Grain Refinement

mpcoppin

Member
After learning that grain refinement and normalizing are not the same thing, I was left with the question of how do I refine grain?

Unlike some of you who know what you are doing, I am confident that I regularly overheat steel when forging. I don't do it on purpose, I'm just not that good at this (I'm still figuring out how to run my forge, etc...). So... If I have abused, burned, and mutilated a piece of steel that I am attempting to turn into a blade of some sort... How do I fix the many problems I have caused inside that mangled piece of steel? Assuming I have an oven and can achieve something similar to good temperature control, what do I do?

As I understand it, heating up to a good normalizing temp (as described in other threads) will sort of homogenize things so that the steel is consistent throughout the piece. Now what?

As I understand it, you don't actually reduce grain size, you just make new grains. And, one of the ways to make new grains form is to heat something up to the point where new austenite is formed. To keep the grain small, you don't heat it more than necessary and you cool it quickly.

The question I have is this... To what point do I heat it? Do I heat to the point where austenite begins to form or the point where the steel is fully austenite (For example... You start getting austenite with 80CrV2 at about 1395 degrees and you should have full austenite at 1415)?

I would assume that I'm shooting for full austenite since I want things to be uniform. But what about the austenite that started forming at 1395? How big are those grains now? Is this why people do multiple heats at decreasing temperatures?
 
I happen to have a side by side of 52100 that was blown up during forging, broken, and then one half of it cycled to dull orange and broken again. Pretty dramatic difference just like Kevin said one refinement cycle can produce. F9F42180-EE0B-42B0-9B04-56791CB73091.jpeg
 
Last edited:
Kevin, awesome information. With grain refinement cycles, so I am clear, once heated to decalecense and past non magnetic you can pull from oven or forge and let cool in still air until it returns to magnetic and that will have cycled grain structure? I was told let it cool to black, which I assume would be magnetic again....then can I go straight back in the forge for one more grain refinement cycle, let cool to black and then righ back into the forge heat to critical temp and quench?
Hope that makes sense.

We need to set up zoom meetings so we can all talk about this stuff...would be alot easier then typing...LOL
 
we call this grain size, grain structure is usually when someone tries to describe patterns in the steel due to carbide banding, which is undesirable.

if you are not sure about temperatures, quenching a few times from dull cherry red after cooling down in still air from bright red, almost orange, is a decent way to make up for lack of temp control. would try it out with a few pieces of steel, to see what gives good results and get your eye in... and snapping, before you temper, after steel is properly cold. i would just nudge it over dull cherry red on the last heat, or not, depending on how well you judge temperatures. there are better ways, but this seems to be easiest and can give good results
 
How does this look? It's 1080 brought to 1485 plus or minus 5 and quenched in Canola. I have some Parks 50 on order, but it's slow coming.

These were done in a small forge. My HT Oven won't be here until October!

20210825_142415.jpg20210825_142420.jpg
 
It can be hard to tell grain size form photos, but that looks darn good to me. I'll bet you break a file or drill bit you'd have a hard time telling the difference in grain. From what I see there is no actual "grain" visible, but just a "greyish" smear which is what is wanted. I'd say write down what you did, and do it again. good job
 
It can be hard to tell grain size form photos, but that looks darn good to me. I'll bet you break a file or drill bit you'd have a hard time telling the difference in grain. From what I see there is no actual "grain" visible, but just a "greyish" smear which is what is wanted. I'd say write down what you did, and do it again. good job

Thanks! Those are Macro pictures so they are pretty zoomed in. It's the best I've ever done. I my past attempts the grain was visible like fine sand.
 
One question that pops into my head is how do you know what temperature the steel was at if you don't have a HT oven? 1080 is a simple steel to heat treat in a forge; it's real forgiving. Just wait for the decalessense to pass and then quench. There's not enough carbon in it or other metals to form a lot of carbides other than cementite which goes into solution easily.

I hope that you get good use out of your HT oven. In the end, there's no such thing as too much temperature control.

Doug
 
One question that pops into my head is how do you know what temperature the steel was at if you don't have a HT oven? 1080 is a simple steel to heat treat in a forge; it's real forgiving. Just wait for the decalessense to pass and then quench. There's not enough carbon in it or other metals to form a lot of carbides other than cementite which goes into solution easily.

I hope that you get good use out of your HT oven. In the end, there's no such thing as too much temperature control.

Doug

I have something similar to this in my forge. I can adjust the gas flow and get the temperature to settle in pretty well. Not perfect, but at least I get some idea of where I'm at. The probe on mine is much better than shown here.

71RwBVOWLVL._SL1500_.jpg
 
Do you have a muffer tube in the forge for HT'ing? That really helps with temp control on the blade. A hole in the muffer tube allows the temperature probe to be inserted close to blade for monitoring temp.
 
Do you have a muffer tube in the forge for HT'ing? That really helps with temp control on the blade. A hole in the muffer tube allows the temperature probe to be inserted close to blade for monitoring temp.

I don't, but it's a really small forge. I'm using this Atlas.

The rear has a small cut out that I slide the entire probe in, similar to this pic of an older model. I have the model showing the front view.

unnamed.jpg

41PsK9qui6L._AC_.jpg
 
Back
Top