Magnetic Alloys for Temperature Control

[T3sl4co1l]

Hey, I found another collection of keywords that is utterly impossible to google.  Yay...

Has anyone seen a supplier for alloys of given Tc (Curie temperature)?

Example, this is what Metcal irons do, as far as I know: the base simply provides power, you select the tip for desired temperature.  It sheds power by reducing permeability at Tc.

If I wanted to make my own tips, or other similar applications (induction heating susceptors), who would I talk to?  Does anyone stock this?  Would I be looking at a custom run of <consult the literature>% from a whole-ass mill?!

Tim

[TimNJ]

I am not of much help but:

https://patents.google.com/patent/US4701587A/en

https://patents.google.com/patent/US5329085A/en

https://patents.google.com/patent/US4695713A/en

Interesting reads. I have no idea what the alloy would be...something nickel-iron? Like invar or something?

[TimNJ]

https://www.totalmateria.com/page.aspx?ID=CheckArticle&site=ktn&LN=IT&NM=374

https://www.totalmateria.com/images/Articles/ktn/Fig374_2.jpg

Invar-36: Tc = 280C
Invar-39: Tc = 340C
Invar-42: Tc = 380C


I don't know if you can even buy the weirdo alloys (pretty sure Invar 36 is by and far the most popular), or if this type of material is suitable at all, just first thing that came to mind.

[sandalcandal]

Crazy idea, maybe you can make your own induction heating crucible and mix your own alloys, Applied Science style.

Edit: With the recent advances in power semiconductors and capacitors, it should be cheaper and easier than ever to make a decent induction heating crucible/forge. Makes me want to make one now that I think about it... it's just a resonant locked LC converter after all...

[moffy]

Where I previously worked, we used INVAR for bases of some laser gyroscopes(very low expansion rate), but it was very hard to get in the form we needed. The purchasing/engineers had to make a special deal with I think a Norwegian foundry for several tons. They were very excited at the time.

[jpanhalt]

I searched on "curie temperature alloys." As expected, lots of hits in various protected sources to which I no longer have access.  Ternary alloys of Fe, Ni, and Co seem to be commonly studied.  One citation studied a Mo and Al alloy.   Some abstracts were so limited, it was hard to tell what alloys were studied.  There was also mention of a well-studied "Cantor alloy."

Here's one such hit for Fe,Ni,Co alloy: https://aip.scitation.org/doi/10.1063/1.1713257
Here's Al-Mo alloy: https://journals.aps.org/prb/abstract/10.1103/PhysRevB.103.064407

[T3sl4co1l]

Crazy idea, maybe you can make your own induction heating crucible and mix your own alloys, Applied Science style.

I mean, there's a few steps missing there, but yes, generally possible...

I've actually wanted to do that for quite a long time.  I was doing amateur foundry in high school, and that's what gave me the idea to make an induction heater.  Alas, I need something other than this apartment to do that in..!

Add to that: grinder, rolling mill, probably a regular mill, annealing furnace (that's at least trivial if I've got the first part down ), shear, and anything else to make finished parts (brake? dies? brazed or spot welded items?).

I wonder how tricky it is to get a magnetically soft alloy.  The amount of carbon and other alloys should be very small, AFAIK.  With such little carbon there's also a very fine line before picking up oxygen instead.  Which dissolves quite well into molten iron, eventually separating as molten FeO floating on top.  Cool fact, all iron oxides melt slightly below the pure metal.  Which looks really interesting when you've got a piece of steel at that temperature in a neutral to oxidizing atmosphere: it starts dripping!

So, inert gas purging may be desirable; add welder and gasses to the list.

But really, that's at or under $10k of equipment, to do small scale production of real materials, not a huge cost as capital goes.

Composition is easier to control, I think, with a large heat.  The big guys can also take samples in the middle of a run, go over and XRF it, and adjust it as needed.  Which, even adding XRF capability isn't a huuuuge investment as they show up relatively cheap from time to time.  But I'd really be wanting to think about finding customers for specialty alloys at that point, and spinning off the business...

