Hi,
I was under impression that nonlinear BH characteristic of powdered cores is more function of the actual material structure than the material. Attached is third page from Unitrode Magnetics Handbook. This is the only reference describing rounding mechanism of BH curve I have found so far.
Best regards,
Jaka Racman
I was under impression that nonlinear BH characteristic of powdered cores is more function of the actual material structure than the material. Attached is third page from Unitrode Magnetics Handbook. This is the only reference describing rounding mechanism of BH curve I have found so far.
Best regards,
Jaka Racman
Attachments
Hmm…of the seven attributes listed in that sentence, my powder core will kick butt on a laminated core in five of them (impurities need some work, and strain will be a by-product of ramming the mould)…I’m just sayin’.
Before I figured out pressure in the mould was key, I was going to build a vacuum chamber with a vibrating plate to sit the mould on and pour a slurry in. THAT method would require a very low viscosity binder.
To use a thick binder, like slow setting epoxy, you have to ”mull” it into the powder. Mulling is a process were you knead the binder into the powder, think bread dough. The key here is determining the precise amount of binder to fill the voids and no more. As you stated, any excess will not be expressed.
Aye,aye Capin’
How do you propose to remove the carbon from the ground up ferrite matrix? I think you will find that with all the impurities present, you will need approximately 1000 times the number of turns for your powdered-core transformer primary to reach the necessary inductance for a hi-fi OPT than an "ordinary" laminated or tape wound FeSi-cored transformer. After all, what you really have is ground up cast iron of varying composition with all kinds of other miscellaneous impurities to boot, not ground up high purity silicon iron magnetic core material.
Three million turns is a lot of wire.
John
Maybe a chemical process might be useful for cleaning the material and for micro-crystalinizing the powder. Like HCl acid, this should react to form FeCl (same stuff as in PC board etching solution). Hopefully the carbon won't react or dissolve in it. Then use some other reagent to precipitate out the iron crystals. Might be able to avoid the powder grinding process altogether. May lose the silicon in the process.
Hmmm, lets see, copper precipitates out the iron from FeCl, thats how the PC board etchant stuff works. Dissolves the copper off the board in the process. This might be a new source for cheap material, check with some PC board houses and see if they dispose of large quantities of used etchant with iron colloidal dust in it.
Don
Hmmm, lets see, copper precipitates out the iron from FeCl, thats how the PC board etchant stuff works. Dissolves the copper off the board in the process. This might be a new source for cheap material, check with some PC board houses and see if they dispose of large quantities of used etchant with iron colloidal dust in it.
Don
Well color me embarrassed….I should have said “ I BELIEVE this is a function of the core material not being simple Fe-Si”. I would still be wrong, but I wouldn’t have sounded so absolute about it
. Thank you for the resource. As you discovered, finding good info on this topic is rather difficult.I’m still not sure how much of this problem with ferrite translates to iron powder cores. As mentioned earlier, iron being MUCH softer than ferrite, it will deform more under pressure, increasing the contact size between particles. This should minimize this effect. How much ? I don’t know (yet).
I’m not sure. I think the key here is that we’re talking “matrix’ not “alloy”. High carbon steel under a microscope shows the carbon forms little squiggly pockets of carbon distributed randomly . It seems to me if the iron is pulverized, the breaking points will be these pockets, as carbon is weaker than iron. That being the case, a good deal of the carbon will separate from the iron, becoming free standing carbon dust. Then it would be a matter of magnetically separating the two.
1000 times the turns !?! An air-core tranny would have about 400 times the turns...this could be bigger than I thought…NEGATIVE PERMEABILITY… I smell a Nobel.
I don't know where you get the figure of 400 times. In a single-ended application, especially the size you're talking about, grain-oriented silicon steel has at least 2-3 thousand times the permeability of air at low frequencies. Cast iron needs approximately 1000 times the magnetizing force of GOSS to reach it's initial intrinsic induction. In other words, for the material you propose to have ANY magnetic flux present in the core you need 1000 times the number of turns as GOSS.
