Connect with us

Ferrite resistivity

Discussion in 'Electronic Design' started by George Herold, Sep 14, 2012.

  1. So following a comment by Mikko that ferrites stop working when cooled
    to liq. helium temperatures. I ‘discovered’ that ferrites have a
    phase change somewhere near 120 K. (at least magnetite does.)
    Google “Verwey transition”
    Wiki is mostly silent on the subject though there is a bit here.
    http://en.wikipedia.org/wiki/Charge_ordering

    So first I’m a ferrite novice. I’ve wound some RF transformers in the
    deep past, but that’s about it.
    I found that I could measure the resistivity of some ferrite beads
    with just my DMM. I’ve got two types of bead the 43 material and the
    73 material. The reported resistivity’s are 43 = 1E5 ohm-cm and 73 is
    100 ohm-cm.

    I made a little jig to squeeze the beads in. The brass screw has a
    cone turned on the end.

    http://bayimg.com/FAcKEAAeE

    So first off there seems to be a huge variation in the piece to piece
    resistivity. At least an order of magnitude in the few pieces I
    looked at.
    Second the resistivity was (most of the time) much higher than the
    reported numbers.

    1E6 to 1E7 for the 43 material and 3E3 to 2E2 for the 73. (The 200
    Ohm-cm for one piece of 73 was about right.)

    And finally the 43 material was some older stuff in my parts box. I’m
    not quite sure of it’s provenance. So I got some new pieces out of
    stock. For the new material I couldn’t measure the resistance with my
    setup. I even biased the bead from a 30 volt supply and used the
    10Meg of the DMM as a voltage divider... I could measure a 1 G-ohm
    resistor that way, but not the beads! resistance greater than 10 G
    ohm or os.

    I’m wondering if anyone has some more in depth knowledge they might
    share. The ferrites look like they might be a ‘model system’ for some
    new solid state experiments. Besides looking at the Curie
    temperature the phase transition at 120K shows a peak in the heat
    capacity, change in resistivity and magnetic properties. What could
    be better!

    Thanks,

    George H.
     
  2. whit3rd

    whit3rd Guest

    [and on resistivity of the ferrite material]
    Probably the resistivity is quoted mainly because it affects inductor Q.
    Ferrites are, alas, composite materials; the magnetic grains (of hopefully
    uniform size, big enough to magnetize) are in a matrix of fired ceramic
    like kaolin clay. So, electric conduction depends on percolation through
    the ceramic matrix, or through contact between grains, or in the boundary
    layer where the kaolin is in contact with the hematite/magnetite.

    It is dubious that resistivity repeats from batch to batch, or from one contact
    point to another. A bit of internal fracture might not change the magnetic
    properties much, but would kill a resistivity measurement. I've done
    impedance measurement on ferrites that had LOTS of internal movement,
    at some frequencies the internal fractures rang like a lossy bell.
     
  3. Bill Sloman

    Bill Sloman Guest

    Zebra strip?

    http://www.fujipoly.com/design-guidelines/89.html

    Farnell don't seem to stock it any more, so you may find it difficult
    get hold of.
    The Kelvin connection scheme is always attractive.
     
  4. I did stick the extreme pieces of 73 on an SRS RCL meter. (1.2k and
    25k) (wire through the bead) The 25k did show less resistance and
    higher Q consitent with the resistivity... but only a 10-20%
    difference.

    So if most of the resistance is because of some random path... I still
    should be able to see any phase change in the resistivity.

    George H.
     
  5. Sure, the resistance changed as I cranked on the screw. But only by
    maybe a factor of 2 or 3. I could tighten things up to within ~20% of
    the cracking point.... after I'd cracked a few beads.
    Oh yeah this was very crude... the spinning of the screw was the worst
    part.

    I tried using some 'gummy' copper tape as a 'gasket', but it spun off
    into a crumple ball.
    Hmmm.. those beads aint so big, making two more (ring?) terminals
    would be real work.

    George H.
     
