Coil improvements

[hobbes_lives]

In addition to that, I suggested that eliminating the contact between the cross section shielding could minimize eddy currents.

That's the part I'm curious about. I'm curious of how much eddy current exists in that direction, given that it is 90 degrees to the coil's lines of flux. There is likely some, due to not being at exactly 90 degrees, but there shouldn't be much. Perhaps an experiment is in order...

Here is a simple test to perform.

Make a single loop of wire 3/16" ID and solder it together to form a conductive ring.

Make another loop the same size but not soldered together and not with their ends touching.

Move each loop under and close the PI coil with a 10us delay and tell us what the response is from each loop.

I'm quite familiar with a PI's response to open vs closed loops. It's the same reason why a closed hoop earring can be detected quite deep, whereas an open hoop can only be detected as very shallow depths. It's all about closing the loop for eddy currents to flow, thus lengthening the decay time of the received signal.

Now, I propose a second part to the experiment. Turn the same closed loop of wire 90 degrees to the face of the coil and see what happens then. What's the PI's response in this case? (I know you already know the answer.)

[bbsailor]

Kyle,

Perhaps I have been unclear. Maybe I should have started another thread with the coil proposal, but I tried to integrate it into an existing topic.

You said: "Perhaps I'm overlooking something obvious, or am not clear about the proposed idea?"

The proposed idea is simply to post coil parameters as I suggested.

In addition to that, I suggested that eliminating the contact between the cross section shielding could minimize eddy currents.

Here is a simple test to perform.

Make a single loop of wire 3/16" ID and solder it together to form a conductive ring.

Make another loop the same size but not soldered together and not with their ends touching.

Move each loop under and close the PI coil with a 10us delay and tell us what the response is from each loop.

bbsailor

[hobbes_lives]

Esteban and All,

If you want to make a coil with the least amont of capacitance and one that does not generate any eddy currents on the shield try this shielding method.

View this as a cross section of the coil.

A circle - the coil bundle
Another circle - slightly larger that represents the coil-to-shiled spacer thickness.
Another circle - directly over the spacer that represents the shield, but has a slight space so there is no connection around the wire bundle circumference in which eddy currents could become generated.

Here is how to make such a coil.

Lets assume that the wire bundle with the spiral wrap spacer is 3/16" diameter viewed across it's cross section. The width of a shield that makes a complete loop around the coil bundle will be 3/16: or .187 X 3.14149 or .587".
You would need to make the width of the shield slightly less than .587 to prevent the ends of the shield from touching. This could be very difficult to achieve, espicially when using a mesh type shield (Scotch 24) that can stretch when applying the shield. Here is a method that will achieve the benefits outlined in the article that you referenced.

Cut the shield exactly .587 wide or even a little more. Using .5" wide electrical tape put .25" over the shield and .25" around the shield spacer to secure the shield around the circumference of the whole coil diameter (10" to 12" typically). Then using a .25" diameter spiral wrap or electrical tape secure the free end of the shield around the cross section of the coil. The tape covering the shield edge prevents the shield from touching itself but allows a slight overlap for full shielding with no gaps. This same method can be used where the coil leads exit the coil winding where a normal slight gap is recommended. Just as long as the shield does not touch around the circumference, the gap will be preserved.

Hi bbsailor,

If I'm understanding your description properly, in addition to the gap that is normally left to prevent eddy currents from flowing in the shielding around the circumference of the coil, you are proposing leaving a gap in the shielding all the way around the inside circumference of the coil? If this is an accurate description of the proposal, what is the expected result?

I can see where it would prevent eddy current flow around the circumference of the wire bundle, but I wouldn't think that would necessarily be an issue, since that would be 90 degrees out of phase to the coil's lines of flux.

Perhaps I'm overlooking something obvious, or am not clear about the proposed idea?

It'd definitely be interesting to compare the performance of a coil shielded in the typical fashion to one shielded in the proposed fashion to see if there is a performance enhancement.

Kyle

[bbsailor]

Proposed Coil Disclosure Idea

Esteban and All,

If you want to make a coil with the least amont of capacitance and one that does not generate any eddy currents on the shield try this shielding method.

View this as a cross section of the coil.

A circle - the coil bundle
Another circle - slightly larger that represents the coil-to-shiled spacer thickness.
Another circle - directly over the spacer that represents the shield, but has a slight space so there is no connection around the wire bundle circumference in which eddy currents could become generated.

Here is how to make such a coil.

