Exact spacers for core gapping?

[cupoftea]

Hi,
We are doing an 2kW LLC converter using PM6249 transformer core from Epcos. We are using an external resonant inductor (and additional external magnetising inductor). We will interleave wind the transformer. However, we will still get some leakage inductance as you know. We don’t want any leakage L as it will add to the external resonant inductor and reduce the upper resonant frequency. So, We have to put up with this, but would therefore like to know exactly how much leakage inductance we are going to get. As such we wish to gap the core (every one), to exactly the same amount…..then we will get the same leakage L from each one, and can then deal with it.

Are there any exact-thickness spacers available for core gapping? We need approx. 0.3mm, and exactly same each time…….even if its 0.32mm etc etc.
We don’t want to machine gap the cores as custom gapping costs a fortune, and adds to lead time.

PM6249

https://www.tdk-electronics.tdk.com/inf/80/db/fer/pm_62_49.pdf

 

[KlausST]

Hi,

don´t know if this is possible: Maybe use a plastics foil, maybe onesided or twosided selfadhesive.
Foil made of various materials should be available in various thickness.

Klaus

 

[cupoftea]

Thanks, thats a good idea, must admit i'd like it to be like those spark plug gap things, but plastic as you say, very precise thickness.

 

[Magnethicc]

As far as I know, gapping does not effect much the leakage inductance. Anyone else care to comment on that?

 

[FvM]

As far as I know, gapping does not effect much the leakage inductance. Anyone else care to comment on that?

You are right.

 

[byjuncal]

As far as I know, gapping does not effect much the leakage inductance. Anyone else care to comment on that?

Umm, I don't completely agree with it. Leakage is directly related with the coupling between windings but there are other effects that modulate the leakage two. When you add a gap between the cores, the flux finds a easier way to go out of the ferrite when it is crossing the gap. In other words,, the effective permeability of the overall core has been reduced and there will be more flux going out.
The effect tis reduced if the gap is only in the center leg.

 

[cupoftea]

Thanks,
this thread discussed it previously

What is the relationship between leakage inductance and core gap size?

Hello, Supposing I have an ETD39, N87, Epcos Ferrite transformer. Ns = 30 Np = 30 I put the ungapped ETD39 core halves into the bobbin and measure the leakage inductance. Now I do the same again, but this time with a 0.2mm centre leg gapped ETD39 core. …Which of the above two leakage...

www.edaboard.com

...but it just seems unbelieveable that an ungapped core with Lp=2mH could be interleave wound and then have a leakage inductance of just 4uH....i usually kind of think "Lp/100" is the figure for leakage inductance.....wrong though i probably am.

--- Updated Apr 1, 2022 ---

Here is the actual problem...

Supposing you have an ungapped PM62/49 core. Its “sandwich” wound Np1:Ns1:Ns2:Np2
Where Np1 = 11 turns and Np2 = 10 turns are in series. Ns1 and Ns2 are separate coils of 9 turns each.
What is the leakage inductance measured at the 21 turn primary with both secondaries shorted?
Each layer of the 4 layers of turns span the entire bobbin length.

--- Updated Apr 1, 2022 ---

PM 62/49

https://www.tdk-electronics.tdk.com/inf/80/db/fer/pm_62_49.pdf

 

[FvM]

Thanks for reminding the previous thread. Modifying my transformer example to a interleaved windings configuration confirms the claim in post #5 and #6 "gapping does not effect much the leakage inductance". It's particularly true for interleaved windings.

I'm copying the old core geometry, but changing 4x17 turns primary+secondary to 2x17p / 4x17s /2x17p. Between ungapped and gapped (0.3 mm inner+outer) core, the about 43 µH leakage inductance varies only by a few 100 nH. Field lines are shown for shown for the load current generated field.

--- Updated Apr 2, 2022 ---

What is the leakage inductance measured at the 21 turn primary with both secondaries shorted?
Each layer of the 4 layers of turns span the entire bobbin length.

2.1 µH with 2 mm wire, 2.4 µH with 1.5 mm.

 

[byjuncal]

Hi FvM, maybe I did not explained properly,
You are right the gap change itselft does not change the leakage directly. What it is clearly changed is the ratio between Llk and Lm.

​	​	​	​	​

I generated this table some time ago running FEM simulations. The only change between simulations is the gap. As you can see, the coupling coefficient goes down when the gap is increased. You can see the huge effect on the ratio Lm/Lr

 

[cupoftea]

2.1 µH with 2 mm wire, 2.4 µH with 1.5 mm.

Thankyou very much, may i ask is that at 100kHz?

 

[FvM]

Counter question, how's the winding made? Litze wire? Generally speaking, leakage inductance isn't much frequency dependent.

 

[cupoftea]

Thanks, yes its Litz wire

 

[cupoftea]

Also, is it the case that whether the primary "sandwiching" is via two outer "bread" layers in series, or two outer "bread" layers in parallel....it makes no difference to the magnitude of the leakage inductance?

 

[FvM]

Outer layer winding is longer and has respectively higher leakage inductance than inner layer. Series circuit has slightly higher effective leakage than parallel circuit.