  6. legg

    legg Guest

    Although this is a useful method to differentiate the two materials,
    it's not going to give meaningful bulk resistivity information.
    The higher resistivity material is sintered in a fairly crystaline
    structure, with insulator/semiconductor type relationships between
    domain groups. You can't count on it functioning as an insulator
    in-circuit, because this characteristic is uncontrolled - but you sure
    can count on the low resistivity stuff causing shorts, if misapplied
    in a higher voltage circuit.

    Some materials are not formed as individual beads, but are machined
    from tubular structures, or finished into their final dimensions by
    grinding, The machined parts provide a more reliable contact to the
    bulk material, but their function is also marginally affected due to
    the reduced bulk impedance at the machined surface.

    It used to be that low resistivity, low frequency parts also had
    noticably lower currie temperatures. This restricts uses of parts in
    places where significant power loss is expected - they become
    loss-self-regulating, much like an NTC resistor - which is no good if
    a loss must be absorbed and the site of application isn't heatsunk in
    some way.

    RL
     
  7. Thanks, If this works I can try looking for better samples.
    The ferrite beads are probably the 'low end' of the production run.
    A low Curie temperature would be attactive. (easier to reach)
    I like this melamine foam which is good to?? (at least 140 C)

    George H.
     
  8. Hmm, I think the first thing I need to get rid of is the twisting
    motion of the screw. I don't quite see how to do that. (And keep the
    large adjustment range.)

    Now here's a question for your four point measurement suggestion.

    If I've got some random walk conduction through the sample...
    (to explain the large resistance)
    Then when I put on my two voltage probes,
    do they make random walks through the sample too?
    In which case the position of the volatge probes becomes only an
    approximate measure of the distance along the conduction channel.


    Hey, can I do some sort of AC measurement that probes shorter range
    conduction?
    Or is that at a very high frequency?

    George H.
     
  9. My ferrite knowledge is about 40 years old, from the Philips plant where
    I worked as ferrite lab engineeer, so there are probably newer
    compositions on the market.

    Ferrites are solid solutions of either manganese or nickel and zinc and
    iron oxides, sintered to form micro crystals around 1 micron size. Each
    crystal is intended to be one magnetic domain, for lowest losses. There
    are minor but well controlled additions of silica and/or calcium oxides
    which act as glassy insulators between the ferrite grains. In the
    manganese-zince ferrites there is also a critical amount of FeO in the
    inter-grain spaces, whereas the grains themselves are MnO2-ZnO-Fe2O3. (
    Magnetite, Fe3O4 is really a solid solution of FeO in Fe2O3, ie its a
    'ferrite' without the manganese, nickel or zinc components ).

    Basically there are two formulations:

    Manganese zinc ferrites
    For low frequency, say 1 KHz to 200 KHz.
    High permeability, around 2000 and up, low resistivity, around 100-1
    kohm-cm or so.
    Curie temp around 150-200 degC, depending on formulation (
    manganese-zinc ratio ).

    Nickel zinc ferrites
    For high frequency, say 500 KHz to 50 MHz.
    Permeability around 50-500, high resistivity, around 1e9 ohm-cm or higher.
    Curie temp 200-500 degC, again depending on formulation ( nickel-zinc
    ratio ).

    Resistivity is not well controlled, and only has a small part in ferrite
    losses. It also has very high negative, and very uncontrolled, temp coeff.

    Ferrites are optimised for work around -20 to about +100 deg C. Below
    that, the permeability falls rapidly, and there may well be phase
    changes, as you already discovered.
     
  10. Yes, we had some offline discussion with George about the
    phenomenon. The microscopic mechanism behing Verwey transition is
    quite intriguiging, and its existence does not seem to be widely known
    - at least I had never heard about it before (well, if its covered in
    Kittel that is just my ignorance, we followed Aschroft-Mermin...). The
    ferrites I have tested in LHe have all been the standard MnZn or NiZn
    spinel types (it seems hard to find other types, like garnets, sold in
    small quantities) and it sounds likely that it was exactly the Verwey
    transition which killed them - assuming the transition also happens in
    mixed spinels, the papers I've found only describe pure magnetite.

