Drum Core Stray Magnetic Field

[Klaus Kragelund]

Hi

I have an application in which I need high stray field from a drum core in order to pass a signal on to another drum core

Anyone ever measured the stray field above a drum core? Someting like this type:

http://image.made-in-china.com/43f34j00PsZtJgRaZiqd/SMD-Unshielded-Power-Inductors.jpg

It is non shielded, which is good for stray field, but it has a small "hat", so that bends the field somewhat

Cheers

Klaus

 

[Klaus Kragelund]

On Saturday, February 28, 2015 at 11:56:04 AM UTC+1, Robert Macy wrote:

On Sat, 28 Feb 2015 03:18:26 -0700, Klaus Kragelund
klauskvik@hotmail.com> wrote:

Hi

I have an application in which I need high stray field from a drum core
in order to pass a signal on to another drum core

Anyone ever measured the stray field above a drum core? Someting like
this type:

http://image.made-in-china.com/43f34j00PsZtJgRaZiqd/SMD-Unshielded-Power-Inductors.jpg

It is non shielded, which is good for stray field, but it has a small
"hat", so that bends the field somewhat

Cheers

Klaus

Kind of the technique used in EMC Open Air Test Sites to calibrate an
antenna:

Use identical coil and a known impedance [spectrum analyzer] I think 'near
field' is valid.

If not, use a tiny shielded loop of known area into your spectrum analyzer.

This only gives you 'single' point measurement. I'd use femm 4.2 to model
the coil and put in those pesky manufacturer variations. The range of
magnitude of the radiating fields may astound you.

Thanks

I was thinking about possibly getting the percentage of the external field versus the coupled internal field. But that requires invasive probing..

Cheers

Klaus

 

[Lasse Langwadt Christense]

Den lřrdag den 28. februar 2015 kl. 11.18.33 UTC+1 skrev Klaus Kragelund:

Hi

I have an application in which I need high stray field from a drum core in order to pass a signal on to another drum core

Anyone ever measured the stray field above a drum core? Someting like this type:

http://image.made-in-china.com/43f34j00PsZtJgRaZiqd/SMD-Unshielded-Power-Inductors.jpg

It is non shielded, which is good for stray field, but it has a small "hat", so that bends the field somewhat

Cheers

Klaus

stick two on a vector analyser and see what they do when mounted like you intend?

some what related: http://www.st.com/web/en/resource/technical/document/application_note/CD00268020.pdf

-Lasse

 

[DecadentLinuxUserNumeroUn]

On Sat, 28 Feb 2015 08:17:28 -0700, RobertMacy wrote:

On Sat, 28 Feb 2015 05:19:36 -0700, Lasse Langwadt Christensen
langwadt@fonz.dk> wrote:

...snip....

stick two on a vector analyser and see what they do when mounted like
you intend?

some what related:
http://www.st.com/web/en/resource/technical/document/application_note/
CD00268020.pdf

-Lasse

Or, a Network Analyzer and note the S21 term.

There is a widget called an H-Field PRobe used to be made by HP.
Calibrated thingy, looks like a doorstop with an SMA connector on the
back to connect to your SA. It's shielded, calibrated, and only measures
magnetic field. Usually used for 'sniffing' fields emanating from PCB
traces, but could be used here.

Buy a ferrite core. Put both windings on it and make a transformer and
put aside all this "local coupling" nonsense, especially if you are trying
to rely on it for some aspect of circuit operation.

Same core equals guaranteed coupling. Use a single turn if all you want
is a sense line.

 

[DecadentLinuxUserNumeroUn]

On Sat, 28 Feb 2015 09:38:23 -0600, Tim Williams wrote:

You'll primarily have field lines coming off the edges of the top/bottom
plates and arcing through space. The field above and below won't be so
much, so that a vertical offset would be inconvenient (but on a planar
PCB, that doesn't matter!), but placing them side by side should give
quite reasonable "weak" coupling.

