Gas Line Installation for Tankless Water Heaters in Frisco, TX

Fuel gas piping can be sized by more than one method, and which method the installer uses can change the answer by a full pipe size. NFPA 54 section 6.2 and the identically structured IFGC section 402.4 both recognise a longest-length method, a branch-length method and a hybrid pressure method. They are all compliant. They are not all equally expensive to build, and most homeowners are never told a choice existed.

Longest-length method

Find the developed length from the meter to the most remote outlet in the house. Then size every section of the system using that single length column, with each section carrying its own connected load. It is the conservative method and the fastest for an inspector to check, because there is only one length to verify.

Branch-length method

Size the sections along the longest run using that run's length, then size each remaining branch using the developed length from the meter to the most remote outlet on that branch specifically. A branch that is physically close to the meter is allowed to be sized as though it is close to the meter, rather than being penalised for a distant appliance it does not feed.

The difference is worth real money on a tankless job. Take a Frisco house where the most remote gas outlet is a furnace at the far end of a 90 foot developed run, and the tankless unit sits 35 feet from the meter. Under the longest-length method, the branch feeding a 199,000 BTU unit has to carry roughly 199 cubic feet per hour at the 90 foot column. Schedule 40 pipe at 90 feet carries about 83 CFH in 3/4 inch and about 157 CFH in 1 inch, so neither works and the branch has to be 1-1/4 inch. Under the branch-length method, that same branch is sized at its own 35 feet, where 1 inch carries roughly 264 CFH. One inch is sufficient, and you have just avoided pulling 1-1/4 inch pipe through 35 feet of structure.

The published capacities assume specific conditions and stop being valid when those conditions change. The numbers below are for Schedule 40 black steel pipe, natural gas at 0.60 specific gravity, an inlet pressure of less than 2 psi, and an allowable pressure drop of 0.5 inch water column. Change any of those and the table changes with it. A 2 psi system, for example, allows a far larger pressure drop and therefore reads off a completely different table, which is the whole reason a 2 psi conversion can rescue a run that will not otherwise fit.

Two things in that table surprise people. The first is how fast capacity collapses with length rather than with diameter: 1 inch pipe loses more than two thirds of its capacity between 10 and 100 feet. The second is that the jump from 1 inch to 1-1/4 inch roughly doubles capacity, because flow scales with something close to the fifth power of internal diameter rather than with the label on the pipe. When a run is genuinely too long for 1 inch, going up one more size usually solves it outright instead of marginally.

Gas pipe is sized on developed length, which is measured along the centreline of the pipe. In an ordinary system the code does not require every elbow to be converted into feet, because the tables already carry a margin. Where a system contains an unusual number of fittings, the equivalent length of those fittings has to be added. A tankless retrofit that snakes around ductwork in an attic is exactly the case the exception was written for.

Work an example on 1 inch pipe. A 40 foot measured run with eight 90 degree elbows, two branch tees and a shutoff adds 8 x 2.6 plus 2 x 5.2 plus 0.6, which is 31.8 feet of equivalent length. The run is not 40 feet for sizing purposes. It is about 72 feet, and 1 inch pipe at 72 feet carries around 178 CFH, which is short of the 199 CFH a full-output unit wants. The elbows just cost you a pipe size, and nothing about the installation looks any different afterwards.

Corrugated stainless steel tubing is fast. It runs through framing like electrical cable, it needs a fraction of the fittings, and on a retrofit through finished space it can be the difference between a one day job and a three day one. It also has a wall thickness on the order of 0.01 inch. Schedule 40 black iron in the same nominal sizes runs about 0.109 inch at 1/2 inch and about 0.133 inch at 1 inch, which is roughly ten times thicker.

That thickness difference is the entire safety argument. When lightning strikes at or near a structure, current spreads through every conductive system in the building and large potential differences appear between them. If an arc jumps from another grounded metal system to the gas piping, the arc energy has to be dissipated where it lands. Thick steel pipe absorbs it. A thin corrugated wall can be perforated by it, and a perforation in a live gas line inside a wall cavity is the failure mode nobody wants to explain afterwards. Following litigation over lightning-related CSST failures, a direct bonding requirement entered the model codes in the 2009 edition cycle and now sits in NFPA 54 section 7.13 alongside the general metal piping bonding rules of NEC 250.104(B). The practical requirement for yellow-jacketed CSST is a bonding conductor no smaller than 6 AWG copper, attached to a rigid metallic section of the gas piping downstream of the meter and run to the electrical service grounding electrode system, with the connection accessible for inspection.

