Simple radio solution for ultra low power low bandwidth

[kuon]

I am working on a medical sensor on which I have a cortex M SMT32 micro controller.

The micro controller is doing all the job and with it's deep sleep I can have a very low power consumption.

That being said, I need a low power, low bandwidth (1kb per minute), low range but high penetration (should be able to penetrate a medical ward wall) wireless solution.

At first, I tried an NRF54820 based bluetooth solution, but I had some issues:

- bluetooth is so damn complicated...
- the SDK and tools are very heavy, and a simple hello world program can be very hard to manage
- the 2.4Ghz band is just too crowded where I would need to deploy those devices
- penetration of the 2.4Ghz band can be problematic, even with thin walls

For this reason, I am looking for a simple transceiver, in the 826Mhz band, SIP based with a simple interface (UART can work but SIP often provides better performances which yield to less time awake and lower power).

So far I found the analog ADF7024.

It seems to be the perfect candidate, it has a 240Bytes buffer and a simple SIP interface.

Now I have two questions:

- Anybody has experience with what I am doing, is this a good solution/chip?
- I have never done antenna design (except for NFC), and I would like to use a PCB trace antenna if possible. I found many different design and vague schema, but no proper schematic with specifications. I also have no idea how to tune the RF matching network highlighted in the datasheet (below image). I have a tinyVNA which I used a few times for NFC antenna designs, but those only have a single tune capacitor, so it is very easy. I realize this is a broad question, but this forum has always been very informative.

[kuon]

I made some researches about the impedance matching, and I think the easiest path is to:


- copy the schematic and layout from  EVAL-ADF7024DB2Z (eval board with 50 Ohm anteanna connector)
- use a chip antenna like W3013 from pulselarsen

In the W3013 datasheet, there is an optional "shunt" between the end of the antenna line and ground, what is its purpose and how to calculate its value?

[radiolistener]

In the W3013 datasheet, there is an optional "shunt" between the end of the antenna line and ground, what is its purpose and how to calculate its value?

this is a matching component (capacitor or inductor). It's value is selected during tuning process.

[voltsandjolts]

Some options:

RF Solutions radio modules

433Mhz might give better wall penetration than higher frequencies, depending on power of course.

[geggi1]

Look at some of the solutions used for RFID.

[rfclown]

I made some researches about the impedance matching, and I think the easiest path is to:


- copy the schematic and layout from  EVAL-ADF7024DB2Z (eval board with 50 Ohm anteanna connector)
- use a chip antenna like W3013 from pulselarsen

In the W3013 datasheet, there is an optional "shunt" between the end of the antenna line and ground, what is its purpose and how to calculate its value?

I would agree in general that the easiest path is to copy schematic and layout from an eval board.

For a different solution, I've used LoRa for low data/power. First time I used it was for someone that wanted a temp sensor that ran for years off a battery that was to be placed in a commercial walk in cooler. I used a module that was pre-certified because they didn't want to have to do FCC certification. I used it a second time for someone who was concerned with range. The processing gain from the chirp modulation lets it receive signals that are in the mud.

[kuon]

In the W3013 datasheet, there is an optional "shunt" between the end of the antenna line and ground, what is its purpose and how to calculate its value?

this is a matching component (capacitor or inductor). It's value is selected during tuning process.

If the antenna is 50Ohm, and the chip is matched to 50Ohm (and the feed line too), why is it needed?

I found a matching network designed for the ADF7024, 0900PC15F0030. If I use it, the antenna line should be 50Ohm, thus I can connect it directly to the chip antenna.

I looked at many modules, but I couldn't find a 433Mhz or 868Mhz module with a built in chip antenna, so if I need to build the antenna network, I don't see the point of the module vs the IC as the RF matching is the "hardest" part.

I picked the ADF7024 because the protocol is very easy to use with the integrated buffer.

[rfclown]

In the W3013 datasheet, there is an optional "shunt" between the end of the antenna line and ground, what is its purpose and how to calculate its value?

this is a matching component (capacitor or inductor). It's value is selected during tuning process.

