Michael A. Terrell wrote:

Winston wrote:

(...)

Thanks for the P-M cite, Bob. I enjoyed hearing from Alan.

At the risk of turning the 'elegant and beautiful' into the
'byzantine and ugly', what would prevent one from designing
a current-mode PWM controller so that electromagnets with
'too thick' wire could be driven optimally, with just the
proper amount of current for maximum attraction yet not so
high as to cause excessive power dissipation?

I don't understand the conflict with using multiple MOTs as
electromagnets. Every second electromagnet could be driven
with opposite polarity so that no repulsion occurs between
them, yes?


You could connect the primaries in series.

We need to know the 120 Hz Xl of each MOT primary to arrive at
the best topology and drive scheme, yes? I vaguely recall
babbling about measuring the L of a modified MOT primary a
couple weeks ago, for that reason. I would be *shocked* to
learn that any two (modified MOT) primaries could be relied upon
to have exactly the same reactance.

Bob's tests showed reasonable performance driving the *secondary*
with fullwave rectified DC at 120 V input, so I expect that
we need to measure reactance to provide adequate performance
and flux matching.

--Winston

Winston wrote:

Michael A. Terrell wrote:

Winston wrote:

(...)

Thanks for the P-M cite, Bob. I enjoyed hearing from Alan.

At the risk of turning the 'elegant and beautiful' into the
'byzantine and ugly', what would prevent one from designing
a current-mode PWM controller so that electromagnets with
'too thick' wire could be driven optimally, with just the
proper amount of current for maximum attraction yet not so
high as to cause excessive power dissipation?

I don't understand the conflict with using multiple MOTs as
electromagnets. Every second electromagnet could be driven
with opposite polarity so that no repulsion occurs between
them, yes?


You could connect the primaries in series.

We need to know the 120 Hz Xl of each MOT primary to arrive at
the best topology and drive scheme, yes? I vaguely recall
babbling about measuring the L of a modified MOT primary a
couple weeks ago, for that reason. I would be *shocked* to
learn that any two (modified MOT) primaries could be relied upon
to have exactly the same reactance.


If they were made for the same magnetron they sould be quite close.


Bob's tests showed reasonable performance driving the *secondary*
with fullwave rectified DC at 120 V input, so I expect that
we need to measure reactance to provide adequate performance
and flux matching.

--Winston


--
You can't fix stupid. You can't even put a band-aid on it, because it's
Teflon coated.

Snip
I believe it is made in Australia's mexico, Melbourne, east of the
rabbit proof fence. That is the address on the video. I would
love to have on but cannot justify the expense.

Melbourne, Australa's Mexico??

The rabbit proof fence ran east to west and it's about a thousand miles
from Melbourne.

On Wed, 2 Feb 2011 22:54:09 +1100, "Grumpy"
wrote:


Snip
I believe it is made in Australia's mexico, Melbourne, east of the
rabbit proof fence.
I should have said Melbourne, capital city of Australa's Mexico

Victoria is " South of the border, down Mexico way" according to my
Queensland friends.



The rabbit proof fence ran east to west

north to south

and there were about 3 fences
and they only delayed the westward movement of the rabbits.
and Victorians !

and it's about a thousand miles
from Melbourne.

You bit ! VBG

Alan

Winston wrote:
....
I don't understand the conflict with using multiple MOTs as
electromagnets. Every second electromagnet could be driven
with opposite polarity so that no repulsion occurs between
them, yes?

Well, the MOT electromagnets have both poles in the same plane, so you
wouldn't have to alternate. Like you would with magnets with poles on
the opposite ends of a bar or cylinder. Bob

Bob Engelhardt wrote:
Winston wrote:
...
I don't understand the conflict with using multiple MOTs as
electromagnets. Every second electromagnet could be driven
with opposite polarity so that no repulsion occurs between
them, yes?

Well, the MOT electromagnets have both poles in the same plane, so you
wouldn't have to alternate. Like you would with magnets with poles on
the opposite ends of a bar or cylinder. Bob

I'm confused.

Let's say we have two identical MOT electromagnets
connected to the same pulse - D.C. source in the same way.

We discover that the outer 'ears' of the E cores are
both at magnetic 'south'.

Aligned side-by-side, wouldn't these magnets repel
each other?

Followup question: Would that affect the ability of
the magnets to attract the clamping bar, for better
or worse?

I can envision both possibilities:
The attraction to the clamping bar would be improved
because the resulting field distortion would tend to
push the lines of flux outwards from each electromagnet.

The attraction to the clamping bar would be hindered
because of magnetic 'phase cancellation' between the
magnets.

I'm tending towards the latter opinion.


--Winston

On 2011-02-04, Winston wrote:
Bob Engelhardt wrote:

[ ... ]

Well, the MOT electromagnets have both poles in the same plane, so you
wouldn't have to alternate. Like you would with magnets with poles on
the opposite ends of a bar or cylinder. Bob

I'm confused.

Let's say we have two identical MOT electromagnets
connected to the same pulse - D.C. source in the same way.

We discover that the outer 'ears' of the E cores are
both at magnetic 'south'.

Aligned side-by-side, wouldn't these magnets repel
each other?

Yes -- but if properly clamped down, that would not do anything
bad.

Followup question: Would that affect the ability of
the magnets to attract the clamping bar, for better
or worse?

It would weaken the clamping attraction in that area I believe.
If I were making a long electromagnet using Microwave Oven Transformers,
I would probably strip off the windings entirely, butt the cores against
each other as follows (letters are core identifiers, '+' are outer
poles, '-' are inner poles, three cores shown here, but I would use as
many as needed for the length to be covered:

+-----+ +----------+ +-----++-----+ +----------+ +-----++-----+ +----------+ +-----+
| | | | | || | | | | || | | | | |
| | | | | || | | | | || | | | | |
| A+ | | A- | | A+ || B+ | | B- | | B+ || C+ | | C- | | C+ |
| | | | | || | | | | || | | | | |
| | | | | || | | | | || | | | | |
+-----+ +----------+ +-----++-----+ +----------+ +-----++-----+ +----------+ +-----+

And would wind them in a figure eight pattern as follows (numbers are
half turns)
1111111 222222222222 11111111111111 222222222222 11111111111111 22222222222 1111111
+-----+1 2+----------+2 1+-----++-----+1 2+----------+2 1+-----++-----+1 2+----------+2 1+-----+1
| | 1 2 | | 2 1 | || | 1 2 | | 2 1 | || | 1 2 | | 2 1 | |1
| | 1 2 | | 2 1 | || | 1 2 | | 2 1 | || | 1 2 | | 2 1 | |1
| A+ | 1 | A- | 1 | A+ || B+ | 1 | B- | 1 | B+ || C+ | 1 | C- | 1 | C+ |2
| | 2 1 | | 1 2 | || | 2 1 | | 1 2 | || | 2 1 | | 1 2 | |2
| | 2 1 | | 1 2 | || | 2 1 | | 1 2 | || | 2 1 | | 1 2 | |2
+-----+2 1+----------+1 2+-----++-----+2 1+----------+1 2+-----++-----+2 1+----------+1 2+-----+2
222222 11111111111111 22222222222222 11111111111111 22222222222222 11111111111 2222222

N S N S N S N

2 becomes 3 and follows 1, then becomes 4 for the trip back, etc. Since
the end poles of a typical core are half the width of the center core
this would make equal area poles except at the very end, and alternating
poles fairly close together for maximum grip. How many turns would be
needed would be fun to calculate, of course. :-)

I've marked below the poles the alternating North and South
poles of a momentary status. (Or, if you use DC, this could be a stable
status as long as the current is flowing.)

