I have walked through a lot of duct shops that buy a perfectly good TDF flange forming machine, run clean flanges all day, and then watch their assemblies fail a leakage test at the corners. The flange is only half of the connection. The corner insert is the other half, and in my experience it causes more field complaints than the flange itself ever will.
A TDF corner insert does three jobs at once. It closes the open end of the flange profile, it squares the joint so the two duct sections sit at 90 degrees, and it carries a big share of the gasket compression load. If the insert is loose, the flange walls spread under pressure and the gasket relaxes. If it is oversized, you crack the flange lip trying to drive it in. Either way, you find out on the test rig, which is the most expensive place to find out.
So let me go through the corner options that actually exist, the numbers that matter when you pick between them, and the mistakes I keep seeing on production floors.
Walk any fitting shop and you will run into four families of TDF and TDC corners: stamped sheet metal corners, corner inserts with a gasket seat, die-cast or formed corner pieces installed by a corner matic, and welded or riveted corners for special jobs. They are not interchangeable, and the wrong choice shows up as either labor cost or leakage.
The classic version is a small triangle or L of 20 to 24 gauge galvanized sheet, folded and slipped over the flange ends, then secured with a screw or a rivet. It is cheap, every shop can make them on a press or buy them by the carton, and they work fine on low-pressure comfort ducting. The weakness is repeatability. A corner installed by hand with a screw gun depends entirely on which operator did it and how they felt about their afternoon. On a 1,000 joint order, you will get 40 different corner installations.
These are formed steel or aluminum pieces that combine a stiff corner with a molded gasket seat. The insert presses into the flange cavity and the gasket rides against the mating flange face. On medium and high pressure work, this is what I specify. The insert does the squaring, the gasket does the sealing, and neither depends on operator skill. The trade-off is price per piece and the need to keep stock of the right size for your flange height. If you fabricate your own inserts, our TDF corner mold and TDC corner mold pages show the profiles each series produces.
A pneumatic corner matic drives the insert into the assembled flange corner in about two seconds, with consistent force every time. If your crew assembles more than roughly 150 corners per day, the machine pays for itself in labor and in avoided rework. I have timed crews doing corners by hand at 20 to 30 seconds each once you include reaching for the piece, positioning, fastening, and checking. A corner matic cuts that to under five seconds including handling, and the joint quality stops depending on the operator.
For kitchen exhaust, fume extraction, or anything above 1,000 Pa that also vibrates, I still see welded corners specified. A tack weld on each flange corner, ground and sealed, beats any mechanical insert for permanence. The cost is a welder, heat distortion on light gauge flanges, and slow cycle time. Riveted corners sit between: stronger than a screw, faster than welding, and they do not need an insert stock at all if you have a rivet gun and a drill.
Before you order a pallet of any corner type, put your own numbers on paper. These are the figures I use when a customer asks me which corner to run.
| Corner type | Install time per corner | Typical pressure class | Sealing method | Relative piece cost | Stock burden |
|---|---|---|---|---|---|
| Stamped sheet corner + screw | 20-30 s (manual) | Up to 500 Pa | Gasket tape on flange face | Lowest | Screws and corners, bulky cartons |
| Gasketed steel insert | 5-10 s | Up to 1,500 Pa | Built-in gasket seat | Medium-high | One SKU per flange height |
| Insert driven by corner matic | 2-5 s | Up to 1,500 Pa | Built-in gasket seat | Medium-high | Same inserts, plus machine |
| Tack-welded corner | 40-60 s | Any, incl. vibrating | Weld + sealant | Labor, not material | None, uses welder |
| Riveted corner | 15-20 s | Up to 1,000 Pa | Gasket tape + rivet clamp | Low | Rivets only |
Read that table against your own order book. If most of your work is 250 Pa commercial comfort duct, the stamped corner is honest economics. If you chase SMACNA Class B or C leakage ratings regularly, the gasketed insert is not optional, and a corner matic stops being a luxury past the volume I mentioned.
The single most common failure I get photos of is an insert that will not seat. Nine times out of ten, the problem is not the insert. It is the flange cavity.
TDF flanges come in 20 mm, 25 mm, 30 mm and 35 mm heights, and the insert for a 25 mm flange will never seat on a 30 mm profile no matter how hard the pneumatic driver hits it. Check the cavity width too. A flange formed from 0.8 mm coil has a slightly different cavity than one formed from 1.2 mm, because the material thickness fills part of the cavity. On my bench, a 25 mm flange formed from 0.8 mm sheet takes an insert nose of about 20.5 mm comfortably. The same flange from 1.2 mm sheet wants a nose closer to 20 mm or the insert fights you all day.
