Looking for manufacturer of huge inductor

[GK]

It was a very similar application, resonant booster, a couple kHz, several hundred watts. I can't remember exactly what the core design was, but it was steel and it wasn't a toroid. There may be some crossover frequency where ferrite is better, but I remember everything we did with ferrite wasn't as efficient. We also wound about 1 toroid and said the heck with that! The advantage of working with a winding house is they'll typically have a large stock of different cores, and will do all the calculations with a very high degree of certainty, assuming you can describe the exact operating conditions. They know their materials and do it every day. That part about knowing your operating conditions is critical.

From this threads opening post:

"Because of low frequency (30-60-ish kHz)....."

You're out with your suggestion by over an order of magnitude.

[dave_j_fan]

want to know
the difference in performance of ferrite EE core in a nice machine winded case as compared to
handwided with turns here and there .Consider frequency 30Khz

[Bud]

Coilcraft makes big current inductors.

[David Hess]

If I am not sure where I am going to end up during developement, I find the largest E or pot core which will fit in the space required, fill the winding window with copper, and then adjust the gap for either the desired inductance or more likely to a point where the copper and core losses are balanced.

To support a given L-I-squared product, for a core of sufficient size, there is a minimum gap size and associated A_L value. The latter determines the number of turns required for the desired inductance value. You can't just fill the window with a (presumably arbitrary) number of turns and then gap for the desired inductance. If you wind too few turns you'll then have to gap for an A_L too high to support the LI² product and if you wind too many turns you'll have to gap for an A_L value below the manufacturers recommended minimum for that particular core, which negatively effects efficiency in other ways.

I think you missed the gist of what I was getting at but my description was brief and I do not think we disagree.  There is an assumption that filling the winding window results in more turns than necessary maximizing the inductance for a given gap.  If that does not allow the flux to be lowered enough to stay within the manufacturer's recommendations at the frequency of operation, then the core is too small unless smaller wire is used to get more turns which raises the copper losses even more.

In practice during development with a larger than necessary core, the flux is lower than the manufacturer's recommendations so core losses are lower than they would be with a properly sized core and wire losses are usually higher although not always; you can always make a multifilar winding or use larger wire if the core was too large to start with.  An optimized design would use a smaller core with a minimum of wire and be more economical.

I just do this during prototyping when I am unlikely to have an optimally sized core available.

Also, could you point to a readily available pot core large enough to support the OP's predicted LI² range?

A pot core?  No, although I am sure they exist.  I have a whole pile of big 3C8 E cores (and smaller pot cores) which probably meet the required frequency and power requirements but they are old (is 3C8 material even made anymore?) and I have no idea where to buy more simply because I have not needed to in 10+ years.

I have always found purchasing bare cores and bobbins to be a frustrating exercise.

[GK]

If I am not sure where I am going to end up during developement, I find the largest E or pot core which will fit in the space required, fill the winding window with copper, and then adjust the gap for either the desired inductance or more likely to a point where the copper and core losses are balanced.

To support a given L-I-squared product, for a core of sufficient size, there is a minimum gap size and associated A_L value. The latter determines the number of turns required for the desired inductance value. You can't just fill the window with a (presumably arbitrary) number of turns and then gap for the desired inductance. If you wind too few turns you'll then have to gap for an A_L too high to support the LI² product and if you wind too many turns you'll have to gap for an A_L value below the manufacturers recommended minimum for that particular core, which negatively effects efficiency in other ways.

I think you missed the gist of what I was getting at but my description was brief and I do not think we disagree.

No. Your post suggested that so long as you pick an oversized core you are safe to wind an arbitrary number of turns and then gap for the desired inductance. I was just pointing out that, without any baseline calculations, this isn't necessarily true.

I think that you are making what is a basic task of engineering by numbers out to be some kind of protracted empirical exercise.   

To minimise core losses you only gap a suitable core down to the A_L value required to support your LI² product and no further. It is then that you compute the necessary number of turns for the desired inductance. Once you know the number of turns required you select the wire gauge so that the winding window is utilised. If at that point your copper losses turn out to be too high you simply move on to a larger core and do the calculations again.

You start by establishing your flux density and the required gap, not your number of turns.

In this context I don't really follow your advice for balancing core and copper losses. Suppose that I have just designed my inductor. I have gapped down to an A_L value to support my computed LI² product.
I cannot make the gap any smaller as then my inductor will not maintain the desired inductance value at the required DC current. So my only option for "tweaking" is to make the gap bigger. If I make the gap bigger my core losses increase. However a larger gap means a reduced A_L value, which means I will require more turns of a lower gauge wire to both achieve the desired value of inductance and utilise the winding window. That means I get increased copper losses along with increased core losses.