Ideally, you would want to cut apart the core from another
transformer and use it to double the size of the end poles from the
above drawing.

I can envision both possibilities:
The attraction to the clamping bar would be improved
because the resulting field distortion would tend to
push the lines of flux outwards from each electromagnet.

The attraction to the clamping bar would be hindered
because of magnetic 'phase cancellation' between the
magnets.

The center of each pole -- or combined pole -- would have less
flux than the edges, so smaller poles are better -- up to the point
where the flux loops back before it significantly penetrates the
workpiece to reach the clamp bar.

Enjoy,
DoN.

--
Remove oil spill source from e-mail
Email: | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

DoN. Nichols wrote:
On 2011-02-04, wrote:
Bob Engelhardt wrote:

[ ... ]

Well, the MOT electromagnets have both poles in the same plane, so you
wouldn't have to alternate. Like you would with magnets with poles on
the opposite ends of a bar or cylinder. Bob

I'm confused.

Let's say we have two identical MOT electromagnets
connected to the same pulse - D.C. source in the same way.

We discover that the outer 'ears' of the E cores are
both at magnetic 'south'.

Aligned side-by-side, wouldn't these magnets repel
each other?

Yes -- but if properly clamped down, that would not do anything
bad.

We both think that there would be significant weakening of the
clamping action, at least in the 'like poles' area, yes?

Followup question: Would that affect the ability of
the magnets to attract the clamping bar, for better
or worse?

It would weaken the clamping attraction in that area I believe.
If I were making a long electromagnet using Microwave Oven Transformers,
I would probably strip off the windings entirely, butt the cores against
each other as follows (letters are core identifiers, '+' are outer
poles, '-' are inner poles, three cores shown here, but I would use as
many as needed for the length to be covered:

+-----+ +----------+ +-----++-----+ +----------+ +-----++-----+ +----------+ +-----+
| | | | | || | | | | || | | | | |
| | | | | || | | | | || | | | | |
| A+ | | A- | | A+ || B+ | | B- | | B+ || C+ | | C- | | C+ |
| | | | | || | | | | || | | | | |
| | | | | || | | | | || | | | | |
+-----+ +----------+ +-----++-----+ +----------+ +-----++-----+ +----------+ +-----+

And would wind them in a figure eight pattern as follows (numbers are
half turns)
1111111 222222222222 11111111111111 222222222222 11111111111111 22222222222 1111111
+-----+1 2+----------+2 1+-----++-----+1 2+----------+2 1+-----++-----+1 2+----------+2 1+-----+1
| | 1 2 | | 2 1 | || | 1 2 | | 2 1 | || | 1 2 | | 2 1 | |1
| | 1 2 | | 2 1 | || | 1 2 | | 2 1 | || | 1 2 | | 2 1 | |1
| A+ | 1 | A- | 1 | A+ || B+ | 1 | B- | 1 | B+ || C+ | 1 | C- | 1 | C+ |2
| | 2 1 | | 1 2 | || | 2 1 | | 1 2 | || | 2 1 | | 1 2 | |2
| | 2 1 | | 1 2 | || | 2 1 | | 1 2 | || | 2 1 | | 1 2 | |2
+-----+2 1+----------+1 2+-----++-----+2 1+----------+1 2+-----++-----+2 1+----------+1 2+-----+2
222222 11111111111111 22222222222222 11111111111111 22222222222222 11111111111 2222222

N S N S N S N

2 becomes 3 and follows 1, then becomes 4 for the trip back, etc. Since
the end poles of a typical core are half the width of the center core
this would make equal area poles except at the very end, and alternating
poles fairly close together for maximum grip. How many turns would be
needed would be fun to calculate, of course. :-)

Not as much fun as designing that coil winding machine!

I'm getting a headache trying to think of a way to make that
inductor more than one layer deep.

(...)
The center of each pole -- or combined pole -- would have less
flux than the edges, so smaller poles are better -- up to the point
where the flux loops back before it significantly penetrates the
workpiece to reach the clamp bar.

Termed 'leakage flux', yes?

--Winston

On Thu, 03 Feb 2011 17:54:24 -0800, Winston
wrote:

Bob Engelhardt wrote:
Winston wrote:
...
I don't understand the conflict with using multiple MOTs as
electromagnets. Every second electromagnet could be driven
with opposite polarity so that no repulsion occurs between
them, yes?

Well, the MOT electromagnets have both poles in the same plane, so you
wouldn't have to alternate. Like you would with magnets with poles on
the opposite ends of a bar or cylinder. Bob

I'm confused.

Let's say we have two identical MOT electromagnets
connected to the same pulse - D.C. source in the same way.

We discover that the outer 'ears' of the E cores are
both at magnetic 'south'.

Aligned side-by-side, wouldn't these magnets repel
each other?

Followup question: Would that affect the ability of
the magnets to attract the clamping bar, for better
or worse?

I can envision both possibilities:
The attraction to the clamping bar would be improved
because the resulting field distortion would tend to
push the lines of flux outwards from each electromagnet.

The attraction to the clamping bar would be hindered
because of magnetic 'phase cancellation' between the
magnets.

I'm tending towards the latter opinion.


--Winston


Yes they would try to repel each other and this is desirable.

If the second south pole was not present, some of the available
flux would be radiated uselessly sideways. The presence of the
second south pole forces this flux back into the wanted
direction.

This is the case if there is a sjgnificant air gap to the work
piece and the two south poles are close together.

If there is no airgap the flux path generated by each coil is
shortcircuited. There means that there is no significant stray
field so the polarity of the next magnet is unimportant.

If there is a large (many initial airgap lengths) separation
between the magnets relative polarity is again unimportant.

Another way of looking at it is that the two closely adjacent
poles is a single magnet of twice the width that you have
refrained from sawing in half.

