Drum Core Stray Magnetic Field

On Sun, 01 Mar 2015 11:37:55 -0700, Tim Williams
<tiwill@seventransistorlabs.com> wrote:

"RobertMacy" <robert.a.macy@gmail.com> wrote in message
news:eek:p.xutcqseu2cx0wh@ajm...
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.

But it doesn't have analytical output, does it?

You could do a plot of given geometry (some fixed ratio of width and
height) vs. distance and get a series of values, then curve fit a similar
equation. Which of course has the same problems as trying to draw up the
approximate equation from first principles, but with a known error bar
and
no need for hefty E&M equations. On the upside, even if you end up with
linear interpolations between data sets, you're doing pretty good.

Tim

No analytical output, only calculated results given a geometry, current,
frequency.

However, femm has LUA script. Using LUA script you can set up a 'sequence'
of analyses, just like you manually putting in values, drawings, etc. That
way, you can 'step through' changes in geometry and frequency. I often use
the script to step through frequency and for each result addend to an
output text file until completely through the spectrum. Then using the
text file as an input use FREE octave, Matlab clone, to plot and curve fit
to the accuracy desired. At least doing that sequence I have a bit more
trust in the formulas.

Also, for physical position, where two objects are moved around in space,
there is a 'trick' to maintaining relative accuracy WITHOUT increasing the
mesh to the point we're talking about a day long set of calculations, but
more like 20 minutes or less. For example, moving an object and changing
the mesh yields around 0.05% accuracy, BUT moving the object but NOT
changing the mesh can yield down into the 1ppm accuracy. but again that's
'relative' accuracy. but makes for much easier curve fitting.
 
On Sun, 01 Mar 2015 10:06:14 -0700, Klaus Kragelund
<klauskvik@hotmail.com> wrote:

...snip.....
I did download it a couple of months ago (due to one of your great posts)

Right now I have another project running, and the spiral experiments are
done on my own time and have pushed off the femm tool due to the
learning curve. But I guess I should get started, will probably pay off
in the long run (the design cycle time is very short compared to
ordering and testing PCBs)


...snip...

To shorten the learning curve. WHICH SMD power inductor? URL to datasheet,
please.

I'll draw up a model that shows you the field around one. That should
shorten the learning curve a bit, much faster to work with an existing
model and poke around rather than try to start from scratch!

Also, will look at modeling your 9 turn spiral. Using femm and lua script
it's easy to create table of induced voltage vs distance etc.


Will check again, but could NOT gain access to my google gmail account.
Google may have upgraded their gmail to the point I can no longer even
access incomding emails. arrrggg! wouldn't be so bad but there's no gain
for any of the 'improvements' they've done so far.
 
On Saturday, February 28, 2015 at 2:18:33 AM UTC-8, Klaus Kragelund wrote:

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

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

The solder pads on this inductor are occluding the field; you will get half or less of the
available flux coupled, just due to the solder pads and the copper footprint they
connect to, if your intent is to use field leakage down into the printed circuit board.
Until you get down to frequencies where the skin depth is larger than the conductor
thickness, conductive material blocks/reflects magnetic field.
 
On Monday, March 2, 2015 at 12:55:18 PM UTC+1, Robert Macy wrote:
On Sun, 01 Mar 2015 10:06:14 -0700, Klaus Kragelund
klauskvik@hotmail.com> wrote:

...snip.....
I did download it a couple of months ago (due to one of your great posts)

Right now I have another project running, and the spiral experiments are
done on my own time and have pushed off the femm tool due to the
learning curve. But I guess I should get started, will probably pay off
in the long run (the design cycle time is very short compared to
ordering and testing PCBs)


...snip...

To shorten the learning curve. WHICH SMD power inductor? URL to datasheet,
please.

I haven't spend much time selecting one yet, but I have a kit from bourns, so could start with the SDR0302-4R7ML:

http://www.mouser.com/ds/2/54/SDR0302-36037.pdf

4.7uH part, SRF 45MHz

I'll draw up a model that shows you the field around one. That should
shorten the learning curve a bit, much faster to work with an existing
model and poke around rather than try to start from scratch!