[T3sl4co1l]

That's closer to the essence of it, but: since inductance is derived from the gap (independent of the core size and material*), whereas flux density is dependent upon the core size and number of turns (independent of gap), you want to calculate turns first, and then the gap.

*So long as the core itself has permeability much higher than the average / equivalent permeability in the final design.  Or, to put it more generally: the gap is what's required after subtracting the core's contribution, which is generally small (and obviously, not variable ).

A typical design process for me: I need an inductor of some capacity (inductance and peak current), and optionally, some amount of loss.  I pick through cores that seem likely, and calculate N = Vpk / (4*Bmax*Ae*F) (for a square wave, or for a sine of V = Vrms, the coefficient is 4.44 instead of 4).  Bmax is determined either from Bsat (~0.3T for most ferrite) or from the loss budget (p_c = P / v_e; requires core v_e; look up B on the loss vs. B, F curves -- not possible if absent).  Finally, the gap is given by l_g = mu_0 * Ae * N^2 / L.  (Or more precisely, that's the total air gap equivalent; to get the physical gap, subtract the core's air equivalent, l_e / mu_r.)

Tim

[GK]

That's closer to the essence of it, but: since inductance is derived from the gap (independent of the core size and material*), whereas flux density is dependent upon the core size and number of turns (independent of gap), you want to calculate turns first, and then the gap.

 

I should have written "LI² product" rather than "flux density". In any case it makes no sense to start by calculating the number of turns. You can't calculate the number of turns unless you know the freaking AL value and the AL value is determined by the gap size. Did ANYONE look at the AL vs LI² core selection charts that I posted earlier in this thread and bother to comprehend them?

You start by determining your AL value based on the LI² product and you gap the core for that AL value. THEN you are free to compute your number of turns.

[T3sl4co1l]

It is more sloppy when working with powder cores, yes.  In which case, you're about as well off looking at empirical graphs.  On the upside, you rarely deal with powder core shapes, and therefore don't have to worry about gaps; in that case, you have no choice but to use the core's characteristics alone to store energy.

You can find gap by energy, too: the energy density of a magnetic field is simply B^2 / (2*mu_0), so you can calculate the gap length if you know the gap area (~= Ae).  Again, subtract the core's equivalent air gap length to find the physical gap.

You didn't seem to have a problem with my equation which calculates N without needing AL, though.

Ed: I could also be more explicit by noting: IF you have a gap at all.  That's your one degree of freedom between flux density and field strength.  If that factor is strictly fixed by your choice of core (e.g., a powder toroid), you can only make your choice through core selection and turns.

Tim

[GK]

We were specifically discussing user-gapped cores. This place is just as bad as an audio forum!

[T3sl4co1l]

This place is just as bad as an audio forum!

That was uncalled for.

But I guess if you're not interested in discussing it rationally, that makes you no better either.  Ooooh! 

Tim

[GK]

This place is just as bad as an audio forum!

That was uncalled for.

But I guess if you're not interested in discussing it rationally, that makes you no better either.  Ooooh! 

Tim

Now you just trolling and giving evidence to my assertion. I think the points (to which your replies have simply been tangential) I have made about designing DC inductors with gapped cores are clear and rational. In fact I haven't deviated from basic text book stuff.

[johansen]

I think you can manage three three ETD59 cores side by side.. but preferably, find a core with a square cross section for the center core.

just two of them is probably sufficient:
basing design off "150-200uH rated at 30A"

for two cores side by side and a custom bobbin.. or none at all..
a 1.5mm physical air gap
24 turns for 150uH inductance.
.3T flux =68mJoules energy stored
80% copper fill factor makes 8.7 watts resistance loss
--if the ferrite saturates at .4T then you can take it all the way to 40 amps, which is 121 milijoules.

increasing the air gap to 2mm physical gap makes about
16 watts copper loss@ 32 turns for 200uH @ 30 amps makes .3T
40 amps makes .4T at 27 watts copper loss at 80% fill factor.