If the antenna is 50Ohm, and the chip is matched to 50Ohm (and the feed line too), why is it needed? ...

The antenna isn't just the chip, it is the chip along with the PCB structure shown on the datasheet. If you duplicate everything on the datasheet PCB structure, you should be ok. But it will then vary with anything near the antenna, like a plastic case. It's good to have a convenent way to tweak it, but if you don't have a way to measure, it's hard to tweak. I was brought into a company where they had implemented a similar 915 MHz transceiver which had a chip antenna in the reference design, but they didn't follow the details of the metal around the antenna (and that "chip antenna" needed a more complicated structure to make it an antenna). They were only getting a few feet of range. (They had no instruments to have any idea of what was going on).

Again, if you follow the datasheet layout exactly you should be ok. Notice it is on the board edge. Notice NO metal on ANY layer under rectangular antenna structure.  If in a plastic case, leave some space between antenna and case. Stuff near the antenna will shift it's resonant frequency (any antenna is like that). I won't kill it, but the more space you can give it, the less it will have an effect.

[kuon]

In the W3013 datasheet, there is an optional "shunt" between the end of the antenna line and ground, what is its purpose and how to calculate its value?

this is a matching component (capacitor or inductor). It's value is selected during tuning process.

If the antenna is 50Ohm, and the chip is matched to 50Ohm (and the feed line too), why is it needed? ...

The antenna isn't just the chip, it is the chip along with the PCB structure shown on the datasheet. If you duplicate everything on the datasheet PCB structure, you should be ok. But it will then vary with anything near the antenna, like a plastic case. It's good to have a convenent way to tweak it, but if you don't have a way to measure, it's hard to tweak. I was brought into a company where they had implemented a similar 915 MHz transceiver which had a chip antenna in the reference design, but they didn't follow the details of the metal around the antenna (and that "chip antenna" needed a more complicated structure to make it an antenna). They were only getting a few feet of range. (They had no instruments to have any idea of what was going on).

Again, if you follow the datasheet layout exactly you should be ok. Notice it is on the board edge. Notice NO metal on ANY layer under rectangular antenna structure.  If in a plastic case, leave some space between antenna and case. Stuff near the antenna will shift it's resonant frequency (any antenna is like that). I won't kill it, but the more space you can give it, the less it will have an effect.

Yes I will replicate the exact dimension in the datasheet.

I have a tinyVNA, but I am not very familiar with the tuning process. Is it like NFC, you just measure the antenna and adjust the tuning capacitor (the shunt component I was asking about, in the datasheet) until the frequency is right? I'll try and see how it goes.

[rfclown]

NFC? National Football Conference? But yes. You'd measure S11 of the feed line input and see that you have a good return loss (low S11) at your frequency of interest. The VNA sends a signal out the port and measures how much bounces back. S11 is the ratio of what bounced back versus what was sent. If it bounces back, it didn't get out the antenna. S11=0 means everying bounced back (open circuit, short circuit). S11= -10dB means than the amount bouncing back was 10 dB below the signal that went out. -10 dB is one tenth of the power, so 90% of the power went somewhere (hopefully radiating nicely into the ether). The only way to see how it is actually radiating is with measurements in a chamber or field (rotate antenna and see how much is picked by a receiver (like spectrum analyzer).

[radiolistener]

If the antenna is 50Ohm, and the chip is matched to 50Ohm (and the feed line too), why is it needed?

actual impedance depends on every nuance, include PCB material properties, copper properties and even PCB mask material properties. It's impossible to predict all these things. Even two identical PCB manufactured on the same factory at different time can have different RF properties. So if you want to get the best performance, you will need to precise tune it with a special equipment. Do not use mask at antenna and transmission line region if you want to minimize that variation.

[kuon]

NFC? National Football Conference?

NFC is near field communication. The only kind of antenna (loop antenna) I already successfully designed and tuned using my tinyVNA.

Thanks for all your feedback. I'll try and see where it goes.