Alternate title: Why Medium Wave frequencies do not work for high speed data transmission
Much ink has already been spilled on the merits of HD Radio or lack thereof. The latest NAB sponsored project is the all digital mode test of AM HD Radio. As of this writing, four stations have volunteered to conduct the all digital (no analog) test on their facilities. I predict no great performance improvement in AM HD Radio system will be demonstrated from these tests.
Medium wave is the radio frequencies between 300-3,000 KHz, which is where the Standard Broadcast or AM broadcasting service in the US exists. There are several reasons why trying to make a 20 KHz wide linear digital signal work in those frequencies posses some technical problems:
- Electrical noise is very prevalent below 1,000 KHz
- Ratio of bandwidth to available frequency is low, therefore low data rates must be used
- Narrow bandwidth of existing antenna systems, particularly directional antennas and phasors
- Existing uses of the frequencies in question
Electrical noise plagues existing Analog AM broadcasting because of the frequencies being used and the type of modulation. Noise floors in urban areas can be very high, preventing any reception. Noise generators include things like high tension power lines, street lights, loose electrical connections, cracked insulators, appliances, etc. Add to that intentional RF generators such as broadband over power line and other power line communications systems and an RF noise cacophony exists. Digital modulation schemes are only immune to this noise as long as the noise floor is low enough to establish an acceptable signal to noise ratio. If the noise floor is too high, data errors cause dropouts and eventually the receiver will mute or blend back to analog. Thus, HD Radio tendency to jump between a higher quality digital audio source and a lower quality (made lower by the presence of IBOC signal) analog audio source, much to the annoyance of listeners. If the analog signal were not there, it would be silent. In short, there is no magic bullet when it comes to RF noise in the AM broadcast band.
The bandwidth available in the AM broadcast band is drastically less than in the FM broadcast band. AM channels are not more than 20 KHz wide (actually 10 KHz without adjacent channel interference), versus FM channels which are 200 KHz wide. This is due in part because AM is the senior service, founded when frequencies above 1,000 KHz were considered junk. It was necessary to use less frequency spectrum because there was less usable spectrum to be had. If AM stations were allowed the same bandwidth as FM station, there would only be room for eight stations in the entire AM broadcast band (520-1,710 KHz in US). These existing narrow channels can only carry so much data; 24 kbps to about 40 kbps without causing too much interference to adjacent channels. Any person can tell you, listening to a 40 kpbs .mp3 is not high fidelity or even a pleasant experience. And yes, iBiquity’s own AM HD Radio CODEC is better than .mp3, but not that much better.
Because the wavelength of Medium Wave AM broadcasting (or medium frequency) is very long compared that of FM, antenna systems tend to be large and inefficient. The AM Broadcast Band wave lengths run from 577 meters (~1,900 feet) at 530 KHz to 175 meters (~575 feet) at 1710 KHz. Very tall structures are needed to radiate this signal effectively and the RF energy needs to be distributed evenly about the center frequency through the entire 20 KHz bandwidth, which is difficult to accomplish. This is due to the physical change in size of an AM radiator over the required bandwidth span of 20 KHz. The becomes more pronounced at the low end of the band, where the required bandwidth becomes a larger portion of the available RF frequency. COFDM requires a linear response for its entire bandwidth, otherwise errors will be induced on an already lossy transmission system. Many existing phasors and antenna systems, particularly directional antenna systems are narrow banded and will not pass the IBOC signal without significant modification, which is expensive.
An example of this, the tower base impedance measurement, WTMN 1430 KHz, Gainesville, Florida:
As clearly demonstrated by the graph, the base impedance of this tower is not symmetrical about the carrier frequency of 1430 KHz. This is a somewhat extreme case, but most AM tower exhibit some asymmetry of sideband energy. The problem can be fixed, but it will be expensive and cash strapped AM radio stations can ill afford that kind of investment particularly of an experimental nature.
In the US, there is an overcrowding issue on the AM band, thanks mostly to the FCC’s issuance of too many licenses for the frequencies available. That is water under the bridge, but it does create an interference prone band. Digital radio signals are not, somehow magically immune to co-channel and adjacent channel broadcast interference. Similar to the situation with electrical noise, interference from other broadcast stations will induce errors on the received COFDM signal, thus causing drop outs.
Thus, it does not matter whether the system is tested in Hybrid digital/analog mode, or all digital mode, it will not be an improvement over plain analog AM broadcasting.