Jim

wrote:
On Thu, 03 Feb 2011 17:54:24 -0800, Winston
wrote:

(...)

I can envision both possibilities:
The attraction to the clamping bar would be improved
because the resulting field distortion would tend to
push the lines of flux outwards from each electromagnet.

The attraction to the clamping bar would be hindered
because of magnetic 'phase cancellation' between the
magnets.

I'm tending towards the latter opinion.


--Winston


Yes they would try to repel each other and this is desirable.

If the second south pole was not present, some of the available
flux would be radiated uselessly sideways. The presence of the
second south pole forces this flux back into the wanted
direction.

(...)

So, the first - prize answer doesn't involve MOTs.
(Unless one is building a *very* narrow bender, of course.)

*Ideally speaking* in terms of raw performance, one
would fabricate a stackup of silicon 'transformer iron'
laminations in the form of a single, very thick 'C-E-C' core,
then secure a single rectangular winding in the gap.

Look, for example at the magnetic pole pieces in the
Real Thing: http://www.magnabend.com/advantages.html

Then one would pot the windings and surface-grind the
face of this electromagnet to hinder swarf from
short-circuiting the magnetic path.

By Jove, I think I've got it!

--Winston -- Prolly less expensive to just buy one.

On Fri, 04 Feb 2011 06:34:22 -0800, Winston
wrote:

wrote:
On Thu, 03 Feb 2011 17:54:24 -0800, Winston
wrote:

(...)

I can envision both possibilities:
The attraction to the clamping bar would be improved
because the resulting field distortion would tend to
push the lines of flux outwards from each electromagnet.

The attraction to the clamping bar would be hindered
because of magnetic 'phase cancellation' between the
magnets.

I'm tending towards the latter opinion.


--Winston


Yes they would try to repel each other and this is desirable.

If the second south pole was not present, some of the available
flux would be radiated uselessly sideways. The presence of the
second south pole forces this flux back into the wanted
direction.

(...)

So, the first - prize answer doesn't involve MOTs.
(Unless one is building a *very* narrow bender, of course.)

*Ideally speaking* in terms of raw performance, one
would fabricate a stackup of silicon 'transformer iron'
laminations in the form of a single, very thick 'C-E-C' core,
then secure a single rectangular winding in the gap.

Look, for example at the magnetic pole pieces in the
Real Thing: http://www.magnabend.com/advantages.html

Then one would pot the windings and surface-grind the
face of this electromagnet to hinder swarf from
short-circuiting the magnetic path.

By Jove, I think I've got it!

--Winston -- Prolly less expensive to just buy one.



If you're building from scatch there is no point in usiing
silicon steel laminations. These are only needed to reduce eddy
current loss when using AC excitation.

With pure DC or rectified AC exicitation this is unnecessary -
solid soft iron or mild steel is just as good.

The simplest arrangement is to mill a long slot all the way along
one side of a rectangular bar of mild steel to leave an long U
section.

It can be excited by a single winding round the bottom of the U
or, slightly more efficiently (shorter mean turn length) by a
pair of windings - one on each vertical limb.

Jim

wrote:

(...)

If you're building from scatch there is no point in usiing
silicon steel laminations. These are only needed to reduce eddy
current loss when using AC excitation.

With pure DC or rectified AC exicitation this is unnecessary -
solid soft iron or mild steel is just as good.

I was thinking in terms of relative permeability.
Steel at ca. 100 isn't quite as nice as silicon steel
at ca. 4000.

http://en.wikipedia.org/wiki/Magnetic_permeability#Values_for_some_common_mater ials


The simplest arrangement is to mill a long slot all the way along
one side of a rectangular bar of mild steel to leave an long U
section.

It can be excited by a single winding round the bottom of the U
or, slightly more efficiently (shorter mean turn length) by a
pair of windings - one on each vertical limb.

I like the double sets of pole pieces as shown in the Magnabend ad.
Lots of flux near the 'bending point' is good. Two concentrations
of flux tends to resist part yaw, too.


--Winston

On Thu, 03 Feb 2011 22:51:36 -0800, Winston wrote:
DoN. Nichols wrote:
On 2011-02-04, Winston wrote:
....
Followup question: Would that affect the ability of the magnets to
attract the clamping bar, for better or worse?

It would weaken the clamping attraction in that area I believe.
If I were making a long electromagnet using Microwave Oven
Transformers, I would probably strip off the windings entirely, butt
the cores against each other as follows (letters are core identifiers,
'+' are outer poles, '-' are inner poles, three cores shown here, but I
would use as many as needed for the length to be covered:

[snip ascii graphic]

Not as much fun as designing that coil winding machine!

I'm getting a headache trying to think of a way to make that inductor
more than one layer deep.
....

Suppose you wanted about 20 layers of wire. Get 20 rolls of wire,
and run a flat layer of 20 at once, weaving in and out from left end
to right end of stack of cores, then back from right end to left end,
etc. Probably not easy to do automatically or well.

--
jiw

On 2011-02-04, Winston wrote:
DoN. Nichols wrote:
On 2011-02-04, wrote:

[ ... ]

Aligned side-by-side, wouldn't these magnets repel
each other?

Yes -- but if properly clamped down, that would not do anything
bad.

We both think that there would be significant weakening of the
clamping action, at least in the 'like poles' area, yes?

Yes. (At least I know that *I* believe it.

[ ... ASCII graphics snipped -- quoting levels distort it ... ]

2 becomes 3 and follows 1, then becomes 4 for the trip back, etc. Since
the end poles of a typical core are half the width of the center core
this would make equal area poles except at the very end, and alternating
poles fairly close together for maximum grip. How many turns would be
needed would be fun to calculate, of course. :-)

Not as much fun as designing that coil winding machine!

I'm getting a headache trying to think of a way to make that
inductor more than one layer deep.

Take a piece of rope the right diameter to fit between the
poles. Thread it in the figure-8 shape though the poles. Cut to length
where the ends overlap. Take it out, and make a wooden form which is
perhaps just large enough to leave a 1/2" gap between the ends. Cut
grooves in the form about every inch.

Wind enough turns on the form to duplicate the diameter of the
rope.

Thread lacing tape through the grooves and tighten lace loosely.

Lift free from the form (perhaps bolt-on flanges to keep it
clustered until time to remove).

Push the bundle of wire to the bottom of a slot near one end,
and work your way down to the end and back and around the final end.

Add more lacing tape to tighten the bundle, and then treat with
the lacquer used for motor windings to keep the wires in place in the
row of poles.