That would simply be great. Then some good whiskey is going your way :)

I could reverse engineer the inductor, number of turns, Al value, RDC etc. Would that be of benefit?

Also, will look at modeling your 9 turn spiral. Using femm and lua script
it's easy to create table of induced voltage vs distance etc.


Will check again, but could NOT gain access to my google gmail account.
Google may have upgraded their gmail to the point I can no longer even
access incomding emails. arrrggg! wouldn't be so bad but there's no gain
for any of the 'improvements' they've done so far.

Hope it works out for you

Thanks

Klaus
 
On Mon, 2 Mar 2015 15:22:23 -0800 (PST), Klaus Kragelund
<klauskvik@hotmail.com> wrote:

On Monday, March 2, 2015 at 12:55:18 PM UTC+1, Robert Macy wrote:
On Sun, 01 Mar 2015 10:06:14 -0700, Klaus Kragelund
klauskvik@hotmail.com> wrote:

...snip.....
I did download it a couple of months ago (due to one of your great posts)

Right now I have another project running, and the spiral experiments are
done on my own time and have pushed off the femm tool due to the
learning curve. But I guess I should get started, will probably pay off
in the long run (the design cycle time is very short compared to
ordering and testing PCBs)


...snip...

To shorten the learning curve. WHICH SMD power inductor? URL to datasheet,
please.


I haven't spend much time selecting one yet, but I have a kit from bourns, so could start with the SDR0302-4R7ML:

http://www.mouser.com/ds/2/54/SDR0302-36037.pdf

4.7uH part, SRF 45MHz

A pair of those, on opposite sides of a board, would be interesting as
a HV isolator. No vias to worry about.

An LED on the top and a photodiode on the other side might work, too.
Shine the light through the FR4, through a hole in the ground plane.


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
 
"whit3rd" <whit3rd@gmail.com> wrote in message
news:212634b1-1574-470d-a85a-569413f024f2@googlegroups.com...
The solder pads on this inductor are occluding the field; you will get
half or less of the
available flux coupled, just due to the solder pads and the copper
footprint they
connect to, if your intent is to use field leakage down into the printed
circuit board.

Less than that. The ferrite plate shields flux in the axial direction,
which is pretty weak to begin with (helical solenoids coupled end to end
have k < 0.1 or so for winding length and end-to-end gap comparable to the
diameter). So expect the coupling to be pretty bad.

If you could get "drum" type inductors that are mounted perpendicular to
the board, it would look better, at least as long as the height between
plates is greater than the PCB thickness. Dogbone or wirewound chip type
parts may not be too bad, given enough length (say, 1206 and up), but they
lack the width that a drum style would have (which would be like having a
horizontal axis >= 1818 wirewound).

Hmm, ferrite chips placed off the ends of a "drive" inductor could act as
pole pieces, improving coupling (making an "H" pattern, in the top-down
view). It would be automated-assembly-friendly...maybe. The necessities
of small pads and close part-to-part spacing won't go over too well,
though.

Tim

--
Seven Transistor Labs
Electrical Engineering Consultation
Website: http://seventransistorlabs.com
 
On a sunny day (Mon, 2 Mar 2015 18:04:56 -0600) it happened "Tim Williams"
<tiwill@seventransistorlabs.com> wrote in <md2t5j$lje$1@dont-email.me>:

Hmm, ferrite chips placed off the ends of a "drive" inductor could act as
pole pieces, improving coupling (making an "H" pattern, in the top-down
view). It would be automated-assembly-friendly...maybe. The necessities
of small pads and close part-to-part spacing won't go over too well,
though.

Tim

2 potcore halves on each side over a small coil should work better:
--- mounting screw
/ || \
----------- || ---------- potcore
| ------- | || | ------ |
| | //// | | || | | //// | |
============== || ================
||

'////' = coil

same on bottom but flipped vertically

Old rotating transformtr.
At least the field lines would sort of link.
 