Lets assume that the wire bundle with the spiral wrap spacer is 3/16" diameter viewed across it's cross section. The width of a shield that makes a complete loop around the coil bundle will be 3/16: or .187 X 3.14149 or .587".
You would need to make the width of the shield slightly less than .587 to prevent the ends of the shield from touching. This could be very difficult to achieve, espicially when using a mesh type shield (Scotch 24) that can stretch when applying the shield. Here is a method that will achieve the benefits outlined in the article that you referenced.

Cut the shield exactly .587 wide or even a little more. Using .5" wide electrical tape put .25" over the shield and .25" around the shield spacer to secure the shield around the circumference of the whole coil diameter (10" to 12" typically). Then using a .25" diameter spiral wrap or electrical tape secure the free end of the shield around the cross section of the coil. The tape covering the shield edge prevents the shield from touching itself but allows a slight overlap for full shielding with no gaps. This same method can be used where the coil leads exit the coil winding where a normal slight gap is recommended. Just as long as the shield does not touch around the circumference, the gap will be preserved.

Proposed Idea! When doing experimenting with coils and shielding, keep good notes and post your results this way so builders have a clear point of reference.

Coil Diameter
Coil number of turns
Mesured or calculated inductance
Wire size (AWG or mm), solid or stranded
Wire OD including insulation thickness
Insulation type, PVC, Kynar, Teflon, Enamel
Coil self resonance of the coil alone
Coil self resonance with the shield added and connected to one coil lead
Measured coil-to-shield capacitance (between one coil lead and the shield)

Based on my tests and the information in the article that you referenced, the full coil-to-shield capacitance as measured is not imposed on the coil to lower it's resonance. Distributed capacitance is a complex calculation but it's effects are very easily measured by comparing the before and after self resonance of adding the shield. I find that only between 20% to 25% of the measured coil-to-shield capacitance is imposed on the coil to lower it's resonance. This is a key point that should be measured by every coil maker to see for themselves.

The reason why I use Scotch 24 is that it is a mesh and provides an adequate shield but has less surface area than a solid shield. The mesh is configured with thin AWG36 wire and will not respond to low delays. I tried it at 7.5us and the response is very, very faint (this is good). When making this test at low delays, move the shield under the coil on a wood stick so you are not picking up your hand.

I suggest that when people talk about their PI coils that they disclose the measurements I suggested above. This will help everyone compare their own coils to what others are making.

Anyone making a PI coil in the range of 300 uH should have a shielded self resonance (no coax connected) near 900KHz. Higher is better. My non-shielded 10.5" ID coil resonance was 1.25Mhz using 19 turns of AWG 30 single strand Teflon insulated wire. Adding the coax brings the self resonance down to between 600KHz to 750KHz depending on the coax type and length.

All, please comment on this proposed idea!

bbsailor

[Esteban]

A paper about the effect of electrostatic screening in Rogowski coils, can be useful for another kinds of coil. Complete article in zip.

IV. THE EFFECT OF ELECTROSTATIC SHIELDING ON THE ROGOWSKI COIL

To improve measurement immunity it is advisable to fit

an electrostatic shield to the coil. Care must be taken to
ensure that the shield does not form a shorted turn around
the Rogowski coil. This will dramatically reduce the
bandwidth of the transducer limiting it to measurement of
frequencies substantially less than a cut-off frequency
determined by the skin depth of the shield. For example the
Rogowski coil of Figure 5. encircled by a copper braid of
thickness 0.22mm at a diameter of 5mm, is estimated to have
a bandwidth of only 7.8kHz. If however the shield is split in
the median plane so as not to form a shorted turn then the
restriction on bandwidth is much less severe. One way of
ensuring that the shield does not form a shorted turn, yet
provides good shielding, is to cover the coil in a close
packed helical winding of thin copper tape ensuring that
each adjacent turn has sufficient gap so as not to touch it’s
neighbour. Figure 8. shows the effect of such a shield on the
CWT30 Rogowski transducer with 300mm coil.

[Aziz]

Hi wyndham,

This was the quote I was wondering if stainless steel wire might be of any benefit
Thanks to bbsailor for taking the time to explain where I was going off track. Wyndham

Do not use any high magnetic permeability materials (µr) like iron, nickel, cobalt, chrome, etc. (also stainless steel) for induction balanced coils. They would disturb the balance.
Particularly the magnetic remenance of stainless steel shield would affect the coil. Copper is quite good for such material (µr ~ 1).

