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.

[SOLVED] Inductor quality value for switching regulators

Status
Not open for further replies.

Zak28

Advanced Member level 2
Joined
Aug 19, 2016
Messages
579
Helped
6
Reputation
12
Reaction score
6
Trophy points
18
Activity points
4,681
Does an inductors quality value also mean its ferrite material will have minimal losses at the frequency at which it has highest quality value?
 

depends on the freq at which the Q is measured, QL can be expressed as wL/R ( w = 2.pi.f ) if the Q is given for a range of chokes at the same freq then you can compare meaningfully - highest Q will give the lowest loss as long as you stay well away from saturating the ferrite, i.e. < B = 40mT at 1200kHz
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
Well the Q value is the ration between reactance to the resistance of the coil, both having unit in Ohm so Q value is unit less.
We can’t directly link Q value to the core losses but indirectly linked.
Suppose identical construction (same core material and conductor) of two toroid inductor having same Q value, if we change the core material of one inductor with very low core loss material then Q value of that inductor will increase because low core loss material get higher inductance or reactance
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
Hi,

I can't remember to ever have seen a "Q" specification in a "storage" inductor for swiching regulators.
But maybe there are.

In my eyes it makes not much sense, because "Q" is usually given for a DC free, sinusoidal waveform.
Typicall for HF filter inductors.

Klaus
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
any switching regulator that uses a choke has an element of AC or ripple current in it - at high freq's Rac and core losses can often dominate over the "DC" losses in the winding. As B is directly proportional to current I, flux swing is directly proportional to I swing ...
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
What are these curve depicting? https://www.digikey.com/product-det.../SLF12575T-221M1R3-PF/445-174412-1-ND/5880186

It seems somewhat related to Q and impedance.


inductors.png
 

they actually show impedance ZL with freq - not true inductance where it curves up ...

true Q curves with freq are very rare these days - as it shows quite clearly how shite most off the shelf chokes are
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
they actually show impedance ZL with freq - not true inductance where it curves up ...

true Q curves with freq are very rare these days - as it shows quite clearly how shite most off the shelf chokes are

Perhaps the manufacturers dont prefer to maintain costly measurement equipment which must be calibrated regularly which must measure a certain quantity from each batch for it to provide some accuracy to the datasheet for a product line. It might be a mechanism to liquidate bad batches if their QC isn't performing well.

Is that 221M inductor compatible with 1.2Mhz? It appears to land right as the curve initiates to increase.
 

You can't tell anything about the losses from those curves

Q is easy to measure, and even +/- 5% Q curves would be of value to a designer - but the cheaper offerings have cheap ( read lossy ) ferrites and the real Q is low

Q curves over freq and over various DC bias - would tell you everything about a choke - this is why they don't do it any more .... the first to do it would look bad ...
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
Much akin to lambda power supplies in which among other things TDK capacitors are utilized. Also TDK is a well known supplier of a vast variety of ferrites. They appear to have much works with ferrite materials which might cause them to use the better type in their less expensive inductors.

Does a high SRF (higher than intended switching frequency) determine whether a core material is decent?
 

No ... not in the slightest ...

- - - Updated - - -

the SRF must be much higher than Fsw by the way ...
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
Is the SRF more likely to give issues to a switching regulator than a lossy core material?

Perhaps a lossy material is fine to use if its not saturated to high losses causing excess heat to the regulator.
 

"more likely" is the wrong question. Both parameters are linked to different circuit issues. Core losses are causing core heating, winding capacitance as manifested in self resonance frequency causes EMI problems and possibly excessive switching losses.

Regarding Q curves, there are applications beyond storage inductor, e.g. filter where Q may be of interest. As this are small signal values, they don't tell anything about storage inductor losses.
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
I have to disagree with FvM, Q measurement applies to power and storage i.e. buck inductors too, a high Q, even at signal level, indicates a low loss construction,

of course Q can be measured at any power level, but if you measure two chokes of the same uH on a proper Q meter, the higher Q one will have the lower losses in a power circuit.
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
Is a 1000uH inductor with Q (from datasheet) of 40 measured at 0.796Mhz very lossy for 1200Khz operation?
 