On Tuesday, March 3, 2015 at 12:32:03 AM UTC+1, John Larkin wrote:
On Mon, 2 Mar 2015 15:22:23 -0800 (PST), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Monday, March 2, 2015 at 12:55:18 PM UTC+1, Robert Macy wrote:
On Sun, 01 Mar 2015 10:06:14 -0700, Klaus Kragelund
klauskvik@hotmail.com> wrote:

...snip.....
I did download it a couple of months ago (due to one of your great posts)

Right now I have another project running, and the spiral experiments are
done on my own time and have pushed off the femm tool due to the
learning curve. But I guess I should get started, will probably pay off
in the long run (the design cycle time is very short compared to
ordering and testing PCBs)


...snip...

To shorten the learning curve. WHICH SMD power inductor? URL to datasheet,
please.


I haven't spend much time selecting one yet, but I have a kit from bourns, so could start with the SDR0302-4R7ML:

http://www.mouser.com/ds/2/54/SDR0302-36037.pdf

4.7uH part, SRF 45MHz

A pair of those, on opposite sides of a board, would be interesting as
a HV isolator. No vias to worry about.

An LED on the top and a photodiode on the other side might work, too.
Shine the light through the FR4, through a hole in the ground plane.

That's nice.

I tried to find numbers for the transparency of the prepeg, but found nothing.

Cheers

Klaus
 
On Mon, 02 Mar 2015 16:22:23 -0700, Klaus Kragelund
<klauskvik@hotmail.com> wrote:

...snip....

I haven't spend much time selecting one yet, but I have a kit from
bourns, so could start with the SDR0302-4R7ML:

http://www.mouser.com/ds/2/54/SDR0302-36037.pdf

4.7uH part, SRF 45MHz

thanks.

I'll draw up a model that shows you the field around one. That should
shorten the learning curve a bit, much faster to work with an existing
model and poke around rather than try to start from scratch!


That would simply be great. Then some good whiskey is going your way :)

I could reverse engineer the inductor, number of turns, Al value, RDC
etc. Would that be of benefit?

Hey, your on! Yes, please reverse Engineer to get N and wire Awg. You
might be able to do with 'examination' WITHOUT destroying, eh?

I'll duplicate your efforts here by trying to replicatee the spec from a
model.

Will do two, one with NO plane and one with 1 oz. copper footprint that
goes under the inductor. The eddy currents play havoc with the field.

I forgot. What frequency are you planning on communicating with?

..snip....
Will check again, but could NOT gain access to my google gmail account.
Google may have upgraded their gmail to the point I can no longer even
access incomding emails. arrrggg! wouldn't be so bad but there's no
gain
for any of the 'improvements' they've done so far.

Hope it works out for you

Thanks

Klaus

Apparently false alarm. back to 'normal' later in the day. WHEW!

Found the gerbers for the spiral coil, but can't find the 'reader' I use.
sigh. Have a 'bad' pdf document that shows ALL the masks. Bad in that
even at 300% you can barely discern what's going on. The good news is that
you can definitely see VERY FEW vias. And figure out how to do such a
layout yourself.

Robert
 
On Tue, 3 Mar 2015 07:02:31 -0800 (PST), Klaus Kragelund
<klauskvik@hotmail.com> wrote:

On Tuesday, March 3, 2015 at 12:32:03 AM UTC+1, John Larkin wrote:
On Mon, 2 Mar 2015 15:22:23 -0800 (PST), Klaus Kragelund
klauskvik@hotmail.com> wrote:

On Monday, March 2, 2015 at 12:55:18 PM UTC+1, Robert Macy wrote:
On Sun, 01 Mar 2015 10:06:14 -0700, Klaus Kragelund
klauskvik@hotmail.com> wrote:

...snip.....
I did download it a couple of months ago (due to one of your great posts)

Right now I have another project running, and the spiral experiments are
done on my own time and have pushed off the femm tool due to the
learning curve. But I guess I should get started, will probably pay off
in the long run (the design cycle time is very short compared to
ordering and testing PCBs)


...snip...

To shorten the learning curve. WHICH SMD power inductor? URL to datasheet,
please.


