Although much of modern leak testing was born out of rocket science, it’s not rocket science as they say. We’ve come a long way since automated leak testing started to become established, Nevertheless, it still sounds so easy – just fill something with air or gas and monitor the pressure change or detect the escaping gas with some sort of gas detector. So what could possibly go wrong?
Well, as with most things there’s a downside risk. Let’s take a look at some common pitfalls for the unwary.
• Using unfiltered air. Unless the air that is used to pressurize the part is clean and dry, there exists the almost certainty that the leak tester will become damaged either eventually or extremely quickly. Some people might be surprised how dirty the air in factories can be before we even begin to notice the taste or irritation. I remember visiting a wheel bearing plant where literally everything touched was covered in a film of oil. I’ve seen leak testers literally dripping with oil when they arrived at the service department. One instrument maker mounted PCBs vertically so that oil mist dripped off and did not pool on the boards.
Metal particles in the air or other trash can also destroy sensitive transducers and precision valves.
Fluid handling is a subject worth having its own specialized category of engineers, so it would be foolish to try and discuss it in detail in a short affair like this blog. However, we can’t simply ignore its importance as it relates to leak detection.
As leak detection specialists we are concerned with dynamic viscosity, the resistance to flow when an external force is applied.
In very simple terms viscosity is a measure of attachment to the neighbouring molecules. The net result is that a more viscous liquid or gas will pass through a hole more slowly than a less viscous one.
Units of Measure
To complicate things, there are different units of measurement for different fields of study. For leak testing geeks our unit of measure is the PASCAL seconds (Pa.s) which may also be expressed as the Newton second per square metre (N s/m2) or as the kilogram per metre second (kg/(ms)). You’ll also come across the poise (P) and centipoise (cP)
1 cP = 10-2 P = 10-3 Pa s
The viscosity of a liquid or gas is determined by the strength of attraction between molecules. The following equation may be used to calculate the ratio between a liquid and gas:
Note: This assumes laminar flow characteristics, which should be verified in practice.
In last week’s blog we discussed the plight of a fuel rail manufacturer trying to work with a very small reject limit. I received several questions about the wisdom of setting up such a test in the first place and so this week I am going to elaborate on leak rate specification.Here are some things to consider when creating a leak test specification.
Is it fit for purpose?
Over the history of modern leak testing, collectively the industry has solved probably tens of thousands of leak testing applications, even more, and a sort of unofficial playbook has emerged for pressure decay leak testing and other methods. Your supplier’s application engineering group is aware of these specifications in one form or another. Here are some examples:
Clearly there is a benefit in selecting a specification that reflects the actual usage conditions. This is where it’s possible that things can become a little difficult. Of course I am referring to the catch all demand that “it must not leak!”. For the record, once again, everything leaks, it’s just a matter of degree. If we corner ourselves by insisting on the tightest specification possible we are inviting trouble and unnecessary expense.
It was the smallest thing but it was a big deal. The customer is assembling fuel rails. The fuel injectors are manually inserted and then the assembly is tested for leaks.
Bob (not his real name) is highly frustrated with the variability of his leak detection data and downright upset with getting a whole lot of false rejects.
Bob’s leak detection system includes a Qualitek mR with a QDIFF differential pressure sensor. The test specification is 77psig with a reject threshold of 0.003psig. Let’s look at it a different way. The reject level is just 0.004%! It’s a very small percentage but nevertheless it is within the capability of the instrument. At this point you might be inclined to say just change the limit up a bit and everything will be fine. The trouble is Bob has to work to a specification and nobody needs high pressure fuel spraying over a hot engine.
In the business of leak detection it’s often easy to create names for things and then those names stick around for a while and eventually become a permanent fixture of our specialized vocabulary.
Great examples of this can be found hiding in plain sight with the manuals and application notes we publish. For example, there are application notes that describe the “valve blow by “ test and the “missing bearing” test. In the medical device industry we have the “occlusion test” and the “creep test” (my personal favorite).
We’ve been asked for these tests so many times that we not only named them and described them but we even went to the extent of creating a test type for some of them and ultimately building a test program for them and including it in the instrument test menu.
Let’s look at one method by way of an example; the occlusion test.
The Occlusion Test.