I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related
to height (plate-to-plate centers distance) and diameter.

Tim

Yes... grab it from the side. I would use an air core for the grabber
though. No core needed on the guy you are trying to grab stray fields
with.

Then, you can make it small, and turn it 90 degrees and place it right up
next to the exposed winding faces.

 

[RobertMacy]

On Sat, 28 Feb 2015 03:18:26 -0700, Klaus Kragelund
<klauskvik@hotmail.com> wrote:

Hi

I have an application in which I need high stray field from a drum core
in order to pass a signal on to another drum core

Anyone ever measured the stray field above a drum core? Someting like
this type:

http://image.made-in-china.com/43f34j00PsZtJgRaZiqd/SMD-Unshielded-Power-Inductors.jpg

It is non shielded, which is good for stray field, but it has a small
"hat", so that bends the field somewhat

Cheers

Klaus

Kind of the technique used in EMC Open Air Test Sites to calibrate an
antenna:

Use identical coil and a known impedance [spectrum analyzer] I think 'near
field' is valid.

If not, use a tiny shielded loop of known area into your spectrum analyzer.

This only gives you 'single' point measurement. I'd use femm 4.2 to model
the coil and put in those pesky manufacturer variations. The range of
magnitude of the radiating fields may astound you.

 

[RobertMacy]

On Sat, 28 Feb 2015 05:19:36 -0700, Lasse Langwadt Christensen
<langwadt@fonz.dk> wrote:

...snip....

stick two on a vector analyser and see what they do when mounted like
you intend?

some what related:
http://www.st.com/web/en/resource/technical/document/application_note/CD00268020.pdf

-Lasse

Or, a Network Analyzer and note the S21 term.

There is a widget called an H-Field PRobe used to be made by HP.
Calibrated thingy, looks like a doorstop with an SMA connector on the back
to connect to your SA. It's shielded, calibrated, and only measures
magnetic field. Usually used for 'sniffing' fields emanating from PCB
traces, but could be used here.

 

[Tim Williams]

You'll primarily have field lines coming off the edges of the top/bottom
plates and arcing through space. The field above and below won't be so
much, so that a vertical offset would be inconvenient (but on a planar
PCB, that doesn't matter!), but placing them side by side should give
quite reasonable "weak" coupling.

I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related to
height (plate-to-plate centers distance) and diameter.

Tim

--
Seven Transistor Labs
Electrical Engineering Consultation
Website: http://seventransistorlabs.com

"Klaus Kragelund" <klauskvik@hotmail.com> wrote in message
news:055ca89d-7afa-4650-ad26-5c681cbef84f@googlegroups.com...

Hi

I have an application in which I need high stray field from a drum core
in order to pass a signal on to another drum core

Anyone ever measured the stray field above a drum core? Someting like
this type:

http://image.made-in-china.com/43f34j00PsZtJgRaZiqd/SMD-Unshielded-Power-Inductors.jpg

It is non shielded, which is good for stray field, but it has a small
"hat", so that bends the field somewhat

Cheers

Klaus

 

[John Larkin]

On Sat, 28 Feb 2015 09:38:23 -0600, "Tim Williams"
<tiwill@seventransistorlabs.com> wrote:

You'll primarily have field lines coming off the edges of the top/bottom
plates and arcing through space. The field above and below won't be so
much, so that a vertical offset would be inconvenient (but on a planar
PCB, that doesn't matter!), but placing them side by side should give
quite reasonable "weak" coupling.

I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related to
height (plate-to-plate centers distance) and diameter.

Is that the near-field approximation? Far-field is 1/d^3.

Tim

I was playing with the idea of mounting a pair of surface-mount drum
cores on opposite sides of a PC board, axially coupled, to make a high
voltage isolated coupler. But the customer went away so I didn't
develop the idea. Side-by-side should work, too.

Things like this are easier to try than to analyze.


