saturable inductor

[Michael]

I thought I posted it already, it did not go through. Grrrrrrrrrrrrr
Suppose I have two toroids sandwiched together. One is powder with
single turn inductance (A1)=96nH. Another is high mu ferrite with
A1=15uH.
I apply 10V to 1 turn. The ferrite will saturate (~200mA) pretty soon
(1/4 us). By that time the inductor will have ~0.26uJ stored in it....
How will current look?
I haven't got samples yet, I cannot try it...

 

[Joerg]

I thought I posted it already, it did not go through. Grrrrrrrrrrrrr
Suppose I have two toroids sandwiched together. One is powder with
single turn inductance (A1)=96nH. Another is high mu ferrite with
A1=15uH.
I apply 10V to 1 turn. The ferrite will saturate (~200mA) pretty soon
(1/4 us). By that time the inductor will have ~0.26uJ stored in it....
How will current look?

There will be that classical shark fin, followed by a loud bang, molten
solder splattering about, stuff flying around, smoke alarms blaring ...

I haven't got samples yet, I cannot try it...

Be careful and make sure there si something that will limit the current.

 

[Michael]

There will be that classical shark fin, followed by a loud bang, molten
solder splattering about, stuff flying around, smoke alarms blaring ...


Be careful and make sure there si something that will limit the current.

Spark, smoke, sirens... I failed to ask why it doesn't happen

 

[Joerg]

Spark, smoke, sirens... I failed to ask why it doesn't happen

Well, if you have a current limit it won't ;-)

 

[Joerg]

Aww, Joerg, you take the excitement out of life.

I had the ongoing amusement one project of kibitzing a theoretically
experienced analog guy (from MIT, no less!) on a switching amplifier.
One of the things that I diagnosed was that mysterious (to him) shark
fin, although it was only one of the reasons that things went BANG on
that particular circuit.

Some of the guys also overlook the fact that there is a rather large
tolerance on ferrite core specs. Later, some folks in purchasing think
that a ferrite is a ferrite is a ferrite. Phssst ... *BANG*.

The part that added the most humor (for me) was that my job title (and
function) at the time was very firmly planted in "software engineer",
although I was being allowed to design a (simple) board.

Don't ya love it when that happens? I once snuck into a mechanical
design review. I wasn't supposed to be there but urgently needed to pass
a message on to one of the guys. Looked at the PowerPoint screen and
just had to ask: "Looks like there is way too little airflow. And what
happens if that fan up there fails?" ... "Uhm ..., well, ... s..t!"
Major redesign.

 

[Michael]

Some of the guys also overlook the fact that there is a rather large
tolerance on ferrite core specs. Later, some folks in purchasing think
that a ferrite is a ferrite is a ferrite. Phssst ... *BANG*.


Don't ya love it when that happens? I once snuck into a mechanical
design review. I wasn't supposed to be there but urgently needed to pass
a message on to one of the guys. Looked at the PowerPoint screen and
just had to ask: "Looks like there is way too little airflow. And what
happens if that fan up there fails?" ... "Uhm ..., well, ... s..t!"
Major redesign.

OK, I kind of guessed (qualitatively) that at first dI/dt is "slow"
then there is a step (?? how long does it take for mu to drop??). The
current steps up by sqrt(mu) and then keeps ramping up mu times as
fast.
Does it make sense?

 

[Joerg]

OK, I kind of guessed (qualitatively) that at first dI/dt is "slow"
then there is a step (?? how long does it take for mu to drop??). The
current steps up by sqrt(mu) and then keeps ramping up mu times as
fast.
Does it make sense?

Depends on the core, it's in the ferrite datasheet. Usually the "grace
period" ends at some point where the inductor basically acts like an air
coil, as if the core wouldn't be there anymore. For most practical
circuits such as switch mode supplies that comes pretty close to a dead
short with all its consequences.

 

[Michael]

Depends on the core, it's in the ferrite datasheet. Usually the "grace
period" ends at some point where the inductor basically acts like an air
coil, as if the core wouldn't be there anymore. For most practical
circuits such as switch mode supplies that comes pretty close to a dead
short with all its consequences.

I am not worried about "dire consequences" - I am sandwiching two
cores actually (ferrite and iron powder), what I need to know is HOW
FAST ferrite "disappears". They should use something like this in
series trigger transformers.

 

[Joerg]

I am not worried about "dire consequences" - I am sandwiching two
cores actually (ferrite and iron powder), what I need to know is HOW
FAST ferrite "disappears". They should use something like this in
series trigger transformers.

All I can tell you is "really fast". For example, it can take just a few
nanoseconds to drive Fair-rite #43 material into complete saturation.
BTDT, got broom and shovel and cleaned up the mess ;-)

The voltage applied to a transformer usually determines that. The higher
it is the faster the primary current will rise and will core saturation
be reached.