[GK]

For anyone out there genuinely curious about designing gapped DC inductors along the lines I have detailed in this thread, here is a link to the application note the AL/LI2 charts were taken from:

http://www.google.com.au/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&cad=rja&uact=8&ved=0CCMQFjAB&url=http%3A%2F%2Fwww.mag-inc.com%2FFile%2520Library%2FProducts%2FFerrite%2FPowerDesign.pdf&ei=cshaVJffJYOwmAWD04HgBw&usg=AFQjCNHMwjsI8gqj_10odVfGPG5dIidKww


Go straight to page 13 "Inductor design" for the technical description for core selection and gapping.   

[David Hess]

...

If I make the gap bigger my core losses increase.

...

Could you explain this statement in more detail?

[GK]

...

If I make the gap bigger my core losses increase.

...

Could you explain this statement in more detail?

Well I don't have any of my textbooks handy right now, but:

"Why are actual core losses larger than calculated?

When calculating the core losses, it is assumed that the structure is homogeneous. In reality, when core halves are mated, there is leakage flux (fringing flux) at the mating surfaces, and the gap losses contribute to the total losses. Gap losses are caused by flux concentration in the core and eddy currents generated in the windings. When a core is gapped, this gap loss can drastically increase overall losses. Additionally, because the cross-sectional area of many core geometries is not uniform, local “hot spots” can develop at points of minimum cross section. This creates localized areas of increased flux density, resulting in higher losses at those points."

From here:

http://www.mag-inc.com/products/ferrite-cores/learn-more-about-ferrites

More gap more fringing flux. The fringing flux can also dramatically increase electromagnetic emissions, BTW.
 

[Smokey]

Magnetics are the real deal.
http://www.mag-inc.com

Their inductor sizing software saves a ton of time.  Of course since high flux core are almost as much black magic as RF, they could just be making it up and I'd have to believe them. 

Really though, they make some killer inductors and materials.

[T3sl4co1l]

Hmmm, gapping should increase losses in local spots (assuming constant V, F applied to the same winding), due to fringing causing partial saturation around the edges of the gaps.  But, because reactive power is higher (proportionally, whereas the core losses are only around the fringing -- a low order effect), the overall Q factor is higher.

Often more significant is the strongly divergent field around the gap, which totally cooks any wires near there.

So, both Q and reactive power go up, but reactive power goes up faster, so you can indeed cook things if you aren't careful!

Tim

[GK]

Ferrite core performance degrades with increasing temperature. A flux density concentrated around the gaps will contribute to core heating and increased overall core losses before the inductor design gets that bad that the wires actually melt.

[megajocke]

It sounds a little like it's not fully clear what parameters are held constant and which are varied in this discussion. 

For a fixed core size and flux density but a variable gap, the largest gap length (lowest Al value) gives the highest energy storage capability at saturation. However, for a fixed inductance a larger gap would require more turns, which means you'll have higher copper losses at a certain current.

For this reason, there is a limit to the longest useful gap, although if the peak current is much higher than the RMS current, things will be different. Put in a huge gap and a certain core will store lots of energy, but it won't work for continuous duty.

So you'd usually not want to gap the core more than neccessary because copper loss would be higher if you did. In the case of the LI² - Al curves, the manufacturer has presented this information as the Al you should use for a certain energy storage to get the lowest copper loss.

If you don't have those curves and want to determine what is needed you could calculate it from scratch. I'd start with determining the amount of linked flux (flux times turns, often denoted by capital greek Psi) which needs to be supported, and this is given by inductance times desired saturation current:

Psi = N * Phi = L * I

where

Phi = B * A (flux equals flux density times cross sectional area)
 
The minimum number of turns for a certain saturation current on a fixed core, fixed saturation flux density, fixed inductance and variable gap length (or variable Al value equivalently) can then be calculated. Yes, it is actually a minimum number when parameters are fixed and varied in this manner. This can be seen by rearranging the equation above:

B = Phi / Ae = (L * I) / (N * A)

As L, I and A are fixed, B increases with decreased N so you'd want to make sure that

N > (L * I) / (Amin * Bsat)

Actually, given that everything is given except the number of turns and the gap length, specifying the minimum number of turns at this point is the same as specifying the maximum allowable Al value. (minimum required gap length)

Now you can determine if the turns will fit given a realistic current density, or alternatively make them fit by choosing thin enough wire and see if it gets too hot. (either by experiment or calculation) This is the lower Al limit (maximum gap limit) in the energy curves.

If you don't require an optimal design, the previously presented method of winding as many turns as you can on a variable-gap core works to get maximum inductance for a certain current by selecting the wire to give allowable temperature rise (or current density by rule of thumb, for example).

