[Closed] Any metallurgists in this evening? Liquid nitrogen, shrink fitting...

https://www.engineeringtoolbox.com/temperature-allowable-stresses-pipes-d_1338.html

That tells me that as you suspect the material will behave plastically at 1050°C and the parts will deform relieving the stress that keeps the parts together.

Not a metallurgist as my now edited post shows...

Does it need to go to 1050°C? If it has been supplied in the 'Solution Treated' state (seems reasonable for easy machining), won't it just require ageing at a lower temperature?

require heat treatment to achieve this condition

Depending on the heat treatment they might need some machining to get to correct size. Probably best to wait and speak to whoever is doing the heat treatment.

As keen as you are to assemble I would wait. Getting those bits apart will require considerable force. If they have not got design features to allow disassembly you might damage them trying.

Why not heat treat before assembly?

Brainflex - just a question of urgency, and wasting the N2.

I suspect as the chaps above suggest, worth doing properly and being late rather than risking ruining the lot. The material has been supplied in H900, so too brittle and needs to be annealed before retreating, hence the high temp.

Cheers,

1050°C is too low for annealing:

https://precisiongrinding.com/steel-plate-services/stress-relieving-and-annealing-steel/

Even at 1050°C you are into stress relieving and it's only the stress that keeps a cryofit together. The whole point of a cryofit is that the two parts are compressed and stretched within their elastic limits and the resistance of the joint depênds on the level of stress achieved.

 annealed

This would definitely relieve some of the stress that goes into creating the interference fit. In fact if you ever wanted to separate the parts again this is how you would probably do it.

I doubt they would fall apart after annealing but the fit will not be as strong as designed.

Its not steel though its stainless steel.
I found this data sheet for 17-4 PH I think your 1050 C is the solution treatment and then it will be aged at some temp lower than that depending on grade required.

You're absolutely building a nuclear weapon aren't you op?

What condition has it been supplied in, and why do you need it in the solution annealed condition?

Because: it's martensitic when treated correctly, so there will be a degree of dilation/distortion associated with going up to austenite and back again (I've been out of stainless for a few years, can't remember if 1050 will take you through that off the top of my head). If it's just taking the precipitates into solution the shape change will be minimal. A spec sheet will give dilation per unit time at temperature which could be used to fag packet a bit of fettling to maintain interference.

If you want condition A for corrosion resistance, especially if stress corrosion is a risk, then any differential cooling between the components would result in a residual stress and increase the risk of stress corrosion cracking.

Edit: just spotted that it's H900 and the issue is toughness. If anything, H900 has slightly higher elongation to failure than Condition A, I'll see if I can find factor toughness data to confirm, but I wouldn't be too worried about H900 being brittle.

Lots of useful information in that doc 5plusn8 but nothing on the stress resistance of the material at high temperatures. They give the properties at room temperature after exposure to high temperatures but they don't give the resistance at high tempertures which is what is of interest if you want to work out if the stress that keeps a cryo joint together will be lost.

I maintain that if you reach a temperature that achieves "stress relief" or "annealing" you have by definition relieved the stress that keeps the joint together.

https://www.engineeringtoolbox.com/metal-temperature-strength-d_1353.html

Brainflex – just a question of urgency, and wasting the N2.

See if they'll deliver more N2 - its the delivery thats the cost and the hassle really rather than the material - if they can leave you with a bigger flask it'll keep for longer.

I needed some for a filming project a few years back - just needed a few litres but the company only delivered to my area every two weeks - so he gave me a huge vat of the stuff it and it sat quite happily in its flask for over a week

I presume I don't need to warn you about the explosion risks and all that 🙂

educator
p8 shows dimensional changes due to hardening which may be useful
p14 and 15 give change in yield and tensile strength from 500c down, bummer its not higher. Surely 1050c will reduce modulus, yield and tensile by large amount?

I have nothing useful to contribute to this thread, but once again I'm massively impressed by the fact that STW does indeed seem to know everything!

Surely 1050c will reduce modulus, yield and tensile by large amount?