I haven't spend much time selecting one yet, but I have a kit from bourns, so could start with the SDR0302-4R7ML:

http://www.mouser.com/ds/2/54/SDR0302-36037.pdf

4.7uH part, SRF 45MHz

A pair of those, on opposite sides of a board, would be interesting as
a HV isolator. No vias to worry about.

An LED on the top and a photodiode on the other side might work, too.
Shine the light through the FR4, through a hole in the ground plane.

That's nice.

I tried to find numbers for the transparency of the prepeg, but found nothing.

Having an old 2-side board and a mini-maglight handy, I'd guess around
5%.


--

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

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
 
On Tuesday, March 3, 2015 at 4:14:25 PM UTC+1, Robert Macy wrote:
On Mon, 02 Mar 2015 16:22:23 -0700, Klaus Kragelund
klauskvik@hotmail.com> wrote:

...snip....

I haven't spend much time selecting one yet, but I have a kit from
bourns, so could start with the SDR0302-4R7ML:

http://www.mouser.com/ds/2/54/SDR0302-36037.pdf

4.7uH part, SRF 45MHz


thanks.

I'll draw up a model that shows you the field around one. That should
shorten the learning curve a bit, much faster to work with an existing
model and poke around rather than try to start from scratch!


That would simply be great. Then some good whiskey is going your way :)

I could reverse engineer the inductor, number of turns, Al value, RDC
etc. Would that be of benefit?

Hey, your on! Yes, please reverse Engineer to get N and wire Awg. You
might be able to do with 'examination' WITHOUT destroying, eh?

I'll duplicate your efforts here by trying to replicatee the spec from a
model.

I only had the SDR0503, 330uH, in stock, so will try that tomorrow and post back here:

https://www.bourns.com/pdfs/sdr0503.pdf



Will do two, one with NO plane and one with 1 oz. copper footprint that
goes under the inductor. The eddy currents play havoc with the field.

I forgot. What frequency are you planning on communicating with?

I don't know yet, I will try different parallel capacitors for the tank and sweep the frequency

Cheers

Klaus
 
On Tue, 03 Mar 2015 15:52:03 -0700, Klaus Kragelund
<klauskvik@hotmail.com> wrote:

....snip....
4.7uH part, SRF 45MHz

Always, happens workon one, and it's NOT the one! Still much applies, hope
you got my models for the 4.7uH


I only had the SDR0503, 330uH, in stock, so will try that tomorrow and
post back here:

https://www.bourns.com/pdfs/sdr0503.pdf



Will do two, one with NO plane and one with 1 oz. copper footprint that
goes under the inductor. The eddy currents play havoc with the field.

I forgot. What frequency are you planning on communicating with?


I don't know yet, I will try different parallel capacitors for the tank
and sweep the frequency

Cheers

Klaus

I coupled two inductors on opposie sides of a 62 mil PCB to find out that
the 4.7uH has K of around 0.084, which is fairly good.
 
On Mon, 02 Mar 2015 16:22:23 -0700, Klaus Kragelund
<klauskvik@hotmail.com> wrote:

....snip...
I haven't spend much time selecting one yet, but I have a kit from
bourns, so could start with the SDR0302-4R7ML:

http://www.mouser.com/ds/2/54/SDR0302-36037.pdf

4.7uH part, SRF 45MHz

I'll draw up a model that shows you the field around one. That should
shorten the learning curve a bit, much faster to work with an existing
model and poke around rather than try to start from scratch!


That would simply be great. Then some good whiskey is going your way :)

Klaus,

I sent you some modeling of that 4.7uH part It appears there are three
layers of 7 turns each layer of 34 Awg wire total of 20.5 turns, half turn
to make the connections to the base. Surprisingly the fields go
EVERYWHERE! Due to the short little profile even a perm of 200 in the core
yields around 5.4 effective perm.

I could NOT get the high frequency operation easily The low Q at 7.96MHz
implies a drop to 2.7uH, or such so quit. But appears that when the core
loses ALL its perm; you still have the AIR core left which provides around
900nH, where self resonance occurs, but too difficult to model that right
now.

