I have an old project which I am going to freshen up and migrate to a PCB and, maybe make it a bit more compact.
The main element is an nRF24L01+ transceiver of the type with a power amplifier and an external antenna. The specific version is similar to this:
I am happy with the range etc. so I don't want to make any big changes but I am now wondering about substituting the supplied antenna with a trace on the PCB to save some space.
My first thought were to have a vertical trace say 3mm wide of the same length as the existing antenna (or maybe two matching traces on both sides of the board) on an area otherwise free of copper. There is no possibility of adjustment of the current antenna because of the restricted size of the box and this works fine, so I could imagine that a fixed trace would be no problem.
Of course, with a PCB it is possible to specify all sorts of antenna layouts, but is the simplest the best ?
There's an old appnote from TI that precisely describes a 2.4GHz trace antenna. It appears to be what's used in the Lolin/Wemos D1 Mini. As far as I've been able to tell, it works extremely well. It works from inside my metal mailbox (the box is a bit leaky, but I was still impressed).
I know almost nothing about antennas but I wonder what happens if you connect a low power PCB antenna to a high power transmitter - does it upset the balance of the transmitter circuit?
Considering the low price of the low power nRF24 modules with PCB antenna I would not consider it worth the trouble trying to down-grade a high power module even if the change would work.
Robin2:
I know almost nothing about antennas but I wonder what happens if you connect a low power PCB antenna to a high power transmitter - does it upset the balance of the transmitter circuit?
Robin2:
I know almost nothing about antennas but I wonder what happens if you connect a low power PCB antenna to a high power transmitter - does it upset the balance of the transmitter circuit?
It's not the power, it's the impedance that matters. A miniature antenna if correctly matched (study "SWR" ) will radiate most of the power (and heat up with what it does not radiate). If not correctly matched, then the transmitter itself will dissipate the balance which it must be designed to withstand.
Good point. a low power transmitter can usually tolerate a badly matched antenna because the heat generated by the reflected power is minimal. So a lot of actual low power antenna designs haven't been optimized to the extent that high power antennas have, just made to get "good enough" VSWR to radiate a reasonable percentage of transmitted RF. High power antennas can't be so casual because of the transmitter dissipation.
My first thought is "copy an existing design" as well. Designing such an antenna doesn't sound like a trivial thing to do, to me.
Note that the ESP12 has two different antenna designs: a few years ago a new design (a simpler looking trace, in a way) was used. That design has supposedly better range. For my application I couldn't tell the difference, they both work fine, but I haven't been looking for the limits.
Paul_KD7HB:
First make tests by gluing a copper wire to your circuit board material and test that. So much easier to adjust.
The trace I see on the latest ESP12 antenna design is branched, and has no constant width. Likewise the design given in the AP linked in #2, which indeed looks a lot like the old ESP12 antenna design (I don't have any such module at hand so can't compare in detail). I suspect you can not simply replace it by a round, constant tickness copper wire.
To make the module smaller I sawed the antenna part of the PCB from a low-power nRF24 module and then soldered on a short length of flexible wire as an antenna. It works fine for the short range I'm interested in. I have not tested its ultimate range.
Thanks for the comments so far.
Indeed the application is a notifier system attached to my mail box. It wakes up periodically, switches on an infra red beam break detector and sends a message.
Anyway, the "PA" part of this module, referenced in the OP, appears to be a "range extender" with these markings:
rfaxis
x2401c
Clearly, NRF24L01 modules without this range extender can have the small "square wave" like PCB trace and many do. For example,
However, the ones with the range extender appear all to have external antennas, probably because the small PC trace is deemed to be insufficient or inappropriate. So I don't think it make sense to copy one of these small ones.
Here is a tear down of an typical 2.4GHz antenna. It is not what it initially appears to be and that is a simple straight piece of wire. It is actually a di-pole construct.
So my current thinking is to create a dipole trace of the correct length on the PCB. I don't thnk with a 2 layer PCB there is much chance of creating the equivalent of a 50 Ohm coax cable running to it, or is there ? So I'll have to put the output of the module physically close to center of the di-pole.
Well in case you want to look at it, my mailbox notifier project is here:
It uses the Lolin D1 Mini, which has the trace antenna, and it works fine, even from inside the mailbox. I also tested it with the D1 Mini Pro, which has a connector for using an external antenna (you have to reposition a zero-ohm jumper for that), and that works fine as well with a standard 2.4GHz wifi antenna. Then I also tried it cutting off the wifi antenna and simply exposing about 30mm of the pigtail coax center conductor, as shown in your picture, and that works fine as well.
My guess is that the trace antenna may actually work better, and be less directional, than a bare wire, but the wire would let you get the antenna outside the metal mailbox, which would be a good thing.
ShermanP:
My guess is that the trace antenna may actually work better, and be less directional, than a bare wire, but the wire would let you get the antenna outside the metal mailbox, which would be a good thing.
That is important. Of course, in some places that would be an invitation to thieves if it's obviously an antenna.
I have compared 2.4Ghz LoRa devices and the ones with PCB antennas were much worse, circa 10dB less effective as radiators, than conventional SMA antennas.
Simple wires were not far behind the SMA antennas, as I recall.
When the ground dries out in the UK, its been raining a lot, I will have another go at the comparisons.