I've done something like this years ago, when rewinding a burnt
out three-speed inverted rotor cap run capstan motor for a tape deck.
Except that I did not have the proper lacquer to finish the job. There
were patterns passing through every other slot (with the secondary
winding passing between the slots already used, then more windings going
through every fourth slot, with the second one going through the ones
mind-way between, and finally windings going through every eighth slot,
and through those half-way between those. *Lots* of magnet wire used in
that project. :-)

(...)
The center of each pole -- or combined pole -- would have less
flux than the edges, so smaller poles are better -- up to the point
where the flux loops back before it significantly penetrates the
workpiece to reach the clamp bar.

Termed 'leakage flux', yes?

That sounds like a good term.

Enjoy,
DoN.

--
Remove oil spill source from e-mail
Email: | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

Bob Engelhardt wrote:
Winston wrote:
...
I don't understand the conflict with using multiple MOTs as
electromagnets. Every second electromagnet could be driven
with opposite polarity so that no repulsion occurs between
them, yes?

Well, the MOT electromagnets have both poles in the same plane, so you
wouldn't have to alternate. Like you would with magnets with poles on
the opposite ends of a bar or cylinder. Bob

There's 2 ways to arrange MOT electromagnets (MOTEs) and it occurs to me
that I'm not sure that we're thinking of the same way.

If the top surface of a MOTE has poles N-S-N, then I'm thinking that
multiple MOTEs would be arranged:

N N N
| | |
S S S
| | |
N N N ...

not

N-S-N N-S-N N-S-N ... (or N-S-N S-N-S N-S-N if you prefer)

The first way puts the front pole very close to the bend axis, the
second one puts it about an inch more away (the thickness of the winding.

Bob

Bob Engelhardt wrote:
Bob Engelhardt wrote:
Winston wrote:
...
I don't understand the conflict with using multiple MOTs as
electromagnets. Every second electromagnet could be driven
with opposite polarity so that no repulsion occurs between
them, yes?

Well, the MOT electromagnets have both poles in the same plane, so you
wouldn't have to alternate. Like you would with magnets with poles on
the opposite ends of a bar or cylinder. Bob

There's 2 ways to arrange MOT electromagnets (MOTEs) and it occurs to me
that I'm not sure that we're thinking of the same way.

If the top surface of a MOTE has poles N-S-N, then I'm thinking that
multiple MOTEs would be arranged:

N N N
| | |
S S S
| | |
N N N ...

not

N-S-N N-S-N N-S-N ... (or N-S-N S-N-S N-S-N if you prefer)

The first way puts the front pole very close to the bend axis, the
second one puts it about an inch more away (the thickness of the winding.

So, by extension:

N N N S S S
| | | | | |
S S S N N N
| | | | | |
N N N S S S ...

Yes?

--Winston

James Waldby wrote:

(...)

Suppose you wanted about 20 layers of wire. Get 20 rolls of wire,
and run a flat layer of 20 at once, weaving in and out from left end
to right end of stack of cores, then back from right end to left end,
etc. Probably not easy to do automatically or well.

It'd make a toroid winder look like a yo-yo in comparison.

As Pentagrid mentioned, it'd be 'way simpler and more effective
to lose the MOTEs and create a long E core with a rectangular
winding. Additionally, I think C cores on the ends would be
beneficial, as we see in the Magnabend literature.

Now where did I leave that slab of silicon steel?

--Winston

On Fri, 04 Feb 2011 13:44:05 -0800, Winston
wrote:

wrote:

(...)

If you're building from scatch there is no point in usiing
silicon steel laminations. These are only needed to reduce eddy
current loss when using AC excitation.

With pure DC or rectified AC exicitation this is unnecessary -
solid soft iron or mild steel is just as good.

I was thinking in terms of relative permeability.
Steel at ca. 100 isn't quite as nice as silicon steel
at ca. 4000.

http://en.wikipedia.org/wiki/Magnetic_permeability#Values_for_some_common_mater ials


The simplest arrangement is to mill a long slot all the way along
one side of a rectangular bar of mild steel to leave an long U
section.

It can be excited by a single winding round the bottom of the U
or, slightly more efficiently (shorter mean turn length) by a
pair of windings - one on each vertical limb.

I like the double sets of pole pieces as shown in the Magnabend ad.
Lots of flux near the 'bending point' is good. Two concentrations
of flux tends to resist part yaw, too.


--Winston


The permeability figures quoted in your reference are for
quenched 0.9% carbon steel. This is not a mild steel but a high
carbon steel - file hard when quenched.

Mild steel is typically 0.1% carbon and magnetically pretty
similar to soft iron.

Jim

Winston wrote:
So, by extension:

N N N S S S
| | | | | |
S S S N N N
| | | | | |
N N N S S S ...

Yes?

Right.

wrote:
On Fri, 04 Feb 2011 13:44:05 -0800, Winston
wrote:

wrote:

(...)

If you're building from scatch there is no point in usiing
silicon steel laminations. These are only needed to reduce eddy
current loss when using AC excitation.

With pure DC or rectified AC exicitation this is unnecessary -
solid soft iron or mild steel is just as good.

I was thinking in terms of relative permeability.
Steel at ca. 100 isn't quite as nice as silicon steel
at ca. 4000.

http://en.wikipedia.org/wiki/Magnetic_permeability#Values_for_some_common_mater ials


The simplest arrangement is to mill a long slot all the way along
one side of a rectangular bar of mild steel to leave an long U
section.

It can be excited by a single winding round the bottom of the U
or, slightly more efficiently (shorter mean turn length) by a
pair of windings - one on each vertical limb.

I like the double sets of pole pieces as shown in the Magnabend ad.
Lots of flux near the 'bending point' is good. Two concentrations
of flux tends to resist part yaw, too.

Er. Make that three concentrations vs two concentrations.


--Winston


The permeability figures quoted in your reference are for
quenched 0.9% carbon steel. This is not a mild steel but a high
carbon steel - file hard when quenched.

Mild steel is typically 0.1% carbon and magnetically pretty
similar to soft iron.

Jim

Ah! Good catch!

Still, at least a 2.5:1 advantage to soft iron WRT mild steel:

http://www.microwaves101.com/encyclo...cmaterials.cfm

2.5 times better performance for only 20% more money is probably
worth it. See for example McMaster 89175K27 at $169.71 each vs
8910K487 at $142.31 each. I agree that paying ca. $680 for core
material for each prototype would be beyond the budget for most
hobbyists, though!

--Winston -- Like me, for example!

On Sat, 05 Feb 2011 07:02:32 -0800, Winston
wrote:

wrote:
On Fri, 04 Feb 2011 13:44:05 -0800, Winston
wrote:

wrote:

(...)

If you're building from scatch there is no point in usiing
silicon steel laminations. These are only needed to reduce eddy
current loss when using AC excitation.

With pure DC or rectified AC exicitation this is unnecessary -
solid soft iron or mild steel is just as good.

