What Interpass Temperature Is Protecting

Why this matters

Interpass temperature is the one number on a welding procedure that a welder can violate without noticing, because nothing on the machine displays it. It is a property of the plate, not of the equipment, and it has two limits that protect two different failures running in opposite directions. Read it as "keep it hot" and you will over-cook a stainless handrail. Read it as "do not overheat" and you will hand a hardenable steel a cracked joint that shows up two days later. The window is two-sided, and which side is about to bite you depends on what is under the arc.

Before you touch the steel to check it: preheated plate holds enough heat to burn through a leather glove, so read it with a contact pyrometer on a handle or a temperature-indicating crayon rather than a bare hand. If you are preheating with a fuel-gas torch, that is combustion in your work area: ventilate, keep a personal carbon monoxide monitor on you in any enclosed or partly enclosed space, and handle cylinders and hose under 29 CFR 1910.253. Everything in this article happens under an arc, which means arc radiation to unprotected eyes and skin at a distance greater than people expect (filter shade for the process and current, plus screens, per 29 CFR 1910.133 and 29 CFR 1910.252(b)(2)) and a fume plume whose constituents follow the base metal and its coating rather than the smoke's appearance: manganese on plain carbon steel, hexavalent chromium on stainless under 29 CFR 1910.1026 in general industry or 29 CFR 1926.1126 in construction, zinc oxide over galvanizing. Those are inhalation routes and they need local exhaust at the arc plus respiratory protection under a written program meeting 29 CFR 1910.134, not gloves.

Preheat and interpass are the same measurement at two moments

Preheat is the temperature of the joint immediately before the arc is struck on the first pass. Interpass is the temperature of the joint immediately before the arc is struck on every pass after that. Same instrument, same location, different moment. Almost every procedure sets a minimum that applies to both, and separately sets a maximum that applies only to interpass, because after the first pass the welding itself is the heat source.

Location matters as much as the number. The code your procedure invokes, most often AWS D1.1 in the edition your project specification or adopted building code names, specifies how far from the point of welding the reading is taken and on which surface. Take it at the toe of the last bead and you are reading the bead, not the joint. Take it thirty seconds after the arc goes out on thin material and you have already missed it.

What the floor is protecting

The floor buys cooling time, and it buys it for two mechanisms at once.

The first is hardness. A hardenable carbon or low-alloy steel that cools quickly from welding temperature transforms to a hard, low-toughness structure. Warm surrounding metal conducts heat away more slowly, so the joint spends longer above the transformation range and comes out softer. The sibling card on what a welded joint changes about the parent metal owns the metallurgy of that band; the point here is that the interpass floor and the heat input the procedure permits are the two field-adjustable controls over it once the joint design and the plate are fixed, and the heat input card owns the second.

The second is hydrogen escape. Diffusible hydrogen leaves warm steel far faster than cold steel, so holding the joint above the floor between passes gives it time to leave before the metal is cold, hard and stressed at the same moment. Hydrogen cracking needs three conditions simultaneously, and the related card on hydrogen owns that model; the floor attacks two of the three.

Direction check: raise the floor and the cooling rate falls, hardness falls, hydrogen has longer to leave, and the heated volume grows, which means more shrinkage and more distortion. Every one of those moves together. There is no setting where preheat is free.

What the ceiling is protecting

The ceiling protects properties that heat destroys, and it is material-specific.

• Quenched and tempered steels got their properties from a controlled heat treatment. Hold the joint above the temperature that treatment was set at and you have re-tempered it in the field, which shows up as lower strength or lower toughness in a band nobody inspects. The steel producer's own data sheet and the engineer own the number.
• Austenitic stainless loses corrosion resistance with time spent in the sensitizing range. A high interpass temperature means the joint re-enters and lingers in that range on every pass. The parent-metal card owns the mechanism and the temperature band; the consequence for you is that the interpass maximum is the operational form of it.
• Any material gets more distortion from a hotter assembly at constant heat input, because the volume that expanded and yielded in compression is larger.

Where a procedure gives a maximum with no minimum, the ceiling is the whole point and waiting is productive work. Where it gives a minimum with no maximum, the reverse.

Where the number comes from, and who owns it

You do not derive an interpass window on the shop floor. It is a function of the steel's carbon equivalent, the combined thickness at the joint (which sets how fast heat leaves), the hydrogen level of the consumable, and the restraint. The WPS owns the number for your job. Behind the WPS, AWS D1.1 in the adopted edition carries minimum preheat and interpass tables by steel group and thickness for structural work, ASME Boiler and Pressure Vessel Code Section IX and the governing piping or vessel code cover pressure work, and the engineer of record owns any departure. Each of those binds through adoption or contract, not on its own.

