored water in the vent line. I understand that I should have a column of water in the tube that is about 2'3" above the vent fitting. I don't get anywhere close to that before it fills to the vent fitting.

I thought maybe I might have a major leak somewhere so I tried the balloon technique and it worked great. Double checked every rivet and seam with out any leaks and the balloon is still inflated after an hour of testing.

I really would like to get the manometer to work but I don't understand what I am doing wrong. Maybe it is just late and I can't think straight or I am having a senior moment.

Any input would be appreciated.
 
I used 3/8" tubing and had no problems getting it filled. I would leave one of the vents on the tank open when you fill the tubing, then seal the tank back up. I can imagine that if the tubing is narrow, the surface tension of the water will capture and hold on to air spaces. You could try flicking the tubing with your fingers to try to dislodge any stubborn air pockets in the tube.

Once the water is filling the bottom of the "U" shape of the tubing, put a pound of air in to move the water column up about 14 inches. Make a mark on the tubing and note the temperature. It will fluctuate several inches up or down depending on how the temperature changes.
 
Also, the volume of air in the tube that your are displacing back into the tank is not enough volume to build the pressure you need to hold the water column, thats why you can't attach the tube then pour in the water. Basicaly the amount of air it takes to bring that 16 gallon tank up 1 psi is more than is in the tube. So, you need a way to add more air to the inside of the tank after the bottom loop of hose is filled with your colored water. One way to do this is to add some sort of valve or other way to introduce air into the tank, either on another fitting, or off a tee on your line right before it attaches to the vent line. You can fill your hose with water before you even attach it to the tank, you can get a good, bubble free fill this way by just using your mouth to suck the water up the tube untill its full the amount you need (sounds like about 7-8 feet ought to work with your setup: 3 feed down, 3 feet back up and a couple of feet above the tank.) Then attach your hose, add pressure to the tank untill the water column rises to where you want it and your done!
 
Adding pressure

cut the valve stem off an old inner tube and press it into a piece of alluminum tubing, add a dab of proseal and your have a great way to pressurize with a bicycle pump. Flare the other end for whatever fitting you are attaching to and your good to go. Stubbing off a short piece of alluminum tubing connected to a piece of hose also works.
 
This is a bit off topic, but I would suggest that you put leak test your tanks by putting fuel in them and seeing what happens.

It is much simpler than the mamometer test and it serves to check the tanks as they will actually be used.
 
Adding Air

If I remember correctly, the test kit you can buy from Van's (I did) has a schrader valve that screws into the tank drain fitting to plug it up. I used it to add (VERY CAREFULLY) a bit of compressed air to get a good solid water column in the tube.
 
you are going to need to put a ball valve on the closed fitting into the tank somewhere besides the manometer connection.

Then you can blow into the 'port' and pressurize the tank and the manometer from the 'back' side.

Get the manometer working first. ie.,, no airbubbles and water at the same level on both sides of the loop.

I wouldn't worry about colored water. You can see plain water just fine in a Tygon tube. Then if you blow it out or other event, you won't have colored water on the floor.
 
This is a bit off topic, but I would suggest that you put leak test your tanks by putting fuel in them and seeing what happens.

It is much simpler than the mamometer test and it serves to check the tanks as they will actually be used.

then he has to figure out what to do with the fuel for the next couple months while everything else gets worked on. I suppose he could use auto gas? and then run it in his car to "dispose" of it? but then there won't be any dye.
 
Air is used instead of fuel so if a leak is found :( then you won't have to restore the inside of the tank by numerous wipes with a solvent in order to apply pro-seal to the offending area.
 
use the manometer

It works just fine. I just pressurize it (after adding water to the tube) by blowing air into the fuel vent. put a short length of tygon on the fuel vent fitting and clamp it with something after blowing into it.

You will probably have to use some grease of something on the tygon to fitting interface so that it will seal. I use a little white lithium on the fitting, then slide on the tygon and secure it with a cable tie. A leak here can fool you into thinking the tank is leaking.

If your tank leaks (at first) its probably the screws in the inspection plate (HOPEFULLY!), in any event you can find really small leaks by spraying a little soapy water on the tank while pressurized. Even a very slow leak with make a soap bubble. I had one that took 5 minutes to make a bubble the size of a marble (really small). I doubt that this would have been easy to find with fuel or water or something in the tank.

Good Luck
 
Paul - the local pressure changes...

..can also affect the manometer levels.