I was thinking in terms of relative permeability.
Steel at ca. 100 isn't quite as nice as silicon steel
at ca. 4000.

http://en.wikipedia.org/wiki/Magnetic_permeability#Values_for_some_common_mater ials


The simplest arrangement is to mill a long slot all the way along
one side of a rectangular bar of mild steel to leave an long U
section.

It can be excited by a single winding round the bottom of the U
or, slightly more efficiently (shorter mean turn length) by a
pair of windings - one on each vertical limb.

I like the double sets of pole pieces as shown in the Magnabend ad.
Lots of flux near the 'bending point' is good. Two concentrations
of flux tends to resist part yaw, too.

Er. Make that three concentrations vs two concentrations.


--Winston


The permeability figures quoted in your reference are for
quenched 0.9% carbon steel. This is not a mild steel but a high
carbon steel - file hard when quenched.

Mild steel is typically 0.1% carbon and magnetically pretty
similar to soft iron.

Jim

Ah! Good catch!

Still, at least a 2.5:1 advantage to soft iron WRT mild steel:

http://www.microwaves101.com/encyclo...cmaterials.cfm

2.5 times better performance for only 20% more money is probably
worth it. See for example McMaster 89175K27 at $169.71 each vs
8910K487 at $142.31 each. I agree that paying ca. $680 for core
material for each prototype would be beyond the budget for most
hobbyists, though!

--Winston -- Like me, for example!


The 2.5:1 permeability difference is only significant at
extremely small air gap and medium flux densities.

With 5"iron length and .005" residual air gap the working
permeability of the iron circuit drops to 877 for silicon iron
and 667 for mild steel.

The above assumes that the iron is working somewhere near its
maximum permeability flux density. Typical electromagnets work at
higher flux densities where the permeability is starting to drop.

Kay and Laby "Physical and Chemical Constants" shows how this
varies for both 3% oriented silicon steel and mild steel.

Excitation Oersteads 10 50 500

Mild steel flux density 14,000 17,000 21,000

Silicon steel flux density 17,800 19,000 20,300

Jim

wrote:
On Sat, 05 Feb 2011 07:02:32 -0800, Winston
wrote:

wrote:
On Fri, 04 Feb 2011 13:44:05 -0800, Winston
wrote:

wrote:

(...)

If you're building from scatch there is no point in usiing
silicon steel laminations. These are only needed to reduce eddy
current loss when using AC excitation.

With pure DC or rectified AC exicitation this is unnecessary -
solid soft iron or mild steel is just as good.

I was thinking in terms of relative permeability.
Steel at ca. 100 isn't quite as nice as silicon steel
at ca. 4000.

http://en.wikipedia.org/wiki/Magnetic_permeability#Values_for_some_common_mater ials


The simplest arrangement is to mill a long slot all the way along
one side of a rectangular bar of mild steel to leave an long U
section.

It can be excited by a single winding round the bottom of the U
or, slightly more efficiently (shorter mean turn length) by a
pair of windings - one on each vertical limb.

I like the double sets of pole pieces as shown in the Magnabend ad.
Lots of flux near the 'bending point' is good. Two concentrations
of flux tends to resist part yaw, too.

Er. Make that three concentrations vs two concentrations.


--Winston


The permeability figures quoted in your reference are for
quenched 0.9% carbon steel. This is not a mild steel but a high
carbon steel - file hard when quenched.

Mild steel is typically 0.1% carbon and magnetically pretty
similar to soft iron.

Jim

Ah! Good catch!

Still, at least a 2.5:1 advantage to soft iron WRT mild steel:

http://www.microwaves101.com/encyclo...cmaterials.cfm

2.5 times better performance for only 20% more money is probably
worth it. See for example McMaster 89175K27 at $169.71 each vs
8910K487 at $142.31 each. I agree that paying ca. $680 for core
material for each prototype would be beyond the budget for most
hobbyists, though!

--Winston-- Like me, for example!


The 2.5:1 permeability difference is only significant at
extremely small air gap and medium flux densities.

With 5"iron length and .005" residual air gap the working
permeability of the iron circuit drops to 877 for silicon iron
and 667 for mild steel.

The above assumes that the iron is working somewhere near its
maximum permeability flux density. Typical electromagnets work at
higher flux densities where the permeability is starting to drop.

Kay and Laby "Physical and Chemical Constants" shows how this
varies for both 3% oriented silicon steel and mild steel.

Excitation Oersteads 10 50 500

Mild steel flux density 14,000 17,000 21,000

Silicon steel flux density 17,800 19,000 20,300

Do you happen to know where 'malleable iron' fits in
this chart?

Good information. Thanks Jim!

--Winston

On Sat, 05 Feb 2011 19:13:04 -0800, Winston
wrote:

Snip
Er. Make that three concentrations vs two concentrations.


--Winston


The permeability figures quoted in your reference are for
quenched 0.9% carbon steel. This is not a mild steel but a high
carbon steel - file hard when quenched.

Mild steel is typically 0.1% carbon and magnetically pretty
similar to soft iron.

Jim

Ah! Good catch!

Still, at least a 2.5:1 advantage to soft iron WRT mild steel:

http://www.microwaves101.com/encyclo...cmaterials.cfm

2.5 times better performance for only 20% more money is probably
worth it. See for example McMaster 89175K27 at $169.71 each vs
8910K487 at $142.31 each. I agree that paying ca. $680 for core
material for each prototype would be beyond the budget for most
hobbyists, though!

--Winston-- Like me, for example!


The 2.5:1 permeability difference is only significant at
extremely small air gap and medium flux densities.

With 5"iron length and .005" residual air gap the working
permeability of the iron circuit drops to 877 for silicon iron
and 667 for mild steel.

The above assumes that the iron is working somewhere near its
maximum permeability flux density. Typical electromagnets work at
higher flux densities where the permeability is starting to drop.

Kay and Laby "Physical and Chemical Constants" shows how this
varies for both 3% oriented silicon steel and mild steel.

Excitation Oersteads 10 50 500

Mild steel flux density 14,000 17,000 21,000

Silicon steel flux density 17,800 19,000 20,300

Do you happen to know where 'malleable iron' fits in
this chart?

Good information. Thanks Jim!

--Winston


Malleable iron is a heat treated variant of cast iron and since
this is similar to an annealing processs I would expect similar
or somewhat better permeablity.

Oersteds 10 50 500

Cast iron `5,000 8,500 14,000

Annealed 1% Carbon Steel 6,500 16,200 20,200

Swedish Soft Iron 14,800 17,000 21,000

Jim

In article ,
Winston wrote:

wrote:
On Fri, 04 Feb 2011 13:44:05 -0800, Winston
wrote:

wrote:

(...)

If you're building from scatch there is no point in usiing
silicon steel laminations. These are only needed to reduce eddy
current loss when using AC excitation.