--

John Larkin Highland Technology, Inc
picosecond timing laser drivers and controllers

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[Tim Williams]

"John Larkin" <jlarkin@highlandtechnology.com> wrote in message
news:hrt3fa5sfcup7n2prrpific56ncghk5v3c@4ax.com...

I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related
to
height (plate-to-plate centers distance) and diameter.

Is that the near-field approximation? Far-field is 1/d^3.

Beats me. Formulas like that show up for simple cases. Although I think
that was for axial rings, not sideways, so it may well be steeper.

As for the far field, come on John, you have whiteboard to scratch it out
on.

I was playing with the idea of mounting a pair of surface-mount drum
cores on opposite sides of a PC board, axially coupled, to make a high
voltage isolated coupler. But the customer went away so I didn't
develop the idea. Side-by-side should work, too.

Axial is probably worse, especially with big copper pads between them.
The end caps tend to shield, and pads don't help.

Wouldn't be too bad with rod cores, but those are big and through hole.

RF spring coils are probably quite good, side by side. Or at funny
angles...
http://seventransistorlabs.com/Radio_20m/Images/RF_Mixer2_sm.jpg
The left hand coils are 4MHz, double tuned and critically coupled.
Remember fields go through angles as they spew off the end of a solenoid,
so the near right angle between the coils is correct with the (admittedly
hard to see) offset axes.

The coupling in this orientation was better than end to end (coil forms
too long) or side by side (not enough k, increased C).

I've taken apart FM and TV tuner boxes that have their coils arranged in
stacks of turns, so they look like one continuous solenoid but made from
separate segments of wire. This allows for taps, secondaries,
input/output coupling, etc. with reasonable k's (> 0.1).

And frequently bent over in goofy ways and gooped up with beeswax, to
adjust the tuning (without spending money on varicaps) then stabilize it
mechanically.

> Things like this are easier to try than to analyze.

Well, the geometry would be pretty easy to set up in FEMM, I expect. But
yeah, it's mostly one of those "suck it and see" situations, whether
simulated or tested. Any analytical formula will be crude at best (hence
my disclaimer).

Tim

--
Seven Transistor Labs
Electrical Engineering Consultation
Website: http://seventransistorlabs.com

 

[John Larkin]

On Sat, 28 Feb 2015 13:10:49 -0600, "Tim Williams"
<tiwill@seventransistorlabs.com> wrote:

"John Larkin" <jlarkin@highlandtechnology.com> wrote in message
news:hrt3fa5sfcup7n2prrpific56ncghk5v3c@4ax.com...
I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related
to
height (plate-to-plate centers distance) and diameter.

Is that the near-field approximation? Far-field is 1/d^3.

Beats me. Formulas like that show up for simple cases. Although I think
that was for axial rings, not sideways, so it may well be steeper.

As for the far field, come on John, you have whiteboard to scratch it out
on.

I was playing with the idea of mounting a pair of surface-mount drum
cores on opposite sides of a PC board, axially coupled, to make a high
voltage isolated coupler. But the customer went away so I didn't
develop the idea. Side-by-side should work, too.

Axial is probably worse, especially with big copper pads between them.
The end caps tend to shield, and pads don't help.

We'd have to punch a hole in the ground plane, for sure. But no
surface creepage to wory about, just 62 mils of epoxy-glass.

Wouldn't be too bad with rod cores, but those are big and through hole.

RF spring coils are probably quite good, side by side. Or at funny
angles...
http://seventransistorlabs.com/Radio_20m/Images/RF_Mixer2_sm.jpg
The left hand coils are 4MHz, double tuned and critically coupled.
Remember fields go through angles as they spew off the end of a solenoid,
so the near right angle between the coils is correct with the (admittedly
hard to see) offset axes.

The coupling in this orientation was better than end to end (coil forms
too long) or side by side (not enough k, increased C).

