[Prev][Next][Index][Thread]

Re: Ferrite chokes & saturation


From: 	Malcolm Watts[SMTP:MALCOLM-at-directorate.wnp.ac.nz]
Sent: 	Thursday, November 20, 1997 2:33 PM
To: 	Tesla List
Subject: 	Re: Ferrite chokes & saturation

Hi Gary,

> From:   Gary Lau  19-Nov-1997 1728[SMTP:lau-at-hdecad.ENET.dec-dot-com]
> Sent:   Wednesday, November 19, 1997 3:39 PM
> To:     tesla-at-pupman-dot-com
> Cc:     lau-at-hdecad.ENET.dec-dot-com
> Subject:    Ferrite chokes & saturation
> 
> Many of us employ ferrite core chokes in our xfmr protection networks.
> While the effectiveness (HF attenuation factor) of an L-C low pass filter
> can be fairly easily calculated, I haven't seen any quantitative
> discussion on whether core saturation is occuring.  If it is, that pretty
> much wipes out the benefit of an inductor.
> 
> I've looked over the Amidon web site for info on saturating ferrite core
> chokes, but the graphs, beyond being illegable, I just didn't know how to
> apply them.
> 
> Does anyone actually know how to determine, either through specs or
> measurement, if a ferrite core choke is likely to saturate in a Tesla xfmr
> protection network?  Is it the 60 Hz current or the HF current that is
> significant, or both?  How would either of these be calculated?

    It is dependent on peak Ampere-turns applied, initial core 
permeability and core volume and area.
    First you have to know or measure the instantaneous peak current 
you will put through the choke windings. Most manufacturers publish 
Hanna curves or ballistic induction (B-H) curves for their materials 
which relate energy storage (0.5LI^2) to peak flux density (Bpk). Bpk 
is a function of effective core area (flux lines/area). The basic 
mechanism is something like: for each flux line generated by applied 
Ampere-turns (N.I) force, there is one less that can "fit" into the 
core. In effect, the generated flux lines "fill" the core up. You can 
see that for a maximum number that can be fitted into the core, the 
process leads to a non-linear reduction in inductance. Typical 200kHz 
switchmode type ferrites have a figure for Bpk of around 400 - 500 
mT. The cores start going badly non-linear around 250mT.
    I have a complete suite of equations I can bring in that will 
allow you to work out the gapping required to support the flux your 
windings will generate knowing only the initial permeability, core 
volume and cross-section diameter or length if you are interested. I 
use these in designing switchmode supply chokes and transformers.

> If it is the HF current that causes saturation, could one then assume
> that if a primary strike occurs, saturation current may then be reached
> and choke effectiveness lost, at the time it is most needed?
>
> And on the topic of xfmr protection networks, assuming safety gaps 
are
> employed at the xfmr outputs, are series chokes and/or R's integral to
> the safety gap's effectiveness?  I'm trying to rationalize why many on
> this list use chokes but no bypass caps, or R's but no chokes.  Could
> these series components be viewed as ballast devices, significant only
> if and when the safety gap fires?

I am leaning towards the use of series resistors and bypass caps only
given the cost and effort of building chokes approaching a Henry or 
so. Chokes with significant Q can generate enormous spikes across the 
small stray capacitances present. I will be testing these ideas 
shortly as I am in a situation where I *have* to use neon transformers
for a job and they *have* to last.

Malcolm