With pure DC or rectified AC exicitation this is unnecessary -
solid soft iron or mild steel is just as good.

I was thinking in terms of relative permeability.
Steel at ca. 100 isn't quite as nice as silicon steel
at ca. 4000.

http://en.wikipedia.org/wiki/Magneti...r_some_common_
materials


The simplest arrangement is to mill a long slot all the way along
one side of a rectangular bar of mild steel to leave an long U
section.

It can be excited by a single winding round the bottom of the U
or, slightly more efficiently (shorter mean turn length) by a
pair of windings - one on each vertical limb.

I like the double sets of pole pieces as shown in the Magnabend ad.
Lots of flux near the 'bending point' is good. Two concentrations
of flux tends to resist part yaw, too.

Er. Make that three concentrations vs two concentrations.


--Winston


The permeability figures quoted in your reference are for
quenched 0.9% carbon steel. This is not a mild steel but a high
carbon steel - file hard when quenched.

Mild steel is typically 0.1% carbon and magnetically pretty
similar to soft iron.

Jim

Ah! Good catch!

Still, at least a 2.5:1 advantage to soft iron WRT mild steel:

http://www.microwaves101.com/encyclo...cmaterials.cfm

2.5 times better performance for only 20% more money is probably
worth it. See for example McMaster 89175K27 at $169.71 each vs
8910K487 at $142.31 each. I agree that paying ca. $680 for core
material for each prototype would be beyond the budget for most
hobbyists, though!

I would be tempted to fabricate the magnetic circuit components by
bolting together four pieces of 1018 steel rectangles: one wide 0.5"
thick piece to form the bottom, two 0.5" thick pieces to form the
outsides, and one 1.0" thick piece to form the center, roughly following
the outline shown in the MagnaBend literature.

This will be a considerable savings on steel material, and on machining
effort.

Bolting-together would be accomplished using hex socket flat head
machine screws through the bottom piece screwed into drilled and tapped
holes in the steel center and side pieces.


Joe Gwinn.

wrote:
On Sat, 05 Feb 2011 19:13:04 -0800, Winston
wrote:

Snip
Er. Make that three concentrations vs two concentrations.


--Winston


The permeability figures quoted in your reference are for
quenched 0.9% carbon steel. This is not a mild steel but a high
carbon steel - file hard when quenched.

Mild steel is typically 0.1% carbon and magnetically pretty
similar to soft iron.

Jim

Ah! Good catch!

Still, at least a 2.5:1 advantage to soft iron WRT mild steel:

http://www.microwaves101.com/encyclo...cmaterials.cfm

2.5 times better performance for only 20% more money is probably
worth it. See for example McMaster 89175K27 at $169.71 each vs
8910K487 at $142.31 each. I agree that paying ca. $680 for core
material for each prototype would be beyond the budget for most
hobbyists, though!

--Winston-- Like me, for example!


The 2.5:1 permeability difference is only significant at
extremely small air gap and medium flux densities.

With 5"iron length and .005" residual air gap the working
permeability of the iron circuit drops to 877 for silicon iron
and 667 for mild steel.

The above assumes that the iron is working somewhere near its
maximum permeability flux density. Typical electromagnets work at
higher flux densities where the permeability is starting to drop.

Kay and Laby "Physical and Chemical Constants" shows how this
varies for both 3% oriented silicon steel and mild steel.

Excitation Oersteads 10 50 500

Mild steel flux density 14,000 17,000 21,000

Silicon steel flux density 17,800 19,000 20,300

Do you happen to know where 'malleable iron' fits in
this chart?

Good information. Thanks Jim!

--Winston


Malleable iron is a heat treated variant of cast iron and since
this is similar to an annealing processs I would expect similar
or somewhat better permeablity.

Oersteds 10 50 500

Cast iron `5,000 8,500 14,000

Annealed 1% Carbon Steel 6,500 16,200 20,200

Swedish Soft Iron 14,800 17,000 21,000


Interesting! I hadn't heard of Swedish Soft Iron before.

Thanks!

--Winston

Joseph Gwinn wrote:

(...)
I would be tempted to fabricate the magnetic circuit components by
bolting together four pieces of 1018 steel rectangles: one wide 0.5"
thick piece to form the bottom, two 0.5" thick pieces to form the
outsides, and one 1.0" thick piece to form the center, roughly following
the outline shown in the MagnaBend literature.

This will be a considerable savings on steel material, and on machining
effort.

Bolting-together would be accomplished using hex socket flat head
machine screws through the bottom piece screwed into drilled and tapped
holes in the steel center and side pieces.

I'm concerned about the three additional gaps:
Between each vertical and the base plate
Between the center pole an the base plate
Magnetic flux falls as the cube root of distance, so anything
we can do to reduce reluctance is good.

I'd be tempted to punch many 'E' laminations and
stack them sideways. That design wouldn't have any
additional gaps.
On the ends, I would attach 'C' laminations.
The laminations would attach together as we see in
some older transformers: a long insulated machine
screw with shoulder washers in 4 corners.

http://www.alliedelec.com/Images/Pro...8482_large.jpg

--Winston

On Sun, 06 Feb 2011 15:55:54 -0500, Joseph Gwinn
wrote:

Snip
I would be tempted to fabricate the magnetic circuit components by
bolting together four pieces of 1018 steel rectangles: one wide 0.5"
thick piece to form the bottom, two 0.5" thick pieces to form the
outsides, and one 1.0" thick piece to form the center, roughly following
the outline shown in the MagnaBend literature.

This will be a considerable savings on steel material, and on machining
effort.

Bolting-together would be accomplished using hex socket flat head
machine screws through the bottom piece screwed into drilled and tapped
holes in the steel center and side pieces.


Joe Gwinn.


Your fabricated construction should be fine. Although there are
two additional residual air gaps in the magnetic circuit these
are hard bolted gaps dropping to zero near every bolt and would
not add significantly to the series of four main residual gaps
that occur between the pole pieces and the work piece.

A further simplification would be to revert to the U
configuration. but with a single coil on the lower bar of the U
(or on the U leg remote from the bend line). This is slighly less
efficient than a coil on each leg because of the longer mean
turn length but this is more than compensated for by the ability
to locate the full sized pole piece where it matters most -
close to the bend line.

The width of the U is an interesting free variable. It makes
little difference to the total reluctance but a longer lower bar
both increases the available winding area and exerts the bending
force at a longer lever arm. On this basis - the longer the
better. For this application a gap in the U about equal to the
limb width looks like a reasonable choice.

There is no special compensating advantage in a longer limb
length so these should only be long enough to give sufficient
winding area.For the above U size, limbs long enough to give an
approximately square winding area would be appropriate.