RF is easy: just resonate both coils. Wideband, specifically pulses,
gets trickier.

I've taken apart FM and TV tuner boxes that have their coils arranged in
stacks of turns, so they look like one continuous solenoid but made from
separate segments of wire. This allows for taps, secondaries,
input/output coupling, etc. with reasonable k's (> 0.1).

And frequently bent over in goofy ways and gooped up with beeswax, to
adjust the tuning (without spending money on varicaps) then stabilize it
mechanically.

Things like this are easier to try than to analyze.

Well, the geometry would be pretty easy to set up in FEMM, I expect. But
yeah, it's mostly one of those "suck it and see" situations, whether
simulated or tested. Any analytical formula will be crude at best (hence
my disclaimer).

Wideband, with ferrites involved, could be bench tested in ~~1% of the
time it would take to set up a sim. Or maybe 0.1%

I got some drum core samples last week, for a different use. Maybe
I'll try a coupled pair. Too bad that customer went away; they piled
on too much crap to make a deal possible.




--

John Larkin Highland Technology, Inc
picosecond timing laser drivers and controllers

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[Klaus Kragelund]

On Saturday, February 28, 2015 at 4:38:40 PM UTC+1, Tim Williams wrote:

You'll primarily have field lines coming off the edges of the top/bottom
plates and arcing through space. The field above and below won't be so
much, so that a vertical offset would be inconvenient (but on a planar
PCB, that doesn't matter!), but placing them side by side should give
quite reasonable "weak" coupling.

I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related to
height (plate-to-plate centers distance) and diameter.

Nice info, thanks

 

[Klaus Kragelund]

On Saturday, February 28, 2015 at 6:20:25 PM UTC+1, John Larkin wrote:

On Sat, 28 Feb 2015 09:38:23 -0600, "Tim Williams"
tiwill@seventransistorlabs.com> wrote:

You'll primarily have field lines coming off the edges of the top/bottom
plates and arcing through space. The field above and below won't be so
much, so that a vertical offset would be inconvenient (but on a planar
PCB, that doesn't matter!), but placing them side by side should give
quite reasonable "weak" coupling.

I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related to
height (plate-to-plate centers distance) and diameter.

Is that the near-field approximation? Far-field is 1/d^3.



Tim

I was playing with the idea of mounting a pair of surface-mount drum
cores on opposite sides of a PC board, axially coupled, to make a high
voltage isolated coupler. But the customer went away so I didn't
develop the idea. Side-by-side should work, too.

That is pretty much the same I am trying out

Other idea is spiral turns, adding a custom ferrite cylinder to reduce the gap, but I do not think that will add much gain, since running in resonance would combat the gap

Cheers

Klaus

 

On Sat, 28 Feb 2015 18:58:38 -0800, John Larkin
<jlarkin@highlandtechnology.com> wrote:

On Sat, 28 Feb 2015 17:07:26 -0800 (PST), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Saturday, February 28, 2015 at 6:20:25 PM UTC+1, John Larkin wrote:
On Sat, 28 Feb 2015 09:38:23 -0600, "Tim Williams"
tiwill@seventransistorlabs.com> wrote:

You'll primarily have field lines coming off the edges of the top/bottom
plates and arcing through space. The field above and below won't be so
much, so that a vertical offset would be inconvenient (but on a planar
PCB, that doesn't matter!), but placing them side by side should give
quite reasonable "weak" coupling.

I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related to
height (plate-to-plate centers distance) and diameter.

Is that the near-field approximation? Far-field is 1/d^3.



Tim

I was playing with the idea of mounting a pair of surface-mount drum
cores on opposite sides of a PC board, axially coupled, to make a high
voltage isolated coupler. But the customer went away so I didn't
develop the idea. Side-by-side should work, too.


