Last month I spent a morning in a sheet metal shop in Sharjah watching a Pittsburgh lock former run 0.8 mm galvanized. The operator had been fighting open seams for two weeks. He had changed the forming speed, swapped the coil supplier twice, and blamed the zinc coating. The real problem was a roll gap of 1.4 mm on a 0.8 mm strip. Almost double what it should have been. We closed it to 0.85 mm, ran a 300 mm test strip, and the seam pull test went from "comes apart with two fingers" to "needs a chisel and a hammer".
That morning is why I stopped trusting the dial on the adjuster and started carrying a feeler gauge set in my bag. The Pittsburgh lock is one of the oldest joints in sheet metal work. It is also one of the least forgiving. Everything about it depends on a handful of dimensions staying inside a window two or three tenths of a millimetre wide, and the machine will run outside that window all day without complaining, without an alarm, and without giving the operator any clue that something is wrong until the ducts arrive on site and the inspector pulls the first joint apart.
A Pittsburgh lock is a folded seam. One edge of the blank is formed into a pocket with a hook folded back inside it, the other edge is formed into a simple upstanding flange. When the two edges meet, the flange is driven into the pocket and the whole thing is closed, either by hand with a hammer and a setting tool or by a seam locker. The finished seam is four layers of metal thick, which is why a Pittsburgh joint is so much stiffer than a simple lap.
Three numbers control whether that works.
The pocket depth is the first. On most machines built for the 0.5 to 1.2 mm range it comes out at 9 to 11 mm. The hook returns about 6 to 7 mm inside the pocket. The upstanding flange on the mating edge is normally 8 to 9 mm tall so that it bottoms out inside the pocket rather than stopping short and leaving a void. If the flange is too short, the seam looks closed but there is no interference fit and the joint will rattle loose within a season of fan vibration.
The second number is the total seam allowance, which is the width of strip the lock consumes along each edge. For a standard Pittsburgh this is 30 to 34 mm on the pocket side and 12 to 14 mm on the flange side. On a 1250 mm wide blank that is roughly 45 mm gone before you have made a single cut. Shops that nest blanks tightly and forget this end up with a duct that is 40 mm narrower than the drawing.
The third is the roll gap, and it is the one that gets ignored.
Every forming station on a lock former is a pair of rolls. The gap between them determines how much the metal is squeezed as it passes. The correct gap is not the material thickness. It is the material thickness plus a clearance of 0.03 to 0.08 mm, and that clearance exists because galvanized strip is not a theoretical plane. It has crown, it has thickness tolerance, and it has a zinc layer that behaves like a soft, slightly sticky lubricant under pressure.
Set the gap to exactly the nominal thickness and you will get three things in sequence. First, zinc pickup on the roll faces, which looks like a grey smear and eventually becomes a build-up that marks every panel. Second, galling on the pocket roll, because the steel is being burnished rather than formed. Third, and worst, a seam that is dimensionally correct on the outside and work-hardened on the inside, so it cracks when the duct is lifted.
Set the gap too wide and nothing dramatic happens. The pocket comes out shallow and rounded instead of square-cornered. The hook does not close fully. The seam holds when you test it by hand on the bench, then opens on the truck. This is the failure mode I see most often, because it is silent.
Take a strip of the actual production material, not a test piece of unknown origin, and run it through one station at a time with the drive in jog mode. Use a feeler gauge between the roll face and the formed metal at the point of maximum contact. You are looking for the gauge leaf that slides with light drag, not the one that will not enter and not the one that falls through.
Do this at both ends of the roll. A roll that is 0.1 mm tighter on the operator side than the drive side will produce a seam that is tight at one end of the panel and loose at the other, and no amount of adjusting the main screw will fix it. That is a shimming job, and it is worth the hour.
Galvanized steel at 0.6 mm springs back about 1.5 to 2 degrees after leaving the last forming pass. Stainless at the same thickness springs back closer to 4 degrees and work-hardens as it goes, which is why 304 on a machine set up for galvanized will produce a pocket that looks formed but will not accept the flange without being hammered in. If you run both materials, keep a written setup sheet per material and per gauge. Memory is not a setup sheet.
The table below is the starting point I use for a standard 1250 mm Pittsburgh lock former with hardened D2 rolls. Treat it as a starting point, not gospel. Coil tolerances vary by mill and by region, and the first strip of a new coil always tells you something the last one did not.
| Strip gauge | Roll gap (mm) | Pocket depth (mm) | Flange height (mm) | Seam allowance (mm) | Forming speed (m/min) |
|---|---|---|---|---|---|
| 0.5 mm | 0.54 - 0.58 | 9.0 | 8.0 | 30 | 18 - 22 |
| 0.6 mm | 0.64 - 0.68 | 9.5 | 8.5 | 32 | 16 - 20 |
| 0.8 mm | 0.84 - 0.88 | 10.0 | 9.0 | 33 | 14 - 18 |
| 1.0 mm | 1.04 - 1.09 | 10.5 | 9.5 | 34 | 12 - 15 |
| 1.2 mm | 1.25 - 1.30 | 11.0 | 10.0 | 36 | 10 - 12 |
| 1.5 mm (single side) | 1.56 - 1.62 | 12.0 | 11.0 | 38 | 8 - 10 |
Two notes on that table. Above 1.2 mm most standard lock formers need either a heavier frame or a two-pass arrangement, because the pocket roll cannot displace that much metal in one bite without deflating the shaft. And the speed column is a range because the real limit is usually how fast the operator can feed a 3 m blank without letting it skew, not what the motor can do.