Jim

wrote:
On Sun, 06 Feb 2011 15:55:54 -0500, Joseph Gwinn
wrote:

Snip
I would be tempted to fabricate the magnetic circuit components by
bolting together four pieces of 1018 steel rectangles: one wide 0.5"
thick piece to form the bottom, two 0.5" thick pieces to form the
outsides, and one 1.0" thick piece to form the center, roughly following
the outline shown in the MagnaBend literature.

This will be a considerable savings on steel material, and on machining
effort.

Bolting-together would be accomplished using hex socket flat head
machine screws through the bottom piece screwed into drilled and tapped
holes in the steel center and side pieces.


Joe Gwinn.


Your fabricated construction should be fine. Although there are
two additional residual air gaps in the magnetic circuit these
are hard bolted gaps dropping to zero near every bolt and would
not add significantly to the series of four main residual gaps
that occur between the pole pieces and the work piece.

A further simplification would be to revert to the U
configuration. but with a single coil on the lower bar of the U
(or on the U leg remote from the bend line). This is slighly less
efficient than a coil on each leg because of the longer mean
turn length but this is more than compensated for by the ability
to locate the full sized pole piece where it matters most -
close to the bend line.

The 'C-E-C' core has another advantage over the 'U' core
aside from reduced reluctance.

http://www.magnabend.com/advantages.html

Notice how the 'C' core features provide clamping force in the
left and right extreme sides of the pole piece.
Those four corners would have much less clamping force in a 'U
core configuration.

--Winston

On Mon, 07 Feb 2011 06:10:17 -0800, Winston
wrote:

wrote:
On Sun, 06 Feb 2011 15:55:54 -0500, Joseph Gwinn
wrote:

Snip
I would be tempted to fabricate the magnetic circuit components by
bolting together four pieces of 1018 steel rectangles: one wide 0.5"
thick piece to form the bottom, two 0.5" thick pieces to form the
outsides, and one 1.0" thick piece to form the center, roughly following
the outline shown in the MagnaBend literature.

This will be a considerable savings on steel material, and on machining
effort.

Bolting-together would be accomplished using hex socket flat head
machine screws through the bottom piece screwed into drilled and tapped
holes in the steel center and side pieces.


Joe Gwinn.


Your fabricated construction should be fine. Although there are
two additional residual air gaps in the magnetic circuit these
are hard bolted gaps dropping to zero near every bolt and would
not add significantly to the series of four main residual gaps
that occur between the pole pieces and the work piece.

A further simplification would be to revert to the U
configuration. but with a single coil on the lower bar of the U
(or on the U leg remote from the bend line). This is slighly less
efficient than a coil on each leg because of the longer mean
turn length but this is more than compensated for by the ability
to locate the full sized pole piece where it matters most -
close to the bend line.

The 'C-E-C' core has another advantage over the 'U' core
aside from reduced reluctance.

http://www.magnabend.com/advantages.html

Notice how the 'C' core features provide clamping force in the
left and right extreme sides of the pole piece.
Those four corners would have much less clamping force in a 'U
core configuration.

--Winston

A correctly proportioned U core has the same reluctance as the
equivalant E core.

The vertical clamping force would be identical.

As discussed in the penultimate paragraph (omitted in your reply)
the bending force depends on the effective lever arm length which
is a free choice determined by the chosen width of the U gap.

With a U core, the flux is delivered by the outer pole pieces.
For the same overall width, the U core delivers a greater bending
moment than an E core because half the E core flux goes to the
centre limb which has a shorter bending moment

These are all second order effects and will make little practical
difference. Ease of manufacture and cost of materials are more
important. The E core construction used by magnabend is
ineresting but I suspect that this was chosen for its mechanical
convenience because it completely surrounds the coil and
protects it from even the most ham fisted mechanic.

Jim

wrote:
On Mon, 07 Feb 2011 06:10:17 -0800, Winston
wrote:

wrote:
On Sun, 06 Feb 2011 15:55:54 -0500, Joseph Gwinn
wrote:

Snip
I would be tempted to fabricate the magnetic circuit components by
bolting together four pieces of 1018 steel rectangles: one wide 0.5"
thick piece to form the bottom, two 0.5" thick pieces to form the
outsides, and one 1.0" thick piece to form the center, roughly following
the outline shown in the MagnaBend literature.

This will be a considerable savings on steel material, and on machining
effort.

Bolting-together would be accomplished using hex socket flat head
machine screws through the bottom piece screwed into drilled and tapped
holes in the steel center and side pieces.


Joe Gwinn.


Your fabricated construction should be fine. Although there are
two additional residual air gaps in the magnetic circuit these
are hard bolted gaps dropping to zero near every bolt and would
not add significantly to the series of four main residual gaps
that occur between the pole pieces and the work piece.

A further simplification would be to revert to the U
configuration. but with a single coil on the lower bar of the U
(or on the U leg remote from the bend line). This is slighly less
efficient than a coil on each leg because of the longer mean
turn length but this is more than compensated for by the ability
to locate the full sized pole piece where it matters most -
close to the bend line.

The 'C-E-C' core has another advantage over the 'U' core
aside from reduced reluctance.

http://www.magnabend.com/advantages.html

Notice how the 'C' core features provide clamping force in the
left and right extreme sides of the pole piece.
Those four corners would have much less clamping force in a 'U
core configuration.

--Winston

A correctly proportioned U core has the same reluctance as the
equivalant E core.

True for an electromagnet assembled using factory fixtures
and laminating techniques. I'm trying to envision how to
build a (bottom coil) U - core electromagnet without relying
on an 'L-I' core assembly and it's additional gap, using
'hobbiest' level tools and techniques. I agree that a
U-core electromagnet assembled with coils on the vertical
arms could be built by a hobbyist without any additional
gaps. Making *that* electromagnet in such a way as to not
interfere with the bending surface on the front of the machine
would be an interesting exercise.

The vertical clamping force would be identical.

I agree. And I also agree that the distribution of the
clamping force in the U electromagnet is superior to
that of the E electromagnet.

As discussed in the penultimate paragraph (omitted in your reply)
the bending force depends on the effective lever arm length which
is a free choice determined by the chosen width of the U gap.

For the relatively short distance from pole - to - pole and
given the equal clamping force, the difference might not be
very significant, particularly for ferrous workpieces.

Where did I put that magnetic FEA tool?


With a U core, the flux is delivered by the outer pole pieces.
For the same overall width, the U core delivers a greater bending
moment than an E core because half the E core flux goes to the
centre limb which has a shorter bending moment

These are all second order effects and will make little practical
difference. Ease of manufacture and cost of materials are more
important. The E core construction used by magnabend is
ineresting but I suspect that this was chosen for its mechanical
convenience because it completely surrounds the coil and
protects it from even the most ham fisted mechanic.