The word ‘occlusion’ is related to ‘occlude’ which means blocked off or obstructed. With that said, we can deduce the Occlusion Test is designed to detect parts that have a pathway blocked or obstructed. So how do we detect this defect? There are several ways to detect occluded parts and below are the most popular test methods we use:
• Pressure Decay Leak Test
The Sprint iQ occlusion tests are performed with the Pressure Decay test method. The difference between them is how we apply the limits and a slight adjustment to the sequence. The limits are applied such that the pressure change in the test step must decay by a minimum amount or the test rejects. So in other words the part must leak in order for it to pass an occlusion test. The sequence is similar to a pressure decay test. The part is filled, isolated from the pressure source, and then the pressure decay is measured. The one difference in the sequence has no stabilize time. There is no time to stabilize because most of the pressure would be lost during the stabilize step and none would be available to measure during the test step.
The Sprint iQ also has a built-in feature that improves the sensitivity of this test. It is called Occlusion Release. This feature allows the coupling port to turn off a signal used to control the sealing of the outlet port of the test part at the beginning of the test step. Another way this can be configured is to leave the seals engaged and open a valve internal to the tester that vents to atmosphere. The outlet port of the test part is then connected to this valve.
I often get asked if the Qualitek mR can do an occlusion test. Yes it can, just not as gracefully since the QmR does not have the built in Occlusion Release features, however it does have the flexibility to apply the limits for detecting a minimum amount of decay to pass.
So you see that there are different ways to discover the defect of an occlusion and the answer to solving the problem depends upon how you approach it.
One approach is to buy an leak test instrument with the test already built in, as is the case with Sprint iQ.
Another approach is to consider the application and using the tools you already have devise a method to detect the defect.
Some days is seems that no two clients have the same application and I think aloud to myself “why can’t people make do with our standard instruments?” or words to that effect . Then when I take a step back and think about it I remind myself once again that, in most cases, that’s what they will end up buying and with our help we’ll teach them how to use it in the most effective way to solve their (sometimes) unique application.
We design instruments like Qualitek mR, Optima vT and Vector to be the ultimate configurable leak testers with special steps, program jumps, programmable logic and customizable pneumatics so that they are capable of solving pretty much any pressure decay leak test or flow measurement application. To use them we first define the defect and then design the leak test, or other test, so that the defects can be segregated from the norm.
Instruments like Sprint iQ are more market focused and have a degree of flexibility but by and large come preloaded with certain test types that are commonly found in that market.
So which comes first, method or application? For me, application is always first – figure out the objective, design the test and validate the results. But if you do that and then discover that someone has already done that why not take advantage of it and buy a leak tester with the test already built in?
More information about Occlusion testing and how to set one up can be found in Chapter 4.1.4 page 4-16 of the Sprint iQ Owner’s Guide. Also an Application Note (Occlusion Testing AN-20.pdf) on this subject can be found on our website at: http://www.uson.com/product-literature.
The Devil in the Detail
On the face of it this looks like a straightforward leak testing application. The part is made of rigid plastic, it’s not too big and the leak rate specification is comparable to that of other fuel system components. The purpose of the carbon canister or automobile evaporative emission control system is to block or capture vaporized hydrocarbons and prevent their release into the atmosphere. The vapors can arise from refueling, changes in temperature affecting the fuel tank, “hot soak” wherein a hot engine continues to evaporate fuel in the system, and permeation occurring when components of the fuel system become saturated with fuel.
Hydrocarbon molecules are attracted to the non-polar surface of the activated carbon within the canister and they are stored within the pores by physical adsorption. Canister filling occurs during temperature changing events (referred to as diurnal events) and refueling. During vehicle operation the canister is purged back to the fuel delivery system.
The carbon filled canister is required to be tested after filling with the carbon element. The carbon moves during the leak test and compresses absorbing the pressure. This makes it appear as if the canister is leaking. Since the leak testing specification is tight due to the limits expected when examining a part for potential fuel leakage, even the smallest change in pressure must be investigated. When testing carbon fuel recovery with testers using air we must consider the following:
– The carbon will absorb the test air just like it absorbs the hydrocarbons so when doing repeatability tests the wait times between tests on the same part may need to be as much as 5 minutes or more.
– We do not recommend using standard leak testing calculators to calculate cycle times. Although the canister measures 2 liters of volume empty it might look more like 4-6 liters of volume to the leak tester. This again is due to the absorption of the air by the carbon. It just does not want to release the air molecules very easily.
