AM Tuners

Our contributor David Smith (not to be confused with David "Anonymous" or David Rich) kicks things off with this essay.

AM Tuners

AM radio has a long, rich history, with the first commercial U.S. broadcasts beginning in the early 1920s. By the 1930s, the early phase of technology was maturing with superheterodyne radios and receivers being the norm and some great sets being made by communications specialists like Hammarlund and Hallicrafters and by radio set makers such as E.H. Scott and Zenith. AM radio served the nation well, but a desire for greater fidelity and immunity from atmospheric interference got experimenters (specifically Edwin Armstrong) working on the FM process. Still, FM wouldn't become a factor until well after WWII and AM sets continued to evolve. The hi-fi boom of the '50s became synonymous with the new LP and FM broadcasting, both being pushed by the new component hi-fi set makers. When FM became popular, for the most part, focus shifted to FM tuner design, and AM sections became a low-cost add-on feature of barely acceptable performance. There were, and still are, exceptions, and these are what will interest us most. But first a bit of background.

AM stands for amplitude modulation. A carrier frequency (the nominal frequency that you tune to such as "1210 on your AM dial") has its amplitude modulated with the intended audio signal. You can actually see the waveform of the audio impressed on top of the carrier. Benefits are that AM was a very easy modulation scheme to use in the early radio days and, more importantly, very easy to receive and demodulate.

This is an important point to make regarding AM: it's actually pretty easy to make a functioning AM tuner. The frequencies are relatively low. RF amplification is plentiful, demodulation can be extremely simple. The AM broadcast band is at the noisy end of the RF spectrum, so phenomenal sensitivity isn't required. A superheterodyne tuner can be achieved with three transistors and a diode, or a single IC. In the tube days, millions of AC/DC radios were sold with four-tube plus rectifier layouts. Usually the performance bar for AM was set pretty low. There are a lot of high-end FM tuners that have the bare bones of an AM section added just so that you can pick up a little talk radio. In fact, why would you want anything more?

The main reason is for the fun of AM. Programming on AM is different from FM with an emphasis on talk radio for sports, news, politics and personal advice. In fact, AM is "funkier" with all manner of oddball programming that wouldn't be mainstream enough for FM. For me, AM also reminds me of my youth, with late nights spent trying to drag in distant stations on a homemade set, or the drive from Cincinnati back to Cleveland late on Sunday night after visiting my wife-to-be at grad school. The old Chevy only had AM, but listening to WLS in Chicago and other stations would keep me awake. Others will find AM essential to pick up a station or information that they can't get on a local FM station. I listen regularly to NPR, unavailable in Canada, from WNED-AM, a Buffalo, New York station. Even though most of these examples are talk-oriented, a better-sounding AM tuner makes the listening more pleasurable. And there are still some worthwhile music stations on AM.

The most useful AM tuner feature. With AM, daytime and nighttime reception are very different cases. Daytime reception is good for local stations, with a range similar or a little further than FM's range. The range is not so great as to cause a lot of interference from stations in distant cities. Nighttime reception becomes very long distance. On any night, I have no problem picking up stations from all over the east coast from my Toronto location, including New York City, Atlanta, Philadelphia, Boston, Cleveland and Chicago. So we have daytime reception which is good for local stations and nighttime reception where huge distances can be covered and the band becomes chockablock with stations. We have a system that isn't immune to interference as FM is (no AM rejection or static rejection, no capture effect). We want wide bandwidth for best fidelity but require narrow bandwidth for rejection of adjacent channels and atmospheric noise. It isn't hard to see why the most useful feature on a "better" AM tuner would be to have multiple IF bandwidths. With dual bandwidths you can have a narrow setting for cutting through interference and a wide bandwidth for maximum fidelity when a station is in the clear. Readers of this site are well aware of the desirability of dual bandwidths in better FM tuners. They probably don't realize that the idea originated in AM, where its effect is much more obvious and essential.

IF bandwidth becomes the central parameter defining the sound of an AM tuner. In fact, assuming no significant audio response shaping after the detector, the IF bandwidth, the IF filter's corner shape and falloff slopes are all directly translated into the audio frequency response of the tuner. Generally, due to the double-sideband nature of the signal, the IF bandwidth divided by two equals the audio bandwidth. Thought of another way, the carrier frequency, tuned to the center of the IF, becomes the new DC frequency, and all sidebands will have audio pitches directly related to their distances from the carrier.

