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.

Register Log in

buck inductor losses

Status
Not open for further replies.
Y

yoosefheidari

Member level 5
Joined
Aug 31, 2013
Messages
91
Helped
0
Reputation
0
Reaction score
1
Trophy points
1,288
Activity points
2,051
hi
i want to calculate core loss in buck inductor(powder core).but this core is in dc bias condition and in datasheet there is "peak ac flux density vs core loss"curve.
i calculate B for inductor by this formula B=N*I/R*A.but can i use that curve to calculate core loss?if not how can i calculate core loss?
 

I use online tools like this. You can input parameters for DC and ripple and get total losses. Even if you're not using coilcraft you may be able to find a similar inductor from them.
https://www.coilcraft.com/apps/power_tools/compare/

Wurth has new tools with some of this as well. Other manufacturers may as well.
 
Calculating core losses under DC magnetization isn't simple. I'm not sure if the mentioned tools do it correctly. According to literature (Core Losses Under the DC Bias Condition Based on Steinmetz Parameters), the magnetization effect is relative low for powder cores.

Coilcraft and Wuerth tools are only covering ferrite materials, I believe. Ferrite has much higher DC effect than powder core.
 
Core losses are invariably larger than estimated - we have designed a lot of buck and similar converters

core losses from off the shelf inductors are actually very poor - we need to oversize significantly to get reasonable losses esp above 50kHz

high AC ripple of current ( and hence flux ) is the big killer here, along with Rdc ...
 
Easy peasy said:
Core losses are invariably larger than estimated - we have designed a lot of buck and similar converters

core losses from off the shelf inductors are actually very poor - we need to oversize significantly to get reasonable losses esp above 50kHz

high AC ripple of current ( and hence flux ) is the big killer here, along with Rdc ...
Click to expand...
for example we have a buck converter with 10amps output current and 0.5amps ripple current at 200Khz.
you mean DC term of 10Amps didn't produce losses in core?(or is negligible)and that 0.5Amps ripple is cause of core losses?
 

Hi,

You could use the core loss density graph presented for the core material. Trace the loss density at your switching frequency and multiply it by the core volume to get the core loss.

What you need to do is to determine the actual material, then look for a datasheet for that material so you could trace the graph to obtain the core loss density value.

- - - Updated - - -

Since you already have the datasheet, then you could post a link to the datasheet here.
 
Akanimo said:
Hi,

You could use the core loss density graph presented for the core material. Trace the loss density at your switching frequency and multiply it by the core volume to get the core loss.

What you need to do is to determine the actual material, then look for a datasheet for that material so you could trace the graph to obtain the core loss density value.

- - - Updated - - -

Since you already have the datasheet, then you could post a link to the datasheet here.
Click to expand...
yes but that graph is "peak ac flux density VS core loss"but what about dc bias condition?i think that graph isn't useful for calculate losses due to dc bias or dc magnetizing.am i right?
 

Please post a link to the datasheet.
 

I couldn't open the pdf and the website is not very good looking.
 

yoosefheidari said:
for example we have a buck converter with 10amps output current and 0.5amps ripple current at 200Khz.
you mean DC term of 10Amps didn't produce losses in core?(or is negligible)and that 0.5Amps ripple is cause of core losses?
Click to expand...

Core losses are AC or ripple current only. DC bias may alter the AC losses however by changing the position on the BH curve.

Playing with the coilcraft tool for a few minutes I can't tell if DC bias has an impact on AC losses or not. Though their loss chart color codes AC vs total losses I don't see them breaking out the numbers individually so its hard to tell.
 

The core loss calculation procedure given for Magnetics powder cores is to determine actual Bmin and Bmax with DC bias according to BH curve and use Bpk = (Bmax-Bmin)/2 for the loss calculation. See 2015-Magnetics-Powder-Core-Catalog for details. They state:
Core losses result only from AC excitation. DC bias applied to any core does not cause any core losses, regardless of the magnitude of the bias.
Click to expand...

Although that's probably not exactly correct, it's apparently a valid estimation for powder cores.

According to literature and experience, the behavior of ferrite cores is considerably different.

- - - Updated - - -

Some literature related to core loss modelling and DC bias effect:

https://en.wikipedia.org/wiki/Steinmetz's_equation

https://www.hpe.ee.ethz.ch/uploads/tx_ethpublications/05936124_01.pdf
 

if DC current has no effect on core loss then what this formula for? B=NI/RA
we use dc current in that formula to calculate B but where is use of B and for what??!
another question:if AC ripple current produce loss in core then can we decrease loss by increasing wire turn resault in increasing inductance and decreasing ripple current?!!
 

DC current affects permeability and actual inductance. You need to take care that the core isn't saturated.
 

FvM said:
DC current affects permeability and actual inductance. You need to take care that the core isn't saturated.
Click to expand...

it means DC current just affects on permeability and actual inductance and has no affect on losses right?
what i understand is :
1-calculate dc magnetizing force to avoid core saturation
2-calculate core loss due to ripple current
am i right?
 

1. Typically you start with a switching frequency, ripple and peak current specification.
2. Calculate required L
3. Select inductor core according to I²L figure
4. Estimate losses, if too high choose larger core
 
yoosefheidari said:
if AC ripple current produce loss in core then can we decrease loss by increasing wire turn resault in increasing inductance and decreasing ripple current?!!
Click to expand...

Not necessarily because additional turns increases the magnetic field change (thus core losses) for a given ripple amplitude. I guess I'm not sure of that exact relationship but increasing turns also adds wire which also increases restive losses. So for a given inductor size there is always a sweet spot for inductor value vs efficiency.
 
asdf44 said:
Not necessarily because additional turns increases the magnetic field change (thus core losses) for a given ripple amplitude. I guess I'm not sure of that exact relationship but increasing turns also adds wire which also increases restive losses. So for a given inductor size there is always a sweet spot for inductor value vs efficiency.
Click to expand...

i read a text from magnetics inc that says my tought is right.by increasing turns and inductance core losses become low and even can goes near zero.but this will resault in increasing copper loss.
so you right there is a sweet spot that both losses are minimum
 

if you just increase the turns, then the B goes up for the same I, IF you have enough headroom the increase in L will give you lower ripple and hence lower core losses BUT, you will have higher R losses - and - if you were too close to saturation before - you may not get higher L but lower L as the B-H curve bends over - and thereby giving much higher peak ripple currents - SO - stay well away from saturation and keep the flux ripple ( aka current ripple ) down to 10% of the max current ( or less ) and make sure you are well away from saturation ... most off the shelf bobbin type inductors use pretty average grades of ferrite and cannot be run at rated without getting very hot ...
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top