That is pretty much the same I am trying out

Other idea is spiral turns, adding a custom ferrite cylinder to reduce the gap, but I do not think that will add much gain, since running in resonance would combat the gap

Cheers

Klaus

I am currently laying out a PCB with a bunch of test circuits, and I'm
going to include some "transformers" just for fun... spiral traces on
parallel PCB layers, some shielded and some not. No ferrites.

One of my coworkers tried it a year or so back. He tried a ferrite
core around the transformer, too. FR4 was too lossy to make it
worthwhile, for power anyway.

 

[John Larkin]

On Sat, 28 Feb 2015 17:07:26 -0800 (PST), Klaus Kragelund
<klauskvik@hotmail.com> wrote:

On Saturday, February 28, 2015 at 6:20:25 PM UTC+1, John Larkin wrote:
On Sat, 28 Feb 2015 09:38:23 -0600, "Tim Williams"
tiwill@seventransistorlabs.com> wrote:

You'll primarily have field lines coming off the edges of the top/bottom
plates and arcing through space. The field above and below won't be so
much, so that a vertical offset would be inconvenient (but on a planar
PCB, that doesn't matter!), but placing them side by side should give
quite reasonable "weak" coupling.

I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related to
height (plate-to-plate centers distance) and diameter.

Is that the near-field approximation? Far-field is 1/d^3.



Tim

I was playing with the idea of mounting a pair of surface-mount drum
cores on opposite sides of a PC board, axially coupled, to make a high
voltage isolated coupler. But the customer went away so I didn't
develop the idea. Side-by-side should work, too.


That is pretty much the same I am trying out

Other idea is spiral turns, adding a custom ferrite cylinder to reduce the gap, but I do not think that will add much gain, since running in resonance would combat the gap

Cheers

Klaus

I am currently laying out a PCB with a bunch of test circuits, and I'm
going to include some "transformers" just for fun... spiral traces on
parallel PCB layers, some shielded and some not. No ferrites.




--

John Larkin Highland Technology, Inc
picosecond timing laser drivers and controllers

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[Klaus Kragelund]

On Sunday, March 1, 2015 at 4:09:31 AM UTC+1, k...@attt.bizz wrote:

On Sat, 28 Feb 2015 18:58:38 -0800, John Larkin
jlarkin@highlandtechnology.com> wrote:

On Sat, 28 Feb 2015 17:07:26 -0800 (PST), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Saturday, February 28, 2015 at 6:20:25 PM UTC+1, John Larkin wrote:
On Sat, 28 Feb 2015 09:38:23 -0600, "Tim Williams"
tiwill@seventransistorlabs.com> wrote:

You'll primarily have field lines coming off the edges of the top/bottom
plates and arcing through space. The field above and below won't be so
much, so that a vertical offset would be inconvenient (but on a planar
PCB, that doesn't matter!), but placing them side by side should give
quite reasonable "weak" coupling.

I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related to
height (plate-to-plate centers distance) and diameter.

Is that the near-field approximation? Far-field is 1/d^3.



Tim

I was playing with the idea of mounting a pair of surface-mount drum
cores on opposite sides of a PC board, axially coupled, to make a high
voltage isolated coupler. But the customer went away so I didn't
develop the idea. Side-by-side should work, too.


That is pretty much the same I am trying out

Other idea is spiral turns, adding a custom ferrite cylinder to reduce the gap, but I do not think that will add much gain, since running in resonance would combat the gap

Cheers

Klaus

I am currently laying out a PCB with a bunch of test circuits, and I'm
going to include some "transformers" just for fun... spiral traces on
parallel PCB layers, some shielded and some not. No ferrites.

One of my coworkers tried it a year or so back. He tried a ferrite
core around the transformer, too. FR4 was too lossy to make it
worthwhile, for power anyway.