This applies to adjacent stations as well. Say you are tuned to a station at 900 kHz and another station of similar strength is on the next available slot above, 910 kHz. With AM reception, there is no inherent rejection of this adjacent undesired station. The highest harmonics of the 910 kHz programming will be closest and, rather than having their pitch defined by their distance from the 910 kHz carrier, they are now defined by their distance from the desired station carrier of 900 kHz. In effect, the audio frequencies for the adjacent station are inverted, with highest frequencies demodulating as lowest and lowest frequencies demodulating as highest. In fact, the 910 kHz carrier will be resolved as a 10 kHz fixed tone. The audible effect will be that the highest frequencies of the adjacent station will intrude as "splatter" if your IF bandwidth is narrow or indecipherable "monkey chatter" when the bandwidth is wide. The adjacent carrier will be a fixed whistle. Communications receivers tend to have very tight filtering with steep skirts to reduce breakthrough from adjacent or alternate channels, but clearly there is a trade-off between this filter steepness and sound quality.

You can tune an analog tuner off the nominal center frequency a moderate amount and both improve the audio response and avoid possible interference on one side, although distortion will rise from imbalance between the two sidebands. One high-tech solution to this increase in distortion is a technique known as synchronous detection. Rather than using peak detection from a simple diode detector, an oscillator is locked in sync with the original carrier but at a high and constant level. This drives the detector harder and reduces the distortion due to full or partial fading of either the carrier or one of the sidebands. A more sophisticated version of this uses cancellation techniques to null out sidebands (and interference) on one side or the other. It is called "selectable sideband" and can really make a difference under the right conditions. For the most part, these advanced techniques have been available only on recent high performance shortwave receivers.

Whistle filters are also a desirable feature. As described above, another station may be located 10 kHz away (9 kHz in Europe) and its carrier would contribute a steady 10 kHz tone or "whistle." Better tuners such as the Scott tube tuners would include a deep and narrow audio notch filter at that frequency to remove the offending tone, hopefully without removing too much audio along with it. Later tuners used this feature less frequently. With steeper ceramic filters, the philosophy was just to roll the audio from 2 kHz and let the IF sharpness kill the 10 kHz whistle. It works, but at the cost of very muffled sound quality.

Some tuners used a combination of multiple IF bandwidths and several audio filtering steps to give more steps of total bandwidth. The McIntosh MR 55 and MR 66 use three IF filter bandwidths for "narrow, medium and wide 1," each with a whistle filter in place. The widest option, "wide 2," foregoes the whistle filter, extending the audio response of wide 1 from 8 kHz to 12 kHz. Fisher also had several tuners that used the same scheme. Another feature to look for would be 3-gang tuning and an added RF amplifier stage. While 4- and 5-gang tuning seems like the entry level to "better" FM tuners, you will rarely find more than 3-gang tuning in an AM tuner. Still, 3-gang tuning meant that a tuner had an RF amplifier ahead of the mixer stage and would have both better RF selectivity, to reduce image response, and a decidedly better weak signal noise performance.

The golden age of AM tuners. Stereo was first available to consumers around 1955, as two-track 7½" reel tapes. The stereo record was released in 1958. It wasn't until June of 1961 that the FCC final approved the Zenith approach to FM multiplex. In the meantime, the FCC allowed broadcasters to fill the void via AM/FM stereo simulcasts. Odd as it seems, the right channel was sent out on an FM channel and the left channel was sent out on AM. You may wonder about the matching of channels and the quality of imaging made in this way, but remember that this was in the very early days of stereo when it wasn't all that unusual to convert to stereo by picking up a second speaker that wasn't even of the same brand as your first speaker. ("Sound from the left and sound from the right - that's stereo!") I don't know how prevalent these broadcasts were, but evidently they were common enough that a number of tuner manufacturers built tuners to cater to them. They are easy to spot in that they are really two independent tuners on one chassis. Look for two tuning knobs, two dial pointers, and complex mode switching for FM, AM and AM/FM stereo.