Yes, even stainless. That's what the graphs in my last link say, to next to nothing compared with strength at room temperature in fact. The Twin Towers didn't collapse because the steel structure melted, it simply reached a temperature of around 1 000°C at which point it lost it's strength and the buliding went down.

Not an expert on this alloy, but I would have expected you would be well into creep territory by 1050°C, so (I would think) your interference fit stresses would relax fairly quickly.

What on earth are you making, btw?

Edit:

1050°C is too low for annealing:

For clarity, I believe that the treatment in question is to bring precipitates back into solid solution prior to ageing.

Yes, even stainless. That’s what the graphs in my last link say, to next to nothing compared with strength at room temperature in fact. The Twin Towers didn’t collapse because the steel structure melted, it simply reached a temperature of around 1 000°C at which point it lost it’s strength and the buliding went down.

Ballpark is 1/2 strength lost at 500°C for structures (ASFP yellow book and the Eurocodes). By 1000c, yes, it's virtually nothing.

Although I'm pretty sure the twin towers were bought down by a covert government operation. Probably the same people that created the covid virus.

You’re absolutely building a nuclear weapon aren’t you op?

OP is le Chiffre and I claim my £5.

I knew I'd get some answers on here! Or at least some pointers towards more questions. What a widely knowledgeable bunch of people you are. Thanks 🙂

Good point on the N2 delivery - I'll get a bigger tub delivered to keep options open. Just been doing the RAMS for that.

We'd asked for H1025 as being a reasonable compromise between strength and toughness. H900 isn't vastly different and we do have some margin on paper, but I'd much prefer a ductile failure than fracture.

What is it? It's a tapered spigot joint which uses the collar with flats to hold the two halves together. The ribbed bits get bonded into tubes so we have a demountable, stiff, strong, modular tube thing for a secret purpose. The shrink fit is to capture the collar between two parts so it can rotate freely, without using circlips or fasteners to make it captive.

I have absolutely no idea what you guys are talking about?!?!?.............
Couple of questions:
I assume you're not making a nuclear weapon, so what are you trying to make in layman's terms?
Could you not just super glue the two pieces together?

LBS should be happy to chase and face it for a packet of biscuits.

Good point on the N2 delivery – I’ll get a bigger tub delivered to keep options open. Just been doing the RAMS for that.

Once you have a vat of Liquid Nitrogen in the building its mandatory to dress like C.A. Rotwang

What is it? It’s a tapered spigot joint which uses the collar with flats to hold the two halves together. The ribbed bits get bonded into tubes so we have a demountable, stiff, strong, modular tube thing for a secret purpose.

If I buy one will STW get a kickback from Lovehoney again?

Fancy looking stunt pegs, your bmx will look sweet.

Mowgli,

Ooh, ooh, ooh, I think I know a bit about this.

No, I wouldn't advise shrink fitting tomorrow. You are right: any temperature high enough to affect the structure will relieve the stress, so undo all the hard work of the precision machining and shrink fit.

You shouldn't have to heat it up to 1050 C to make the material tougher. Doing that would be unnecessary and might increase the risk of distorting it (I can't remember if that is hot enough for transformation to austenite, with the associated volume change). H900 is precipitation hardened for 4 hours at 900 F (482 C), H1025 is precipitation hardened for 4 hours at 1025 F (551 C), so to change the material condition from H900 to H1025, it should just be a matter of heating to 551 C for a while (might have to reduce the time a little bit to allow for the time already at 482 C, the heat treater may already know if and how much).

I think you are right that H1025 is a better balance between strength and toughness than H900: 0.2% yield strength only drops from 198 to 168 ksi, but impact strength rockets from 16 to 40 ft.-lb.

Heat treatment of 17-4 PH isn't quite like the classic plain carbon steel harden and temper that a lot of people are used to:
Solution annealing: heated to 1900 F (1038 C), held for 1/2 hour after the core reaches temperature (This dissolves the alloying elements in the matrix) then air cooled or oil quenched to 90 F (32 C) or less (this locks everything in place before it has time to move).
Precipitation hardening (hence the PH in the name): heat to 900, 1025, 1075 or 1150 F and hold for 4 hours (this lets mostly the copper, diffuse into clumps that pin grain boundaries and I think dislocations in place, making the material harder. A higher precipitation temperature lets the clumps grow bigger, but there are less of them, so there is less pinning: the material is lower strength, but it can distort around the tip of cracks, spreading the load, giving greater toughness. I think there is also some tempering of the martensite, which increases toughness). Air cool gives a gives a repeatable thermal profile so everything is repeatable.