Edit: With the recent advances in power semiconductors and capacitors, it should be cheaper and easier than ever to make a decent induction heating crucible/forge. Makes me want to make one now that I think about it... it's just a resonant locked LC converter after all...

It's really not so hard, a PLL gets resonance tracking by default, the output of which can be pulled to control current, voltage, power, whatever.  Just add on those sensors (typically a voltage divider, current transformer, and active rectifier and filter to get the envelope of them), and error amps, and you've got it.  Don't forget fault circuitry, a peak current detect is worthwhile, and desat protection for the inverter can save literal buckets of transistors.

Compensation is rather difficult in this method, as the complex pole in the loop response corresponds to the difference between driven and resonant frequencies.  Consider taking the envelope of this and trying to control it:



Stimulus is FSK (open loop, fixed frequency drive), with the lower amplitude being higher above resonance.  Evidently, 3.3kHz above, and the higher amplitude, close to 1kHz.  Resonance looks to be about 21kHz then.

The alternative is always locking to resonance, and modulating the supply -- so you need two converters.  Still easy enough, just more stuff to do, and a bit less efficiency.

Note that you can even do real-time impedance measurement, given a voltage and current sensor.  You can solve for the network components in real time, and infer properties of the work; maybe even temperature or distance or cross section.  (Or with more refinement: eddy current testing is used to detect variations or defects in metals.)

If you're interested enough to do experiments, I recommend a series resonant load.  The resistance is usually quite low, making it easy to match with a one-turn toroidal transformer, or maybe two.  (Note that there are better ways, with lower leakage inductance, but this is pretty easy to construct.)  Also, as a simple series circuit with a known component (capacitor), you can solve for L and R in real time, if you like.


But I digress; given a load with fixed Tc, and modest coupling, control may not even be needed.  I think this could be useful for certain dedicated, well-defined loads -- certainly, Metcal found one such case.  I suppose for Tc nearly room temperature, it could be used for just keeping things warm, and up from there, anything like: preheating; cooking, frying or baking food; soldering, etc.  If such materials exist, it would be a boon for development of applications in these spaces.

Tim

[sandalcandal]

Crazy idea, maybe you can make your own induction heating crucible and mix your own alloys, Applied Science style.

I mean, there's a few steps missing there, but yes, generally possible...

I've actually wanted to do that for quite a long time.  I was doing amateur foundry in high school, and that's what gave me the idea to make an induction heater.  Alas, I need something other than this apartment to do that in..!

Add to that: grinder, rolling mill, probably a regular mill, annealing furnace (that's at least trivial if I've got the first part down ), shear, and anything else to make finished parts (brake? dies? brazed or spot welded items?).
[...]
But really, that's at or under $10k of equipment, to do small scale production of real materials, not a huge cost as capital goes.

Composition is easier to control, I think, with a large heat.  The big guys can also take samples in the middle of a run, go over and XRF it, and adjust it as needed.  Which, even adding XRF capability isn't a huuuuge investment as they show up relatively cheap from time to time.  But I'd really be wanting to think about finding customers for specialty alloys at that point, and spinning off the business...

I was thinking in the scope of just getting a cast puck of material with \$ x\$ °C Curie temperature to begin with but it sounds like you're thinking of making a legit product and starting a small business 

Inert gas purging shouldn't be difficult or expensive. There's plenty of fairly affordable and accessible equipment and gas available as arc welding supplies. Making a non-pressurised air-tight enough "bowl" is also not particularly hard with welding or even just brazing/soldering. [On second thought it might be smarter to use something non-conductive at least for the wall which the coil passes through and aluminium tape to seal the joint between the nonconductive wall and the rest of the bowl. Similar tape sealed setups are used when welding on sensitive materials]

Those XRF guns do look very neat and not overly expensive https://youtu.be/KdfHVcU8U7U

As for not having the space, you could consider looking for a local Maker space that might be able to accommodate the project. There are constantly more and more around; the communities tend to be pretty accommodating too.