Everything above assumes the meter can supply what the piping can carry. Frisco is served by Atmos Energy, and residential service is delivered through a diaphragm meter with a stamped capacity rating on its badge. Common residential meters are rated in the range of roughly 250 CFH for a small older set and roughly 425 CFH for a larger one, measured at a half inch water column differential. Those ratings are printed on the meter. They are not a matter of opinion and you can go and read yours. Now add the load. A 199,000 BTU tankless unit is about 199 CFH on its own. Add an 80,000 BTU furnace at 80 CFH, a second furnace at 60 CFH, a 65,000 BTU range at 65, a 22,000 BTU dryer at 22 and a 40,000 BTU fireplace at 40, and the connected load is 466 CFH. That exceeds a 425 CFH meter and comfortably exceeds a 250. Diversity arguments get made here, and there is some truth in them because not everything fires at once, but the sizing convention is connected load and an inspector is entitled to hold you to it.

• Read the capacity stamped on the meter badge before anything else
• Total the input rating from every gas appliance nameplate in the house
• Include anything planned but not yet installed - pool heater, generator, outdoor kitchen
• Where the total exceeds the meter rating, the utility sets the meter, not the plumber
• Utility meter changes run on the utility's schedule, which is measured in weeks rather than days
• The service regulator and the buried service line can also be the limiting item, not just the meter

New or altered gas piping gets tested before it is put into service, and the test is one of the few parts of the job with a stopwatch attached. IFGC 406.4.1 sets the test pressure at not less than 1.5 times the proposed maximum working pressure and never less than 3 psig. IFGC 406.4.2 sets duration by system volume, with a minimum of half an hour for a system under 10 cubic feet of pipe volume. Local amendment can raise both, and the authority having jurisdiction is the final word, so we confirm the current Frisco requirement rather than assuming last year's.

1. 1

   Isolate the appliances before pressurising anything

   Appliance regulators are not built to survive test pressure. Where the test pressure exceeds roughly 1/2 psi, appliances have to be disconnected and the piping capped, not merely shut off at their valve. A test that blows the diaphragm in a furnace regulator is an expensive way to pass a plumbing inspection.

2. 2

   Use a gauge with resolution proportional to the test

   The code requires the test gauge to be able to resolve a meaningful fraction of the test pressure. A 100 psi dial marked in 5 psi steps cannot demonstrate that a 3 psi test held. Low range gauges and mechanical or digital recorders are what belong on the test tee.

3. 3

   Charge with air or inert gas, never with the fuel itself

   The test medium is air, nitrogen or carbon dioxide. Testing with natural gas is both prohibited and self-defeating, since the point of the exercise is to prove tightness without a fuel present.

4. 4

   Let temperature settle before you start timing

   Compressing air heats it. A system charged and read immediately will show a falling gauge purely from cooling, and an installer who does not know that will chase a leak that does not exist. Let it equalise, then mark the start.

5. 5

   Hold, and leave the gauge on for the inspector

   The inspector wants to see the gauge holding, not a photograph of it holding. We leave the test in place through the inspection rather than restoring gas and asking anyone to take our word for it.

6. 6

   Leak check every joint again after gas is restored

   A pressure test proves the system as a whole. It does not prove each joint under working conditions with the appliance drawing. Every joint gets checked with an electronic combustible gas detector and with solution after the system is back in service.

One honest caveat about the test. A pressure test proves tightness at test pressure. It says nothing about whether the pipe is large enough. Those are separate failures with separate symptoms, and a system can pass its test perfectly and still starve the appliance under load. The measurement that answers the sizing question is inlet pressure at the appliance with the unit at full fire and every other gas appliance in the house running, which is why we do that as a distinct step rather than treating the inspection sticker as proof of performance.