# ISM band bandwidth compared to other bands for theoretical WLAN use, performance?

As a purely theoretical question, what would it be like if wireless computer networking equipment was redesigned and allowed to operate at some other band than the unlicensed ISM band which is presently used? Some other band means in this context a "better" band, as in some licensed band considered to be "better" than the ISM band. E.g. the cellular band used for cell phones, which are allowed to transmit at higher power levels.

This is not a question regarding the politics or cost involved in licensing these bands and it should be ignored completely for this question to try keep things more simple.

By wireless networking I mean the typical 802.11* variants, which of course would have to be redesigned at the physical layer to work with a different band.

By "what it would be like" I mean:

• A) What would the end result in difference in theoretical bandwidth be?
• B) How would it affect power savings in terms of better SNR?
• C) How is MIMO affected? Is MIMO more easy to implement and exploit on some radio bands than others?

For A) I'm assuming that using the Shannon-Hartley theorem would work:

$C = Blog_2(1 + \dfrac{S}{N})$

If only I knew what bandwidth $B$ to pick and what the electrical engineering considerations to take into account would be. I'm assuming that you could use a very wide arbitrary band just to improve your theoretical results, but it would be costly engineering wise? Just to answer a theoretical question like this, how would you even determine what frequency range would be "realistic" in this context? Would you simply have to pick some arbitrary range, which also renders the question itself meaningless? If this is the case, then simply assume that the cellular frequencies were suddenly available for use, or even all of the UHF band if that is easier to answer.

As for the S/N ratio, I'm assuming that when moving away from the ISM band, it becomes more safe (without getting into the debate regarding possible harm from cell phones), and simply increasing the transmission power would also allow for higher bandwidth. However it would also cause more noise for everyone else?

B) I'm assuming this is directly related to A, as some frequency ranges have different properties in terms of transmittance, absorption and reflectance? In addition there would probably be interference from other devices on some bands which are strictly speaking not intended as radio transmitters, e.g. microwave ovens for the ISM band.

C) For this, I don't know. I know MIMO is a big and essential part of the physical layer for modern WLANs, but I don't know how it relates to different frequency bands. Is part of the reason MIMO is used simply that the transmission power is so limited on ISM, and that the band has other limitations/issues to deal with?

You don't need to ask this as a theoretical question, because it already exists. The 13-centimeter band overlaps the 2.4 GHz ISM band used by Wi-Fi. A licensed amateur can operate Wi-Fi equipment on the standard channels at a higher power, or operate on channels outside of the ISM band (but still in the amateur band allocation) if the equipment supports it. Consumer equipment usually does not (without hardware modifications), but a lot of professional hardware does, as there are commercial licenses that grant RF privileges beyond the ISM limitations also.

Plenty of people have done exactly this. See HSMM.

As for what effect this has, the Shannon-Hartley theorem is indeed the right tool to model the ideal behavior, which isn't far off from real behavior. Operating at a higher power increases the $S/N$ term (by increasing $S$), allowing a higher channel capacity to be sustained over a long distance, or being more able to overpower noise (be it natural or man-made).

Moving outside of the ISM band also increases the $S/N$ term, but instead by reducing $N$, simply because there are so many unlicensed Wi-Fi APs, cordless phones, and microwave ovens operating in the ISM band. Moving out of the ISM band, you avoid a whole ton of man-made interference.

The bandwidth ($B$) refers to the channel bandwidth, and this will be defined by the modulation. For example, Wi-Fi uses 20 MHz or 40 MHz channels depending on device configuration. PSK31 requires about 31 Hz, SSB about 4 kHz, and the FM used on amateur VHF about 15 kHz.

Regulations permitting, you can make the bandwidth as large as you want. If you can maintain the same signal-to-noise ratio, then the Shannon-Hartley theorem says that you can have increased channel capacity. However, increasing the channel bandwidth also increases noise power. If you double the bandwidth, then there is twice as much noise power within that bandwidth. If you don't also double your transmit power, you are decreasing the signal-to-noise ratio.

If you make the bandwidth wide enough, you will need to significantly redesign the antenna, filters, and other components. There are a class of modulations that work in this mode. Check out ultra-wideband (UWB) and spread-spectrum (SS). You will also have to solve the problem of how multiple users can share the wide swatch of spectrum you are now monopolizing, and if you read about UWB and SS you will see they each have different techniques for accomplishing that.

MIMO is a complex collection of techniques, and as such it's difficult to point to any one reason why it's used. It's certainly not the case that MIMO is useful only in power-limited ISM bands. As a counter-example, modern radar uses phased arrays instead of mechanically steered antennas, and these are some of the highest power transmitters in the world that don't qualify as weapons.

The techniques in MIMO are also applicable to any frequency. Shifting wireless networking to any other frequency would only affect MIMO to the extent that it affects radio communication in general. For example, if frequency is high enough, Earth's atmosphere becomes quite absorbent, and radio communication becomes difficult.

• I should have made it more clear, for A) when I ask about end result in terms of bandwidth, I really am wondering what the number of bits per second would be (C in Shannon-Hartley), i.e. how big the difference would be. How many orders of magnitude are we talking, is it possible to ball park it somehow? I'm just wondering how "bad" the ISM band really is, compared to other alternatives. – AttributedTensorField Jan 5 '15 at 8:15
• @AttributedTensorField Sure, get a field strength meter and measure $N$ for the ISM band. Then measure it for some other band. Then put the values into the Shannon-Hartley equation and do the math. The improvement might be anywhere from none (rural area with very little man-made interference) to extreme (dense apartment buildings where everyone has a microwave and a 2.4 GHz Wi-Fi access point, pushing the noise floor 40dB higher). – Phil Frost - W8II Jan 5 '15 at 12:56