Regards,
Aziz

[wyndham]

In my hast to digest all this, It looks as if I have combined several threads from bbsailor and Aziz, Stefan and Tinker. I have created a coil that has parts of all , at least in my mind. I need to get back to that correspondence course on brain surgery

Hi Aziz,

I have been using method "C" for about 12 years. I make the shield material using enamelled 0.2 millimeter enamelled copper. This shield material is optimum for a PI and there is no coupling of the magnetic field through the shield. I wind the shield onto the coil so that the wires are at 90 deg to the coil winding. I also keep a spacer between shield and coil winding to minimise capacitance effects..

Regards,

Stefan

This was the quote I was wondering if stainless steel wire might be of any benefit
Thanks to bbsailor for taking the time to explain where I was going off track. Wyndham

[bbsailor]

Wyndham

You said: "the cross section would be something like the core with 30awg enameled copper then the scotch 24 shield then an insulator then a 1/4" spaced copper wrapped around the coil, another layer of insulating then a 1/8 " spaced wrap around the coil".

It is like this: The words in italics above are out of sequence. The words underlined are not needed. The sequence is as follows: (1) Coil wire bundle, (2) spacer, one or two layers of spiral wrap, (3) shield, (4) shield cover to prevent the shield from moving relative to the coil.

1. The core bundle Outside Diameter wire will have a core diameter about 5X the outside diameter of an individual strand (including the insulation). Enter this wire bundel diameter as the diameter of the center conductor in the Mogami web site (listed in my previous post) coax capacitance calculator.

2. The thickness of the insulator between the coil winding and the shield is the shield spacer. Enter it's thickness in the calculator.

3. The dielectric constant of the shield spacer varies according to the material used as the spacer. Teflon and Polyethylene (PE) are good. PVC is not as good with a higher dielectric constant. Enter various materials in the Mogami calaculator and see the difference between the coil to shield capacitance.

If you use a solid shield like a thin foil and compare that against using Scotch 24 mesh, you will find that the Scotch 24 has less capacitance because it has less surface area than a solid shield.

Thin wire bundles have less capacitance than thicker wire bundles. Use 19 turns of AWG30 single strand Teflon insulated wire with a bundle OD size of .024" X 5. = 0.12" for a good low capacitance mono coil.

All of the techniques that I have outlined in my article combine to minimize capacitance and make a potentially faster coil.

The benefits of making a fast mono coil with the techniques outlined above can be lost when you choose to use 7 feet of coax to make a hip mounted control box. This is why it is better to mount the control box on the coil shaft near the coil 30 to 34" away (depending on your height). Then, plan on mounting the heavy battery pack on a body mount pack with a headphone jack on the battery pack so you only have one wire between your body and the control box.

Does this help?

bbsailor

[wyndham]

If I'm understanding what has been posted about the coil in question, the cross section would be something like the core with 30awg enameled copper then the scotch 24 shield then an insulator then a 1/4" spaced copper wrapped around the coil, another layer of insulating then a 1/8 " spaced wrap around the coil.
I'm remembering something I found several years ago and wondered if this might fit in. The website was commenting on different materials and the inductance values and possibly capaticence values as well and stainless steel was mentioned as a preferred material.
If the SS wire was used instead of the copper in the wrapping around the scotch shield, who this be of any benefit?
Thanks Wyndham

[bbsailor]

bbsailor,
I used your suggestion of putting the scotch 24 over the 1/8" spiral warp and the covering the shield with the 1/4" spiral wrap. Worked like a dream! I was concerned that the coil/shield capacitance would be greater, but effects are not noticeable. The coil (7.125", 23 turns, 298uH) with 1/8" spiral wrap had a resonant freq of ~975 kHz. I could barely see any decrease with shield added. With shield and 30" mogami dual coax the resonant freq is 950 kHz.

I cut the scotch 24 so that it was only 1 layer thickness and ~1/2" wide (I actually cut it 1/4"- 5/16" from the folded edge). I also used clothespins as you suggested every 1 1/2" to hold the scotch 24 while wrapping with the 1/4" spiral wrap.

Thanks again for you words of wisdom!


Regards,
J. L. King

JL King

Based on the before and after resonance measurements (above), I estimate that the total coil-to-shield capacitance measured between one coil wire and the shield is approximately 23 pf (+/-3pf). This is a very low number and should contribute to making a fast coil.

Can you share with the forum members the key measurement on some of the other coils that you made? Be sure to include the approximate cross section diameter of the coil wire bundle, inductance, and self resonance (before and after the shield) but no coax connected.

Nice Job!

bbsailor

[hobbes_lives]

I also used clothespins as you suggested every 1 1/2" to hold the scotch 24 while wrapping with the 1/4" spiral wrap.

Great suggestion bbsailor, and thanks for posting it KingJL. I wish I knew that trick when I was struggling with Scotch 24 on a mono coil a couple of weeks ago.