Regarding coupling between two: One ontop a board, one on the bottom
couple ataround -.084 which doesn't mean much drop between them. putting
to the side willdrop, 'lifting' off the PCB will drop the signal but
pretty much a lo goes through


I could reverse engineer the inductor, number of turns, Al value, RDC
etc. Would that be of benefit?

..snip...

yes, take a look at the top layer and tell me if you can measure the Awg
AND count the number of turns in the layer.

Robert
 
On Wednesday, March 4, 2015 at 7:08:02 AM UTC+1, Robert Macy wrote:
On Mon, 02 Mar 2015 16:22:23 -0700, Klaus Kragelund
klauskvik@hotmail.com> wrote:

....snip...
I haven't spend much time selecting one yet, but I have a kit from
bourns, so could start with the SDR0302-4R7ML:

http://www.mouser.com/ds/2/54/SDR0302-36037.pdf

4.7uH part, SRF 45MHz

I'll draw up a model that shows you the field around one. That should
shorten the learning curve a bit, much faster to work with an existing
model and poke around rather than try to start from scratch!


That would simply be great. Then some good whiskey is going your way :)


Klaus,

I sent you some modeling of that 4.7uH part It appears there are three
layers of 7 turns each layer of 34 Awg wire total of 20.5 turns, half turn
to make the connections to the base. Surprisingly the fields go
EVERYWHERE! Due to the short little profile even a perm of 200 in the core
yields around 5.4 effective perm.

I could NOT get the high frequency operation easily The low Q at 7.96MHz
implies a drop to 2.7uH, or such so quit. But appears that when the core
loses ALL its perm; you still have the AIR core left which provides around
900nH, where self resonance occurs, but too difficult to model that right
now.

Regarding coupling between two: One ontop a board, one on the bottom
couple ataround -.084 which doesn't mean much drop between them. putting
to the side willdrop, 'lifting' off the PCB will drop the signal but
pretty much a lo goes through

That sounds promishing. I had the SDR0503 in stock, but I will order the SDR0302 4.7uH, so I will measure on the same as you simulated. But, it means it will take a couple of days before I report back.

Please let me know if there are any other special measurements that would be prudent, then I will add those

Cheers

Klaus
 
On Wed, 04 Mar 2015 01:42:11 -0700, Klaus Kragelund
<klauskvik@hotmail.com> wrote:

...snip...
That sounds promishing. I had the SDR0503 in stock, but I will order the
SDR0302 4.7uH, so I will measure on the same as you simulated. But, it
means it will take a couple of days before I report back.

Please let me know if there are any other special measurements that
would be prudent, then I will add those

Cheers

Klaus

Is ok. Here is the modeling technique, so you can do it yourself on the
330uH coil is:

First work at LOW frequency, Rdc and that Inductance (at 1kHz). Take the
physical dimensions. Try to find the 'window' size of the coil. Take a
Wire Gauge Table of diameters/Areas. Assume a Stacking Factor of anywhere
from 0.5 to 0.7 For small coils the SF can be fairly good, even 0.7 Also,
assume a slightly 'less' diameter caused by stretching the wire. The
tension placed when using these automatic winders slightly thins the wire,
shifting the whole manufacturing range to smaller diameter and slightly
higher resistance

Given the Rdc, find the length that makes that resistance for each gauge
of wire near your estimation of wire gauges. Then wrap each length around
the core, any core, These open cores are so poor you can use perms
anywhere from 100 to 2000 and get the SAME inductance. Adjust the coil
window to accomodate. Check to see if you 'filled' the window. If your
guess fills the window, makes the right inductance, and makes physical
sense as it lays on the bobbin, you're done with the basics. You have N
and wire gauge.

Now you'll find you get close to the inductance, but what threw me off for
a while, until I realized the logistics of making a connection to the
coil, was that there had to be a 'half' turn. once I applied an
'adjustment', in this case, ((N-0.5)/N)^2 times the inductance of 21 THEN
I got better than 1 % accuracy to match both the Rdc and the L, and used
the bobbin up completely. My guess is Bournes just takes a bobbin and
fills it with varying sizes of wire, and that's what makes up the family
of available parts.

