Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

SMPS transformer air gap

Status
Not open for further replies.

peterthegrate

Newbie level 2
Joined
Oct 18, 2018
Messages
2
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
37

Enlightenment requestedI've designed many switch-mode power supply transformers in the past and have not questioned the presumption that the ferrite core air gap is best located on the core centre limb, because that's how they are supplied. But I'm working on a new project where extreme high efficiency is a principle objective. On taking measurements on prototype transformers, I've found that the transformer loss is lower when the gap is distributed across all three E-core limbs - ie using a spacer instead of a centre limb only gap. This has been a consistent finding on a number of samples (gapped for same inductance in all cases). Google hasn't found much in the way of confirmation of this, although there are a few hints that this could be the case. The primary winding is full bobbin width foil with secondaries either side of standard magnet wire (high turns ratio). So I'm wondering if anyone has experienced the same or has any observations or comments that may shed light?
Thanks in anticipation . . .
 

The greater the aspect ratio between the cross-sectional area of the core and the height of the gap, the smaller the magnetic field dispersion. However, by spreading the gap at several points, especially at the periphery of the ferrite, this tends to create more 'sources' of EMI, which may now be closer to neighboring components.
 

Thanks for responding, I understand the EMI issue & have an overall outer copper foil wrap which does not appear to have a significant effect on efficiency. The transformer is also located within its own screening can. The assembly has been tested & is EMC compliant.

I'm not sure of exactly what you mean by 'magnetic field dispersion' & why that would affect the losses? Are you suggesting that the different field distribution would somehow change the losses due to induced currents in the windings; similar to the proximity effect, or maybe some localised loss in the ferrite? The losses due to leakage inductance, which is very small anyway, would not account for difference between the two gapping methods.
FEA software would be useful for this, but too expensive for me!
 

By 'magnetic field dispersion' I mean the geometry of the magnetic field lines not totally confined within the gap. At first a gap should affect the losses little if it is designed to only slightly increase the energy stored in the transformer, but depending on the case, if a significant part of the energy of the magnetic field is stored in the 'air', and moving away the surfaces of the core , the edges of the ferrite will radiate beyond the small volume of the gap.
 

the answer here is that for a winding wound close to the centre gap, having only a centre gap creates more eddy current loss in the wire or foil due to the presence of mag field fringing out from the gap and cutting the wires, inducing current

Thus spreading the gap reduces the effect at the centre limb - this is well known to designers of HF, 400kHz chokes for resonant converters - and others ...
 

Enlightenment requested

Consider a EE core; the centre leg is fat because it has to carry twice the flux (both the outer limbs taken together). Air gap has a high reluctance and this tends to reduce the saturation point. Now If the centre leg has twice the cross section (of the each of the side limbs taken individually), So if you reduce the centre gap by half, you must put equal gap on each of the side limbs to keep the total reluctance the same.

peterthegrate is right in this count.

This is a first order approximation because the flux lines bulge near the gap (and bigger the gap the bigger is the bulge) and a smaller gap (equal on all the three legs) is more desirable. But you need to keep the gap accurate because small changes in the gap can cause large changes in the total reluctance of the magnetic path.

Core loss per cycle is proportional to the area of the hysteresis loop but I do not know how much the transformer loss will be affected if you shift the gap from the centre leg to all the three legs. Because you have done the real life measurements, can you please let us know how much is the effect?

If the reduction in the core loss is 1-2%, then perhaps it can be safely ignored.
 

yes Easy Pasy is right.......in fact, you can get even better efficiency by using an integrated gap core.
When i worked at Panasonic, we took an offline 40w flyback transformer apart........we saw no gap!...not in centre leg or anywhere!
But these were special integratred gap ferrites...so the gap is distributed throughout the core.
 

special integratred gap ferrites...so the gap is distributed throughout the core...

But a low-mu core will also appear to be a high-mu core with a gap. You can use the same ferrite powder (I guess) and change the amount of the binder and indifferent powder. But you can figure out the amount of the gap by looking at the overall weight (for the same size, the weight will be less).
 

When we used to design really low loss 500kHz chokes at 20A pk we used RM14 with ferrite discs and spacers in the centre pole - so 2mm ferrite, 1mm gap, etc for 15mm or so, this kept the internal fringing field down and reduced wire losses in the 2000/48 litz wire ...
The RM14i give a measure of self shielding externally.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top