A word about purity which is little understood. For some materials to perform the way they are intended, there is a threshold of purity that needs to be crossed. Often, especially in electronics, 99% purity won't work, and even 99.9% or 99.99% purity will not have the desired properties. This is true for a number of things including tungsten for filaments, nickel for cathodes and plates, dielectric material for capacitors, copper wire, and magnetic materials. The chief difference between Permalloy and Supermalloy is the care of manufacture and the resulting purity which results in three times the permeability.
I have read a number of advanced texts on the properties of magnetic materials over the years and remain totally amazed at the lengths taken to manufacture these materials such that they will perform as intended. They aren't expensive because the makers are out to gouge DIY'ers.
That's all,
John
Maybe, if there is anybody with access to a lab and/or a chemistry prof that would be able to ask/investigate this let me know. I am more than ignorant when it comes to chemistry.
According to the chart on page 208 of the Radiotron Handbook, 4th edition, 4% Silicon Steel has an initial permeability of 450.
http://www.oldradioz.com/manuals/rdh4/CHAPTR05.PDF
I have not actually wound an air-core audio transformer, but I have wound air core chokes. None had 1000 times the windings of their iron cored brethren.
I am having a hard time wrapping my head around a 10% decrease in purity resulting in a 99.9% decrease in permeability. The carbon, binder, and whatever other crap floating in the mix aren’t actually contaminating the SiFe alloy they are contributing to the distributed air gap.
Casey Brown
jlsem said:
the 400 number seems a lot more realistic to me. the perm of say 2500 is for the material, but you then have to add the length of the airgap into the magnetic path to get the effective perm.
why not look at it from the other direction? Rather than cry about all of the perm you lose from an airgap, why not start with an aircored trannie and thank the magnetic gods for all the extra L the powder gives you.
making your transformer coil with a 1X2 inch cross section with a 5 inch diameter will allow you to put the whole thing in a 12 inch round pan of shavings and hope for a perm of 50. you could pour your binder / powder mix into the "mold" and completely encase the winding. you could even use a large fluxes (both DC and AC) to shake the core into allignement. i have to wonder how a few days at 20KGdc+ 20KGac would compare to a multi-ton jack.
dave
Nuke it!
Perhaps we need to apply some unusual innovative engineering ideas to this material to get improved results. I suggest doing the whole process in one single step. We would surround the brake shavings with multiple explosives in such a way as to produce a converging shock wave. These explosives would require precision timing to trigger them, perhaps someone can come up with some expendable NOS vacuum tube circuitry for this.
This intense converging shock wave should produce an extreme pressure condition approaching the environment at the center of the earth, where iron is hypothesized to exist in a unique quantum magnetic state with ideal properties for making tube output transformers. There may be some risk however that the resulting material will be too hard to machine into any useful shape. I guess if all else fails we can blow it up again to retrieve the diamonds formed from the carbon impurities.
Don
Perhaps we need to apply some unusual innovative engineering ideas to this material to get improved results. I suggest doing the whole process in one single step. We would surround the brake shavings with multiple explosives in such a way as to produce a converging shock wave. These explosives would require precision timing to trigger them, perhaps someone can come up with some expendable NOS vacuum tube circuitry for this.
This intense converging shock wave should produce an extreme pressure condition approaching the environment at the center of the earth, where iron is hypothesized to exist in a unique quantum magnetic state with ideal properties for making tube output transformers. There may be some risk however that the resulting material will be too hard to machine into any useful shape. I guess if all else fails we can blow it up again to retrieve the diamonds formed from the carbon impurities.
Don
True, but with some DC bias, permeability (at low frequencies) increases exponentially to almost 50,000 near the knee of the magnetization curve. I am still talking about M-6, which is what nearly all decent OPT's use. I don't have any charts for materials with DC bias, so I can only guess, but the DC magnetization curve for M-6 shows a permeability of 30,000 @ .1 Oersteds and 3000 Gausses (about 2/3 the way up the curve).
John
John, I think I came off stronger than I intended when I said:
In my sardonic brain I thought this was funny, after re-reading, it comes off as snotty...my apologies
Pretty much how I feel about.
Well...I'm trying to be a little more sophisticated than that. My very loose target is an effective permeability of around 200, which would translate into a transformer roughly twice as big as a "normal" one.
I did think about something along those lines. I decided that an amorphous pattern would be preferable to an aligned one. Not to mention the added complexity.