Worn flange rolls open up the cavity. A profile that measured 20.8 mm when the tooling was new can drift to 21.5 mm after a couple of years, and then every insert goes in loose. The joint still looks assembled, but the gasket seat floats and leakage creeps up. I recommend measuring the cavity on a formed scrap piece once a month with a caliper, and re-cutting or replacing rolls when the drift passes roughly half a millimeter. This costs nothing compared to a site callback.
On an auto duct line, the corners are usually sheared by the same machine that forms the flange, and they come out square. On manual lines, corner cuts made with a nibbler or a rough shear leave burrs. The burr holds the insert out of full seat by a fraction of a millimeter, the driver ram marks the flange face, and the gasket never compresses evenly. Deburr the corner or punch a clean notch, and half of the insert complaints disappear.
If you buy a pneumatic corner matic, three setup details decide whether operators love it or quietly go back to screw guns.
First, air pressure. Most corner matics want 0.5 to 0.6 MPa at the inlet with a real flow, not just a gauge reading on a dead-end line. A 6 mm air hose run 15 meters from the compressor will sag to 0.4 MPa at the tool and every stroke feels weak. Put a small receiver tank near the machine if your shop air is marginal.
Second, insert orientation. The gasket seat must face the mating flange. It sounds obvious, but on the driver table, an insert dropped in upside down goes in backwards, and the operator may not notice until a whole pallet of duct is cornered wrong. I have seen shops add a simple painted dot or a spring-loaded gate in the feed tray so an upside-down insert physically cannot sit in the pocket. It costs almost nothing and it removes the only real defect mode of the machine.
Third, ram height per flange series. If you switch between 25 mm and 35 mm flange work in the same shift, set and lock a physical stop for each series, do not rely on the operator remembering to adjust the ram. A full-depth stroke on a 25 mm flange bottoms out, marks the flange, and occasionally bends the corner of the duct itself.
When a duct section fails a leakage test, crews instinctively re-tape the seams. In my experience, on TDF systems, more than half of the leakage is at corners and at the flange joint itself, not the seams. The numbers from a test I ran on a 600 x 300 mm section at 750 Pa make the point: with stamped corners and tape, the section leaked around 0.9 liters per second per square meter of duct surface; swapping only the corners for gasketed inserts, same tape, same crew, dropped it to around 0.35. That is the difference between scraping through and passing with margin.
The mechanism is simple. A loose corner lets the flange walls breathe under pressure pulsation, and every breathing cycle works the gasket tape loose at the flange. Fix the corner and the flange face stays clamped, so the tape does its job for years instead of months.
Let me put rough numbers on a mid-size shop doing 200 TDF corners per day, five days a week.
Manual stamped corners with screws: about 25 seconds each means roughly 1.4 labor hours per day just for corners, plus fastener cost. Over a year, that is over 350 labor hours. A basic corner matic costs less than most shops think and drives that down to maybe 0.3 labor hours, while the inserts themselves cost more per piece than a bent sheet corner. Depending on your local labor rate, the machine typically returns its cost in 6 to 14 months, and the consistency bonus is free. Shops that want the inserts produced in-house rather than bought add a dedicated TDF corner production line, which makes sense only at volumes most shops never reach, so treat it as a separate project, not a line accessory.
Against that, if your work is mostly low-pressure and your crews are already fast, do not buy the machine just because it is nice. Spend the money on a better flange forming machine instead, because a square, dimensionally stable flange forgives more corner sins than anything else.
Before you commit to an insert series or a corner machine, run through this list. It is short on purpose.
One last note from experience: whichever corner system you choose, train two people, not one, on the correct installation. The week your only trained operator is on leave is the week quality quietly drops, and you will find out from a customer instead of from your own floor.
Steel inserts survive one or two careful removals, but the gasket seat deforms on the first drive-in and rarely seals a second time. For permanent installations reuse does not matter. For mock-ups and temporary test sections, plan on new inserts for each final assembly.
Most models run at 0.5 to 0.6 MPa measured at the machine inlet with the ram moving, and they consume roughly 0.1 to 0.15 cubic meters of free air per stroke. If your line pressure is stable at 0.7 MPa but the machine feels weak, look at hose diameter and length before blaming the machine.
No. TDC flange profiles and corner geometry differ from TDF, and the inserts are not interchangeable. Buying a combined TDF/TDC corner machine or matched mold sets is the clean solution if you produce both systems, which many export-oriented shops do.
Loose fit almost always traces back to flange cavity drift from worn forming rolls or to a cavity formed from thicker sheet than the insert was designed for. Measure the cavity, compare with the insert specification, and if the fit is borderline, a light crimp with pliers on the flange lip at the corner restores interference without damaging the profile.