 

[Joerg]

I'm kinda side-stepping Joerg's whole thread with this reply -- what he
has to say makes oodles of sense in a switching-supply frame of
reference, but I doubt that that's what you're thinking of (or you are
trying to make a switched supply with your oddball stacked cores, which
means you may be cracked -- go look in the mirror and ask yourself if
you're sane as a check).

The total voltage on the coil is equal to the IR drop of the wire plus
the back-emf on the coil. The back-emf of the coil is proportional to
the the rate of change of the product of the current and the
permeability of the core material -- and while the current (and
magneto-motive force) is going up, the permeability of the core material
is going down.

So initially your inductor will have high inductance and the current
will ramp up at a more or less constant rate commensurate with the
zero-current inductance. At some point the ferrite in the core will get
saturated with flux and won't accept any more than air does and your
current will start ramping up at a more or less constant rate
commensurate with whatever else you have in your coil -- if it's just
air, then it'll ramp up pretty darn fast, if that iron powder coil
hasn't saturated yet it'll ramp up at a rate commensurate with the iron
powder coil (until it saturates).

I _don't_ know how fast the transition will be between the two straight
ramps, or how straight they'll really be (the coil resistance will
affect that to some extent anyway).

You ought to be able to make an estimate of the transition if you can
get your hands on a B-H curve of the material -- the sharper the
transition between the sloped part around 0 H and the nearly level part
around high H, the sharper the transition between the two 'modes' of the
core.

Good explanation. "Pretty darn fast" means just that, it gets down to
nanohenries and the current (hopefully) pegs against some limiter.
Michael should keep in mind that B-H curves and such are just typical
data. There can be huge tolerances in core behavior, I have seen in
excess of 30%. Us analog/RF dudes usually stay a good 40% below the
"pretty darn fast" range.

 

[Vladimir Vassilevsky]

About that same time (grad school) I had a direct-conversion 40 meter
radio whose receive frequency would shift a few hundred Hz in the
presence of a strong magnet outside the case.

I've always wanted to build a radio that tunes the LO that way on
purpose -- just an iron powder coil, a rare earth magnet, and a
mechanism to bring them together in a controlled way. Of course I have
No Clue how well it would work, but you could sure have fun with writing
some marketing hype to go with the technique!

BTW, the magnetically controlled inductance is the old trick for
electronic tuning of the LC tanks while keeping good Q and linearity.
The parameters attainable are better then what can be done with varactors.


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

 

[Joerg]

BTW, the magnetically controlled inductance is the old trick for
electronic tuning of the LC tanks while keeping good Q and linearity.
The parameters attainable are better then what can be done with varactors.

It was also a nice method to achieve a mostly frequency-linear tuning
range for oscillators. IIRC Drake and Collins used that. The downside
was that the worm gear needed constant attention. So the usual setup was
the receiver or tranceiver itself with a screwdriver (no, not the drink
....) and a bottle of rifle oil sitting next to it. In Germany we used
the Ballistol brand which supposedly was also good to cure other
ailments such as saddle sores after too much horseback riding. I don't
know about that and would not do it but the stuff sure is quite
universal in use:
http://www.ballistol.com/uses.htm

 

[Joerg]

Yes, although varactors sometimes win if you need to pass high power (say,

Nah, real men use those:

http://www.ph.utexas.edu/~phy-demo/demo-txt/5c10-11.html

 

[legg]

I thought I posted it already, it did not go through. Grrrrrrrrrrrrr
Suppose I have two toroids sandwiched together. One is powder with
single turn inductance (A1)=96nH. Another is high mu ferrite with
A1=15uH.
I apply 10V to 1 turn. The ferrite will saturate (~200mA) pretty soon
(1/4 us). By that time the inductor will have ~0.26uJ stored in it....
How will current look?
I haven't got samples yet, I cannot try it...

Composite cores are available from a few mfrs, inluding some parts
that are basically two toroids glued together. The most common
application is in a swinging choke, used to reduce minimum continuous
load current levels in power supply unregulated auxiliary outputs.

A part stressed over the full flux range (+/-) will show a change of
permeability at fairly low flux levels - a monitored current waveform
produced by a fixed alternating voltage will show 2 different current
di/dt slopes, before saturation.

RL

 

[Vladimir Vassilevsky]

It was also a nice method to achieve a mostly frequency-linear tuning
range for oscillators. IIRC Drake and Collins used that. The downside
was that the worm gear needed constant attention.

I meant pure electronic solution by affecting the permeability of the
core by the magnetic field of the external solenoid. This variant can
have much better linearity and Q then attainable with varactors, which
can be important for LO and input filters. AFAIK some early FM
transmitters used this idea, too.

So the usual setup was
the receiver or tranceiver itself with a screwdriver (no, not the drink
...) and a bottle of rifle oil sitting next to it.