This last weekend we had pretty low pressure with the winds - that returned to normal yesterday.

http://www.wunderground.com/weather...SO48&day=25&year=2010&month=5&graphspan=month

This was about a 0.5% change in the effective volume of the air in your tank - throw some temperature changes in too and that water level can easily change without any leaks...:)

I had the same trouble this last weekend....:(
 
Is the manometer a better method than soap and water?

Here we go again with yet another new thing I have never heard of being done on an RV fuel tank. Never ceases to amaze me but obvioulsy I still am learning and have a long way to go.
What advantages does the manometer method have over the dish soap method? I am guessing you can discover small slow leaks better?
I moved my tanks from inside the hangar, temps about 70 to outside in the sun to do the soap test and my balloon, (actually, it was a similar device made of latex that I got out of my sock drawer, achhummm,) and it blew up so fast from the pressure change that it popped! It's not easy to pop those things.
Can you keep a controlled enough environment (temp and baro) for the manometer to perform well?
 
Thanks for the input

Thanks for all the input. I will try the manometer again once I have completed balloon testing of the tanks.

The first tank showed no sign of any leaks with the bubble test. The balloon remained inflated with only a slight reduction in size over a 20 hour period. Looks good to me.

The second tank showed a very small leak at the outboard baffle/rib seam. That is where the outboard z-bracket is riveted on. I removed a few 1/8" rivets, spread and cleaned the seam and applied new proseal. This is one of the easiest places to fix. Since the leak was very small I was able to bubble test all the other joints, seams and rivets and no other leaks were detected. I will repeat the bubble and balloon test in a few days once the proseal is cured.

I am very optimistic that the tanks will be done in a few days. I will also try the manometer on one tank just to see how it works out.

Thanks for the input.
 
I also recommend the manometer method, once you figure it out, it's very easy to use, and it gives a more quantified measure.

But the comments about local "weather" changes are very true. Soon after I hooked up my first tank, the water level dropped about 1.5-2 inches and I thought I had a leak. But it stabilized at that position, then started to climb back up! It climbed to about 2 inches above its original position. Then it started to drop again. This all took place over a couple of hours, at most. I realized that it was simply responding to the ambient air temperature change caused by the nearby furnance cycling on and off. So I left it hooked up for two full days, and just watched it go up and down the whole time. The range of motion never changed, so I concluded that I had no leaks.

John
 
I also recommend the manometer method, once you figure it out, it's very easy to use, and it gives a more quantified measure.

But the comments about local "weather" changes are very true. Soon after I hooked up my first tank, the water level dropped about 1.5-2 inches and I thought I had a leak. But it stabilized at that position, then started to climb back up! It climbed to about 2 inches above its original position. Then it started to drop again. This all took place over a couple of hours, at most. I realized that it was simply responding to the ambient air temperature change caused by the nearby furnance cycling on and off. So I left it hooked up for two full days, and just watched it go up and down the whole time. The range of motion never changed, so I concluded that I had no leaks.

John


I had the same experience - I connected my manometer in my workshop one evening (about 90 degrees) and left it overnight. I checked it the next morning (about 65 degrees) and was heartbroken to see that it had dropped A LOT. I finally figured it out when later that day it had risen again, so I put it in the house for temperature stability, left it to settle for a couple hours, then marked the fluid level and noted the altimeter setting at that point. I watched it for two days taking into account the local altimeter, that made it easy.
 
Two things to consider

...1 psi is only about 27.7 (let's call it 28) inches of water. This means that the water level in the left tube of the U will be 28 inches different from the water level in the right tube of the U. So, starting from neutral, one side will go down 14 inches which will push the level in the other side up 14 inches.

Temperature will affect the level by roughly .82 inches of water per degree F. So, say you start the test at 70 degrees and pump up the tank (using the aforementioned schrader valve) to 28 inches, then wait a couple hours. If the temperature has gone up to 71 degrees, you should expect to see 28 + 0.82 = 28.82 inches of water.

More info here.

Good luck.
 
Tank may have leak but can't find it!

Not sure which thread to resurrect but I picked this one. I completed my tanks before Christmas, but have yet to definitively conclude whether they are leak free.

I used a water manometer and soap solution. After using packing tape and fuel lube to seal the filler caps, and some fuel lube on the Vans bicycle valve in the fuel drain hole (leaks I discovered with soap solution), I got the pressure to hold well. I did 3-day tests on each tank and found that althought the water level goes up and down like a yo-yo, after taking account of atmospheric pressure and temperature variations there was no evidence of a leak. However I then decided to do a longer test on the left tank, and after almost 2 weeks the water level in the manometer was about 1.25 inches lower than it should have been. (I did a regression in Excel of the water column level against the time, pressure and temperature.)