In my sardonic brain I thought this was funny, after re-reading, it comes off as snotty...my apologies
Pretty much how I feel about.
Well...I'm trying to be a little more sophisticated than that. My very loose target is an effective permeability of around 200, which would translate into a transformer roughly twice as big as a "normal" one.
I did think about something along those lines. I decided that an amorphous pattern would be preferable to an aligned one. Not to mention the added complexity.
jlsem said:
i think you are still looking at the wrong picture. The BH loops you see are for minimally gapped materials. if you add an intentional gap, the loop shears over and you need to rework your numbers. Many of the loops are from real world samples (exisitng core geometries) but others are from ring samples and in both cases neither apply to anything that carries DC with any "added" gap.
valvituede said:
does it work that way? you are starting with a very lossy idea, to mention amorphous seems like an opposite to me. i'm just excited about your cup being half full
dave
Dave,
QUOTE]does it work that way? you are starting with a very lossy idea, to mention amorphous seems like an opposite to me. i'm just excited about your cup being half full [/QUOTE]
I wish I was sure about any of it, alas, I'm just cautiously optimistic. Because of all the unknowns, B-H linearity, gap losses, effects of impurities, so on and so forth, I don’t have any numbers to play with. At this point, I’m broad brushing everything. The only empirical evidence I have that amorphous MIGHT be better than aligned, is the experience I had with tape heads. Back around ’82 or ’83 I couldn’t get an OEM permalloy head for a customers tape deck, but I had access to a generic powdered sendust head. After I installed it and tweaked everything I was impressed enough with how it sounded on my “utilitarian” bench system, I hooked it up in the stores sound room. I was then impressed enough to summon the store owner. Long story short, we then offered this mod/upgrade to our customers. I remember looking up anything I could find about these style of head, and I remember a big deal being made about the amorphous nature of the random grain pattern. A few things to consider, 1) I don’t know if the grain pattern really had anything to do with their magic, or if it was marketing hype. 2)Obviously there was no dc in the coil. 3) I was in my early twenties then, and much recreational consumption was done in the name of a good time, thus the portion of grey matter assigned to archiving this info may or may not be functioning properly. It “feels” right though.
Casey
QUOTE]does it work that way? you are starting with a very lossy idea, to mention amorphous seems like an opposite to me. i'm just excited about your cup being half full
[/QUOTE] I wish I was sure about any of it, alas, I'm just cautiously optimistic. Because of all the unknowns, B-H linearity, gap losses, effects of impurities, so on and so forth, I don’t have any numbers to play with. At this point, I’m broad brushing everything. The only empirical evidence I have that amorphous MIGHT be better than aligned, is the experience I had with tape heads. Back around ’82 or ’83 I couldn’t get an OEM permalloy head for a customers tape deck, but I had access to a generic powdered sendust head. After I installed it and tweaked everything I was impressed enough with how it sounded on my “utilitarian” bench system, I hooked it up in the stores sound room. I was then impressed enough to summon the store owner. Long story short, we then offered this mod/upgrade to our customers. I remember looking up anything I could find about these style of head, and I remember a big deal being made about the amorphous nature of the random grain pattern. A few things to consider, 1) I don’t know if the grain pattern really had anything to do with their magic, or if it was marketing hype. 2)Obviously there was no dc in the coil. 3) I was in my early twenties then, and much recreational consumption was done in the name of a good time, thus the portion of grey matter assigned to archiving this info may or may not be functioning properly. It “feels” right though.
Casey
Acid treatment will not work. Carbon will not dissolve, but the iron will entirely, leaving FeCl2 (NOT FeCl3, which requires a stronger oxidizer to form) in solution. This can be cementated with a more reactive metal, such as zinc, BUT the zinc has to be as fine as the powdered metal you want in the first place. Useless.
Other common powders such as aluminum, magnesium or an alloy (magnalium) are too reactive and will react with the water solution preferentialy to the salt (FeCl2), doing nothing.
Whether or not the alloy will harden by quenching the powder in air depends on the particle size and the composition. If you read up on cast iron, you'll note there are a few grades of grey (i.e. secondary (eutectic formed) flake graphite) cast iron. The weakest has an entirely ferrite matrix around the graphite flakes. It is soft and weak (if the flakes were rounded, as in ductile iron's nodular graphite, it would have high elongation, comparable to mild steel), and as all flake gray irons, brittle (due to the flakes essentially being sharp porosity, becoming stress raisers).