Nine times out of ten this is gap, not material. The tenth time it is flange height. Check the seam cross-section by cutting a sample with a hacksaw and looking at it end-on. You should see four layers in close contact, with the hook curled back inside the pocket and no daylight between any of them. If you see a gap between the hook and the pocket wall, the hook roll is not closing far enough and the fix is in the last station, not the first.
A crack at the outside of the pocket bend means the bend radius is too tight for the gauge, or the strip edge has a burr from the slitter that is acting as a stress riser. Run your thumb along the coil edge. If it catches, the slitter needs attention at the coil supplier's end, and no amount of machine adjustment will save you. On 1.2 mm and above, a 1 mm radius on the pocket roll corner makes a measurable difference to cracking.
If the finished panel is a propeller rather than a rectangle, the entry guide is out of square with the roll line. Set a machinist square against the guide face and the first roll, and check both sides. A 0.5 mm skew over 1250 mm of travel is enough to put a 3 m panel 2 mm out of flat, which then shows up as a twisted duct that will not sit on the hanger.
This is the flange going into the pocket at an angle rather than square. Usually caused by the operator feeding the second edge by eye because the machine has no rear support. On blanks longer than 2 m, a simple roller stand at the outfeed, set 1 mm high, removes most of it.
Too much pressure somewhere in the train, usually the first idler rather than a forming roll. Wipe the rolls with a solvent, inspect for a bright polished band, and back that station off by a quarter turn. Flaking zinc on the outside of a duct is a cosmetic complaint. Flaking zinc inside the pocket is a corrosion site that will be the first place the duct fails in a coastal installation.
A lot of buyers assume the lock former also closes the seam. It does not. It forms the two profiles. Closing is a separate operation, done with a Pittsburgh seam locker that runs along the joint and hammers the flange home, or by hand on small work. On a 1.5 m duct the hand method takes about 90 seconds per seam and leaves the operator's shoulder in a bad state by Friday. A pneumatic seam locker does the same joint in 15 to 20 seconds and, more importantly, does it consistently.
This is why the two machines are usually bought as a pair. A shop that buys only the former ends up with a growing pile of formed panels waiting for someone to close them, which is the most expensive kind of work in progress because the labour is invisible in the costing.
Second-hand units are everywhere, and most of them are fine. Three things to check before money changes hands.
Run a strip and measure the pocket depth at five points along the length. If it varies by more than 0.3 mm, the shafts are bent or the bearings are gone, and that is a rebuild, not an adjustment. Look at the roll faces under a light: a mirror finish with visible grinding marks is normal wear, deep concentric grooves mean the machine has been run with no lubrication and the rolls are past saving. And check the adjuster screws for backlash. If you can turn a handwheel a quarter turn before anything moves, every setup you do on that machine will be approximate.
Rolls are consumable. A hardened D2 set running 0.8 mm galvanized in a single-shift shop will typically need regrinding after 400 to 600 tonnes of strip, which for a mid-size duct shop is somewhere between three and five years. Budget for it. A replacement set is a meaningful fraction of what a used machine costs, and a seller who knows that will price accordingly.
If you are comparing machine formats, the Pittsburgh lock former page on our site lists the bench, floor-standing and duplex configurations we build, and the duplex arrangement is worth a look if you also need a beader on the same footprint.
Keep a test strip of every gauge you run, cut to 300 mm, hanging on a nail beside the machine. When a seam starts behaving oddly, run the test strip before touching anything. It separates material problems from machine problems in about forty seconds.
Write the gap setting on the machine in pencil, not on a piece of paper in the office. Operators change shifts. Paper does not travel.
And clean the rolls at the end of every shift, not every week. Zinc build-up is much easier to wipe off while it is a film than after it has been burnished into the roll surface by eight hours of running.
Yes, within limits. The roll train has enough adjustment range on most 1250 mm machines, but the pocket geometry that suits 0.5 mm is not ideal for 1.2 mm. Expect to keep two setup sheets and to spend 20 to 30 minutes changing over, mostly on the pocket and hook stations. If you regularly run both extremes in the same week, a second machine usually pays for itself in changeover time alone.
Because a Pittsburgh lock is a mechanical joint, not a seal. Even a perfectly formed seam has a capillary path along the fold. Ducts on the discharge side of a fan above roughly 500 Pa need sealant in the pocket before closing, or a gasket tape on the mating flange, or both. The lock quality determines whether the seam holds together. It does not determine whether it is airtight.
Base it on tonnage, not calendar time. Typical figures are 400 to 600 tonnes for hardened D2 on galvanized, less for stainless because of work hardening, more for aluminium because it is softer. Practically, watch for the point where you can no longer achieve the specified pocket depth with the gap inside the correct window. That is the signal.
They solve different problems. Pittsburgh is a longitudinal seam for the duct body and needs no separate corner pieces. TDF is a transverse flange system for joining duct sections, and it is faster on site but needs a dedicated flange former and corner inserts. Most shops making rectangular duct above 400 mm use both, on different axes of the same duct. Choosing between them is not an either-or decision.
If you are laying out a shop from scratch, start with the duct machinery range and work backwards from the duct sizes you actually sell, not the ones you would like to sell. The machine list follows from the order book, never the other way round.