Variable inductor

[John Larkin]

I'm trying to remember how much variability coils with *copper* slugs
had (or were they brass?)

I've haven't seen them since the early days of TV sets (mid '50's).

...Jim Thompson

Brass, and the slug reduces L a bit. Ferrite increases it a lot more.

Old fashioned variometers had a huge range of adjustability. There
were two concentric windings, one of which rotated on a shaft, with
the windings wired in series or parallel. If they bucked, L was very
low, and when they add, you get the N^2 thing going. Some had
quasi-spherical windings on shiny black bakelite, quite beautiful
components. They commonly achieved 10:1 inductance range or more.

John

 

[Ken Smith]

I'm trying to remember how much variability coils with *copper* slugs
had (or were they brass?)

IIRC it was about 25%-30% of the maximum. They were good for trimming the
frequency only.

 

[Mark]

See TOKO Coils

Mark

 

[Ethan]

One interesting idea to consider is using DC currents in
the winding to magnetically saturate the core, reducing
its permeability. This method can easily exceed 4:1
ratios, but it's best done inside a servo loop where
stability doesn't matter.

What kind of stability do you need?

I was reading this thread, and was just about to suggest saturating the
core with a second winding. This way you can control the inductance
electronicly.

Recently I have been playing around with something kind of like this.
I have been characterizing inductors to find out where and how they
saturate, by putting a DC current in one winding and measuring the
inductance on a second winding. The complication is the impedance of
the DC current source appears on the measurement winding. To get a
decent measurement I have to make sure the DC source impedance is very
high relative to the indutance I am trying to measure.

The inductance of the parts I am looking at varies from about 1.5mH to
less than 50uH at 100kHz. (About 30-1 inductance change)

There are some tricks to make the DC source impedance less of an issue.
These are old magnetic amplifier techniques. You can use matched
inductors in series with the DC windings coupled in opposite directions
so the coupling cancels out. Something else I have been meaning to try
is to put two windings on a flat core on quadrature axis so they don't
couple. This is kind of like the variometer with the inner winding
turned to 90 deg.

I would also suggest a core with a lumpy shape rather than a torroid.
An E core or something. If the core saturates unevenly the inductance
will roll off more gradually and be easier to control.

This is all probably more complicated to impelment then what you are
looking for, but hopefully is still interesting.


Ethan

 

[John Perry]

I was reading this thread, and was just about to suggest saturating the
core with a second winding. This way you can control the inductance
electronicly.

Recently I have been playing around with something kind of like this.
I have been characterizing inductors to find out where and how they
saturate, by putting a DC current in one winding and measuring the
inductance on a second winding. The complication is the impedance of
the DC current source appears on the measurement winding. To get a
decent measurement I have to make sure the DC source impedance is very
high relative to the indutance I am trying to measure.

The inductance of the parts I am looking at varies from about 1.5mH to
less than 50uH at 100kHz. (About 30-1 inductance change)

The only trick really needed is to make the second coil orthogonal to
the inductor coil. For instance, your controlled coil would be wound as
usual, through the hole, then the controlling coil wound over the entire
toroid. This arrangement allows you to have the two coils almost
completely insensitive to each other.

Another method uses two cores canceling each other, with the controlled
coil again wound through the holes of each core, but in opposite
directions; then the controlling coil wound through the pair. This way,
the controlled coil's current causes fields opposing each other in the
controlling coil, and vice versa.

John Perry

 

[Ken Smith]

I was reading this thread, and was just about to suggest saturating the
core with a second winding. This way you can control the inductance
electronicly.

It also makes a dandy microphone, thermometer, phase of moon and customer
nearby detector. I'd avoid this method if you can. It can be real
trouble.

[...]

Recently I have been playing around with something kind of like this.
I have been characterizing inductors to find out where and how they
saturate, by putting a DC current in one winding and measuring the
inductance on a second winding. The complication is the impedance of
the DC current source appears on the measurement winding. To get a
decent measurement I have to make sure the DC source impedance is very
high relative to the indutance I am trying to measure.