[KingJL]

bbsailor,
I used your suggestion of putting the scotch 24 over the 1/8" spiral warp and the covering the shield with the 1/4" spiral wrap. Worked like a dream! I was concerned that the coil/shield capacitance would be greater, but effects are not noticeable. The coil (7.125", 23 turns, 298uH) with 1/8" spiral wrap had a resonant freq of ~975 kHz. I could barely see any decrease with shield added. With shield and 30" mogami dual coax the resonant freq is 950 kHz.

I cut the scotch 24 so that it was only 1 layer thickness and ~1/2" wide (I actually cut it 1/4"- 5/16" from the folded edge). I also used clothespins as you suggested every 1 1/2" to hold the scotch 24 while wrapping with the 1/4" spiral wrap.

Thanks again for you words of wisdom!

Regards,
J. L. King

[bbsailor]

Hi bbsailor,

Nice to have you helping with this thread. As yet, nobody makes faster coils than bbsailor.

Here is the way I see the problem. My view might not be correct, so please correct me.
For every increase of diameter, the circumference increases pi or about 3,14. This means that the coil to shield area increased that much and adds that much more capacitance.
When we add a spacer around the coil winding, the area increases much more, however, the distance between the coil and the shield also increases, diminishing the capacitance.
The matter that fills the space between coil and shield is therefore very important. Air is best.
At some given distance, the relation between the distance and "capacitive area" is ideal.
What is this distance? How do I calculate this relationship?
Tinkerer

Tinkerer and All,

Go to this web link:http://www.mogami.com/e/cad/electrical.html

The Mogami cable capacitance calculator can be very useful in calculating the PI coil to shield capacitance. The inner conductor is separated from the shield by some thickness of dielectric material that defines the capacitance between the two conductors.

With this calaculator you don't need to wind that physical coils to calculate coil-to-shield capacitance. Just use the cross section diameter of the PI coil diameter as the conductor diameter. Use the spacer thickness and dielectric constant to calculate the effective capacitance of the coil circumference.

The whole value of the coil-to-shield capacitance is not imposed on the coil to lower the coil's self resonance frequency. The question is: What percentage of the total coil-to-shield capacitance is actually imposed on reducing the coil's self resonant frequency?

If I tell you, no learning will occur!

Tinkerer: Yes, thinner cross section coils have less capacitance!

bbsailor

[Stefan]

Hi Aziz,

I have been using method "C" for about 12 years. I make the shield material using enamelled 0.2 millimeter enamelled copper. This shield material is optimum for a PI and there is no coupling of the magnetic field through the shield. I wind the shield onto the coil so that the wires are at 90 deg to the coil winding. I also keep a spacer between shield and coil winding to minimise capacitance effects..

Regards,

Stefan


The sepearationI

Hi Tinkerer,

there is still improvement to squeeze out:

1. Coil capacitance and effectivity:
Coils making not too much wider. Instead, one could wind in the height direction too. So some distance ring could be used for this. The result will be a square cross-section coil. The wire intercapacitance can be kept low to make PI coils faster.

2. Shielding:
a) Shielding should not be directly made on the coil wires. Some distance 1-2 mm should be used, to lower the shield capacitance.

b) Using low eddy-current shielding:
High impedance (high resistance) coil shielding can be used. Such as with graphite shielding.

or
c) Using very thin enamelled wire wound across the wires. This wire goes from the coax shielding (soldered to shielding) to half of the coil. Another wire is wound on the other half. They are not connected together - they are only connected on one point of the coax shielding. The windings could be spaced on every 1-2 mm. It depends on the wave length, which EM fields you want to screen out. The closer the windings, the higher EM's could be screened out. But the closer the windings, the more capacitance it will be.
c) is probably not used yet. But this gives much improvements on PI coils and one should make some tests on it.

Regards,
Aziz

[Tinkerer]

Which coil current? I have no idea, what you mean by that.

The second fact of this super shield is, that only middle part is connected to coax shield. Other endpoints of the shielding has no potential to any electronics nor to each other (because not connected).

The shield acts as a very low eddy current causing EMF shield.
Aziz

Aziz,

The RX coil is inductively coupled with the TX coil. At any time a current flows in the TX coil,there is also a current in the RX coil.
With the standard PI it is the endeavor to have these currents decay as fast as possible in order to measure the secondary currents generated by the fields from the target in the RX coil.

I have observed that the amplitude of the currents in the RX coil is much influenced by the size of the damping resistor that is across the RX coil.

A conductive spiral that is wound around the RX coil will be inductively coupled to the currents running in the coil.

Tinkerer