P L U M B I N G

A First Look

(This page is still being updated and revised. Content subject to change)

Pipes and Tubes

Know the distinction between pipes and tubes. We are familiar with pipe since they run through our homes, supplying running water and natural gas on demand. Despite this, we often make use of tubing for plumbing liquid rocket systems. Despite being a hollow cylinder for flow, pipes and tubes are quite distinct in their design and application.

Pipes follow known specifications tied to their nominal diameters, which are closer, but rarely equal, to their inner diameters. A 1/2 pipe will have an inner diameter of around 1/2”. Their outer diameters are still fixed to standards such as National Pipe Taper (NPT), but their outside surface are frequently crude and unfinished compared to its often smooth inner finish. Pipes are a basic conveyance, and are usually supplanted by tubes and tube fittings in aerospace application.

Tubes are more direct and specific in its dimensions, and are strictly governed by its outer diameter. The inner diameter is not as important, so long as it exceeds wall thickness requirements, and some cheaper tubes even have a weld bead on the inside. These weld beads rough up the flow and disrupt flared tubing ends, so seamless tubing are typical at added cost. The outer diameter is important for connecting to most tube fittings, which clamp or seal along the outer surface, and so the outer surface must be clean and unblemished.

Parallel and Tapered Threads

When it comes to fittings, threads and their relationship to sealing properties are quite often misunderstood. It can be alarming how often poor seals are attributed to wholly inappropriate installation of fittings. Threaded fittings can be widely divided into two types: Parallel threaded fittings and tapered pipe fittings.

Tapered pipe fittings are commonly found in household plumbing, and a engineer will probably find pipe threads on sensors and valves. Water pipes are usually threaded with tapered threads on the outer surface of a pipe end, since they do not rely on additional sealing surfaces. The standard tapered fitting in the US is NPT, or National Pipe Taper. The “T” stands for taper, not thread. This is to distinguish itself from NPS, or National Pipe Straight. Because the thread diameter tapers down from a larger diameter, at some point, the threads will rub snug against the edges of the threads. This also means that the depth in which an NPT fitting seals at is uncontrolled, and may result in different plumbing lengths in a low tolerance fit environment like a rocket. In this case, one or more fitting may appear to be tightened, but only against the rest of the plumbing, and not against the sealing surface.

In a perfect tapered thread mate, there is a continuous metal-metal seal formed by the thread surfaces filling up all gaps. In practice, this rarely happens without sound and high precision threads, such as those used for NPTF, National Pipe Taper Fuel. Instead, these threads are typically gasketed with PTFE pipe sealant. The pipe sealant, typically in the form of pipe tape, fills up any gaps left behind. 5 wraps of pipe tape in the direction opposite to the direction the fitting will be threaded (to avoid unravelling the tape) and some fluid compatible lubricant will create a cheap and sometimes effective seal. Keep in mind though, that loosening the NPT fitting even slightly will compromise the PFTE seal. Other examples of tapered threads include BSPT, or British Standard Pipe Taper.

Parallel fittings are more closely associated with tube fittings, such as Army-Navy (AN, and mostly interchangeable with JIC 37 degree flare), -lok Double Compression fittings (Such as Swagelok), and O-Ring Boss (ORB), as well as their pipe counterparts National Pipe Straight (NPS) and British Standard Pipe Parallel (BSPP). Importantly, each of these parallel fittings only use their parallel threads for mechanical fastening, and not for sealing. Sometimes the sharp edges of pipe threads are jammed into similarly shaped parallel threads, damaging both and sealing nothing. In a proper parallel fitting, the threads close in a sealing surface such as the conical flare and optional copper conical seal of an AN fitting, the precise tapered ferrule and conical sealing surface of Swagelok, the O-ring and reamed Milspec countersunk port of ORB, the integral conical flares of NPS, or the bonded O-ring of BSPP. Unlike tapered fittings, the depth is more or less consistent. Some variation may occur open installing Swagelok fittings for the first time, where the swaging brings the tube in closer, or when adding conical seals to AN. The orientation of these fittings can sometimes be controlled as well, making them better suited for precision plumbing systems. The independence of the threads from the thickness of the tube/pipe also means they can be used with far thinner tubes, while a pipe requires a certain wall thickness to support threads.

Lightly galled aluminum fitting

Galling

Don’t discover galling the hard way if you can avoid it. If your threads gall, they essentially fuse together and tear itself apart when moving. This can destroy threads, and while tapered pipe threads are the most susceptible, this may happen to any applicable thread. To minimize the risk of this happening, it is important to take steps in prevention and mitigation, because when it happens, it’s already too late.