To extend the comments a bit: You need two inductors or some funny
windings.

* *
---------))))))---------------))))))----------- Inductance
======== =======
---------))))))----- ----)))))))----
* ! ! * !
! ---------------!------- Control
-------------------


OR:

Get an "E" core and wind like this:

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!! !!! !!!
!!! !!! !!!
A------------ C--------- E---------
----- ----- -----
----- ----- -----
B------- !!! D---- !!! F---- !!!
!!! !!! !!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!


The (C-D) winding is the inductor. A is wired to F. B-E is the control.

 

[Joel Kolstad]

Ken,

Get an "E" core and wind like this:

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!! !!! !!!
!!! !!! !!!
A------------ C--------- E---------
----- ----- -----
----- ----- -----
B------- !!! D---- !!! F---- !!!
!!! !!! !!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

The (C-D) winding is the inductor. A is wired to F. B-E is the control.

Don't you want A wired to E? As drawn, via the right-hand rule, the left and
right coils produce flux in the "downward" direction, no?

 

[Terry Given]

Ken,




Don't you want A wired to E? As drawn, via the right-hand rule, the left and
right coils produce flux in the "downward" direction, no?

I'm not sure about that (OTTOMH Ken's right, a) because he usually is,
and b) because that looks right), but here are a few books on saturable
inductors:

Geyger, Nonlinear Magnetic Devices
Storm, Magnetic Amplifiers
Attura, Magnetic Amplifier Engineering
LaFuze, Magnetic Amplifier Analysis
Geyger, Magnetic Amplifier Circuits
Platt, Magnetic Amplifiers: Theory and Application
Reyner, The Magnetic Amplifier
Neeteson, Square Loop Ferrite Core Switching
Quartly, Square Loop Ferrite Circuitry


The first two are IMHO the best.

Cheers
Terry

 

[Ken Smith]

Ken,




Don't you want A wired to E? As drawn, via the right-hand rule, the left and
right coils produce flux in the "downward" direction, no?

Oooops! Yes, I got it backwards. The (ABEF) combination needs to make
the field flow around the outside of the E and not through the middle leg.

 

[Ken Smith]

I'm not sure about that (OTTOMH Ken's right, a) because he usually is,

No, it looks like I messed up in this case. The current in the (A,B)
winding and the (E,F) winding should cause fields that cancel in the
middle leg. This way, there is no coupling to the winding on the middle
leg.

Actually "no coupling" is a bit of an overstatement. They won't cancel
exactly and you will see the 2nd harmonic of the signal on the (C,D)
winding on the control winding.

 

[ian field]

I was reading this thread, and was just about to suggest saturating the
core with a second winding. This way you can control the inductance
electronicly.

Recently I have been playing around with something kind of like this.
I have been characterizing inductors to find out where and how they
saturate, by putting a DC current in one winding and measuring the
inductance on a second winding. The complication is the impedance of
the DC current source appears on the measurement winding. To get a
decent measurement I have to make sure the DC source impedance is very
high relative to the indutance I am trying to measure.

The inductance of the parts I am looking at varies from about 1.5mH to
less than 50uH at 100kHz. (About 30-1 inductance change)

There are some tricks to make the DC source impedance less of an issue.
These are old magnetic amplifier techniques. You can use matched
inductors in series with the DC windings coupled in opposite directions
so the coupling cancels out. Something else I have been meaning to try
is to put two windings on a flat core on quadrature axis so they don't
couple. This is kind of like the variometer with the inner winding
turned to 90 deg.

I would also suggest a core with a lumpy shape rather than a torroid.
An E core or something. If the core saturates unevenly the inductance
will roll off more gradually and be easier to control.

This is all probably more complicated to impelment then what you are
looking for, but hopefully is still interesting.


Ethan

It isn't as easy as just winding a DC coil to saturate the core, a properly
constructed saturable reactor has double E ferrite cores one winding is on
the centre pole and the other is shared between the side limbs, I forget the
exact details but its done to eliminate polarisation and prevent asymmetric
change in inductance as the signal adds/subtracts to/from the saturating
field.