If the WPS does not state an interpass maximum and you are welding something whose properties came from a heat treatment, that is a question for the engineer before the job, not an assumption on the floor.

The instrument is part of the number

An infrared thermometer reads emitted radiation and converts it using an assumed emissivity. Ground bright steel and mill scale have very different emissivities, so the same plate at the same temperature reads differently depending on where you point. That error is a fixed systematic offset for a given surface, not a random spread, and it runs low on shiny metal, which is the flattering direction on a floor check and the dangerous direction on a ceiling check. Use a contact pyrometer or temperature-indicating crayons on the joint, and if you use infrared, calibrate it against a contact reading on that surface first and treat the difference as an offset you carry.

Worked example: one window, two jobs, opposite corrections

A shop runs two jobs in one shift with the same contact pyrometer.

Job A, hardenable low-alloy plate, 1 in thick, ambient 40 degrees F. The WPS on file states a 250 degrees F minimum preheat and interpass, no maximum, six passes. The welder preheats, runs three passes, and takes a 20 minute break.

• Reading before pass 4, contact pyrometer at the joint: 175 degrees F.
• Criterion: 250 degrees F minimum. Deficit 75 degrees F. Below the floor. Stop.
• Correction applied: reheat to 250 degrees F, confirm with the contact pyrometer, resume. Reheat took about 12 minutes.
• Readings before passes 5 and 6: 290 and 310 degrees F, both above 250, both accepted with no maximum stated.
• Count: six passes, six readings, one below the floor, one reheat.

Striking pass 4 at 175 degrees F on 1 in restrained plate does not produce a visible defect. It produces a faster-cooled, harder band and a joint that has not vented its hydrogen, and the crack arrives a day or two later at the toe or under the bead.

Job B, austenitic stainless handrail, 1/4 in wall. The WPS states a 350 degrees F interpass maximum and no minimum, five passes. The welder is running back-to-back to finish before lunch.

• Infrared reading at the joint on the ground bright surface before pass 3: 300 degrees F. Under the ceiling, apparently fine.
• Contact pyrometer at the same spot, same moment: 420 degrees F.
• Correction, printed: the infrared figure is not a temperature until its emissivity matches the surface. On this ground stainless it read 120 degrees F low, a systematic offset in the flattering direction. The governing number is the contact reading, 420 degrees F.
• Criterion: 350 degrees F maximum. Over by 70 degrees F. Above the ceiling. Wait.
• Cooling to 350 degrees F in still shop air took about 4 minutes. Five passes, five readings, one over the ceiling, one wait.

Same shop, same shift, same instrument problem, opposite instruction: on Job A the correction adds heat, on Job B it removes heat. A welder carrying one habit between them gets one of the two wrong every time.

What would flip Job A. If that plate were thin section rather than 1 in, the combined thickness at the joint is smaller, heat leaves more slowly per unit of arc energy, and the tabled minimum drops. Thickness is in the preheat table for exactly this reason, so a habit learned on heavy plate is not transferable downward without checking the table your WPS points at.

Check it against the rules the siblings state

• Floor honoured with a figure: Job A pass 4 read 175 degrees F against a 250 degrees F minimum, 75 degrees F short, reheated before striking. Passes 5 and 6 read 290 and 310 degrees F, both above 250.
• Ceiling honoured with a figure: Job B pass 3 read 420 degrees F against a 350 degrees F maximum, 70 degrees F over, cooled 4 minutes before striking.
• Instrument basis stated: the 300 degrees F infrared reading was rejected as a low-emissivity offset of 120 degrees F on ground stainless, and the 420 degrees F contact reading governs. No arithmetic in this example uses the infrared figure.
• No number introduced late: 250, 350, six passes and five passes all come from the two WPS documents named before the example opens.
• Rounding direction: nothing in this example was rounded toward acceptance. The 420 degrees F reading was taken at face value rather than averaged with the infrared figure.

References

• AWS D1.1, structural welding code for steel, in the edition invoked by your project specification or the adopted building code, for minimum preheat and interpass tables by steel group and thickness and for the location at which the reading is taken
• ASME Boiler and Pressure Vessel Code Section IX and the governing piping or vessel code, as adopted by your jurisdiction, for pressure work
• 29 CFR 1910.252(b)(2) and 29 CFR 1910.133 for arc radiation and eye protection; 29 CFR 1910.1026 (general industry) and 29 CFR 1926.1126 (construction) for hexavalent chromium; 29 CFR 1910.134 for respiratory protection; 29 CFR 1910.253 for fuel gas used in preheating
• See related: What a Welded Joint Changes About the Parent Metal; Why Hydrogen Is the Enemy in a Weld; How to Read a Welding Procedure Specification