If the ground is directly behind the coil it will almost completely stop it working. Magnetic fields only penetrate about 30 microns into copper at 13.56 MHz, and the ground plane will vastly reduce the magnetic field of the coil. The field lines will all have to fit between the edges of the coil and the ground, and driving them through here at a high flux density will use up all the field strength, resulting in low magnetic fields, low inductance and poor coupling range.

If the ground is not close behind the coil I have no idea why it isn't working.

Malcolm White The coil is separated from the copper ground by the paper substrate (thickness of about 0.68mm), so it is not directly behind the coil. Although I have resimulated with a diminished copper thickness and even removed the copper ground as a last resort with close to no difference in results.

0.68 mm is almost no separation. Currents in the ground that close will almost completely cancel the field of the coil. You need any large metal or conductor to be about the same distance away from the coil as the size of the coil, and then it will have only a small effect on the range.

I am not sure why removing the ground has not changed the results. Are there any other conducting parts in your model? Such as a PEC boundary behind the ground, for instance? That will have the same effect as a ground. On the other hand, a PMC boundary close behind the coil will nearly double the field.

Malcolm White First of all, thank you for your responses and your patience, since I am fairly new to CST! To answer your question that are no other conductors in the environment. After re-assessing the paper I have found that the problem might be in the simulation setup, since they use an E-Field pattern instead of RCS.

So I switched to a Time Domain Solver and added a E-Field source at 13.56MHz, and I got the attached results. I should clarify here that my objective is to simply obtain the S parameters of the tag. Should I add ports to the tag and use Frequency Solver? If so what kind and where (I know this might be obvious but the placement of ports at the spiral ends has gotten me some warnings/errors in the past)?

This antenna will not interact much with electric fields at 13.56 MHz. It is a magnetic coil and it generated a mainly magnetic field close to the antenna. This is usually only useful for things like RFID if the tag is closer than about 3 coil radii, but it can be used further away. The fields from this are almost entirely magnetic. They are also near-fields, and do not radiate. The coupling between coils is like a transformer, but with a very small coupling coefficient. RCS is defined in the far-field, and will be extremely low for this kind of coil at this frequency as the radiated field is very small, and the near-field falls off as 1/r^2 or 1/r^3 and vanishes in RCS calculations.

I would run a track, at a different level, from one of the ends of your coil across all the loops to close to the other end, so that you can have a port that is between two wires close together. A discrete port will be suitable for this.

A time domain solver may be ok for this, but a frequency domain solver may be better. CST has a frequency domain solver, so you may be able to just choose it.

Your E-field plot is from 0 to 100 GHz. I suspect that is a mistake. You should be looking at 1 to 100 MHz, or 0 to 30 MHz.

Malcolm White I see. I removed all previous field sources and ports, and ran a new discrete port across both ends (picture attached below) with a Frequency Solver and got a readable S-parameter result (attached aswell). Although it is not really defined and below 10 MHz (although I believe this to be since I do not have the tuning capacitor added yet), any way to make this more defined?

Your S11 should be 1, or slightly below, if there are losses, because the power in the coil cannot go anywhere unless there is something for it to couple to. That is why magnetic chargers are efficient - no power goes out if there is nothing there to charge. You either need to add another coil, or to plot the magnetic field in the places you want it to be able to read.

It is hard for the time domain solver to get the S11 exactly right in this structure because the fields bounce about in it for a long time, and add to (or subtract from) the S11 every time they bounce out of the port. The frequency domain solver doesn't have this problem.

However, the magnetic fields may be fairly accurate even if S11 is not quite right.

Malcolm White I added a H field source and monitor along with a probe at Z=20mm, and got the following result. Although I had to change the Mesh mode in the Frequency Solver to Tetrahedral so it can measure Field Monitors, and this changed the S Parameters. Since I'm working with the Student Version I had to downgrade the mesh cells to be under the 20000 limit (settings attached as well). The magnetic field doesn't seem to be displaying correct results as well.

The S11 looks very good, at just below 1. Make sure you read the scales, it only falls to 0,003 below 1 over the whole range.

I don't know what magnetic field you expect. It is saying about 40 A/m at 13 MHz.

Where are you measuring it? At the coil centre? This is about 50 microtesla, which might be enough if it is at the tag position. It is about the same size as the earth's magnetic field.

Check my numbers - I might be wrong!

Malcolm White I'm measuring the coil 20mm right above the center (attached). Although it's a good sign that there isn't any loss in the S parameters (attached a new version with 100MHz range), since the aim of the coil is to display a deep notch at 13.56MHz, or any frequency since it's not tuned yet, (same as the paper, also attached), I'm not sure the simulation is still correctly setup.

Interestingly, the current at the discrete ports, displays exactly that notch (final attachment). I may be looking at the wrong results.

The notch is because the coil is resonant. It needs to be resonant so that the current in the coil is much larger than the drive current, and so that the fields are much larger than the fields due to the drive current. When the coil is resonant, the currents grow to the level so that all the power is lost in the coil resistance. The power taken by the tag is very low in comparison to this, but can just be detected.

The coil can be self-resonant, but then this would show for all ports, but have you put a tuning capacitor in the discrete port to make it resonant?

I was expecting the coil to be self-resonant as it is, not exactly at 13.56 MHz (since the coil itself has some capacitance), and then I would tune it to the desired frequency dependent on the inductance simulated, by adding that capacitor. So no I haven't put a capacitor in the discrete port yet.

Is it possible it is self-resonant but at a frequency not displayed? I'm trying to place the capacitor, although the schematic view will not let me add further ports or connections (attached), do I need to add more ports in the 3D view?

(I'd like to thank you again, at this point you could deserve a credit in my thesis)

I think you will be able to add a lumped element to your model, to put the capacitance in the circuit.