The impact of putting an antenna in a box

IoT design a usually a matter of antenna as already seen in different previous blog post. Antenna performance is the assurance of your capacity to deploy your object in larger zone and a way to save energy by reducing transmission power.

As we will see, if you get a hardware component and simply put it in a box its radio behavior will be totally different as the box is impacting the transmission.

This post will practically show you the impact of a box on a device radio quality.

Let’s start by introducing SWR concept : you are transmitting a signal from a component (a emitter) to an antenna. If they are correctly match, in theory, all the power is transmitted to the antenna. We have a SWR equal to 1.0:1

In the reality the matching is not perfect and de transmitter impedance is not exactly wire impedance and not exactly antenna impedance. As a consequence every time a part of the power is not transmitted to the antenna but returned to the source.

As this part is not transmitted to the antenna, this energy is lost for the transmission.

The SWR is not the only factor impacting, once the power is transmitted to the antenna, the antenna itself has its own radiation capacity : the received power is not transmitted equally in all the directions and we can have antenna with or without gain. An antenna with gain will potentially radiate more power than received (but usually not in all directions). An antenna with no gain will usually radiate less power than received.

At the end your device transmission power will be something like :

Initial power

minus Power Lost in the connectivity to the antenna

minus Power Lost in the antenna itself.

Basically many things are impacting your antenna matching, first of all the wire between your emitter and the antenna, you can see this post on antenna wire design for more details. The next important thing is the box where you put your device and it’s direct environment.

It means as soon as you change the environment around the antenna you get a different impedance and a different performance. As we will see this performance is really different and this is the reason why the operators in LPWA request you to certify your device even if you just change the box or just put a certified device in a box.

So let’s take the example of a device communicating on Sigfox with this basic response over the air without any box:

We see we have a pic indicating where the loss is minimal. The tiny white arrow is corresponding to 868,2Mhz ; the central Sigfox frequency. This indicates the antenna matching is quite good. The RL is -12dB.

The following table is indicating the corresponding Power Loss … less than a 1dB. This is really great and indicate the impact of the wiring between transmitter and antenna is really low.

SWR / Rl / Power loss equivalence

As a consequence the radio performance is really good as we can see this report from sigfox backend ( You need to trust me as this has no comparison reference other than my experience for this location )

Now we are going to put the device (same one) in a standard ABS plastic box and close the box :

As you can see, just by changing the environment (not the location, not the power) we are really impacting the transmission. Basically we have a loss of 10dB by adding the box. As the transmission power is 14dB originally it’s a lot and in percentage of the power it means going from 25mW to 5mW equivalent. So 90% of the provided energy has been lost in the transmission.

Now let’s see what we get from the VNA analysis once in the box :

As you can see the signal is different and not centered anymore. The white markers on the right are indicating 868,2Mhz and now the antenna is centered on 810Mhz. The SWR value given by the VNA is about 11.0:1 corresponding to a theoretical loss of 80% about 6-7dB.

To solve this problem we can add a matching circuit between the emitter and the antenna. This is a couple of inductor and capacity to filter the signal and modify the inductance. My post on the use of a VNA will help you to understand how to make it and the link to the website I use to calculate the values. I did not had the right inductor/capacity but use some of the more appropriate. The result is quite impressive:

The horizontal scale on this capture is not the same as previously, here the pic is at 840Mhz, It’s not yet as perfect as the beginning but better. The SWR is now corresponding to a loss of 2dB equivalent to a loss of 36%.

This have an impact on the radio performance as in this transmission result :

For sure this is really less than the initial transmission, free in the air. But we can see the saving of the 4-5 dB (from 6-7dB of loss to 2dB of loss as given by the VNA). In fact you have to trust me a little bit as the rssi is not a stable value and you need to get more transfer for having a strong opinion.

If you have follow all what I’ve written until know you may ask me why we have about 10dB lost in the box when VNA says 6-7dB and why once matching corrected we still have something like about 6dB lost when we are expecting 2dB.

So ! this is a good question ! why this box is impacting the transmission about 4dB whatever the antenna matching is.

If any of you can give me some input on this point I’ll be happy to know !! let’s comment 😉

Putting the antenna in a box is not the only thing impacting the transmission : the place the box will be installed also impact the performance. In the following example you will see the difference for exactly the same object, in the same box, in the same place. The difference is : in the first communication the object are nothing around the 20cm, in the second configuration a metal plate is magnetized on it back (normal use case for this device). Check the results

Antenna RSSI First (free) Rssi Snd (magnetized)
3D16 -137dB Not received
04CB -119dB -130dB
0AE1 -107dB -124dB
0DC0 -127dB -121dB
3D14 -118dB -118dB
0A8B -132dB -134dB

In the concerned box it already have a large metal, magnetic area in the same place so it is really related to the environmental change more that the presence of a metallic plate blocking the signal (even if this is also playing a role) . The environmental change is important to consider, particularly when the object is transported by a human moving and putting it in its hand or a pocket. The technology of the antenna and its choice can positively or negatively impact these environmental interactions.

This entry was posted in Hardware, IoT and tagged , , , . Bookmark the permalink.

5 Responses to The impact of putting an antenna in a box

  1. David says:

    Attenuation of box is caused by dielectric properties of matherial from which box is made in combination with thickness of wall. Thickenss of wall doesn’t necessarily mean thinner is better. Abs is not the best matherial to cover antennas and standartly used abs could include some metal oxides and other impurities.

  2. Armon Chojnacki says:

    The reason you experience a loss with your antenna in the box is either due to the housing material (which judging by the loss you described it is some form of polymer and should have negligible effect on attenuation). However this being said any material inside the near field of any rf emitter/ antenna (d@NF = wavelength/2pi) will experience severe signal modification as this will create eddy currents in surrounding media which will contribute parts of opposing phase, hence lowering the total intensity of the detected signal. This is due to the fact that EM fields are vectorial whose scalar intensity in a media are calculated from the medium emmisivity and permitivity which are calculated from complex number expressions.
    In short try to keep your antenna as far away from any passive/conductive components in your device.

    Also make sure that you don’t bend the antenna from its original form. In reality when you do this it is not that you are losing energy you are actually changing the smith chart of the antenna and hence just changing its emission pattern.

    • Paul says:

      Thank you Armon for your comment, I need to investigate now 😉
      Putting the antenna something about 5cm (868Mhz wl / 2pi) from everything is not easy when designing IoT where we try to reduce the global size.
      Do you know if there are some rules like being in a distance equal to a 1/(2,4,8,16…) of the wavelength would make result better if it can’t be perfect ?

Leave a Reply

Your email address will not be published. Required fields are marked *