Skip to content 
The order
A purchase order lands in my inbox.
500 weld neck flanges. ASTM A105. Impact tested at minus 46°C. For a customer in the United States.
I scroll down. Halfway down the page, the customer has written their own chemistry.
Manganese: from 0.60 percent.
My eyes pass over it. A buyer this size has engineers. If they wrote 0.60, someone has already done the thinking.
I don’t stop.
I should have stopped.
The certificate
The billets arrive from the foundry with their certificate.
I run a finger down the columns. Carbon, inside the range. Manganese, inside the range. Everything inside the range.
Then I notice where inside.
Almost every element is sitting at the bottom. Manganese is just above the floor the customer set. Not a decimal more than it had to be.
Still within spec. Still exactly what the order says.
I sign.
A105 allows manganese anywhere from 0.60 to 1.05 percent. Both ends are legal. I don’t know it yet, but only one of them is safe at minus 46.
The hammer
The furnace door opens. The billets come out orange.
They go under the six tonne hammer.
It rises. It falls. The floor jumps under your feet. Again. Again.
Scale sprays across the concrete. A block of glowing steel flattens, spreads, and becomes a flange.
One after another, they come out perfect. Right size. Clean shape.
Nothing looks wrong. Nothing sounds wrong.
That’s the thing about weak steel. Under the hammer, it behaves exactly like strong steel. It keeps its secret until someone freezes it and hits it.
That moment is weeks away.
But someone on my floor is about to see it coming.
The warning
The flanges are lined up for heat treatment. My quality control lead is ready to send them in.
The heat treater picks up the certificate.
He reads it once. Then again, slower.
He puts it down and looks at me.
“Sirji, isme manganese bahut kam hai. Aap iske baad impact test karenge to fail ho jayega.”
Sir, the manganese in this is too low. If you impact test it after this, it will fail.
Silence.
He has no lab. No degree. Just years of watching steel go into his furnace and come out tough, or not.
He knows what manganese does. It’s what keeps steel tough in the cold. It strengthens it without making it brittle, and it holds sulphur in check. Near the minimum, the steel has almost no margin.
I don’t want him to be right. The steel is paid for. The flanges are forged. If he’s right, we start over.
So I go looking for a second opinion.
35 percent
My office. The heat treatment simulator open on the screen.
I type in every element from the certificate. One by one. I set the heat treatment cycle. I set the test: Charpy, minus 46°C.
Run.
The screen thinks.
Probability of passing: 35 percent.
A man and a machine. Two warnings. Same answer.
There are two roads in front of me. Stop, and buy new steel. Or trust the furnace.
Thirty five is not zero.
I trust the furnace.
The first verdict
Two tonnes of flanges go in.
Normalize. Austenitize at 870 to 925°C. The doors open, the steel goes into the water, and the quench tank boils around it. Then temper, slowly, at 595 to 665°C.
Done right, this is how A105 earns its toughness.
The test bars go to the lab.
This is where the second hammer enters the story.
A small notched bar of steel sits in a bath at minus 46°C. A technician lifts it out and sets it on the anvil. Above it, a pendulum is locked high.
He releases it.
It swings down, breaks the bar, and the needle settles on a number.
12 August 2026. The report reaches me.
22 joules. 10 joules. 28 joules.
Average 20. Requirement 27.
Ten joules. That bar barely resisted.
You’d think I’d stop here.
The second gamble
I don’t.
I decide the cycle is the problem, not the steel. If the chemistry is lean, maybe the right temperatures can carry it.
So we turn the dials. New austenitizing temperature. New tempering temperature. Every setting moved in the chemistry’s favour.
The second batch goes in. Into the water. Into the temper. Into the cold.
The pendulum falls again.
It fails. Same margin.
I sit with that second report for a long time.
And finally I see what the heat treater saw on day one. A furnace can only bring out what is already inside the steel. It can refine the grain. It can reshape the carbon. It cannot put in manganese that was never there. It cannot pull out sulphur or phosphorus.
Chemistry sets the ceiling. Heat treatment only decides how close you get.
The room
One room. My quality control lead. The heat treater. The directors.
Nobody is looking for someone to blame. The trail leads to my signature, and everyone knows it.
