G'day and welcome to this week's episode of The Factory. This week, we're interviewing Peter, who's been working on an RFID reader.

How are you going, Peter?

I'm doing great.

Awesome, let's get stuck into it. So yeah, we're talking about RFID. Firstly, Peter, what is RFID? I know some people might have heard of RFID in the context of, you know, maybe they know their cat has RFID in it. What sort of problem is this solving?

Well, this solves the problem of not having to put wires into your cat. Yes, that's a good thing, yeah. So you can read tags that you can embed in your cat or any device that you might have, tags like this. An RFID allows you to read these tags, get the data off them without actually touching them. And that can be really useful if you've got some sort of something in the way between the two devices, or you want to keep something waterproof. Wherever you can't have an electrical, physical electrical connection.

Yeah, that's right. Electrical connections can be tricky in the first place. So this solves that problem. So that's really cool. We've got some applications for RFID. How does this work? How do we get communication without a physical wire?

Well, it all starts with the RFID chip and it generates a high frequency signal which goes into a coil, which is embedded in the printed circuit board of this board. That coil radiates a magnetic field that can be picked up by a reading card, picks up the signal and energy harvest that signal. So there's enough power in that signal to start the chip and for the chip to read whatever information's coming from the RFID reader and then respond accordingly and send data back to the RFID.Reader: So that's really cool. We've got this working board. How do we get data back to the RFID reader? We've got this board. How do we get there? There's lots of prototypes on the board. What were the design goals? What were the challenges? What was the process for getting there?

All right, so the challenge here is to get as much power as possible from our driver chip and into our tag that we're reading. So, and bridging that air gap, getting that power through enough to actually start the chip and allow it to- Well, this thing actually has to turn on, doesn't it?

That's right. It's gotta be able to actually run digital logic. The driver IC chip is, outputs its power via like an inbuilt H bridge and that's a very digital type of circuit. It's gotta be lots of harmonics coming out of that.

That's right. So it's rough. It's just on or off and it's just pushing and pulling at this coil that we've got here. And the first thing that we need to do in order to suppress the electromagnetic interference is to put a low pass filter in, to strop stray components above that 13.56 MHz that we're actually interested in.

Yeah, cool. So we have to design a low pass filter. After we've done our low pass filter, then we need to do impedance matching from the low impedance driver IC to the high impedance or reasonably high impedance in comparison. Relatively high impedance, yeah.

Impedance of that antenna. And so we have to build a matching circuit using passive capacitors and inductors. And also make sure we don't burn out the chip.

And that's right, yeah. We don't want to make smoke. It's not smoke generator. So we want to get as much power transfer as possible.We can, but we don't want to just burn the whole thing out. Yeah, cool.

So Peter, what's the design process? How do we get this low pass filter there? How do we work out the matching circuit?

Okay, so the design process is fun. What we have to do is the first thing that we need to do is to have a look at our requirements for PicoDev. The PiicoDev has a very specific dimensions. So the first thing to do is just design a coil that matches. So we're not even designing a matching circuit first. You're choosing your coil.

No, you start with the coil. And so you design a coil that you'd go, this coil looks like it might just work. Not with any confidence, but it might just work. So design the coil.

So could you say it was an educated guess? We have the dimensions, we guessed. What went into that educated guess? What made us choose a particular dimensions for our coil? Or a particular number of turns?

Well, okay, so more turns means that you get potentially, well, bigger inductance can allow you to get better coupling to the device. So you're thinking like stronger magnetic field with more turns, yeah. More turns equals more of an inductor. More is better, right? Yeah. We just throw everything at this and see if it works. So we started off with this one here. This one has three turns on the top and three turns on the bottom. So six turns total. And great, that's a great starting point. Do you remember the trace thickness on that one? About 0.2 millimetres roughly. Pretty thin traces, yeah. Exactly.

Then we started doing our measurements and found that the inductance was fine. The resistance was just too high. Yes, well, basically we made this matching circuit.Circuit for it and the bandwidth was massive. Yeah. You know, we had this thing operating at 13.56 MHz and the bandwidth on our matching circuit was like five MHz or something. We could barely see any energy being absorbed at all.

Yes. The Q was just way too low. And so the problem with the large bandwidth is that there's less energy to actually be captured by the receiving coil. Yeah, lots of energy going into heating up the coil, not very much energy actually turning the tag on. So, didn't get very far with that one.

So we did that first iteration and it was great because we learned a lot about going through the process of making, we still made a matching circuit, just not very effective. Just didn't work, but we still learned how to measure inductance with an oscilloscope. That was basically just putting a resistor in series within the inductance and picking a frequency, picking several frequencies and doing a phaser division of the voltage and the current to work out what the impedance was of the whole circuit and pulling the inductance out of that impedance calculation.

So what's the actual design process for this matching circuit? We need some component values. How do we get component values? Well, what you do is once you've actually measured your inductance of your coil, then you go to the data sheet and just look at all the formulas and just start plugging values in that we know about, put them into the formulas, calculate capacitor values and inductor values, keeping in mind that you want to be within certain parameters. You want your inductor to be more of an inductor than anything else. More inductive than resistive is aGood start, I think. That's right. More inductive than capacitive, got to stay well below its resonant frequency. Yeah, all of those things.

Yeah, really just turn the handles, take a measurement, turn the handles on the maths, get some component values, see if they work. Lock them on the board.

So first revision, had a crack at it, didn't really work. What did you change? Why do you think it helped?

Okay, the first thing that happened is, okay, we're going to get this resistance down. Let's make the coil shorter. So we took out 50% of the coil by just putting the coil on one side of the board. The other thing that we did is double the trace width or maybe more than double. Just make it bigger, just get rid of those copper losses. So we did those two things and then we then had to then redo all of the calculations again. Thanks spreadsheets.

And the first thing we had to do was to measure the inductor, the new inductor that we just created using the process that was just described, you know, silico, resistor, phase changes, all of that sort of stuff. Lots of maths. Lots of maths.

And then we had to do more maths to get these capacitor values, plonk them on the board. And well, then it worked. What happened? It just worked. Yeah. We got lucky, it worked. Get those values right. And then suddenly it just brings to life. Gotta love reading the manual. Yeah.

So we've got the current working design revision here, Peter. One final question though, why is the coil on the back? I'm not seeing any traces on the top side of this PCB.

Well, we've designed this so that people might mount this on their projects so that it's just flush with the outside of theEnclosure like this. So this is a simulation of the enclosure. We've got enclosure, big EDEV device. Our coil is as close to that as possible. So you get minimum interference. Yeah, nice. Just lets you really, just literally glue it to the inside of the box. Yeah. And hopefully the other side is close enough to the coil that it will still read. Yes. That's really cool.

Well, thanks for that discussion about the RFID project, Peter. You'll be seeing this eventually. Well, we've still got some design work to go. So in the next couple of months, maybe you'll be able to grab one of these and play with it yourself. Yeah, that's right. Thanks a lot. Cheers.