Litz wire - what to get?

[Carl-NC]

First, keep in mind that the magnetic field is proportional to ampere-turns, I*N. At DC, Litz wire isn't going to do anything for you because the current splits evenly amongst the wires and the ampere-turns is the same as normal wire, assuming the overall DC resistances are the same.

Litz starts picking up steam when the current is AC. Skin effect in ordinary wire increases the AC resistance, but to a much lesser extent in Litz. However, for metal detectors you have to get into the 10's of kHz to see a practical benefit. Ferinstance, at typical VLF frequencies of ~15kHz, skin depth is about 0.5mm (copper wire) which is close to the radius of 18awg wire. So as long as you keep the coil wire smaller than, say, 21awg, Litz won't help.

PI detectors are a little different with their fast edges, and getting down to a 5us sample time is like moving into the 50-100kHz frequency realm where skin depth is more like 0.2mm. So Litz can make things faster, but only as long as the coil is already the limiting factor. That is, everything else is really fast.

- Carl

[Alexismex]

Thank you Wm6 for the tip,

[bklein]

First, keep in mind that the magnetic field is proportional to ampere-turns, I*N. At DC, Litz wire isn't going to do anything for you because the current splits evenly amongst the wires and the ampere-turns is the same as normal wire, assuming the overall DC resistances are the same.

Litz starts picking up steam when the current is AC. Skin effect in ordinary wire increases the AC resistance, but to a much lesser extent in Litz. However, for metal detectors you have to get into the 10's of kHz to see a practical benefit. Ferinstance, at typical VLF frequencies of ~15kHz, skin depth is about 0.5mm (copper wire) which is close to the radius of 18awg wire. So as long as you keep the coil wire smaller than, say, 21awg, Litz won't help.

PI detectors are a little different with their fast edges, and getting down to a 5us sample time is like moving into the 50-100kHz frequency realm where skin depth is more like 0.2mm. So Litz can make things faster, but only as long as the coil is already the limiting factor. That is, everything else is really fast.

- Carl

Hi Carl,
I'm mainly focusing on PI detectors and specifically the RX function - this is where I am having a problem just getting a conceptual approach to the engineering involved. I realize there is a whole design challenge in regards to the TX/RX switchover (when a RX signal becomes usable and compatible with highgain preamps etc. following the coil). Sooner you can work with the signal sooner you can see small and gold signals. I am interested to better characterize this signal off the small gold target. How long does it last? Does each target have some sort of resonant characteristic? So I am interested in taking test equipment - I have access to some pretty expensive stuff at work - and doing so, but I don't know exactly what will work or how best to use it. Have you done work in this area? I'm wondering if targets would have resonant characteristics where they may be sensitive to the time inbetween TX pulses.
I have access to spectrum analyzers and VNA's etc. by Agilent. Typically used to characterize USB3.0/SATA/FW800 cables and signal paths.
I though have been mainly working on qualifying DC/DC converters so I use different equipment (although I've worked with emi stuff a few years ago). In my work with power products I came across this "frequency response analyzer" that Ridley Engineering sells:
http://www.ridleyengineering.com/analyzer.htm
I'm wondering if it would be helpful to better understand actual coil designs. I'm just still clueless though what to look for or design to using such equipment, when dealing with pulse response. Thus, I'm wondering if targets have a resonance then once I know that resonance frequency I can optimize a RX coil to sense it. PI theory always describes the target as emitting a decaying signal but I'm wondering if it is more basically that the target reflects all frequencies to some level but has its own resonant frequency based upon its material and dimensional form. Does material overshadow dimensional form or the other way around?
Another thing while I've got ya... How much is gained by placing the TX power source (maybe supercap with series of MLCC's) and the switching fet at the TX coil itself?

Barry

[Carl-NC]

Targets don't really have a resonant frequency, and thickness, shape, and size has as much or more effect on response than metal. If you want to look at early sampling then simply build a good induction-balanced coil. With that you can sample as early as 1us.