I've seen quite a few DIY induction heaters capable of melting steel over the years and given your expertise I don't think it should prove too challenging to make something that works (as long as you don't go over complicating it  )


(Also check out other videos on the guy's channel, he's been doing continued work on induction heaters for quite a few years) Project pages for his mcu controlled PLL induction heater http://www.mindchallenger.com/inductionheater/ (and forum thread for the project https://4hv.org/e107_plugins/forum/forum_viewtopic.php?p=1&id=85701) Apparently it has issues properly locking onto the resonant point once stuff melted and went past the Curie temperature but he might have fixed it since then, I haven't follow closely.

You can get decent results with something as rudimentary as the Mazilli Driver (A Royer Oscillator with MOSFETs)

Thread for the above https://4hv.org/e107_plugins/forum/forum_viewtopic.php?p=1&id=122354 and blog page. A 4KW version here: https://4hv.org/e107_plugins/forum/forum_viewtopic.php?p=1&id=129196 (see attached for the schematic). Electroboom actually made some sensible improvements to the basic circuit https://www.electroboom.com/?p=1198

Another one using an mcu and PLL https://4hv.org/e107_plugins/forum/forum_viewtopic.php?id=140297


@madsbarnkob was more active on that forum so he might recall things a bit better if he sees this or you could go ask on his forum. 4hv.org is unfortunately pretty much dead.

[T3sl4co1l]

I was thinking in the scope of just getting a cast puck of material with \$ x\$ °C Curie temperature to begin with but it sounds like you're thinking of making a legit product and starting a small business 

I mean, if I'm investing in that much equipment (and the time to gain skill in its use), for purposes of meeting an application, and that application goes into mainstream production -- it'll need to be done at the same scale.  And perhaps there are other applications for such material, I have no idea.

Just that, that would be the natural sequence in that case.  And, big "if"s at that (if it takes off, if the material works, etc. etc.).

And conversely, it's a lot of work to do, just to prove some product is feasible; that would be going backwards from how such a product should be developed.  It would be nice if such materials are already available.

And I don't much care about doing a one-off, there are substitutes (just measure the temperature directly and use a controller).  It's an application specific thing.  In particular, if it's just dumb metal strips of mostly iron -- it could be quite a bit cheaper than using a thermocouple and controller.  Like, disposably cheap, maybe.

I've seen quite a few DIY induction heaters capable of melting steel over the years and given your expertise I don't think it should prove too challenging to make something that works (as long as you don't go over complicating it  )

You're forgetting the OG.



This isn't an academic exercise, I've done it before, I know how it works, heh.  I just don't know if metal with this particular property (given Tc) is available anywhere...

I made that diversion in case anyone else was interested -- or if you like (or anyone else), we could split the thread and discuss induction separately.

Tim

[sandalcandal]

I was thinking in the scope of just getting a cast puck of material with \$ x\$ °C Curie temperature to begin with but it sounds like you're thinking of making a legit product and starting a small business 

I mean, if I'm investing in that much equipment (and the time to gain skill in its use), for purposes of meeting an application, and that application goes into mainstream production -- it'll need to be done at the same scale.  And perhaps there are other applications for such material, I have no idea.

Just that, that would be the natural sequence in that case.  And, big "if"s at that (if it takes off, if the material works, etc. etc.).

And conversely, it's a lot of work to do, just to prove some product is feasible; that would be going backwards from how such a product should be developed.  It would be nice if such materials are already available.

And I don't much care about doing a one-off, there are substitutes (just measure the temperature directly and use a controller).  It's an application specific thing.  In particular, if it's just dumb metal strips of mostly iron -- it could be quite a bit cheaper than using a thermocouple and controller.  Like, disposably cheap, maybe.