A higher grade is partially pearlitic, which means as the iron cools through the transformation temperature, the austenite, saturated with carbon, deposits it not as an additon to the graphite flakes but as an eutectoid (layered) cementite and ferrite structure.
The strongest grade is fully pearlitic, where two major phases are present: flake graphite and the pearlite eutectoid. The ferrite is still present, but layered with strong, hard cementite.
Now, where particle size comes in: if you are milling this to say 200 micron, you might have some free graphite, but most will still be trapped inside the matrix of ferrite and/or pearlite. Down around 50 microns, you'll have most of the graphite free, but any pearlite present will still be agglomerated. To free that up, you need it down around 1 micron.
The one nice thing is that almost all cast iron contains a useful percentage of silicon (the exception being wear-resistant, hard white cast iron, and its derivative, the now outdated malleable iron), which being insoluble in carbon and not present as a seperate phase, is fully in solid solution with the ferrite.
Oh, back to my original mention of particle size: if you have material small enough that it has no graphite present in the grains, then heating above the transformation temperature (1600øF) and quenching will not harden it, as there is no source of carbon.
Passing through air *will* generously decarburize it, as well as oxidize the surface. (Noting that it's already reasonably decarburized if it is a soft grade of gray iron and milled fine enough.) Carbon is capable of very quickly diffusing to great depth in iron; case hardening can go 1/16" deep in a matter of hours at a reasonable temperature (yellow heat or so). Such very fine particles will also pour "up", given the increase in temperature (due to high surface area) and the resulting bouyancy.
Uh, seperating C from Fe is simple enough, do a flotation seperation. Or heck, could even go magnetic.
In reply to the posts a few weeks ago (so it would seem... stop posting so much!), sintering isn't that far off. I've sintered pottery in my furnace before. (Clay is nothing more than a very fine oxide powder produced by natural weathering; it has certain properties which allow it to bind itself during firing.) 2000øF is a rather cool temperature in the grand scheme of things, I hope you weren't finding that figure particularly toasty or anything???
As for sintering iron, now that you mention it it may produce too conductive a material, however under steam, CO2 or oxygen (slightly lean mixture in the furnace, as opposed to H2 and CO atmosphere), an oxide layer might be maintained up to 2000-2500øF when the oxide surfaces will diffuse together.
On the other hand, I can burn through a tin can at only 1800øF, powder is certainly dead burned by a mere red-orange heat!
Methinks if you're doing any pyrological processing on this, it's going to be making ferrites, not powdered iron, as the result. Whichever works.
Tim
Other common powders such as aluminum, magnesium or an alloy (magnalium) are too reactive and will react with the water solution preferentialy to the salt (FeCl2), doing nothing.
Whether or not the alloy will harden by quenching the powder in air depends on the particle size and the composition. If you read up on cast iron, you'll note there are a few grades of grey (i.e. secondary (eutectic formed) flake graphite) cast iron. The weakest has an entirely ferrite matrix around the graphite flakes. It is soft and weak (if the flakes were rounded, as in ductile iron's nodular graphite, it would have high elongation, comparable to mild steel), and as all flake gray irons, brittle (due to the flakes essentially being sharp porosity, becoming stress raisers).
A higher grade is partially pearlitic, which means as the iron cools through the transformation temperature, the austenite, saturated with carbon, deposits it not as an additon to the graphite flakes but as an eutectoid (layered) cementite and ferrite structure.
The strongest grade is fully pearlitic, where two major phases are present: flake graphite and the pearlite eutectoid. The ferrite is still present, but layered with strong, hard cementite.
Now, where particle size comes in: if you are milling this to say 200 micron, you might have some free graphite, but most will still be trapped inside the matrix of ferrite and/or pearlite. Down around 50 microns, you'll have most of the graphite free, but any pearlite present will still be agglomerated. To free that up, you need it down around 1 micron.
The one nice thing is that almost all cast iron contains a useful percentage of silicon (the exception being wear-resistant, hard white cast iron, and its derivative, the now outdated malleable iron), which being insoluble in carbon and not present as a seperate phase, is fully in solid solution with the ferrite.