    Anyone knows an easy small-quantity source for microwave ferrites,
    by the way?

    Regards,
    Mikko
     
  11. Robert Macy

    Robert Macy Guest

    I read a paper where someone 'measured' the precession rate of the
    'available magnetic molecules and it was in the range of 1 to 2 GHz,
    which the author used to explain why there are no microwave ferrites.

    In my experience, 1 GHz was about the upper limit, air core above that
    was far more effective, and for EMI graphite..
     
  12. Hi Robert, thanks for your insight. Do you happen to remember
    a reference to the paper?

    To me it looks that Mini-Circuits transformers do use magnetic
    cores,
    see e.g.
    http://217.34.103.131/pdfs/TC1.5-1X+.pdf
    http://217.34.103.131/pdfs/TCM2-33X+.pdf
    http://217.34.103.131/pdfs/TC4-14+.pdf

    There are also all sorts of tuneable microvawe thingies which make
    use of YIG garnets.

    Then there are circulators available for a wide variety of
    frequencies,
    those must utilize ferrite cores, no?

    Regards,
    Mikko
     
  13. Minicircuits makes both TLT:s and ordinary ones. Eg.
    http://217.34.103.131/pdfs/TC1.5-1X+.pdf is an ordinary transformer
    (judging by the 'Config D' in the data sheet), and the in the
    photograph
    it looks to me like having a magnetic core.

    An example of their TLT's is
    http://217.34.103.131/pdfs/TC4-25+.pdf
    which is clearly indicated as 'Config. H' in the datasheet.

    There are a large variety of companies like
    http://www.temex-ceramics.com/site/en/microwave-ferrites-cermatmenu-29.html
    but I still haven't found one with a small-quantity webstore. My first
    guess was RELL but I wasn't able to find the stuff there.

    Regards,
    Mikko
     
  14. legg

    legg Guest

    Adrian - great to hear something from the horses mouth for a change.
    Copied to my magnetics article index, with your permission, I hope.

    RL
     
  15. Yes, I have made superconducting TLTs in the past using the now-
    obsolete
    Amidon #6 (yellow-grey) and I'd expect their other iron powder cores
    to work, too.
    But their ferrites freeze, I've tried. Coincidentally, your colleague
    R. Koch has
    also constructed superconducting TLTs, but I don't know which core
    material he
    used. Amidon, I would guess.

    But as you say, something else is needed for higher frequencies.

    Must go now, the fridge just reached the 8mK base ...

    Regards,
    Mikko
     
  16. About what temperature do they work down to?

    Best regards,
    Spehro Pefhany
     
  17. whit3rd

    whit3rd Guest

    Well, if you have a square-section bead... make a jig with a central pin (to fit
    the central hole of the toroid) and four pogo pins (spring-loaded
    contacts) that engage the flat face of the bead at 0/90/180/270 degrees.
    Inject current at 0 and 90, sense voltage at 180 and 270.

    It's not the usual four-point measurement (for sheets, there's a square-array
    technique that's similar) but it should work well enough to track any
    change in properties.
     
  18. I have used them in LHe only.

    Regards,
    Mikko
     
  19. Thanks for the offer. However, I dug out their paper and the
    material *is*
    mentioned there. They used T30-6 carbonyl-iron cores from Micrometals
    Inc.
    Actually, this sounds suspiciously similar to Amidon #6, I wonder if
    there
    is an universal numbering system for core materials? For instance
    Amidon
    #43 ferrite looks quite similar to Fair-Rite #43 (from memory, the
    Fair-Rite web
    site is down so I cannot check).

    I mis-rememberd the # of the yellow-gray material, by the way, the
    now-
    obsolete stuff I've used was #4, not #6.

    Regards,
    Mikko
     
  20. Thank you, There's nothing wrong with old knowledge.

    George H.
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-