Speaking of the motor tuning:

http://www.armyradio.com/publish/Articles/William_Howard_Russian/Pictures/R-155A-01.jpg

IIRC this R-155 receiver has 8-section variable capacitor bank tuned by
the electric motor. The input section is built using tubes (total of 6),
and the rest of the schematics is made from the discrete BJTs and
crystal filters. In the old times, they knew how to do the real stuff.

In Germany we used
the Ballistol brand which supposedly was also good to cure other
ailments such as saddle sores after too much horseback riding. I don't
know about that and would not do it but the stuff sure is quite
universal in use:
http://www.ballistol.com/uses.htm

I put a notice on that. If there is PITA for whatever reason, the
ballistol should help, is that right?


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

 

[Joerg]

I meant pure electronic solution by affecting the permeability of the
core by the magnetic field of the external solenoid. This variant can
have much better linearity and Q then attainable with varactors, which
can be important for LO and input filters. AFAIK some early FM
transmitters used this idea, too.


Speaking of the motor tuning:

http://www.armyradio.com/publish/Articles/William_Howard_Russian/Pictures/R-155A-01.jpg


IIRC this R-155 receiver has 8-section variable capacitor bank tuned by
the electric motor. The input section is built using tubes (total of 6),
and the rest of the schematics is made from the discrete BJTs and
crystal filters. In the old times, they knew how to do the real stuff.

They sure did but they were blissfully ignorant when it came to power
consumption. Power just came out the wall outlet and was free ;-)

I put a notice on that. If there is PITA for whatever reason, the
ballistol should help, is that right?

According to the movie "My Big Fat Greek Wedding" Windex would be the
answer. Very worthwhile to watch BTW, great fun.

 

[Vladimir Vassilevsky]

They sure did but they were blissfully ignorant when it came to power
consumption. Power just came out the wall outlet and was free ;-)

At the old times, it was the different engineering paradigm: design for
the best performance whatever it takes. With the use of the well
designed stuff we defeat all enemies so the power in the wall outlet
will be completely free.

According to the movie "My Big Fat Greek Wedding" Windex would be the
answer. Very worthwhile to watch BTW, great fun.

I don't watch TV or movies, and I am not subscribed for the cable.


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

 

[Joerg]

At the old times, it was the different engineering paradigm: design for
the best performance whatever it takes. With the use of the well
designed stuff we defeat all enemies so the power in the wall outlet
will be completely free.

That goes back all the way to Caesar and the Romans. They clear-cut
pretty much all of what's now Italy to build one armada after the other,
thinking that'll make them own the world. It didn't and now the soil on
all those mountains has eroded. Nothing big will grow there anymore.

I don't watch TV or movies, and I am not subscribed for the cable.

This one's worth it. You can rent it at video places and watch it with
friends that have a VCR. No idea whether it made it onto DVD. We have
friends who said their (Greek Orthodox) wedding was an almost verbatim
repeat of that story and we all laughed really hard because all of us
had seen that movie. To this day the husband has refused to show us
pictures of his baptism right before the wedding. Too embarrassing, he says.

 

[legg]

On Jul 28, 3:05 pm, Joerg <[email protected]>
wrote:

I am not worried about "dire consequences" - I am sandwiching two
cores actually (ferrite and iron powder), what I need to know is HOW
FAST ferrite "disappears". They should use something like this in
series trigger transformers.

A quick estimate would assume that during the first interval, only the
ferrite path stores energy. As it's ungapped permeability is a number
or orders of magnitude higher than the compounded material, this is
valid.

This duration can be derived from the B= Vt/NA equation.

deltaB = ( V . t ) / ( N . A )

deltaB= flux density change in Teslas ( 1T = 10,000gauss)
V= Voltage applied across the inductor winding (volts)
t= period of applied voltage (seconds)
N= number turns in the coil
A= physical cross-section (ferrite only) of the magnetic medium
(meters^2)

If the applied voltage is unipolar and reset to ~zero, the period to
saturation corresponds to a B value of between 0.3 and 0.4 T (.35
nominal), depending on material grade and temperature.

If the applied voltage is AC, the period of the ferrite operation will
be for double this value, as the flux will swing from + to -
saturation levels.

RL

 

[whit3rd]

About that same time (grad school) I had a direct-conversion 40 meter
radio whose receive frequency would shift a few hundred Hz in the
presence of a strong magnet outside the case.

I've always wanted to build a radio that tunes the LO that way on
purpose

My favorite old keyboard (it was a Burroughs, which should
give a hint of its age) had keyswitches that moved a small
magnet to saturate a core, which had two one-turn windings
(looked like staples), one to excite and one to sense.
It had great key feel...no mechanical damping or hysteresis
like a buckling switch. Hall switches exist, too, but
this was ... elegant.