Oh, back to my original mention of particle size: if you have material small enough that it has no graphite present in the grains, then heating above the transformation temperature (1600øF) and quenching will not harden it, as there is no source of carbon.
Passing through air *will* generously decarburize it, as well as oxidize the surface. (Noting that it's already reasonably decarburized if it is a soft grade of gray iron and milled fine enough.) Carbon is capable of very quickly diffusing to great depth in iron; case hardening can go 1/16" deep in a matter of hours at a reasonable temperature (yellow heat or so). Such very fine particles will also pour "up", given the increase in temperature (due to high surface area) and the resulting bouyancy.
Uh, seperating C from Fe is simple enough, do a flotation seperation. Or heck, could even go magnetic.
In reply to the posts a few weeks ago (so it would seem... stop posting so much!), sintering isn't that far off. I've sintered pottery in my furnace before. (Clay is nothing more than a very fine oxide powder produced by natural weathering; it has certain properties which allow it to bind itself during firing.) 2000øF is a rather cool temperature in the grand scheme of things, I hope you weren't finding that figure particularly toasty or anything???
As for sintering iron, now that you mention it it may produce too conductive a material, however under steam, CO2 or oxygen (slightly lean mixture in the furnace, as opposed to H2 and CO atmosphere), an oxide layer might be maintained up to 2000-2500øF when the oxide surfaces will diffuse together.
On the other hand, I can burn through a tin can at only 1800øF, powder is certainly dead burned by a mere red-orange heat!
Methinks if you're doing any pyrological processing on this, it's going to be making ferrites, not powdered iron, as the result. Whichever works.
Tim
I understand that. But because the core is carrying DC, it is "biased" up toward the region of the magnetization curve where mu becomes very high. It's the gap that lowers the mu and my guess of a factor of ten is very conservative, and leaves a mu in the operating region of the curve of about 3000. The error in my previous post was that I meant to say I haven't any charts for gapped cores, but I do have a representative graph where a reduction by a factor of ten is an extreme case. Again, I am talking about low (60 hz) frequency where high inductance is important. The low figure of initial permeability given in RDH4 is probably for a lower grade of steel at the bottom of the curve at an intermediate frequency.
John
Tim,
I've read your post, and I THINK your saying my idea could work if I rework it a little.
This is what I propose..please tell me if I'm going in the right direction.
1) Heat un-crushed chips to 1600 deg., pour through air. This does a) fully harden chips to facilitate easy rendering to powder b) pulls a percentage of carbon out of the metal to fom the oxide layer.
2) Pulverize in ball mill until powder is slightly larger than 50 microns.
3) "Float" seperate the carbon (scale) that was formed in step 1, and knocked off in step 2.
4)Re-heat to 1600 deg, stir in pot to ensure some O2 is mixed in, cover and let cool slowly, this will a) fully anneal the iron to its "soft" state, and b) form a new oxide layer with the remaining carbon.
garnish and serve.
So.. what do you think ?
Casey
I've read your post, and I THINK your saying my idea could work if I rework it a little.
This is what I propose..please tell me if I'm going in the right direction.
1) Heat un-crushed chips to 1600 deg., pour through air. This does a) fully harden chips to facilitate easy rendering to powder b) pulls a percentage of carbon out of the metal to fom the oxide layer.
2) Pulverize in ball mill until powder is slightly larger than 50 microns.
3) "Float" seperate the carbon (scale) that was formed in step 1, and knocked off in step 2.
4)Re-heat to 1600 deg, stir in pot to ensure some O2 is mixed in, cover and let cool slowly, this will a) fully anneal the iron to its "soft" state, and b) form a new oxide layer with the remaining carbon.
garnish and serve.
So.. what do you think ?
Casey
jlsem said:
when you get into DC bias, you need to look at the incramental permeability, a completely different animal.
you cannot determine perm with a DC bias from an AC loop. You need to look at the DC loop, apply your bias to find the new origin, and then run a new loop for the voltage and frequency in question. the slope of the line drawn through the points of the new loop is your incramental perm. Without an airgap, the incramental perm, will vary quite a bit dependign on a number of factors, but as you increase the gap size the DC loop shears over and narrows bringing the variability of incramental perm down considerably.
dave
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.