– Due to the geometry of most canisters the external pressure and air currents in the room will change internal the pressures slightly during a test causing repeatability nightmares or false rejects. Hint: Keep the part under test protected from fans or air conditioner vents from blowing air on them, also try to avoid placing the test stand next to doors opening and closing.
These considerations are from years of experience in providing leak test solutions for carbon canisters dating back to the mid 1980’s when the carbon canisters began to grow in size and began to be made with plastic housings. In the early 1970’s (the point when they first appeared in automobiles) they were the small metal cans that looked a lot like coffee cans under the hood.
If you are struggling with testing a carbon canister we’d be happy to help, just call our applications engineering group on 281 671 2000 or use the contact form on this web site.
In the last two blogs we have discussed “Comp & Cal procedures and how to avoid ten common mistakes. If you read those posts you’ll recall that the compensation feature allows us to add back the pressure change that occurs when a good part is tested. The pressure change is caused by the adiabatic effect and using the compensation feature lets us shorten the test cycle because we don’t have to wait for the system to settle down before we start the test phase.
In this post I’m going to introduce another type of compensation – it’s the ability to automatically adjust the leak tester compensation value that we created in the Comp and Cal procedure. Now hold on, I can hear you say we just took great care to identify the correct compensation value and now you’re going to start changing it already? Why would we want to do that?
The answer is pretty clear when you think about why we compensated in the first place. It was because the temperature change caused by the adiabatic effect increased the pressure and created what would appear as a leak as the temperature cooled and the pressure dropped.
Last week we looked at how leak tester comp and call works and we established its value in countering the adiabatic effect. This week let’s look at the “don’ts of Comp and Cal” so that you can be a “Comp and Cal Hero”!
Undeniably it’s a very useful item in our leak tester toolbox but like most tools, care must be exercised when using it. Here are ten situations which must be avoided when using Comp & Cal on your leak tester.:
1. Never perform Comp & Cal procedure without first verifying that you are using the correct master part. This sounds obvious but it’s easy to pick up the wrong part and it happens often.
2. Do not run Comp & Cal without making sure that you have the correct Leak Master or calibrated leak on hand. This is a very common mistake, so beware. Leak Masters look the same and it’s easy to get them mixed up especially when the pressure is on to get the production line back up and running.
In leak testing you’ll frequently come across the expression “compensation and calibration “or “comp and cal”. This week we will explain what we mean by it and why it is important.
One of the challenges in leak testing using air to fill or pressurize a part is the “adiabatic effect”. An adiabatic effect is one in which heat is not transferred into or out of a system yet a change in temperature occurs. The change is due to work being done on the system’s surroundings. A good example is that of a piston compressing a gas in a cylinder. Did you notice how your bicycle tire pump gets hot when you inflate the tire?
The adiabatic effect occurs when the air pressure in a volume is increased. Remember the calculation (P1xV1)/T1 = (P2xV2)/T2. As the pressure is increased, the number of molecular collisions also increases because the distance between the molecules decreases. The collisions create friction and the result is heat. We learn early in life that heating the air in a closed vessel increases the pressure. When the pressure ceases to increase or the work stops the air molecules slow down and the number of collisions decreases. This in turns leads to cooling and a drop in pressure. If we start to measure the leak rate immediately when we stop filling the part, the small drop in pressure from the reduced molecular activity will resemble a leak and we might get a false result.
A complete leak test system will require some means of holding the part in place during the test. In fully automated systems, it is common to use pneumatic cylinders to operate the clamps or, in heavy duty situations, hydraulic clamps might be used.
Many leak test systems are manual or semi-automatic and require some sort of manual clamps to keep the part in place. A type of clamp frequently used in this type of situation is the toggle clamp. In this blog post we are going to discuss the selection and use of toggle clamps for leak testing applications.
Toggle clamps are simple mechanical devices that derive their strength through the arrangement of linkages and levers. They come in several different configurations but they all use the same principle.
When the linkage pivots are all in alignment, the toggle clamp is exerting the most force on the object to which it has been applied. However it would be dangerous to leave the clamp in this position because it could quite easily be shaken loose by vibration or shock. Not necessarily dangerous when holding carpentry together but potentially lethal while clamping a pressurized component. Therefore the design permits a little travel beyond the straight line of the pivot points with a hard stop to limit further travel.