Metal Substrate

No Weld Defects on Big Metal Substrates – How We Chase Perfection (Mostly)

Big metal substrates are a whole different headache. Little ones? No problem. Stack 'em, braze 'em, they come out fine.

But when you're making honeycomb two feet wide and three feet long? That's when shit goes sideways. Heat doesn't spread even. Different parts expand at different rates. Gaps open up. Voids happen. One bad spot and the whole thing can fall apart in service.

We make big ones for industrial engines, generators, marine scrubbers. Took us years to get close to zero weld defects. Still not perfect, but we're a lot better than we used to be.

Here's what we figured out – most of it the hard way.

Why Big Parts Suck

Small parts heat up quick. Everything gets to brazing temp at the same time. Filler flows nice and even.

Big part? Edges heat first. Middle lags behind. By the time the center is hot enough, the edges have been cooking for a while. Filler runs out, burns off, or just doesn't go where it's supposed to.

Also, long foil layers want to warp. They don't stay flat. Gaps open up between layers. Those gaps don't braze. Then you got a weak spot that'll crack later.

And the weight. Big substrate is heavy. It can sag in the furnace. Sag means misaligned cells, which means bad flow.

You can't just scale up a small‑part process. You gotta change everything.

Step 1 – Fixturing That Doesn't Suck

First, you need a fixture to hold the stack together under its own weight. We use heavy steel plates top and bottom, with bolts to squish it.

But not too much. Squish too hard, foil buckles. Not enough, gaps. We spent months finding the right torque. Now it's 35 Nm. Not 30, not 40.

Also, the fixture has to let heat circulate. Solid plates? No. We drill holes. Lots of holes. Let the hot gas move through.

We had a batch of big substrates with cold spots right in the middle. Filler never melted. Scrapped the whole batch. Turned out our fixture had solid end plates. Switched to perforated plates. Problem solved.

Step 2 – Furnace Ramp – Slow as Molasses

Small parts, you can ramp fast. 20°C a minute? Fine.

Big parts? Slow. 5°C a minute. Sometimes less. You gotta let the heat soak all the way through before the filler starts to melt.

We put thermocouples inside the stack – front, middle, back. Not just on the outside. If the temperature difference is more than 30°C, we slow the ramp.

One time we had a 10°C difference between center and edge. Thought that was fine. But the filler still didn't flow in the middle. Turns out the furnace was colder in the center. Repositioned the heating elements. Next batch was even.

Step 3 – Filler Placement – Paste, Not Just Sheets

For small parts, you put a thin sheet of brazing filler between layers. Works fine.

Big parts? The sheet can wrinkle or bunch up. Gaps. So we switched to paste. Roll it on thin and even.

But paste can dry out. So we keep the stacks in a humidity‑controlled room before brazing. 50% RH. Too dry, paste cracks. Too wet, filler won't stick.

At the edges, we add extra filler. Edges cool first, so they need more material to flow into the joint.

Step 4 – Heat Soak – Let It Sit

Once the furnace hits temperature, you gotta hold it. Not a few minutes. For big parts, we soak for an hour. Sometimes two.

That gives the filler time to get into every gap. Also lets the whole stack equalize. No hot spots.

We learned this after cutting open a big substrate that looked fine on the outside. Inside, the middle layers had no braze. Filler melted, ran to the edges before the center was hot enough. Extended the soak. Fixed it.

Step 5 – Cooling – Slow or It Cracks

When the braze is done, cool it slow. Too fast, metal shrinks uneven. Warps. Cracks at the joints.

We cool at 3°C a minute down to 500°C. Then faster after that. The first 200°C drop is the most critical. Faster than 5°C a minute? We see micro‑cracks under the microscope.

One batch we tried to speed up cooling to save time. Scrapped half of them. Not worth it.

How We Check for Zero Defects

You can't see inside a brazed joint. So you test.

Peel test. Sacrifice one substrate from every batch. Peel layers apart. No voids? Good. Voids? Whole batch suspect.

Ultrasonic testing. For expensive big ones, we use C‑scan. Sound waves bounce off voids. Shows exactly where bad spots are. Takes time, costs money, cheaper than a field failure.

X‑ray. For critical stuff – military, nuclear, marine. X‑ray sees through metal. You can see filler flow pattern. Uneven? Scrap it.

Visual after cut. Sometimes we just cut a part open and look. Crude, but honest. You see everything.

Real Example – Marine Scrubber

We made a big substrate for a ship exhaust scrubber. 24 inches across, 36 inches long. Stainless, 200 cpsi. Customer wanted zero weld defects. No voids, no cracks.

We did everything slow. Slow ramp, long soak, slow cool. Tested with ultrasonic. Found a small void at one edge. Not big, but it was there.

Scrap it? That's a $10,000 part. We ran a destructive test on another from same batch. Peel showed no other voids. That one edge void was from a wrinkle in the filler sheet. Switched to paste for next batch. No voids. Customer took it.

Real Example – Generator Substrate

A generator maker needed big substrates for landfill gas. Nasty fuel, lots of heat cycles. They had cracking at the edges after a year.

We redesigned the edge joint – more filler, changed foil overlap. Also switched to a slower cooling cycle. New ones been running 3 years. No cracks.

What We Still Screw Up

Sometimes we get a bad batch. Furnace drifts. Bad filler batch. Operator mistake.

When that happens, we scrap the whole batch. No patching. No shipping and hoping.

It hurts. But it's better than a customer calling a year later with a pile of failed parts.

Zero weld defects on big metal substrates ain't easy. But it's possible. Good fixturing. Slow ramp. Even heat. Long soak. Slow cool. Thorough testing. Skip any of that, and you'll have voids, cracks, delamination.