Oxide layers play a central role in corrosion resistance, but are also a key factor in the galling of stainless steel and aluminum. Stainless steel protects itself using a self regenerating layer of chromium oxide, and aluminum forms a layer of aluminum oxide. Unlike steel and iron oxide, chromium and aluminum oxide doesn’t allow oxygen to penetrate, which means they won’t readily rust through like iron will. This is a useful side note, but not important to the subject of galling.

When the threads are tightened and friction increases, especially with tapered threads, the metal will generate large amounts of heat. You can notice this when tightening or tapping a thread. Galling occurs when the heat of friction abrades the soft metal surfaces at the microscopic level and it starts to fuse. Heat makes metal even softer still, and increases the friction heat and likelihood of fusion.

Oxide layers are important to understand since they form very quickly on stainless steel and aluminum, but they are also very thin. The oxides are also harder than the base metal and have slightly less friction when fastening. When this thinner oxide layer is worn away due to friction and heat, it starts to expose the soft and squishy base metal at a rapid pace. Suddenly all the force that was driving the threads while they were hard are now applied to the softer base material, and by then it is too late. Everything was going fine until it wasn’t.

To prevent this occurring, the engineer can mix and match metals when making pipe seals. Stainless can be used with aluminum to better effect than like materials, but galling is always possible when enough heat and friction are involved. Brass and stainless 304/316 go very well together. Make sure to continue reading into the material compatibility section for more.

To reduce the chance of galling, you need to manage friction and heat. When working with high friction fittings like NPT, you need to use a lubricant. The PTFE tape may not always be enough, and it usually helps to apply a compatible lubricant to the threads, such as Krytox oxygen compatible lubricants. Also make sure to tighten fittings slowly and carefully, letting them cool down when they get too hot.

The final and most important thing to remember when dealing with fittings and galling, is to have spare fittings. When it happens, it’s already too late. Sometimes you can repair the threads with NPT taps and dies, but this doesn’t always work well.

Material Compatibility

Some metal combinations are subject to long term galvanic corrosion. When dissimilar metals are joined, ions are exchanged, corroding metals. This effect is influenced by the differences in the anodic index of the adjoining metal alloys. Some metals, like brass and 304/316 stainless, work well together in the long term, whereas aluminum will corrode when connected to stainless or brass for a long period of time.

Some materials are incompatible with certain fluids, and the rocket engineer must absolutely understand this before designing. Oxygen should not be used with plain or zinc coated steel, and acids have an even narrower range of compatible materials.

Similar metals are preferred with cryogenic systems, where the uniform thermal expansion of like metal fittings can make all the difference in maintaining a good seal under thermal load. In this case, mitigation is necessary.

O-ring seals are also very particular. PTFE is suitable for a wide range of applications like cryo and chemical, but provide inferior sealing qualities. Buna, or nitrile rubber, is good for kerosene, but may swell with alcohol. Silicone is not very suitable for kerosene either, and should be used for alcohol.

Valves

Selecting the right valve for the job requires some understanding of how the valve works. Many rocket groups make the mistake of buying cryo rated valves without understanding what makes them cryo rated, and how more ordinary valves find regular use in cyrogenic rocket systems.

The typical ball valve has three sealing surfaces, the input and output ball seats and the stem packing. The stem connects the ball to the valve handle, and needs to be sealed to prevent the interior from leaking. The packing seal is secured against its sealing surface using a packing nut around the stem. Depending on the kind of valve you buy, this packing nut may be separate or shared with the valve handle. In either case, the nut must be fastened and maintained as well as the rest of the valve.

The ball of ball valve is usually described as full or standard port. Standard port balls have a smaller flow path, typically the same as the pipe ID it’s designed for. It shouldn’t constrict flow when open, and is fine for most applications. Full port ball valves are often preferred for main propellant valves, or run valves. These have as large a flow path as is allowed, and is as wide the the pipe threads that connect it, making it wider than the ID of the rest of the plumbing. This helps to allow for full flow as soon as possible, as the flow area can rise quicker as the valve opens. There are also V shaped ports to do the opposite. The V shape allow for a gradual increase in flow area as the valve opens, which makes it better for flow control such as with throttle.

The interior of the valve is also a closed volume which must be managed when dealing with cryogens such as LOX. When a ball valve is closed, it will trap a small but powerful volume of LOX inside. As it warms up, it will pressurize the interior to over 20 ksi and burst through the weakest part in it’s way, whether that be the seals or the body itself. To use a ball valve for cryogenic use, most ball valves can be modified with a tiny vent hole in the ball upstream of flow when closed. This allows trapped volumes to escape backwards, and provides an opportunity to fully clean the valve for use with oxygen systems. It is important to orient the ball correctly, since orienting it downstream will cause the pressurized fluid to flow right through the vent.