The Wiki on 17-4 PH links to a wayback machine cached copy of an outokumpu data sheet, and a North American Stainless data sheet: they give some explanation. I think the alloy was invented by an American company, which is why the names for conditions reflect the precipitation temperature in F: I have to make a point of not getting confused between F and C when thinking of the temperatures!

he Twin Towers didn’t collapse because the steel structure melted, it simply reached a temperature of around 1 000°C at which point it lost it’s strength and the buliding went down.

Alot lower temp than that. Jet fuel burns at about 850 degs C in open air and steel looses half its strength at temps as low as 650 legs C. That's why the 9/11 conspiracy theorists say how could jet fuel melt steel, but as you say the steel didn't need to melt or get anywhere near its melting temperature before it lost its structural integrity.

So probably not a good idea to do the interference fit before heat treatment.

Jingle, that's great, thanks for putting in a way I can understand!

Can't help much other than to remind you to pour some N2 on a carpeted floor because it makes a cool noise.

  H900 is precipitation hardened for 4 hours at 900 F (482 C), H1025 is precipitation hardened for 4 hours at 1025 F (551 C), so to change the material condition from H900 to H1025, it should just be a matter of heating to 551 C for a while

Unfortunately, I don't think it works like that because nucleation and growth are separate things (usually in precipitation hardening).

A higher precipitation temperature lets the clumps grow bigger, but there are less of them....

iIf you've already aged at 482°C, the precipitates will have nucleated in a way that reflects the lower temperature (small and lots of them). Again, not an expert in this alloy, but going by general precipitation hardening behaviour, further ageing will only grow the existing precipitates - it won't give the same distribution you would get if they were first nucleated at 551°C (which would be, as you say, fewer and larger precipitates).  I strongly suspect that it would need heating up to bring the precipitates back into solution followed by ageing at the appropriate temperature.

And this, folks, if why welding aluminium bike frames is such a thorny topic - 6000 series aluminium achieves its strength by this same mechanism.

Yes, having spoken to the heat treatment people this morning, they do need to solution anneal at 1038 C and then do the fresh treatment from scratch at 1025F for 4 hours. We'll have to wait and see how much they change shape by.

Thanks again all for sharing your knowledge!

Mowgli,

Looks tillydog understands behaviour of the precipitates much better than me (thanks tillydog).

Just holding at 551 C should increase the toughness significantly without decreasing the strength too much, but unless you / the heat treatment people can find results from someone who has done it before as a rework procedure, the best time and the effect are a bit of a guess. (Unless there was anything heat treated with your parts that you could use as a test piece, but tensile and impact testing would cost extra, and generally only happens in hindsight - bitter experience going into the heat treatment department when the furnace is up to temperature to see how the run is going, and finding the test pieces left on the desk).

I think your decision between 'just precipitation' and 're-solution then precipitate' could depend on the risk of distortion by re-heating to solution temperature (plus your attitude to that risk versus the risk of the parts being too brittle), and I'm afraid I have forgotten that part and don't have access to the data.

Please let us know how it goes (recently made redundant, and this reminds me of meaningful work).

best question in a while. 🙂

Great question and answers! It strange that having never heard of 17-4ph before everything seems to be made of the stuff..

We recently made a wind tunnel model out of this as it needed to be compliant with pressure regulations. (Fracture toughness and elongation requirements) We ended up with h1150 condition although some good billets of h900 achieved sufficient fracture toughness when cold.

I am not sure how one decides how much fracture toughness is eneough though. One thought can you get hold of the billet certificates the components were machined from? The properties vary billet to billet they just must reach certain levels to be sold to certain spec. You may be lucky and your h900 billet meets your spec?