I did it once with spiral turns, two layer, got 70% efficiency (at tuned frequency)

But, I am looking another way, since the spiral turns requires 4 layer board and blind vias or some special way to get the inner spiral end out to the rest of the circuit

It's fun. Kind of strange it is not seen often (I have never seen it in a product). It way cheaper than any other method, except for isolation via FR4 capacitances

Cheers

Klaus

 

On Sunday, March 1, 2015 at 11:11:09 AM UTC, Klaus Kragelund wrote:

On Sunday, March 1, 2015 at 4:09:31 AM UTC+1, k...@attt.bizz wrote:
On Sat, 28 Feb 2015 18:58:38 -0800, John Larkin
jlarkin@highlandtechnology.com> wrote:

On Sat, 28 Feb 2015 17:07:26 -0800 (PST), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Saturday, February 28, 2015 at 6:20:25 PM UTC+1, John Larkin wrote:
On Sat, 28 Feb 2015 09:38:23 -0600, "Tim Williams"
tiwill@seventransistorlabs.com> wrote:

You'll primarily have field lines coming off the edges of the top/bottom
plates and arcing through space. The field above and below won't be so
much, so that a vertical offset would be inconvenient (but on a planar
PCB, that doesn't matter!), but placing them side by side should give
quite reasonable "weak" coupling.

I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related to
height (plate-to-plate centers distance) and diameter.

Is that the near-field approximation? Far-field is 1/d^3.



Tim

I was playing with the idea of mounting a pair of surface-mount drum
cores on opposite sides of a PC board, axially coupled, to make a high
voltage isolated coupler. But the customer went away so I didn't
develop the idea. Side-by-side should work, too.


That is pretty much the same I am trying out

Other idea is spiral turns, adding a custom ferrite cylinder to reduce the gap, but I do not think that will add much gain, since running in resonance would combat the gap

Cheers

Klaus

I am currently laying out a PCB with a bunch of test circuits, and I'm
going to include some "transformers" just for fun... spiral traces on
parallel PCB layers, some shielded and some not. No ferrites.

One of my coworkers tried it a year or so back. He tried a ferrite
core around the transformer, too. FR4 was too lossy to make it
worthwhile, for power anyway.

I did it once with spiral turns, two layer, got 70% efficiency (at tuned frequency)

But, I am looking another way, since the spiral turns requires 4 layer board and blind vias or some special way to get the inner spiral end out to the rest of the circuit

like a wire link? Just give the inner ends the required clearances.


NT

 

[Wim Ton]

On Sunday, 1 March 2015 12:11:09 UTC+1, Klaus Kragelund wrote:

On Sunday, March 1, 2015 at 4:09:31 AM UTC+1, k...@attt.bizz wrote:
On Sat, 28 Feb 2015 18:58:38 -0800, John Larkin
jlarkin@highlandtechnology.com> wrote:

On Sat, 28 Feb 2015 17:07:26 -0800 (PST), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Saturday, February 28, 2015 at 6:20:25 PM UTC+1, John Larkin wrote:
On Sat, 28 Feb 2015 09:38:23 -0600, "Tim Williams"
tiwill@seventransistorlabs.com> wrote:

You'll primarily have field lines coming off the edges of the top/bottom
plates and arcing through space. The field above and below won't be so
much, so that a vertical offset would be inconvenient (but on a planar
PCB, that doesn't matter!), but placing them side by side should give
quite reasonable "weak" coupling.

I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related to
height (plate-to-plate centers distance) and diameter.

Is that the near-field approximation? Far-field is 1/d^3.



Tim

I was playing with the idea of mounting a pair of surface-mount drum
cores on opposite sides of a PC board, axially coupled, to make a high
voltage isolated coupler. But the customer went away so I didn't
develop the idea. Side-by-side should work, too.


That is pretty much the same I am trying out

Other idea is spiral turns, adding a custom ferrite cylinder to reduce the gap, but I do not think that will add much gain, since running in resonance would combat the gap

Cheers

Klaus

I am currently laying out a PCB with a bunch of test circuits, and I'm
going to include some "transformers" just for fun... spiral traces on
parallel PCB layers, some shielded and some not. No ferrites.