In my eyes it makes not much sense, because "Q" is usually given for a DC free, sinusoidal waveform.

You are right.

An inductor does not have a Q value because it has only one definite parameter: L (inductance).

An ideal inductor is free from resistance or capacitance. It does not have a time constant or associated Q value.

We can define a time constant if the inductor is coupled with a capacitor or a resistor. An ideal LC pair has a very high Q because there is no dissipation. The absorption spectrum of a LC circuit is a simple vertical line.

If the inductor is coupled with a resistor, the Q is low, because of losses and the absorption spectrum will have a broad peak.

Q of an inductor does not make any sense whatsoever.

- - - Updated - - -

of course Q can be measured at any power level...

I am not sure I understand that...

Q is a measure of the quality of the absorption band; narrow bands have higher Q.

The absorption is measured by scanning a frequency and measuring the power absorbed by the sample circuit. You get a graph of absorbance vs frequency. The absorbance is dimensionless (log scale, e.g., db is common) and used on the Y axis.

The excitation power must be as small as possible to reduce saturation. Q has no meaning if you are using high excitation power.
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
Does an inductors quality value also mean its ferrite material will have minimal losses at the frequency at which it has highest quality value?

We cannot conclude on this just like that. It would depend more on what the dominant losses are for the inductor. Sometime eddy current and proximity losses dominate but they are not necessarily core losses.

Overall, it can be concluded that the total losses is least at that frequency relative to the inductive reactance. However, to conclude on core loss behavior specifically, a little bit of analysis has to be done that would have to account also for other losses like eddy current and proximity.
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
A Q of 40 at 800kHz, D = dissipation factor = 1 / Q

(the fundamental definition of Q = 2.pi * peak energy stored / energy dissipated per cycle )

For a cap dissipation = Vac(rms) x Iac (rms) * D = power dissipated

And I assume the same is true for an inductor

So for Q = 40, D = 1/40, this implies if the was an ac voltage across the L of 1V rms say and 300mA rms @ 800kHz, the power dissipated would be 0.3/40 = 7.5mW

However in a ferrite core, losses are proportional to around Bpk^1.7 & Freq^2 so as excitation goes up ( current swing or flux swing ) the losses go up more than proportionally

so the Q would need to be measured at or near a typical operating point - this can in fact be done in a resonance tester with sufficient driving power ( to supply the losses ) as the bulk of the current simply swings back and forward between the cap and the L it is in parallel with ...
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
Can core losses be mitigated for 1200khz by avoiding shielded or partially shielded inductors and use higher resistance parts instead to avoid core losses, or does winding resistance incur greater losses?

Perhaps there is an inductor build type which is overall better with high frequencies, the unshielded bobbin types appear to be with greatest winding resistances.
 

Hi,

So for Q = 40, D = 1/40, this implies if the was an ac voltage across the L of 1V rms say and 300mA rms @ 800kHz, the power dissipated would be 0.3/40 = 7.5mW

You defined the current with "800kHz 300mA RMS"
And you think this is true for
* DC free sinusoidal waveform 800kHz 300mA RMS?
* 0A / 1A square wave with 9% duty cycle, which also is 800kHz, 300mA RMS?
* or a 0.105A / 0.9A square wave with 9% duty cycle, which also is 800kHz, 300mA RMS?
* Or the more triangle shaped current waveforms in CCM, ... with 800kHz, 300mA RMS?
* Or the more triangle shaped with variable long 0A sections current waveforms in DCM, .... 800kHz, 300mA RMS?

I don't expect it.

Klaus
 
  • Like
Reactions: Zak28

    Zak28

    Points: 2
    Helpful Answer Positive Rating
Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top