Good luck!

bitter experience going into the heat treatment department when the furnace is up to temperature to see how the run is going, and finding the test pieces left on the desk

I used to work in a place that made *really* thick (circa 300mm per pane) glass windows for nuclear processing cells.

These needed to be annealed incredibly slowly to avoid them cracking - something like 4-6 weeks of gradual cooling. There was an array of 14 top-hat kilns that the blocks went into after casting. After a spate of breakages, we went way OTT with rules, checklists, monitoring, cross checks, threats of sacking, etc. etc.

The next block to be made was the top item on the agenda at the production meeting every day - every wiggle on the temperature trace having to be explained by the technical boffins, every hourly check filled in and countersigned - the full 9 yards.

When the day came to lift the kiln, nobody wanted to do it. Eventually my mate Arthur said he'd do it. As the hood of kiln 12 was lifted off, people peered anxiously underneath to see if the block was in one piece...

The kiln was empty!

It turned out that the block had gone into kiln 13 by mistake, and been left to its own devices to crash-cool while a few m^3 of fresh air was carefully annealed in kiln 12.

It still makes me laugh today (and it was probably 20 years since it happened - RIP Arthur.)

....solution anneal at 1038 C and then do the fresh treatment from scratch at 1025F....

Units, for goodness' sake!!!!!!

Alot lower temp than that. Jet fuel burns at about 850 degs C in open air and steel looses half its strength at temps as low as 650 legs C. That’s why the 9/11 conspiracy theorists say how could jet fuel melt steel, but as you say the steel didn’t need to melt or get anywhere near its melting temperature before it lost its structural integrity.

Yeah I made a bonfire sculpture a while back - just a 6mm plywood cladding and a spritz of paraffin over a box section steel armature. The cladding burnt away in a few minutes but that alone was warm enough for the steelwork inside to collapse under its own weight

Today's the day. Parts have been solution annealed and re-treated to the correct condition, and then re-machined to get them back in tolerance - it was surprising how much they grew by. Luckily the spigots grew by more than the sockets, so we can still get the correct amount of interference. They've also gone a rather fetching colour - the parts which were H900 to start with have gone purple, and some which were H1150D have gone green.

Ooh! an update - thanks 🙂

I hope you've got a load of things lined up to dip in the liquid nitrogen while it's there...

I'm wondering if it would make sausages too brittle for hammering - please report back.

That’s why the 9/11 conspiracy theorists say how could jet fuel melt steel, but as you say the steel didn’t need to melt or get anywhere near its melting temperature before it lost its structural integrity.

I overloaded a small portable barbecue this summer. Just the weight of the coals over the presumably red heat of the grille was enough to bow it. And then when I came to bend it back it was soft as putty due to having been annealed.

That's not enough Nitrogen... you want a big tank-full, boiling away...

edited - video wouldn't play... grrr...

Not good so far I'm afraid chaps. The theoretical temperature difference for free assembly was 75 degrees (H7/s6), but even with nearly 300 degrees, it wouldn't go. Tried using the big press, which was a bad idea. The two parts galled together and needed 8 tons to separate (with a bang). Now back in the machine shop (again) to skim off the damage and a bit more off for less interference.

Been quite a learning experience this one! In particular, choose a lighter fit for stainless, and don't try to press it!

Probably too late but I would shrink the insert in liquid nitrogen but not heat the other part. All that happens when you try heating the other bit is the they react with each other cold part pulls the heat from heated part and heated part heats up the other bit. Depending on how much the interference is liquid nitrogen will shrink steel bt a few thou

Put bit in nitrogen and leave until it stops bubbling

Hurrah! H7/n6 dropped together like a charm. Test loaded it up to 1 ton and it's not shifting so happy with that.

What's the maximum working load likely to be? A 1 tonne test seems low for such big components.

It's mostly bending, not much axial force.

Cant' work out the scale of these, are you making the world's most niche single speed for travel and these are some home-made S&S couplers? 😉

I am going for backyard ICBM. Why else would he be worrying about temps of over 1000 degrees Celsius. I suspect he has plans for a North Korean candle. If he starts looking for an supply of Uranaium 235 i am shopping him.

Whatever it is, it's 37.22mm in diameter...