So I got out the soap solution and found.... nothing.

Now I am a bit stumped. It is possible that some of the change in level is due to evaporation from the open top of the manometer tube - my shop has a dehumidifier and is typically 21C and 40% humidity. I am going blind staring at soap-covered seams and proseal blobs looking for bubbles.

Three questions for the collective intelligence:

1) Exactly how sensitive is the soap test? I found a leak at the bicycle valve fairly easily due to seeing bubbles actually growing in front of my eyes. But if there was a really slow leak, would the soap test be expected to find it?

2) Is there anything else I could try short of filling the tanks with avgas? I do have a hot-tub but that seems like a method which could be a bit risky, especially when you consider how hard it must be to hold a tank underwater!

3) Anyone know what the evaporation rate would be for water out of the end of a tube in an indoor warm and dry environment?

Thanks in advance.
 
set up a control?

...another manometer not connected to anything, same diameter tubing, in same room. As a guy who maintains a humidor in a heated space in winter, I can tell you that water loss to the ambient air can be surprising. over an inch of column in two weeks doesn't sound excessive for evaporation at all to me.

I'm still puzzling the relationship of headspace air in the tank and manometer tubing to reliability of readings - doing thought-experiments while waiting for the coffee to kick in. Something seems amiss with the entire notion, but can't finger just what, yet.
 
I just tested my tanks as well and had the same concerns. I saw no leaks with soap water test. I concluded that if there were leaks too small to detect with soap water, then they are probably also too small to be of any concern with fuel in the tanks. Is this an accurate assumption?
 
Tank testing

I tested my tanks side by side with two manometers , watched them both go up and down with temp and pressure change for a week ,they stayed level with each other the entire time .
 
baro and temp effect readings

..can also affect the manometer levels.
Those both need to be accounted for and, be warned: If you pressurize the tank when it's cold outside and then it warms up a lot, there is a real danger that you can overpressurize your tank. The only thing that saved me was that the open end of the manometer was low enough that the water spilled out when the pressure in the tank got too high. So when you run this test, make sure the open end of your manometer is no higher than the maximum pressure your tank can withstand.
 
Remember PV=nRT from high school chemistry?

You can account for temp/press factors yourself using the formula P1/T1 = P2/T2 where P and T are both in Absolute units. (The "V" term drops out because the Volume stays constant.) Add 459.67 degs to convert the Fahrenheit temp to absolute temperature and add the barometric pressure to the manometer reading (converted to Inches Hg) to get the absolute pressure inside the tank. Be sure to use the barometric pressure rather than the "altimeter setting". Conversion factors: 1" Hg = 13.6" H20. 1 PSI = 27.68 "H20.

Let's say you pressurized your tank to 13.6" H20 differential on the manometer (which is roughly 1/2 PSI). If the reported altimeter setting at your airport was 29.92" and your airport was at 1,000' elevation, then the barometric pressure would be ~28.92" Hg. (using the approximate 1,000' elev. = 1" Hg. conversion factor). The absolute pressure in the tank would then be 28.92 + 1 (13.6? H2O/13.6) or 29.92" Hg. If, during the test, the outside barometric pressure dropped to 28.00, and since the number of molecules in the tank remains constant (assuming no leaks), the weight of water in the manometer column would have to increase to hold a constant 29.92? to satisfy the equation. Thus the water in the column would rise to 26.11? H20 differential (1.92? Hg x 13.6) to make up for the decreased barometric pressure holding the air the tank. This all assumes there were no leaks and the temperature stayed the same.

So you definitely want to pressure test your tank when the barometric pressure isn?t changing rapidly.

Disclaimer: My coffee hasn't kicked in yet so please check my math and methodology before using any of this.
 
One sure way to find out.

Pour about a gallon and a half fuel in it. You can leave tank baffle side down and fuel will cover the source of most leaks. Leave in garage or outside over night. If there's a leak you will see or feel it. You can stand tank on either end too. May not require a whole gallon and a half just enough to submerge the large seams by the baffle.
 
To the rescue!

Reviving an old thread because it led to my ah-HA moment. Had the same trouble with conducting the manometer test as the OP. Was using 1/4" vinyl tube connected to the vent line and assuming the head pressure of filling the tubing would be sufficient to create the required internal tank pressure of 1 PSI. Couldn't figure out why every time I went to raise the water column, the water would pour first into the tank before achieving any delta height.

I then used the alternate hand pump, balloon, and soapy water test with no obvious bubbles anywhere. So i decided to research, then try to manometer again.