Engineering is compromise. The 'C-E-C' core allows one to assemble
the electromagnet (with square sides and bottom) and with
adequate clamping force. The U core has superior flux distribution
due to it's longer magnetic lever arm and might enable a higher
capacity bender.

Another possibility is a U core with a single winding on the back
vertical so that the front pole could define the entire front lip
and corners of the bender without interference.
*That* design has no additional gaps and could be made by a
hobbyist.

--Winston

Winston wrote:
True for an electromagnet assembled using factory fixtures
and laminating techniques. I'm trying to envision how to
build a (bottom coil) U - core electromagnet without relying
on an 'L-I' core assembly and it's additional gap, using
'hobbiest' level tools and techniques. I agree that a
U-core electromagnet assembled with coils on the vertical
arms could be built by a hobbyist without any additional
gaps. Making *that* electromagnet in such a way as to not
interfere with the bending surface on the front of the machine
would be an interesting exercise.


If we don't take into "too much" account the efficiency and
cost of wire, The entire coil could be on the rear leg ie.
fill the space "inside" with the turns.
I don't quite get this C E C configuration you are talking
about. I know all about I E laminations, I've built a few
transformers in past years (when I was doing Ham Radio bit)
and even "Pot Cores".
...Lew...

Lewis Hartswick wrote:
Winston wrote:
True for an electromagnet assembled using factory fixtures
and laminating techniques. I'm trying to envision how to
build a (bottom coil) U - core electromagnet without relying
on an 'L-I' core assembly and it's additional gap, using
'hobbiest' level tools and techniques. I agree that a
U-core electromagnet assembled with coils on the vertical
arms could be built by a hobbyist without any additional
gaps. Making *that* electromagnet in such a way as to not
interfere with the bending surface on the front of the machine
would be an interesting exercise.


If we don't take into "too much" account the efficiency and
cost of wire, The entire coil could be on the rear leg ie.
fill the space "inside" with the turns.

Two great minds, Lew.

At the end of my recent "Novel Post", I conclude the same thing.
See the last paragraph, starting "Another possibility..".

I don't quite get this C E C configuration you are talking
about. I know all about I E laminations, I've built a few
transformers in past years (when I was doing Ham Radio bit)
and even "Pot Cores".

I remember "Pot Cores"! Where did I put the shim?

Please see:
http://www.magnabend.com/media/images/advantages4.gif

Looking at the pole piece, it's clear that the entire
middle of the electromagnet is built on an 'E' core.
Note that the middle leg of the 'E' does not extend to the
leftmost edge of the electromagnet. That is where the
winding makes it's turn.
The outer vertical legs *do* extend to the edges of the
electromagnet, though.

*That* extension has the cross-sectional resemblance to
the letter 'C' so that is what I'm on about, because I
would have totally puzzled some by referring to it as a
"[" lamination.

--Winston

Winston wrote:

At the end of my recent "Novel Post", I conclude the same thing.
See the last paragraph, starting "Another possibility..".

I don't quite get this C E C configuration you are talking
about. I know all about I E laminations, I've built a few
transformers in past years (when I was doing Ham Radio bit)
and even "Pot Cores".

I remember "Pot Cores"! Where did I put the shim?

Please see:
http://www.magnabend.com/media/images/advantages4.gif

Looking at the pole piece, it's clear that the entire
middle of the electromagnet is built on an 'E' core.
Note that the middle leg of the 'E' does not extend to the
leftmost edge of the electromagnet. That is where the
winding makes it's turn.
The outer vertical legs *do* extend to the edges of the
electromagnet, though.

*That* extension has the cross-sectional resemblance to
the letter 'C' so that is what I'm on about, because I
would have totally puzzled some by referring to it as a
"[" lamination.

--Winston
Aw So. Now I understand. Thanks.
...lew...

Lewis Hartswick wrote:

(...)

Aw So. Now I understand. Thanks.

Don't mention it.
Please post pictures of your prototype to the
dropbox when available?

--Winston

In article ,
wrote:

On Sun, 06 Feb 2011 15:55:54 -0500, Joseph Gwinn
wrote:

Snip
I would be tempted to fabricate the magnetic circuit components by
bolting together four pieces of 1018 steel rectangles: one wide 0.5"
thick piece to form the bottom, two 0.5" thick pieces to form the
outsides, and one 1.0" thick piece to form the center, roughly following
the outline shown in the MagnaBend literature.

This will be a considerable savings on steel material, and on machining
effort.

Bolting-together would be accomplished using hex socket flat head
machine screws through the bottom piece screwed into drilled and tapped
holes in the steel center and side pieces.


Joe Gwinn.


Your fabricated construction should be fine. Although there are
two additional residual air gaps in the magnetic circuit these
are hard bolted gaps dropping to zero near every bolt and would
not add significantly to the series of four main residual gaps
that occur between the pole pieces and the work piece.

This was my thinking as well. I suppose we could machine the interface
surfaces flat to reduce the residual gaps, but I would start with
as-rolled 1018 rectangles. And it's easier to use too many fasteners
than to machine a flat 24" long.

For the obsessive, one can use HiSpot blue and a scraper to achieve
near-perfect mating of the surfaces, as perfect as patience permits.

The middle bar would be shorter than the two side bars, so the winding
has space to make the turns at the two ends.


Do we know if the real MagnaBend uses a fabricated magnetic assembly, or
machined, or what? It is hard to tell from the pictures, but it kinda
looked fabricated to me. Which wouldn't surprise me, given the cost
difference between fabrication (perhaps by welding versus bolting)
compared to plowing two deep grooves in a 3" by 4" steel rectangle on a
large horizontal mill.


A further simplification would be to revert to the U
configuration. but with a single coil on the lower bar of the U
(or on the U leg remote from the bend line). This is slighly less
efficient than a coil on each leg because of the longer mean
turn length but this is more than compensated for by the ability
to locate the full sized pole piece where it matters most -
close to the bend line.

The width of the U is an interesting free variable. It makes
little difference to the total reluctance but a longer lower bar
both increases the available winding area and exerts the bending
force at a longer lever arm. On this basis - the longer the
better. For this application a gap in the U about equal to the
limb width looks like a reasonable choice.

There is no special compensating advantage in a longer limb
length so these should only be long enough to give sufficient
winding area.For the above U size, limbs long enough to give an
approximately square winding area would be appropriate.

The original MagnaBend patent shows the U configuration described above,
but the production units all seem to use the E configuration, so I
assumed that the E was superior in some respect, not necessarily
magnetic. For one thing, the E configuration with aluminum plates
filling the space between the arms of the E (and steel conduit to the
electronics box) offers near complete protection of the windings from
shop accidents.

Joe Gwinn