Catalytic Converter

How We Design Honeycomb Structures for Catalytic Converters – What Works and What Doesn't

If you've ever held a catalytic converter substrate in your hand, you know it looks simple. Just a metal cylinder with a bunch of little holes. What's so hard about that?

But spend a few years making these things, and you learn that every little detail matters. How many cells per square inch. How thick the walls are. Whether the cells line up straight. Even the shape of the inlet face. All of it changes how the converter works.

Here's what goes into designing a honeycomb that actually does its job.

Why a Honeycomb in the First Place

You could use other shapes. Some early converters did – pellets, beads, random packed beds. Those worked okay but had problems. They rattled. They settled. The exhaust found paths around the edges.

A honeycomb is simple. Straight channels. The exhaust goes in one end, travels down a tube, comes out the other. Every cell gets about the same flow. No shortcuts.

The honeycomb also gives you surface area. That's the whole point. The chemical reactions happen on the walls. More walls means more places for the bad stuff to get converted. But you can't just add walls forever – they block airflow. So you have to balance.

That balance is what substrate design is all about.

Cell Density – The 400 cpsi Question

Most cars today use 400 cells per square inch. That's been the standard for decades.

Why 400? Because it works. It gives you enough wall surface to get the job done without choking the engine. At 400 cpsi, the open area is still high – around 80 or 85 percent. The exhaust flows through easily.

Some applications go lower. 300 cpsi flows even better, but you lose surface area. That might be okay for a big diesel where flow is more important than light‑off. Or for an off‑road machine that doesn't have to meet tight emissions.

Some go higher. 600 cpsi gives you more surface area, which helps the converter light off faster and clean better. But the walls get thinner. The whole thing is more fragile. And backpressure creeps up.

I've seen 900 cpsi substrates in some high‑performance applications. They work great when they're new. But they're delicate. One backfire and the whole thing cracks.

Wall Thickness – Thin Is Good Until It's Not

The walls between the cells have to be thick enough to hold the structure together but thin enough to not block flow.

Standard automotive foil is about 0.05 mm thick. That's thinner than a sheet of paper. You can see light through it if you hold it up.

Thinner walls have less thermal mass. That means they heat up faster. Faster light‑off means less time running rich on cold start. That's good for emissions and fuel economy.

But thinner walls are weaker. They can dent during handling. They can melt if the converter overheats. They can vibrate and crack over time.

We've made substrates with 0.03 mm foil for some racing applications. They light off instantly. But you have to be careful installing them. Drop one and it's junk.

For heavy‑duty trucks, we go thicker. 0.08 or 0.1 mm. The converter is heavier. It takes longer to warm up. But it survives a million miles of vibration and regen cycles.

Cell Shape – Round, Square, or Something Else

Most people assume honeycomb cells are hexagonal. That's where the name comes from.

But a lot of catalytic converters don't use hexagons. They use squares. Or rectangles. Or sinusoids – wavy channels that nest together.

Why? Because it's easier to manufacture. You can take a flat sheet, corrugate it into waves, and stack it with a flat sheet. That makes square or rectangular cells. No need to cut hexagons.

The shape of the cell affects flow and surface area. Square cells have slightly less surface area per open area than hexagons. But the difference is small. For most applications, it doesn't matter.

What matters more is that the cells are straight. If the channels wander, the exhaust doesn't flow evenly. Some cells get all the flow. Some get none. That kills efficiency.

We check cell straightness on every batch. A quick visual – shine a light through the substrate. If you see shadows or dark spots, the cells are misaligned. That part doesn't ship.

Inlet and Outlet Face Design

The ends of the substrate aren't just flat. Some have chamfers. Some have rounded edges.

Why? Because the exhaust doesn't hit the converter evenly. The pipe enters from one side. The gas has to spread out across the whole face. A sharp edge can cause turbulence or uneven flow.

We've worked with customers who needed specific inlet angles to match their exhaust manifold. The substrate face had to be cut at 15 degrees. That's not standard. But that's what the CFD model showed would work best.

Cutting a substrate at an angle is tricky. The foil wants to tear. The cells can collapse. We had to build special fixturing to hold everything in place during the cut.

But when it works, the flow distribution is perfect. No hot spots. No dead zones.

Thermal Mass and Light‑Off Time

This is the trade‑off that never goes away.

Thin walls and low cell density mean low thermal mass. The substrate heats up fast. Light‑off happens sooner. That's good.

But low thermal mass also means the substrate cools down fast. If the engine idles for a while, the converter temperature drops. Then when you step on the gas, it has to heat up again.

Thicker walls and higher cell density hold heat longer. The converter stays hot during idle. But it takes longer to get hot in the first place.

There's no perfect answer. It depends on the application. A city delivery truck that stops and starts all day needs fast light‑off. A long‑haul truck that runs steady needs heat retention.

We let the customer's data drive the design. Give us your duty cycle. Give us your target light‑off time. We'll pick the cell density and wall thickness that fits.

How We Test Designs

You can model all you want. But eventually you have to test.

We cut substrates in half and look at the flow distribution. We run them on an engine dyno and measure backpressure. We put them in a lab oven and cycle them from cold to 800 degrees a hundred times.

Sometimes the test says the design works. Sometimes it doesn't.

I remember a substrate we designed with an oval shape and a weird cell orientation. The CFD looked great. But on the dyno, the backpressure was 20% higher than predicted. Turns out the cells were slightly misaligned because of the oval shape. We had to redesign the stacking fixture.

That's how it goes. You try something. It doesn't work. You fix it. Eventually you get it right.

What's Coming Next

Higher cell density is the trend. 600 cpsi is becoming common. 800 and 900 are out there.

But thinner walls mean you need better foil. Better brazing. Better handling. Not every manufacturer can do it.

We're also seeing more complex shapes. D‑shaped substrates to fit around steering racks. Oval substrates for low‑profile exhaust systems. Substrates with variable cell density – more cells in the center where the flow is highest, fewer on the edges.

The basic honeycomb isn't going away. It's too good at what it does. But the details keep getting finer.

Bottom Line

Designing a honeycomb substrate isn't rocket science. But it's not trivial either.

You have to pick the right cell density and wall thickness for the application. You have to make sure the cells are straight and the flow is even. You have to test the design before you ship it.

Get it right, and the converter lasts for years. Get it wrong, and you get high backpressure, slow light‑off, or a cracked substrate.

Most of the time, 400 cpsi with 0.05 mm walls works fine. But when it doesn't, you have to know what to change. That's what experience teaches you.