Nobody wins a duct project because of corner pieces, and plenty of crews lose sleep because of them. A TDF or TDC flange system is only as airtight as the fit between the corner and the two flange rails it bridges, so the corner is the component where the tolerance of two different factories meets in one joint. When the corner is right, installation crews click pieces together and move on. When it is wrong, you get the calls: corners that fall out in transit, corners that need a hammer to seat, gaskets that never seal no matter how much mastic goes in.
Most contractors buy corners from the same supplier as the flange former, or from a local fabricator with a corner press. Either way, before you sign a supply agreement worth thousands of units, it is worth an hour of structured auditing. I have sat through both outcomes: supplier audits that caught a bad mold before the first shipment, and warranty fights over a container of corners that fit nobody's flange. The second outcome costs about fifty times the price of the first.
After enough field complaints, the failure list is short and repetitive. Understanding it tells you what to audit.
Flange depth out of tolerance. TDF corners are nominally 35 mm deep, TDC typically 30 or 35 mm depending on the class, and the tolerance that matters is plus or minus half a millimeter on the overall depth. A mold worn by 0.8 mm produces corners that sit proud of the rail face, and every gasket under them compresses unevenly. This is the classic cause of "these corners leak and the old ones do not."
Burr and flash at the bend line. Corner presses with worn die edges leave a burr along the outer face of the bend. The burr cuts the gasket surface when the corner is driven in, and it also scratches installers, which turns into a workplace complaint long before it turns into a leak report. Run a finger along the sample; you will know in two seconds.
Corner angle drift. A 90-degree corner that is actually 92 degrees still clicks in, but it spreads the flange rails apart at the joint, and the gasket gap shows up as a slow leak that only fails the leakage test. The angle comes from the corner mold, so it drifts as the die wears, and suppliers who never check angle are suppliers whose die maintenance is cosmetic.
Material substitution. Corners are usually galvanized steel around 1.2 to 1.5 mm, and the price of the part is small enough that nobody weighs them on delivery. Which is exactly how a shipment of 0.9 mm corners slips in: they bend during seating and the clip legs lose their spring. Weighing a sample of ten on a kitchen scale catches this in one minute.
Inconsistent clip geometry. The spring legs that grip the flange rails have a tolerance window narrower than most buyers assume. Too loose and the corner rattles out under vibration; too tight and installation crews mangle the flange edge forcing them in. The geometry is set by the mold and the press setup, and it can drift between batches if the supplier does not check.
You do not need ISO theatre for a part this simple. What follows is the audit checklist I use, and it fits in a morning visit or, for overseas suppliers, a video call plus a sample shipment.
First, ask which machine makes the corners: a dedicated TDF corner production line with automatic feed, or a generic press brake with an operator setting stops by hand. The answer predicts batch consistency better than any certificate. Second, ask how often the corner mold gets measured for wear, and ask to see the last measurement record, not a template of what the record should look like. Third, ask what happens when a batch fails their own check: do they sort and ship anyway because the tolerance is "close enough," or does the batch get reworked? The honest answer to the third question tells you more than the first two.
Take ten corners from production, not from a polished sample bag. Then measure the things that matter, and none of this needs a metrology lab:
| Check | Method | Pass criterion (TDF example) | What failure predicts |
|---|---|---|---|
| Flange depth | Calipers, both legs, 3 of 10 pieces | 35 mm ±0.5 mm | Worn mold, uneven gasket compression |
| Corner angle | Try against a machinist square, feeler for gap | Gap under 0.3 mm at the tip | Die wear, spreading rails, slow leaks |
| Clip grip force | Seat on a known-good flange rail, pull by hand | Holds firm, removable without deforming the flange | Batch-to-batch inconsistency |
| Material thickness | Scale: weigh 10 corners, divide; plus caliper on cut edge if available | 1.2–1.5 mm galvanized, weight within ±8% of nominal | Material substitution under price pressure |
| Surface and burrs | Finger along bend line, visual for zinc damage | No perceptible burr, zinc intact at bends | Worn die edges, scratched gaskets, installer complaints |
| Interchangeability | Seat the same corner type from two different batches on one rail | Same seating force both times | Process drift between production runs |
That last check, pulling corners from two separate batches, is the one suppliers fail most often, and it is invisible if you only audit one sample bag. A supplier with real process control can pass it; a supplier pressing corners between other jobs on a shared press brake usually cannot.
Take one corner and one flange rail from your own stock, seat them, and run a gasket over the joint the way your crews do it on site. The lab fit and the field fit are not the same thing: site crews drive corners with a few taps of a mallet, and a corner that is technically in tolerance but at the far end of it can still seat crooked in tired hands at the end of a shift. Ten minutes of simulation on a workbench has saved me from two supplier choices that looked perfect on paper.
Five more minutes on packaging closes the loop. Corners should ship boxed upright in counts that match your order multiples, not loose in a bulk carton, because a corner that arrives with bent clip legs is a defect nobody recorded. Ask how the cartons are marked: a production date or batch code on every box is what makes a future complaint traceable, and a supplier who cannot tell you which press run a carton came from is a supplier you will be arguing with blind. While you are at it, ask for one photo of the batch being produced on their corner making machine for your file. That photo, plus the sealed reference samples, is usually enough evidence to resolve any later dispute in an afternoon instead of a quarter.
An audit only protects you for one day. The supply agreement carries the protection forward, and three clauses do most of the work.
Specify the acceptance criteria as measurable numbers, not brand names: flange depth tolerance, angle tolerance, minimum material thickness, and a reference weight per ten pieces for the size mix you order. Put the reference samples in a sealed bag, both parties signed across the seal, so any future dispute has a physical arbiter. And agree on batch traceability: every carton marked with a production date or batch code, so when a complaint arrives eighteen months later you can tell one worn mold month from a generally good supply.
Price matters here too, and I will say the uncomfortable part: corners are a commodity where a 15% price drop almost always comes out of material thickness or mold maintenance. If a quote lands far below the others, audit that supplier twice as hard rather than celebrating. The cost difference across a year is trivial compared to a single failed leakage test on a hospital job with the consultant watching.
There is a reason the best corner suppliers are fabricators who also run duct lines or flange machinery. They use the corners they sell, they hear the installer complaints directly, and their mold maintenance is not theoretical. When auditing, ask to see the corner machine running if you can, or ask for a short video of a production batch being pressed. Watching the same operator produce fifty corners tells you about consistency in a way no document can, and the good suppliers will send that video without being asked twice.
Should we just make corners in-house on our own corner machine? If you run a duct shop with press capacity and steady volume, in-house production gives you control of material and geometry for a modest capital outlay, and a corner line pays back reasonably above roughly 20,000 corners a year. Many shops make their common sizes in-house and buy the low-volume sizes; that split keeps the audit problem small.
What tolerance should we actually write into a PO? For TDF: depth 35 mm +0.5/-0.5 mm, angle within 0.3 mm gap against a square, material 1.2 mm minimum galvanized. For TDC the depth follows your flange class, but keep the same angle and thickness logic. Numbers you can check with a caliper are worth more than a certificate full of brand claims.
Corners arrived and 10% are slightly bent. Accept and sort, or reject the batch? Reject. Bent corners mean handling or packaging problems that will repeat, and sorted batches always leak a few into installation. Ask for the batch to be replaced and have the packaging conversation in the same email; corners need to be boxed upright, not dumped loose into a carton.
How often should we re-audit a good supplier? A light re-audit every twelve months: ten fresh corners, the same caliper checks, ten minutes. Suppliers change presses, operators and molds, and the batch consistency check across two production dates is the single most valuable ten minutes in the whole routine.