NFC antenna design and tuning, feel like I'm missing some basics

[newto]

As one of many many side projects I'm working on, I've realized adding an NFC sticker and reader to our products would make lives a lot easier. Essentially the goal is that the user can attach an accessory to the main unit, and automatically transfer the data like model, calibration date, and calibration data (the accessories are completely passive). Less room for user entry error, fewer steps for the user, and I can greatly simplify the UI on the main unit.

But I feel like I'm missing something (or several things) from how to design and tune NFC antennas.

Things I know from various application notes and youtube videos:

1) I want my antenna to be between 0.5uH and 4uH, and reasonably close to 1uH
2) I need an NanoVNC/LiteVNC to make measurements
3) I need a matching circuit and EMC filter
4) I will probably be using an NXP reader chip, likely MFRC631?
5) The complications below will make any estimation of the inductance of my coil way off, but I have to start somewhere

Things I'm confused about:

1) Some application notes say I want a symmetrical Smith Chart, while others seem to be okay with sweep being mostly capacitive, but pokes into inductive at the target impedance.
2) Do I want the target impedance to be on the resistive line, or in the inductive side (some ANs have it on the line, others have it inductive)
3) What is my target impedance? I'm seeing 17, 20, 25, and 50, so I'm not sure what my actual goal is.
4) If I want to add a 50R 20cm sma/u.FL extension to the antenna to save space, where is best to put it? And do I need to match 50R on both sides of the extension?

Some added complications:

1) The antenna needs to fit inside a donut shaped area, the PCB has a max diameter of 50mm, and there's a 25mm diameter ground plane with non-nfc related components in the middle.
2) In operation, the NFC antenna will be inserted about 2cm inside the end of a 54mm diameter aluminum cylinder
3) The NFC sticker will be ~5mm from the coil, and not in the middle (since it can't block the components in the middle). The sticker will be oriented correctly with the coil at least.

The complications may make the whole thing impossible, but I at least want to try it out.

I've attached my first poor attempt at matching a coil I made in kicad with this to figure out number of coils, diameter, width, etc: https://www.translatorscafe.com/unit-converter/en-US/calculator/planar-coil-inductance/

I understand it could be 100% wrong, so feel free to make fun of it. And is there's a best location to add the extension with as few components on the PCB with the antenna? Do I need multiple matching circuits to do that? Space is very tight already in the ground plane area

[tszaboo]

You can buy NFC antennas, and tags. NXP has an NFC evaluation board, with different FR4 and Flexible PCB antennas. I'm sure others do as well, probably ST is another big supplier. So try one antenna in your physical arrangement, see if it works. If you stick an LED to a simple NFC EEPROM power pin, you can make a crude field detector. Or, I've seen it even without the EEPROM, just the LED on the coil. There are also standard size PCB tags, stick to one if you can. Ah, and check if your scope supports NFC decoding, the Keysight MSOX3104T for example has it as an option, could be useful.
The frequency of NFC is so low, you could do a lot without a VNA. Unless you get the budget for it, then go for it 

[newto]

Yeah, the gist of what I've been reading is that the best way to design an antenna is to just build it and figure out the parameters from testing, and then do the matching and tuning off that.

I just want to know if I'm on the right track so far, and what I'm missing (especially how to add a cable extension)

[TC]

I completed a similar project this year but with stainless steel vs. aluminum. But we had a custom tag that matched the shape/coil-design of the reader antenna. It works fine with short cables (15cm) and long cables (10M). We didn't do much testing in-between.

You may need ferrite sheet material backing your reader and tag antennas. Material is readily available at distributors like Digikey and Mouser. They will indicate that they are designed for NFC or NFC frequencies.

You are correct... build it and test it. Spend your time on the bench vs. modeling and simulation. Have a test fixture that lets you align the reader antenna and tag and control distances. That's necessary to get repeatable test results.

Your schematic is an adaptation of a differential filter/matcing-network. Convert that to a single-ended design.

[newto]

Your schematic is an adaptation of a differential filter/matcing-network. Convert that to a single-ended design.

Goes to show how little I know, "single ended antenna" was *exactly* the words I needed to know order to make this whole thing 10x easier. ST has an antenna matching tool that has a single ended with cable option that I can copy to get myself started, and the ST25R chips seem to have a simple couple of settings to configure for single ended config.

But we had a custom tag that matched the shape/coil-design of the reader antenna.

By matched, is that just a coil with the same general shape? ie if my reader coil is 46mm diameter, I make another PCB with a 46mm coil matching the specs required for an NFC tag IC, and that will probably work better than a tiny off center sticker? More expensive that way, but from rough guesstimates, 1$ vs 0.10$ for PCB vs sticker, and it would look a lot more professional when installed...

[TC]

By matched, is that just a coil with the same general shape? ie if my reader coil is 46mm diameter, I make another PCB with a 46mm coil matching the specs required for an NFC tag IC, and that will probably work better than a tiny off center sticker? More expensive that way, but from rough guesstimates, 1$ vs 0.10$ for PCB vs sticker, and it would look a lot more professional when installed...

Yes. Basically the same antenna but with just a TAG IC vs. an RF connector and tuning components on the reader antenna.

NFC is a near-field technology. So it is basically electromagnetic. Think of the two coils as inductive windings of a transformer that is air-coupled. They have a mutual inductance when coupled which depends on how well they are aligned, the spacing between them, etc. So tuning depends on all of this.