The benefit here is not that any AM/FM simulcasts exist today but that, for these simulcasts, the manufacturers gave their best shot to making AM sections approach the sound quality of their FM sections. I won't say that they would fool you into believing that they were FM, but during daytime hours and using the widest bandwidths the AM sections did come passably close to FM quality. Typical specs and performance would be a frequency response of 20 to 8000 Hz (or wider with whistle filtering bypassed), with 0.5% distortion and a signal-to-noise ratio of about 45 to 50 dB. Most of these tuners come from the 1958-to-1961 era and include such notable sets as the Fisher 800C receiver, the McIntosh MR 66, the Scott 330 and 331, and several Heathkit and Sherwood models. Other sets from just before or just after this period had similar AM section properties, but without the fully independent tuner sections. For our purposes these would be equally desirable and would include the Mac MR 55, and the Fisher AM80 and FM90T variations. In fact, just about any AM/FM tuner from this era with 3-gang tuning and at least two positions of AM bandwidth is worth a try, and likely to give better AM performance than a lot of modern sets.

After the vacuum tube tuner era, quality AM tuners became less common. Many of the flagship tuners would forego AM altogether, while the same companies left AM as an extra feature for their all-encompassing receivers. As ICs became more common in the late '60s, it became typical to have a single IC for the AM section. A double transformer/single ceramic resonator became the most usual IF selectivity device, and high selectivity seemed to win out over wide bandwidth as the overriding concern. By the "supertuner" era of the late '70s, the Leonard Feldman reviews in Audio magazine would praise any tuner that had response out to 5 kHz as being well above the norm!

Modern trends and AM Stereo

As AM demographics declined and broadcasters looked for something to give them an edge against FM, AM Stereo was offered up as something to bring back listeners. A number of systems were proposed by various radio equipment manufacturers such as Harris and Motorola. The FCC began testing systems in earnest in 1984. In their typical wisdom, though, they gathered all the data from laboratory and field tests, yet declined to pick a winner. Their attitude was to "let the market decide," and a number of systems went on the air. Radio manufacturers could choose to back a particular horse, to make multistandard receivers or, more typically, to stay out of the AM Stereo arena entirely. The FCC finally did decide in favor of Motorola and the C-QUAM system in 1993, but by then consumer desire for AM Stereo had faded. A continuing shift of AM to talk and sports formats made stereo a moot point. A few hardy souls still broadcast in AM Stereo and might provide an interesting DX possibility.

One benefit of the AM Stereo debacle was that frequency response standards were brought forward as an issue. The NRSC, an industry standards body, pushed for the adoption of standardized frequency response for transmitting and reception. It seems that although many receivers had audio response severely rolled off from 2 kHz, many broadcasters worked at keeping their transmitters flat to 10 kHz or higher. AM Stereo, as with FM Stereo, worked best with linear phase IF filtering which usually has a Gaussian or humped IF shape rather than a flat-topped shape. This would tend to roll off the AM audio at a gradual rate. Broadcasters were encouraged to use a pre-emphasis that mirrored this rolloff for a flat overall response. It was also felt that this pre-emphasis would roughly compensate for unsophisticated sets with narrow response.

Whether you have local broadcasters transmitting in AM Stereo or not, the few sets made by Sony, Carver and others might be worth tracking down for their better AM performance.

IBOC

The latest technical shift for AM is the trend to IBOC (in band, on channel). This is a digital technique that adds multiple digital audio sidebands to the AM signal. The program content is carried both as a conventional AM signal and in duplicate as a 4½ second delayed digital audio signal. The thinking is that the consumer with an appropriate new IBOC-equipped radio will tune in to his favorite AM station and his radio will automatically switch over to a quiet and interference-free digital signal (when it is present). Some see this as the salvation of AM radio and others see it as totally ruinous. No one seems to be neutral in the matter except, perhaps, for the typical consumer who doesn't seem to care much at all. Two issues that should concern radio enthusiasts are that the digital sidebands seem to have little degrading effect on that channel' AM modulation, but can severely degrade the listenability of weaker adjacent channels. Secondly, any station that wants to convert to IBOC transmission must curtail its AM signal at 5 kHz, thus limiting its potential quality. The jury is still out about whether IBOC will flourish or go the way of AM Stereo. An optimistic view would be that, as with AM Stereo, a few receivers will be built with better-quality AM sections to showcase the technology. Time will tell.

Tuner Recommendations

As time permits, I'll add some capsule reviews of a few of my favorite AM tuners including my McIntosh MR 55 and MR 66, and Fisher R200B. Here is a listing of a number of sets that have a reputation for better AM performance. I'll leave it to other readers to give their impressions of these notable AM tuners and suggest their own favorites.