Power amplifier for remote controller
I'm designing remote controller based on NRF24L01 and STM32. I want to use SMA antenna. My question is regarding power amplifier (PA). When is it necessary to use one? The range I'm aiming for is 20+ meters. Every commercially available module with NRF24 and SMA antenna has PA, but I would like to save on cost and space on PCB. The NRF24 datasheet says that recommended load impedance is 50 Ω. Is it therefore possible to use SMA connector and antenna with 50 Ω impedance?
EDIT:
Is there a way to calculate the approximate range of this remote controller? I have seen some advanced software that can simulate antennas (too advanced for me), but I want to know what will be the approximate range. Or is it just easier to test this and then add PA later on if it is necessary.
3 answers
You are accessing this answer with a direct link, so it's being shown above all other answers regardless of its score. You can return to the normal view.
#1 before anything else is to do your homework. The Nordic semi web site says: "nRF24 Series Not recommended for new designs". That means forget about using this part! They are phasing it out of production and it will probably go EOL within a couple of years. Also be careful not to use some silicon vendor who has a habit to discontinue their whole product line every second year; I have never used Nordic so I have no experience about them in that regard.
Also in general don't do electronics projects by first picking a part and then try to figure out what to use it for. Write a specification then after that track down suitable candidate parts that fulfil it. That way you automatically don't end up with something obsolete or otherwise unsuitable. Also in these days of massive logistics incompetence among the silicon vendors, it is very important to check that the part can actually be purchased in the real world as well. Even previously reliable vendors fail to deliver right now.
Modern RFIC always come with an internal PA and it's common that the same manufacturer has a line of similar products but with a different PA. Some parts with +4dBm and some with +8dBm, so that's a different internal PA. And yes the whole point of having that is to drive the antenna without an external PA.
You will have to match this internal PA with passives as per the manufacturer's recommendation. How to do that depends on a lot of things: is it just Tx, just Rx or semi duplex? Will there be a "direct tie" to the antenna or will there be an antenna switch between Tx and Rx paths? Will you use a SAW bandpass filter for Rx? Lowpass "Pi" filters are also common. And so on.
50Ω is by far the most common industry standard.
How to calculate your expected range is no trivial thing. The most important aspects tend to be receiver selectivity and sensitivity. And of course Tx output power, which depends on the PA + matching, how much losses you have in the Tx path and how well the antenna works. But also about regulations on the band you intend to use. The 2.4GHz is pretty much the Wild West so you probably don't have to worry too much about that. You have to be very concerned about other interfering radios however. Also there are product standards for Bluetooth, Zigbee etc that will state output powers.
20 meters open sight might be possible even at 0 dBm, though it would take a good design to keep up 20 meters reliably. And just don't expect 2.4GHz technology to work well through walls and similar.
1 comment thread
The data sheet gives a perfectly good example of using the nRF24L01 without an external PA. It has an internal PA capable of driving an antenna <-- just look at the design example on page 63 with PCB layouts on following pages and sure, you can use any suitable antenna. $$$$
Is there a way to calculate the approximate range of this remote controller?
You can use the Friis transmission equation to calculate the free-space link loss between isotropic transmit and receive antennas. It's based on carrier frequency and distance. For a simple antenna like a monopole, you have an extra gain of 2 dB at each end. Anyway the formula for isotropic antennas (null gain) is: - $$$$ Link loss (dB) = 32.4 dB + 20log(MHz) + 20log(kilometres) $$$$ Putting your figures in (and assuming a 1 GHz carrier) we get LL = 32.4 dB + 60 dB + (-34 dB). That's a link loss of 58 dB in free-space. But, we should probably add another 20 dB for fade-margin bringing us to 78 dB of link loss. Then, because I'm assuming monopole antennas, this drops to 74 dB.
You then might ask whether your receiver is capable of working with this amount of attenuation. After all, if only 1 mW is transmitted, your receiver is only going to see -74 dBm and, is this enough?
Well, there's another generally used formula that ascertains how much power a receiver needs based on noise statistics, ambient temperature and receiver bandwidth. Bandwidth is usually swapped with bit rate. Formula: - $$$$ Minimum power = -154 dBm +10log(bit rate) $$$$ So, if your bit rate is 100 kHz then the minimum receive power should be greater than -104 dBm. And, if my numbers match your true numbers then 20 metres should be a breeze.
According to the datasheet that Andy linked to, the chip can be configured for 0 dBm output. With the proper antenna, that is probably good enough for 20 m, but you'll have to test it yourself to know for sure. If your environment is noisy, or there are obstructions causing path loss or multi-path interference, then it's hard to say how much power is required. The relative alignment of the antennas also matter. This is due to both radiation pattern and polarization.
Another issue with a separate power amplifier is regulatory compliance. You need to look up what the maximum allowed transmitter power is in the band you are using and for your purpose. If that's 0 dBm, then any additional amplification would be illegal, for example.
Is there a way to calculate the approximate range of this remote controller?
In theory, for infinitely open space, yes. This would assume particular antennas at particular orientations. However, that's generally not very useful.
The real environment is too complicated to quantify easily, and usually unpredictable and changing anyway. There is no substitute for testing in the field.
That said, my gut feel is that 0 dBm at your frequency with properly matched and oriented antennas that have dipole-like radiation patterns really should work over 20 meters. But, if there's a wall or other stuff between the antennas, or one device is hand-held so that antenna orientation is arbitrary, or there's interference in your band, then you just don't know until you try it.
I would start with the 0 dBm output, then only add an amplifier if tests show you need it. Again, make sure you know the legal power limit for your band and purpose, and be sure to stay within it.
1 comment thread