If you have considerable AC losses things get more difficult of course, but the bulk core losses will be lower for a fixed core size and inductance if you choose a large gap and a large number of turns. The same AC voltage over higher number of turns gives a lower AC flux density.

One thing that can be useful to keep in mind I think is that if you have a fixed AC voltage then core loss is only dependent on the number of turns for a fixed core and not the inductance and gap length, at least as a first order approximation. Increasing the number of turns decreases the core loss in this case, typically at the expense of increased copper loss. This also applies to cores where you can't change the gap, including powder toroids. Sometimes optimizing inductor losses and size can be more important than having a certain inductance.

Copper losses (eddy currents) from the gap fringing field I find more difficult to make general statements about as all dependencies are interlinked and scaling exponents are unintuitive. It also makes all the difference exactly which parameters you hold constant and which you vary. These losses can sadly be significant in many cases, especially if there are many layers of thick wire...

[GK]

For a fixed core size and flux density but a variable gap, the largest gap length (lowest Al value) gives the highest energy storage capability at saturation. However, for a fixed inductance a larger gap would require more turns, which means you'll have higher copper losses at a certain current.

That's pretty much what I said in reply #34.

Copper losses (eddy currents) from the gap fringing field I find more difficult to make general statements about as all dependencies are interlinked and scaling exponents are unintuitive. It also makes all the difference exactly which parameters you hold constant and which you vary. These losses can sadly be significant in many cases, especially if there are many layers of thick wire...

Well, I've always approached DC-carrying inductor design with the idea that  core gaping is a necessary evil to prevent saturation. If you are concerned with efficiency, then you only gap the necessary amount and no more. The suggestion that (for a desired inductance value and DC current) the core gap (and presumably the number of turns) can be adjusted for a happy or optimal medium between "core" and copper losses doesn't make any general sense to me.

[T3sl4co1l]

We're all talking about the same thing here, just from different directions (so, I don't get the "audio forum" epithet).  The graphs are everything tabulated against core type, so you don't have to run through iterations manually.  Energy is stored in the gap, but you're limited by how much you can supply from copper wire, under limits of power dissipation, or Q for a given rating or physical size.  More turns than necessary for desired core performance will only increase copper losses (which requires more gap for the same inductance).

Which reminds me, I want to crank some numbers and find out where the ideal permeability comes from.  Something to do with copper resistivity and core geometry (you'll need some rough inputs on Ae/l_e), and it seems to lie in the 10-60 range (which is what most powdered iron materials come to).

Tim

[johansen]

We're all talking about the same thing here, just from different directions (so, I don't get the "audio forum" epithet).  The graphs are everything tabulated against core type, so you don't have to run through iterations manually.  Energy is stored in the gap, but you're limited by how much you can supply from copper wire, under limits of power dissipation, or Q for a given rating or physical size.  More turns than necessary for desired core performance will only increase copper losses (which requires more gap for the same inductance).

Which reminds me, I want to crank some numbers and find out where the ideal permeability comes from.  Something to do with copper resistivity and core geometry (you'll need some rough inputs on Ae/l_e), and it seems to lie in the 10-60 range (which is what most powdered iron materials come to).

Tim

my spreadsheet might help.
http://johansense.com/bulk/spreadsheets/chokecalculator.ods

ideal permeability isn't really a thing... without lots of other somewhat arbitrary limitations.
for example, the larger the core, the higher the eddy current.. there is an old as dirt application note i can't recall right now, you've probably read it, and i've seen it posted in this forum i believe, that covers the maximum Q as a function of frequency (assuming sufficiently ideal litz wire)--basically the larger the core, the lower the frequency for maximum Q--this is primarily a function of eddy current partly follows core dimensions squared, and skin effect as a function of frequency sets the core utilization.. as of yet, i'm not aware of laminated powdered iron cores... but you can stack smaller cores, the geometry just becomes not very ideal..

but a stack of microwave oven transformer cores 2 feet long still gets you better power density Watts/kilogram, than a stack 1 foot long.. but only by about 3%.
but the same kilograms of iron as a toroidal core, filled with its weight in copper, would deliver probably deliver 2-3 times as much power.. while also being intrinsically more efficient due to fact that steel is better when grain aligned.

given that neigher ferrite nor powered iron suffer from manufacturing limitations.. it would not surprise me to find out that the ideal inductor topology is the pot core.. assuming that copper costs more than powered iron or ferrite.. which doesn't seem to be the case.
when copper is cheap compared to ferrite the ideal core is probably a donut shaped toroid..
btw, i have not yet begun to calculate the gain from O shaped toroid cores versus square ones, there must be some improvement you'd think, but my initial work has shown that the larger the core the better, because power density scales faster than volume.

given the assumption that best efficiency is when eddy and hysteresis core loss equals copper loss, then the ideal permeability can be calculated rather simply.. however it is still frequency dependent.

if the eddy current loss is also a function of permeability, then of course the slope of that curve dictates the ideal copper loss...