The skin effect was almost enough to destroy the Q of the coil going up in
frequency, but not quite. The Rwinding went from 0.15 to over 4.5 ohms. I
didn't see an easy way to model the perm of the core to get 'something' at
1kHz down to AIR at SRF of 45MHz. [I tried making the core slightly
conductive which almost works] but since the Q was so bad at 7.96MHz, I
assumed that by 45MHz the core was completely gone, so I used the 900nH
air core value to calculate the 13.4pF parallel capacitance. As I said,
got the basics can work from there.

Anyway, that's the history of modeling the 4.7uH, 0.15ohm inductor. Now,
you can do the same for your 330uH coil.


What really amazed me was how 'leaky' the fields were around that bobbin!
To line up two cores that are only 2.5 mm tall by 3 mm diameter and
separated by 2 mm of PCB and still get 0.084 coupling is pretty good. Or
if NOT wanted, pretty bad. To prevent EMC problems, where that fields can
so easily reach around your 'line filter'; I would NEVER use an open core.
Well, unless you're willing to make your own shield, but why not just buy
a closed core? I assumed NO conductive planes inside the PCB, because you
wanted to have two cores communicate. You can go back and model the
mounting WITH 1 oz copper layers. and see how much, at what frequency,
gets blocked. You'll see the inductance drops, the Rwinding goes up and
the coupling drops. You can write LUA script to do a series of analyses
that will create a table of coupling values vs distance along the axis.
You'll find you get a function that up close is proportional to inverse
square of distance, then as you get further away its more like inverse
cube of distance. Rule of thumb, at 3 diameters distance [or 3 of longest
dimension] the fields are less than 1% of strongest field. Plus you can
replace any complex structure with a simple magnetic dipole. In other
words *if* you're far enough away, you can't tell anything about the
magnetic structure, they all look alike.


Extra measurements? Try to characterize the part from low tohigh frequency
Use whatever steps are easiest, but I prefer log steps, like, 1kHz, 3kHz
10kHz, etc finer or more coarse. At each data point characterize in terms
of XL(f) and R(f) Later you can curve fit to smooth the data using
octave, FREE Matlab clone. Then overlay plots using the FREE femm at the
same frequencies. Working back and forth, get VERY close.
 
On Wednesday, March 4, 2015 at 1:47:10 PM UTC+1, Robert Macy wrote:
On Wed, 04 Mar 2015 01:42:11 -0700, Klaus Kragelund
klauskvik@hotmail.com> wrote:

...snip...
That sounds promishing. I had the SDR0503 in stock, but I will order the
SDR0302 4.7uH, so I will measure on the same as you simulated. But, it
means it will take a couple of days before I report back.

Please let me know if there are any other special measurements that
would be prudent, then I will add those

Cheers

Klaus


Is ok. Here is the modeling technique, so you can do it yourself on the
330uH coil is:

First work at LOW frequency, Rdc and that Inductance (at 1kHz). Take the
physical dimensions. Try to find the 'window' size of the coil. Take a
Wire Gauge Table of diameters/Areas. Assume a Stacking Factor of anywhere
from 0.5 to 0.7 For small coils the SF can be fairly good, even 0.7 Also,
assume a slightly 'less' diameter caused by stretching the wire. The
tension placed when using these automatic winders slightly thins the wire,
shifting the whole manufacturing range to smaller diameter and slightly
higher resistance

Given the Rdc, find the length that makes that resistance for each gauge
of wire near your estimation of wire gauges. Then wrap each length around
the core, any core, These open cores are so poor you can use perms
anywhere from 100 to 2000 and get the SAME inductance. Adjust the coil
window to accomodate. Check to see if you 'filled' the window. If your
guess fills the window, makes the right inductance, and makes physical
sense as it lays on the bobbin, you're done with the basics. You have N
and wire gauge.