One of my coworkers tried it a year or so back. He tried a ferrite
core around the transformer, too. FR4 was too lossy to make it
worthwhile, for power anyway.

I did it once with spiral turns, two layer, got 70% efficiency (at tuned frequency)

But, I am looking another way, since the spiral turns requires 4 layer board and blind vias or some special way to get the inner spiral end out to the rest of the circuit

It's fun. Kind of strange it is not seen often (I have never seen it in a product). It way cheaper than any other method, except for isolation via FR4 capacitances

Cheers

Klaus

http://www.elektor.de/professional-lab-power-supply-130234-3

 

[RobertMacy]

On Sat, 28 Feb 2015 18:00:14 -0700, Klaus Kragelund
<klauskvik@hotmail.com> wrote:

On Saturday, February 28, 2015 at 4:38:40 PM UTC+1, Tim Williams wrote:
...snip...
I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related
to
height (plate-to-plate centers distance) and diameter.


Nice info, thanks

Once you have coupling on the axis, you can 'estimate' the coupling to the
side as about 6 dB down.

Still say, get a copy of femm 4.2 and model two cores in axisymmetric
mode. Both to get good estimates and to gain understanding of just how
insidious magnetic fields can be!

You can even 'recreate' Tim's formula.

 

[Klaus Kragelund]

On Sunday, March 1, 2015 at 1:09:38 PM UTC+1, meow...@care2.com wrote:

On Sunday, March 1, 2015 at 11:11:09 AM UTC, Klaus Kragelund wrote:
On Sunday, March 1, 2015 at 4:09:31 AM UTC+1, k...@attt.bizz wrote:
On Sat, 28 Feb 2015 18:58:38 -0800, John Larkin
jlarkin@highlandtechnology.com> wrote:

On Sat, 28 Feb 2015 17:07:26 -0800 (PST), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Saturday, February 28, 2015 at 6:20:25 PM UTC+1, John Larkin wrote:
On Sat, 28 Feb 2015 09:38:23 -0600, "Tim Williams"
tiwill@seventransistorlabs.com> wrote:

You'll primarily have field lines coming off the edges of the top/bottom
plates and arcing through space. The field above and below won't be so
much, so that a vertical offset would be inconvenient (but on a planar
PCB, that doesn't matter!), but placing them side by side should give
quite reasonable "weak" coupling.

I would very, very roughly guess that the radial dropoff has a (x^2 +
a^2)^(-3/2) sort of function, with x being distance and a being related to
height (plate-to-plate centers distance) and diameter.

Is that the near-field approximation? Far-field is 1/d^3.



Tim

I was playing with the idea of mounting a pair of surface-mount drum
cores on opposite sides of a PC board, axially coupled, to make a high
voltage isolated coupler. But the customer went away so I didn't
develop the idea. Side-by-side should work, too.


That is pretty much the same I am trying out

Other idea is spiral turns, adding a custom ferrite cylinder to reduce the gap, but I do not think that will add much gain, since running in resonance would combat the gap

Cheers

Klaus

I am currently laying out a PCB with a bunch of test circuits, and I'm
going to include some "transformers" just for fun... spiral traces on
parallel PCB layers, some shielded and some not. No ferrites.

One of my coworkers tried it a year or so back. He tried a ferrite
core around the transformer, too. FR4 was too lossy to make it
worthwhile, for power anyway.

I did it once with spiral turns, two layer, got 70% efficiency (at tuned frequency)

But, I am looking another way, since the spiral turns requires 4 layer board and blind vias or some special way to get the inner spiral end out to the rest of the circuit

like a wire link? Just give the inner ends the required clearances.

That defeats the purpose. Required creepages are from 8mm to 12mm on outer layers depending on which standard and working voltage is relevant, so the spiral would need to be very large

Cheers

Klaus