Fair enough. I wasn't sure of the context or the goals; if this was a hobby/experimental project or attempting something more serious.

You're forgetting the OG.

HA! My recollection from 10 years back is bit fuzzy. Didn't realise you were among the first. [Found your blog posts too https://www.seventransistorlabs.com/tmoranwms/Elec_IndHeat1.html]

I'm not lacking the space or access to tools to do an induction heater project but I'm unfortunately very much lacking the time currently Trying to revist the decade old designs with some of the newer tech particularly with new WBG semiconductors and MLCC does still seem very interesting though.

[jh15]

I always liked the Weller bench Iron I got in the 70's. The bottom of each tip had a single number stamped into the Curie element. 7 for 700F, etc.
When it started clicking, I knew it was up to temperature. And the clicking would remind one that it was still on if you wern't in view of the pilot light.
     Unfortunately not good for modern sensitive electronics, It makes a big spike when switching, and I don't think the leakage spec is good..
Still, I use it for a lot of routine stuff.

[jbeng]

I always liked the Weller bench Iron I got in the 70's. The bottom of each tip had a single number stamped into the Curie element. 7 for 700F, etc.

That was the first thing I thought of when I saw this thread.

Unfortunately not good for modern sensitive electronics, It makes a big spike when switching, and I don't think the leakage spec is good..

When I worked in the EE lab, we used to call that the "Weller pulse" and it would be all over the signal in the preamps we used in our systems.   Want to measure the preamp performance?  All Wellers = OFF.

Still, I use it for a lot of routine stuff.

I still have one and use it often.

[T3sl4co1l]

Ah, bimetallic strip?  Yeah, not quite like that...

Or, I suppose you could do it sideways kinda, use a permanent magnet to pull on a nickel (or alloy) strip, when it heats up it lets go.

To be clear, I want the AC field, no moving parts, version.

Tim

[jbeng]

Ah, bimetallic strip? ...

Nope, no bimetallic strip involved in that iron

To be clear, I want the AC field, no moving parts, version.

I understand. This iron does have moving parts though

... use a permanent magnet to pull on a nickel (or alloy) strip, when it heats up it lets go.

Actually, that is closer to how it operates, see the attached graphic, an excerpt from the Weller WTCPN Tech sheet.

The solder tip itself has a metal slug in the end, which when it's temp is below the stamped setpoint, attracts a magnet in the temperature control system.  This moves a set of switch contacts to close and energize the heating element.  When the tip's metal slug reaches the curie point of the alloy, the magnet is no longer attracted to it and the contacts open, regulating the iron temperature.

It's pretty slick, really.

[T3sl4co1l]

Aha, second guess got it right.  Yeah, likely the same material would work, unless it's a ferrite of course.

Which, is an interesting idea as ferrite Tc's are lower than iron or nickel, but doesn't have the AC losses so would have to be built in such a way to deliver flux to a conductor (perhaps a nonmagnetic stainless instead?).  Would be a lot harder to mass produce though.

Tim

[magicsmoke]

I have no idea if they carry the type of alloy that you are looking for but the first two places that I would call are:

Carpenter Technologies (https://www.carpentertechnology.com/alloy-finder)

or

VAC (https://vacuumschmelze.com/)

[LaserSteve]

I have access to a large amount of  0.495" dia. invar rods. How much do you need?

[T3sl4co1l]

Can you roll that into sheets and tune the Tc for 100-200°C?

I don't actually need it, but I can see applications if it were available.

Tim

[LaserSteve]

Tough to machine. I do not have a drop hammer
and forge to make a blank to roll.  It work hardens, is tough on lathe  tooling, good surface finish is hard to come by.  Probably with correct annealing it becomes easier to work,but I have not tried that yet.  I have a few long  rods made for laser resonators and I just happen to know where there is about 2000 pounds or more of them, factory surplus.

Steve