Now you'll find you get close to the inductance, but what threw me off for
a while, until I realized the logistics of making a connection to the
coil, was that there had to be a 'half' turn. once I applied an
'adjustment', in this case, ((N-0.5)/N)^2 times the inductance of 21 THEN
I got better than 1 % accuracy to match both the Rdc and the L, and used
the bobbin up completely. My guess is Bournes just takes a bobbin and
fills it with varying sizes of wire, and that's what makes up the family
of available parts.

The skin effect was almost enough to destroy the Q of the coil going up in
frequency, but not quite. The Rwinding went from 0.15 to over 4.5 ohms. I
didn't see an easy way to model the perm of the core to get 'something' at
1kHz down to AIR at SRF of 45MHz. [I tried making the core slightly
conductive which almost works] but since the Q was so bad at 7.96MHz, I
assumed that by 45MHz the core was completely gone, so I used the 900nH
air core value to calculate the 13.4pF parallel capacitance. As I said,
got the basics can work from there.

Anyway, that's the history of modeling the 4.7uH, 0.15ohm inductor. Now,
you can do the same for your 330uH coil.


What really amazed me was how 'leaky' the fields were around that bobbin!
To line up two cores that are only 2.5 mm tall by 3 mm diameter and
separated by 2 mm of PCB and still get 0.084 coupling is pretty good. Or
if NOT wanted, pretty bad. To prevent EMC problems, where that fields can
so easily reach around your 'line filter'; I would NEVER use an open core.
Well, unless you're willing to make your own shield, but why not just buy
a closed core? I assumed NO conductive planes inside the PCB, because you
wanted to have two cores communicate. You can go back and model the
mounting WITH 1 oz copper layers. and see how much, at what frequency,
gets blocked. You'll see the inductance drops, the Rwinding goes up and
the coupling drops. You can write LUA script to do a series of analyses
that will create a table of coupling values vs distance along the axis.
You'll find you get a function that up close is proportional to inverse
square of distance, then as you get further away its more like inverse
cube of distance. Rule of thumb, at 3 diameters distance [or 3 of longest
dimension] the fields are less than 1% of strongest field. Plus you can
replace any complex structure with a simple magnetic dipole. In other
words *if* you're far enough away, you can't tell anything about the
magnetic structure, they all look alike.


Extra measurements? Try to characterize the part from low tohigh frequency
Use whatever steps are easiest, but I prefer log steps, like, 1kHz, 3kHz
10kHz, etc finer or more coarse. At each data point characterize in terms
of XL(f) and R(f) Later you can curve fit to smooth the data using
octave, FREE Matlab clone. Then overlay plots using the FREE femm at the
same frequencies. Working back and forth, get VERY close.

Very nice explanation, thanks.

I will have to find a quiet spot, no children, and give it a try

Cheers

Klaus
 
On Tuesday, March 3, 2015 at 11:52:08 PM UTC+1, Klaus Kragelund wrote:
On Tuesday, March 3, 2015 at 4:14:25 PM UTC+1, Robert Macy wrote:
On Mon, 02 Mar 2015 16:22:23 -0700, Klaus Kragelund
klauskvik@hotmail.com> wrote:

...snip....

I haven't spend much time selecting one yet, but I have a kit from
bourns, so could start with the SDR0302-4R7ML:

http://www.mouser.com/ds/2/54/SDR0302-36037.pdf

4.7uH part, SRF 45MHz


thanks.

I'll draw up a model that shows you the field around one. That should
shorten the learning curve a bit, much faster to work with an existing
model and poke around rather than try to start from scratch!


That would simply be great. Then some good whiskey is going your way :)

I could reverse engineer the inductor, number of turns, Al value, RDC
etc. Would that be of benefit?

Hey, your on! Yes, please reverse Engineer to get N and wire Awg. You
might be able to do with 'examination' WITHOUT destroying, eh?

I'll duplicate your efforts here by trying to replicatee the spec from a
model.


I only had the SDR0503, 330uH, in stock, so will try that tomorrow and post back here:

https://www.bourns.com/pdfs/sdr0503.pdf



Will do two, one with NO plane and one with 1 oz. copper footprint that
goes under the inductor. The eddy currents play havoc with the field.

