wire gauge for inductor: skin effect question

[MTong]

What is the procedure to select wire gauge for inductor?

I have been told and seen that first I need to calculate the skin depth based on the switching frequency and then select a wire whose radius is equal or smaller than calculated skin depth. Then, calculate number of strands which equals the rms of inductor current divided by rms current rating of selected wire. However, through this process an amount of current ripple in inductor is not considered.

What I want to say is that let's say that inductor current ripple is 5% of dc current or even let's say 0.5% (just an example) which means inductor current is almost dc and I don't need to be worry about skin effect (I think, I am not sure) because the ac current in inductor in negligible. But if I follow the mentioned procedure I may end up in litz wire which is expensive while I may not actually need it. Therefore, the amount of current ripple should be considered somewhere in the procedure of wire selection for inductor.

For transformer wire selection, the mentioned procedure absolutely makes sense for me since the currents in winding are all ac (for example full bridge dc-dc) and changing the direction completely in each switching cycle.

Please let me know your opinion.

Thank you

[Gallymimu]

Your logic is sound.

what frequencies are you switching at and what kind of power are you pumping through the inductor?

The recommended design approach for skin depth/litz wire selection you mentioned is pretty conservative and you are correct that the DC component of the inductor current is not affected by skin depth.  That said in MOST switcher designs the ripple current in the inductor is 30-50% of the DC value.  Otherwise you probably have way more inductance than you need.  Also low ripple can adversely impact control stability, for instance a peak current mode SMPS design (which is one of the best ways to control the inner current loop) is going to work better with a decent inductor ripple.  I realize you were stating low ripple just as an example though for the purposes of your question.

[zab]

this may help you in this regard.
http://www.diyaudio.com/forums/power-supplies/170340-wire-diameter-smps-transformer-confusion.html

[Gallymimu]

this may help you in this regard.
http://www.diyaudio.com/forums/power-supplies/170340-wire-diameter-smps-transformer-confusion.html

I'd say that table is still way too conservative.  For Rac/Rdc values of 1.0000 WHO CARES!  I'd personally not start worrying until it was above 1.05 or 1.1 and that's for high AC currents where the inductor currents might really be large and make the cost of Litz worthwhile.

Posted on: September 14, 2013, 10:40:15 22:40 - Automerged
Cool little table though for eyeballing skin effect.

[MTong]

Thanks for the replies,
I am just asking in general. This was a question in my mind for a while.

Also, what about the proximity effect due to using litz wire (multiple strands in general) vs using a single conductor? I believe it is a trade off between losses due to skin effect and losses due to proximity effect.  Any general guidelines about this?

Thank you all

[zab]

Also, what about the proximity effect due to using litz wire (multiple strands in general) vs using a single conductor? I believe it is a trade off between losses due to skin effect and losses due to proximity effect.  Any general guidelines about this? 

This is what you are thinking of
 one point of view see here
 http://ep2000.com/uploads/EP_WhitePaper_SkinEffect.pdf
and   
the other point of view given at some where in litzwire.com

Proximity effect is the tendency for current to flow in loops or concentrated distributions due to the presence of magnetic fields generated by nearby conductors. In transformers and inductors, proximity effect losses are generally more significant than skin effect losses. In Litz wire windings, proximity effect may be sub-divided into internal proximity effect (the effect of other currents within the bundle) and outer proximity effect ( the effect of the current in other bundles). The reason for twisting or weaving Litz wire, rather than just grouping fine conductors together, is to ensure that the strand currents are equal. Simple twisted bunched conductor wire can accomplish this adequately where proximity effect would be the only significant problem with solid wire. Where skin effect would also be a problem, more complex Litz wire constructions can be used to ensure equal strand currents. Therefore, in a well-designed construction, strand currents are nearly equal.

For further detail search through that site.

[MTong]

Thank you guys for your replies and sorry for taking so long before getting back here.

There is plenty of info regarding all of these aspects on internet; I do not know how is this considered in industry:
designing inductors for a xx% current ripple, how is skin effect and proximity considered in the design, that is, what is a generic procedure to follow including both this aspects?

Thank you