[T3sl4co1l]

Not sure, I've seen Micrometals' curves of Q versus core selection and windings though.

Skin effect is funny, in that it affects anything with loss, not just if it's bulk conductive.  So you still get the same effect in ferrite and powdered iron; the limitation is around an inch or so width at 100kHz.  This limits the construction of high power transformers (which are built from ferrite bricks instead) and high frequency structures.

I was reading a paper on loaded transmission lines recently; they calculated the dimensions required of ferrite and dielectric for a given loading (impedance) and frequency response.  The ideal coax structure is an alternating stack of dielectric and ferrite washers: the thin widths and gaps between ferrites allow magnetic field to penetrate the ferrite, while the dielectric has some loading effect itself and should be chosen for a low dielectric constant.

Another data point from experience: if you observe the impedance vs. frequency plots of various ferrite products, in particular ferrite beads and chips, you notice something interesting: the curves rarely bear any relation to the published material property.  There's a geometric aspect due to the fields propagating through the ferrite (which, because it has high permeability, occurs at a sizable fraction of the speed of light), and the radial and longitudinal standing waves contribute to peaks and dips in the impedance curve.

Tim

[megajocke]

Well, I've always approached DC-carrying inductor design with the idea that  core gaping is a necessary evil to prevent saturation. If you are concerned with efficiency, then you only gap the necessary amount and no more. The suggestion that (for a desired inductance value and DC current) the core gap (and presumably the number of turns) can be adjusted for a happy or optimal medium between "core" and copper losses doesn't make any general sense to me.

Yes, if the current is pure DC enough, core losses will be negligible even in a minimum gap/minimum turns design like you say. In that case minimum gap is certainly the configuration which will give minimum total loss. Many authors call this a "saturation limited desgin" or something to that matter. If you have for example +-20% of ripple current and design for a peak (AC+DC) flux density of 300 mT you'll only have +-50 mT of AC flux density. (300 mT / 1.20 gives 250 mT of average DC flux density) Even the crappiest power ferrites have quite low loss up to 200 kHz or so for this ripple level.

With high enough AC current amplitude and frequency however, core losses would dominate over winding losses if the core is used all the way to its saturation flux density. In this case it is likely possible to achieve lower total loss by increasing the gap length and number of turns (keeping the same core and inductance) as this would decrease the AC flux density. This is likely to be needed in for example converters operating in discontinuous mode from 100 kHz or so and up (for ferrite cores) and the resonant inductors in quasi-resonant converters.

A bit off topic-maybe, but toroidal mains transformers using grain oriented steel also tend to fall into the saturation-limited category. Even if driven very close to saturation the frequency is so low that core losses are still almost negligible. Here you would also go for the least amount of turns that won't saturate at the expected highest input voltage and lowest frequency.

[johansen]

Quote from: GK on November 08, 2014, 11:49:02 AM

    Well, I've always approached DC-carrying inductor design with the idea that  core gaping is a necessary evil to prevent saturation. If you are concerned with efficiency, then you only gap the necessary amount and no more. The suggestion that (for a desired inductance value and DC current) the core gap (and presumably the number of turns) can be adjusted for a happy or optimal medium between "core" and copper losses doesn't make any general sense to me.

this is the commonly taught and fundamentaly backwards way of looking at the matter.

given that core losses follow flux squared and frequency squared, inductor energy storage per kilogram of inductor must decrease with frequency to keep the same Q
also, energy storage follows flux density squared.

the gap length, or alternatively, the core permeability, sets the copper losses required to achieve a certain flux density.

you can later come to the conclusion that energy storage is proportional to the inverse of permeability keeping flux constant.
a common mode choke core with a permeability (datasheet says it, it must be true!) of 10,000! can only store litterally something like single digit microjoules per cubic cc of core before it is saturated.

regardless of core material, energy storage is proportional to the square root of copper losses keeping flux density constant.
likewise copper loss squared equals energy storage when keeping flux constant