I forgot. What frequency are you planning on communicating with?


I don't know yet, I will try different parallel capacitors for the tank and sweep the frequency

The first tests:

http://www.electronicsdesign.dk/tmp/SDR0503_331uH test.xlsx

These are done on the SDR0503

When the SDR0302 arrives, I will do more test

Notice that the coupling gets a lot better at lower frequency. I would have thought that the results would have been good at 200kHz, since thats still way off the SRF of 7MHz

Side to side and top to top are good, bottom to bottom (mounted on each side of 1.6mm PCB) is not as good, but not way off

Cheers

Klaus
 
On Thursday, March 5, 2015 at 11:43:32 AM UTC+1, Robert Macy wrote:
On Thu, 05 Mar 2015 01:56:27 -0700, Klaus Kragelund
klauskvik@hotmail.com> wrote:

...snip...
The first tests:

http://www.electronicsdesign.dk/tmp/SDR0503_331uH test.xlsx

...snip....


bad link. did you put the whole line inside <>?

Old excel format:

http://www.electronicsdesign.dk/tmp/SDR0503_331uH test.xls

Cheers

Klaus
 
On Thursday, March 5, 2015 at 11:41:13 AM UTC+1, Robert Macy wrote:
On Thu, 05 Mar 2015 01:51:18 -0700, Klaus Kragelund
klauskvik@hotmail.com> wrote:

....snip....

Very nice explanation, thanks.

I will have to find a quiet spot, no children, and give it a try

Cheers

Klaus

Bad news! I can't get ANY model to fit the data sheet! within the
constraints of physically realizable components.!! I drew the expected
structure, filled the window with 38 AWG wire, wire is getting small
already, got the 330uH, but got 3.8 ohm resistance! too high and there's
no room for 'fatter' wire! Consdiering this component is 'early' in their
product offering you got me what's going on! For higher values of
inductors they have to use smaller wire. I know credit card size modems
contain telco transformer that use 54 Awg wire. maybe they just go right
on up.

If you get one of these parts, measure the 'lid' thickness, the heighth of
the coil window, and the heighth of the bottom 'lid'. They should add to 3
mm. If you have a microscope, see if you can discern the wire gauge.

You got me, I don't know. Maybe the 'cut' into this bobbin is huge. But
even with that, trade-offs stop achieving much.

FWIW, years ago during tenure in the Aerospace Industry, we tested the

Maybe it's a high perm core?

I will measure what you asked for and post back

Cheers

Klaus
 
On Thu, 05 Mar 2015 01:51:18 -0700, Klaus Kragelund
<klauskvik@hotmail.com> wrote:

....snip....

Very nice explanation, thanks.

I will have to find a quiet spot, no children, and give it a try

Cheers

Klaus

Bad news! I can't get ANY model to fit the data sheet! within the
constraints of physically realizable components.!! I drew the expected
structure, filled the window with 38 AWG wire, wire is getting small
already, got the 330uH, but got 3.8 ohm resistance! too high and there's
no room for 'fatter' wire! Consdiering this component is 'early' in their
product offering you got me what's going on! For higher values of
inductors they have to use smaller wire. I know credit card size modems
contain telco transformer that use 54 Awg wire. maybe they just go right
on up.

If you get one of these parts, measure the 'lid' thickness, the heighth of
the coil window, and the heighth of the bottom 'lid'. They should add to 3
mm. If you have a microscope, see if you can discern the wire gauge.

You got me, I don't know. Maybe the 'cut' into this bobbin is huge. But
even with that, trade-offs stop achieving much.

FWIW, years ago during tenure in the Aerospace Industry, we tested the
reliability of little transformers that used little wire. Simple temp
cycling test. Ten cycles, full range temperature cycle over twice a day,
test took five days. If the wire was too small the rransformers failed.
Conclusion: Don't use ANY transformer with wire size smaller than 44 Awg.
The expansion/contraction of the temp cycling and the physical structures
having different rates of expansion, literally broke the wires.
 

Welcome to EDABoard.com

Sponsor

Back
Top