 

[Joel Kolstad]

Ken,

To extend the comments a bit: You need two inductors or some funny
windings.

* *
---------))))))---------------))))))----------- Inductance
======== =======
---------))))))----- ----)))))))----
* ! ! * !
! ---------------!------- Control
-------------------

I realize I actually have a fundamental misunderstanding of how this is
supposed to work. You feed current into the control winding, and since the
magnetic field generated in each one points in the opposite direction, don't
you end up with zero net flux *everywhere* throughout the coil and therefore
haven't changed the net inductance the controlled coil sees?

Or isn't the schematic above intended to produce a variable inductance?

Thanks,
---Joel

 

[Tim Williams]

don't you end up with zero net flux *everywhere* throughout the coil
and therefore haven't changed the net inductance the controlled coil sees?

Not with seperated cores. You do develop a nasty voltage on the control
windings (and thus capacitive loading), though, which is why cancelling flux
is a better idea than cancelling voltage.

Tim

 

[Chris Jones]

I'm looking for a variable inductor (approx 10mm size) that has a range of
at least 2:1.
This is for use in a 50-100MHz tuned circuit, and I've calculated that
Imin=185nH, and Imax=366nH.
Does anyone know if such a range is possible?
Thanks.

What's it for anyway? Maybe there is a way to avoid the requirement for
this thing.

Chris

 

[Phil Allison]

"Chris Jones"

What's it for anyway? Maybe there is a way to avoid the requirement for
this thing.

** Now the debate has come full circle.


My original query:

" So are variable caps with a 4:1 range.

What sort ot of dumb wank is this ? "


did not extract a satisfactory reply out of the OP..

But did reveal his character.




......... Phil

 

[Grumps]

What's it for anyway? Maybe there is a way to avoid the requirement
for this thing.

It's for reactively coupling a 200 Ohm output impedance amplifier, into a 50
Ohm load.
Although many interesting comments have been made (and thanks for all of you
who've contributed), the acceptable solution is to change the inductor for
different frequency ranges.

 

[Mike T]

"Chris Jones"

Good question.

** Now the debate has come full circle.


My original query:

" So are variable caps with a 4:1 range.

What sort ot of dumb wank is this ? "


did not extract a satisfactory reply out of the OP..

Probably due to the offensive nature of your posting style.

But did reveal his character.

Seems pretty normal to me.
Certainly revealed (confirmed) your character.

 

[Phil Allison]

"Mike Tosser "

Probably due to the offensive nature of your posting style.

** Not near as offensive as YOURS - fuckhead.

Seems pretty normal to me.

** Well, it would - wouldn't it ?

One fuckhead is just like another.




........ Phil

 

[Chris Quayle]

Grumps wrote...

2:1 variable L is readily available in 10mm gapped pot
cores, although 4:1 is not, SFAICT.

There are several methods to get up to 10:1 inductance
ratios in a bit more space, e.g., moving large cores
into air coils, rotating concentric series-connected
coil sections (variometer), and rolling inductors.

One interesting idea to consider is using DC currents in
the winding to magnetically saturate the core, reducing
its permeability. This method can easily exceed 4:1
ratios, but it's best done inside a servo loop where
stability doesn't matter.

What kind of stability do you need?

The Plessey PR155 series from the sixties used core saturation for rough
tuning of a 37.3 - 67.3 Mhz synthesised first local oscillator + varicap
for fine tuning. Interesting design in that it used the ssb phasing
method in the syntheser and was one of the first (1965) hf receivers to
use a hot carrier diode quad for the first mixer, upconverting to a 37.3
Mhz first if.

You could use a small toroid - conventional turns for the oscillator,
with turns wound across the outside of the core for saturation
control...

Chris

 

[Fred Bloggs]

** Well, it would - wouldn't it ?

One fuckhead is just like another.

There is an uncanny degree of homogeneity among certain mentalities on
NG- very dull.