Gapped potcore/ E core AsubLs

While looking through the Ferroxcube Ferrites databook, I see a specific size core with gaps for several different materials. These materials have a nongapped AsubL ranging from 4600 to 6950. If the gap in all the materials is 270um the AsubL drops to 1000. With the same size gap, even though they start with different AsubLs, the now all have the same AsubL. Example: 3F3 non-gapped AsubL=4600 with a 240um gap AsubL = 1000.

3C81 non-gapped AsubL=6950 with a 240um gap AsubL = 1000. Why doesn't gapped AsubL change in proportion with the original AsubL?

Mikek

Reply to
amdx
Loading thread data ...

The permeability of the core material is high compared to the air gap, so the magnetic field is almost all in the air gap. Thus, the gap almost fully determines the AsubL.

--
mikko
Reply to
Mikko OH2HVJ

Equivalent permeability goes as: mu_eq = l_e / [l_g + (l_e / mu_r)]

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Design 
 Click to see the full signature
Reply to
Tim Williams

It does, but not "in proportion". It's like resistances in series.

Somebody is being lazy.

Reply to
jlarkin

No, not being lazy, just no math skills and trying to understand ferrites.

My real problem at the moment is trying to identify the material of an E core I have with a 0.225" gap. Even with that huge gap, it has an AsubL of 3500. Just now realized that is so high, maybe I did something wrong to get that number. (all four times!)

With 337 turns inductance equals 401mH @10kHz AL=3530

With 387 turns inductance equals 528mH @10kHz AL=3525

With 724 turns inductance equals 1819mH @10kHz AL=3470

With 100 turns inductance equals 29mH @10kHz AL=2900

See any error with my AsubL calculation. Seems huge with a large gap. Am I off by a factor of 100?

The point where U' and u" cross is about 365kHz.

E core size is 55.2/27.6/20.7 (hmm wanted post a dropbox image of core and dimensions, but copy link won't highlight)

Mikek

Reply to
amdx

I think I just solved my dilemma. I have a large E core with a wide gap, I have two windings, put on the oute r legs, one winding is 401mH, other 528mH, series aiding 1819mH. My thinking is in series aiding the magnetic flux is circulating mostly in the outer legs, the gap is just like a high value parallel resistor adding the the flux, but not much. To bring the gap into the picture I need to con nect the two coils series opposing. Mikek

Reply to
amdx62

I didn't mean you, but the Ferroxcube people.

Reply to
John Larkin

I have a large E core with a wide gap, I have two windings, put on the outer legs, one winding is 401mH, other 528mH, series aiding 1819mH. My thinking is in series aiding the magnetic flux is circulating mostly in the outer legs, the gap is just like a high value parallel resistor adding the the flux, but not much. To bring the gap into the picture I need to connect the two coils series opposing.

Oh yeah, for the two windings outside and the gap in the middle, that's quite correct -- note that you need to use a different l_e to calculate mu_r, since the path is all the way around, not through the middle.

But that is a measurement you can use now.

And to find the core material cutoff frequency, you can plot Q(F), but mind you'll probably need to do so with many fewer turns to get meaningful results.

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Design 
 Click to see the full signature
Reply to
Tim Williams

Ah, not sure that's true, I might not have presented what you aren't seeing. First, I don't know who made my core, (junkbox find). Second, it had a strange gap, it was only cut through about 80%,

20% of the center leg was not gapped. Until I chipped it, then I ground it all out. I don't know if I have mentioned it here, in this large gap I'm sliding a toroid in as far as I can, with enough turns to make it a 120uh. Then applying DC to the E core to saturate the toroid. So far all is working well, except. I want a high frequency material for my toroid, (will order) and I don't know what my E core is, so I can't order until I figure it out. The one characteristic I think I gave a handle on is where u' and u" cross on the permeability graph, My E core is about 365kHz. I'm happy with this material because it does not seem to have any remanence. My first core did. (3 iron washers with a slot cut) Mikek
Reply to
amdx

Term definition verification. l_e is that effective length. Mu_r is that relative permeability u'

By cutoff frequency do you mean where u' and u" cross? If yes, I got about 365kHz. using 6 turns. I'm not sure what Q(F) is. Would that be plotting core Q and looking for a matching permeability graphs?

Thanks, Mikek

Reply to
amdx62

For l_e would I use the length around the core but centered inside the core. Should I just ignore the center leg for this measurement.

FWIW. I put the 5000 permeability toroid in the gap, it's tight. Then I measured the inductance to calculate the AsubL, I did this with the coils connected series opposing so the center leg is being used. I got the same AsubL as all my other measurements about 3500. Making me think I could use the published l_e 0f 124mm.

Thanks for your help, Mikek

Reply to
amdx

Yes. Close. mu_r is |u|. Use the vector sum, technically. Since air has mu'' = 0.

Yes, "Q as a function of frequency". Q = 1 being u' = u''.

Yes. l_e appears to be pretty close to midline in the core, maybe rounding off the corners a little bit.

Same thing with v_e, in case you're wondering. Typically very close to A_e

  • l_e. There could be dead weight on a core, I guess, but it also won't be active volume, so it makes sense to be nearby. (You probably won't need to know this until you're calculating total core losses.)

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Design 
 Click to see the full signature
Reply to
Tim Williams

Ok, I'm satisfied I have those right.

Google has not been my friend for this.

No luck finding Q(F)

Reply to
amdx

Not lazy, when one of them has a 10^-7 exponent, to dominate in the numerator of a fraction.

RL

Reply to
legg

If connected in series, the center core leg will be invisible.

If connected in parallel, it's a mess.

Operating independantly, it's a can of worms.

RL

Reply to
legg

I still don't understand where they were lazy. But, there are a lot of things I'd like to see on the Ferrite datasheet that isn't there. Lately it is a remanence number so I could pick a material on the low end of the scale. One of the calcs I did with the 10^-7 also had a 10^6 in it so I ended up with 1.256 as a result, in the denominator. If I was 14 again, I would start a study of math and physics/electronics and practice math some more. I'm 50 years late in knowing what I want to be when I grow up. Mikek

Reply to
amdx

It will be connected series opposing with equal turns on each leg. This puts maximum flux on center leg.

Not sure why if equal turns and opposing. Might look into this, lowers operating voltage by 1/2.

No need to explore that.

Mikek

Reply to
amdx

Mikek

With the same number of turns, you should read ~ same L value.

Series opposing should indicate a leakage inductance.

RL

Reply to
legg

Agreed, I don't know how I got that wrong by 50 turns, I'm setting up a counter for the next time. I have a couple of nice Veeder-Root counters.

I have no knowledge to support that.

I have about 100 turns on each leg. L1 = 29.47mH L2 = 29.47mH L1//L2 = 28.69mH Parallel L1\\L2 = 745.9uH Microhenries antiparallel L1+L2 = 114.7mH114.7mH Series L1-L2 = 2.982mH Anti-series

Large gap in center leg of E core.

Mikek

Reply to
amdx

Try measuring the inductance of one, with the other shorted and compare this to your list of measurements. It will closely resemble one of them.

RL

Reply to
legg

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.