The only question on the table: how do we make sure this never happens again?
We walk out with one principle. Whatever chemistry the customer writes on the order, choosing the actual chemistry is our job. It is the most important decision we make before the first blow of the hammer.
Then we make it a rule. Every heat gets simulated before we buy it.
We only order chemistry with an 80 percent or higher chance of passing the customer’s Charpy test. If 80 percent costs more, we pay it.
Thirty five had cost us two batches.
Eighty became the floor.
Where steel is born?
A rule is useless if you can’t buy the steel. So we go back to where steel is made.
Most steel starts in one of three furnaces. A Basic Oxygen Furnace blows oxygen through liquid iron and refines it hard. An Electric Arc Furnace melts scrap and sponge iron and refines it too, as long as the scrap is clean. An Induction Furnace is fast and cheap, but it mostly just melts. Whatever sulphur and phosphorus go in, come out.
The best mills add two more steps. Ladle refining (LRF) pulls the sulphur out. Vacuum degassing (VD) draws out dissolved gases. That’s where steel earns the right to go into the cold.
Three melt routes to a forging bar. Only ladle refining removes sulphur.
We map which suppliers can truly remove sulphur and phosphorus, and which can only melt.
We change suppliers overnight.
The third verdict
The new steel arrives from RINL, the Visakhapatnam Steel Plant in Vizag. Ladle refined. Sulphur at 0.007 percent. It cleared our 80 percent rule before we ordered a single tonne.
On paper, it’s right.
But the first steel looked right on paper too.
The furnace door opens. The billets come out orange. The six tonne hammer rises and falls. Same floor. Same shake. Same heat treater. The same kind of cycle that failed twice.
The test bars go to the lab.
A bar comes out of the bath at minus 46°C. Onto the anvil. The pendulum is locked high.
Released.
The report reaches me. I read the number twice, the same way I read ten.
143 joules.
More than five times the requirement. Fourteen times the bar that broke at ten.
Same flange. Same hammer. Same furnace. Same heat treater.
Only the steel had changed.
Two hammers
I used to think the six-tonne hammer was where the real work happened. It’s loud. It’s heavy. It shapes the flange.
But the small hammer in the lab is the one that tells the truth. And its verdict is written long before the first blow, on the day you sign for the steel.
The customer’s range told me what was allowed. It never told me what would pass.
That was always my job.
From 10 joules to 143. The flange never changed. The way we buy steel did.
If you’re forging for cold service
ASTM A105 allows manganese from 0.60 to 1.05%. Meeting that range means the steel complies. It doesn’t mean it will be tough at -46°C. A heat at the low end of manganese, with sulphur and phosphorus left in by a melt route that can’t refine them, can fail a low-temperature Charpy test however well you heat treat it. Simulate the chemistry before buying, accept only heats with a high predicted pass probability, and buy from mills with ladle refining and vacuum degassing.
Frequently asked questions
Why do A105 flanges fail Charpy impact tests at -46°C?
Most often because the steel’s chemistry leaves too little margin: manganese at the low end of the 0.60 to 1.05% range, and sulphur or phosphorus left in by a melt route that cannot refine them. Heat treatment cannot make up for a lean or dirty heat.
Does meeting the A105 chemistry range guarantee a pass at -46°C?
No. The range defines what is permitted, not what is tough in the cold. Two heats can both comply and give very different Charpy results.
Can heat treatment fix low manganese or high sulphur?
No. Normalizing, quenching and tempering refine grain and microstructure, but they cannot add manganese or remove sulphur and phosphorus. Those are set at the steel mill.
How do you check a chemistry before buying it?
Run the heat’s full chemistry and planned heat treatment through a simulator that predicts Charpy performance. I only order steel that shows an 80% or higher probability of passing.
Which steelmaking route is best for impact tested forgings?
BOF or EAF steel, followed by ladle refining (LRF) and vacuum degassing (VD). Induction furnace steel melts but does not refine, so its sulphur and phosphorus depend entirely on the raw material.
About the author: Siddh Mashruwala is CEO of Kayon Forging Corporation LLP (KAFCO), a Mumbai based manufacturer and exporter of forged flanges, buttweld fittings and specialty alloy products for the Gulf, European and US markets.