After reading this and seeing the light that there wasn't near enough air volume in the tube to pressurize the tank, I filled the tube with water and used the tank test shrader valve to manually inflate the tank and create the 27" of column height. Now I'm playing the waiting game....

I had to share this Eureka moment because I was convinced for a day that my tanks were hopelessly porous.
 
Manometer

Reviving an old thread because it led to my ah-HA moment. Had the same trouble with conducting the manometer test as the OP. Was using 1/4" vinyl tube connected to the vent line and assuming the head pressure of filling the tubing would be sufficient to create the required internal tank pressure of 1 PSI. Couldn't figure out why every time I went to raise the water column, the water would pour first into the tank before achieving any delta height.

I then used the alternate hand pump, balloon, and soapy water test with no obvious bubbles anywhere. So i decided to research, then try to manometer again.

After reading this and seeing the light that there wasn't near enough air volume in the tube to pressurize the tank, I filled the tube with water and used the tank test shrader valve to manually inflate the tank and create the 27" of column height. Now I'm playing the waiting game....

I had to share this Eureka moment because I was convinced for a day that my tanks were hopelessly porous.

Just to be clear, you want a combined 27" of change. I recommend 48" of tube each side so the pressure can't blow out. One level goes down, the other up and equal amount from equalibrium. 13.5" per side is 1 psi. Mark the tubes a d take a temp measurement. The levels will change with temp of the tank. About 3/4 (.821")" total per degree.
 
Last edited:
Correct, I'm measuring 27" of delta height between the water levels on each side of the manometer (closed and open end). As the tank pressure slowly rises, the closed end level gets pushed to the floor and the free end rises up.

Update: Anyone have a great solution for sealing the fuel cap? Seems like that's the only source of leakage i can find right now and I've tried packing tape and gorilla tape over top of the installed cap but neither has been able to completely seal the opening. I've read other use EZ turn to lubricate the sealing ring on the cap but my fuel caps from Van's are designed to operate with a dry ring (per the instructions).
 
Last edited:
fuel cap leaks

I used some bo-lube on my o-rings. All I want is for the seal to hold for a couple days for leak test. It won't hurt the o-ring.
But what ever they can't leak for a leak down/bubble test.
Art
 
Fuel Lube (now called EZ turn) on the gas cap O rings for the leak test. You'll be needing it at some point for other applications anyway. A great product.

Different note: I leak tested mine with UV dye added to ethanol free car gas before the back plate was installed and again afterward. There's no mistaking a leak or it origin. Didn't have any real ones, thank goodness. Be warned, the fuel will wick along the skins. In the photo, it did so to the unriveted backplate holes. Darn near *&^$ myself before I realized what had happened. Good luck.
 

Attachments

  • Tank UV leak check.jpg
    Tank UV leak check.jpg
    291.1 KB · Views: 208
Fuel Lube (now called EZ turn) on the gas cap O rings for the leak test. You'll be needing it at some point for other applications anyway. A great product.

Different note: I leak tested mine with UV dye added to ethanol free car gas before the back plate was installed and again afterward. There's no mistaking a leak or it origin. Didn't have any real ones, thank goodness. Be warned, the fuel will wick along the skins. In the photo, it did so to the unriveted backplate holes. Darn near *&^$ myself before I realized what had happened. Good luck.

Don't use EZ-Turn if you plan on being in any cold weather. You'll cuss trying to get the cap off, and then you'll cuss even more when you try to remove all the sticky muck from the o-rings. It is a great product, but not for this particular application. Clear packaging tape will seal the cap for the leak tests.

Or, use a soap mixture to look for leaks, much more sensitive than a pressure decay method.
 
Don't use EZ-Turn if you plan on being in any cold weather. You'll cuss trying to get the cap off, and then you'll cuss even more when you try to remove all the sticky muck from the o-rings. It is a great product, but not for this particular application. Clear packaging tape will seal the cap for the leak tests.

Or, use a soap mixture to look for leaks, much more sensitive than a pressure decay method.

Interesting. I tested with fuel versus air so really not much choice regarding the fuel caps. The O-rings alone might have been enough but this surely sealed that part. It was a temporary application of just a very light film on the receptacle side and cleaned up quite easily.; albeit Florida weather. It was leak tight for a week during each leak test. The UV "stain" would have been quite evident.

Anyway. Another possible option. There is zero related guess work.
 
Big Man-O-meter

We use one at my day job for the KC10 Extender (355,000 Lbs. of fuel) Its over 10 feet tall..Big One!
 

Attachments

  • Manometer.jpg
    Manometer.jpg
    346.7 KB · Views: 151
Back
Top