By Bruce McCalley
The alignment of superheterodyne radios has been a mystery to more technicians in the radio/television field than almost any other servicing procedure. Finding a repair shop that really knew how to do the job was (and perhaps still is) all but impossible, even though many shops had the proper, or at least usable equipment.
During the days of radio, alignment was not really so critical, but with the advent of FM and Television, its importance became paramount. Still, most shops either ignored the procedure, or farmed the work out to someone who knew how.
FM and television alignment requires sweep frequency signal generators, and oscilloscopes to do the job properly, but even the most complicated AM radio can be aligned fairly well (although perhaps not optimally) with an accurate standard signal generator. Note the word "accurate." Of the many typical radio repair generators (Hickok, Supreme, and the like) I have used over the years, NONE have been accurate. Close, but not "right on." Most of us got along by aligning the intermediate frequency (IF) stages by setting the generator's dial to 455, 456, 460 or whatever, and then peaking the IF stages. To be fair, the actual IF frequency is not that critical, but the RF alignment is.
RF alignment was done using broadcast stations instead of the generator. The average customer would never know that the IF stages which might have been designed for, say, 456kHz, were actually 461kHz, but they sure could tell if KFI at 640 came in at 650 on the dial!
Those of us in the profession who really cared, used some sort of frequency standard to calibrate our equipment. For example, I used crystal-controlled "standards" and the National Bureau of Standards station WWV (by means of a short-wave receiver, of course) to check them.
Why is the accuracy of the IF so important? Sometimes it's not, but on those sets that use a cut-plate tuning capacitor, the tracking of the dial will be off. In other words, you can set the RF sections to be right on at, say, 1400, but the other end of the dial will be off. This was due to the oscillator not tracking with the RF stage(s) to generate the correct IF frequency across the dial. This was not very critical on the larger sets which used tuning capacitors with all sections alike and "padder" capacitors in the oscillator circuits, which adjusted the tracking at the low end of the dial, since the padder can usually make up for the IF error.
Before getting into some of the "tricks" in alignment, perhaps we should describe the procedure in general. IF alignment is performed first. Using a signal generator at the prescribed frequency connected to the mixer grid circuit (and the RF stage disabled so as to prevent some broadcast station from interfering), and some form of an output meter; each stage is adjusted for maximum output. Be careful to use the lowest signal level that will give a steady output indication; too much can overload the circuits, and too little might not give a reliable output indication. Generally you use the lowest level that will give an accurate reading so that the automatic valume control (AVC) circuit will be inoperative. (Too high a level will cause the AVC to effect the output, making it difficult, if not impossible, to get an accurate reading as the circuits are peaked because the AVC will reduce the output as the signal is increased.)
With this done, the signal generator is loosely coupled to the input of the receiver, and set to, say, 1400kHz. (We used kilocycles but these are hard to find today. KiloHertz can be substituted if one is careful.) The oscillator is set to bring the 1400 kHz in at 1400 on the dial, and then the RF sections are adjusted to also peak on the output meter. The generator is now set to about 600, and hopefully the signal will come in at 600 on the dial. If it doesn't, and the receiver has an oscillator padder, the padder is adjusted to bring the signal in at 600. Since this effects the high end, you must now go back and reset the 1400 oscillator trimmer to track (don't touch the already-aligned RF trimmers), and then go back to 600; repeating as often as necessary to bring both ends of the dial into tracking.
Those sets with cut-plate oscillator sections have no adjustment for the oscillator at the low end. What do you do if the dial is off? Adjust the IF frequency! This takes a bit of experience but it's not really difficult. The easiest way is as follows (and we presume you have pre-aligned the IF at some frequency as described above). Let's presume the dial reads 590 when you are receiving a signal at 600kHz. Detune the dial towards 600 but not so far that you lose the signal. Adjust the IF trimmers to maximum, then reset the oscillator at 1400 as described above. Go back to the low end again and recheck, and repeat as necessary. If you find the error is greater, try adjusting the dial a bit lower than 590 (in this example) and repeat. With experience, the procedure becomes simple. On most of the 5-tube sets we serviced, we never used a signal generator. We set the oscillator and RF at 1400, checked the low end, and peaked the IFs using broadcast stations at the low end of the dial.
If the dial tracks at the low and high ends, but is off in the middle, you have a problem. On those sets with oscillator padders, adjusting the IF frequency, and then the oscillator tracking might correct the problem. More often than not, though, the dial itself is not accurate due to manufacturing tolerances of the variable condenser. If the tuning capacitors have cut slots in their rotor plates, bending the sections of the oscillator plates that are meshed with the stator plates might do the trick. You can do this with the RF sections as well but it is seldom worth the effort since these circuits are relatively broad in tuning (not as selective).
Some receivers have adjustable cores in the RF sections, in addition to the trimmer capacitors. These cores are always adjusted at the low frequency end of the dial, and the trimmer capacitors are always adjusted at the high end. Each affects the other. Set the capacitor first, then the core, and repeat as often as necessary until tracking is accurate. The core in the oscillator coil acts much the same as the padder capacitor. (If there is an oscillator padder and an adjustable core, the manufacturer may have a suggested procedure for proper alignment.)
On receivers which use some form of bandwidth control (called 'selectivity' or 'fidelity' or similar), the control must be set to the sharpest setting (least fidelity or bandwidth or whatever) before IF alignment. There were a number of schemes used for bandwidth adjustment, among them being: (a) mechanical variable coupling (Philco, Hammarlund, etc.) in which the primary and secondary of the IF transformers were moved mechanically closer for greater bandwidth; (b) detuning (Scott, etc.) where a variable capacitor shifts the resonant frequency of the primary and secondary; and (c) a switch-operated system in which coupling is altered by switching in a coil or coils, or by detuning, or a combination of both (RCA, Hallicrafters, etc.). All of these must be set at minimum bandwidth for alignment. (HiFi AM radios are one case where sweep frequency alignment can be used to advantage, but AM sweep generators for broadcast frequencies are rare indeed!)
Output meters are generally of two types: one is an indicator across the speaker voice coil (which might even by your ear!); and the other is an indicator across the AVC line at the detector. The indicator might be a meter, an oscilloscope, a tuning eye, or whatever. The best is the one on the AVC line. AVC voltage increases with signal level and is therefore a positive indication of a peak in adjustment. The output at the speaker, and particularly just listening for maximum, is far less accurate. Why? Because if the set has AVC, the increasing AVC voltage tends to reduce the audio output (that's why it's there; for Automatic Volume Control!) so the "peak" is apparently broader. On early sets with no AVC, of course, there is no choice. Incidentally, using a fixed bias on the AVC line, such as a battery, or sometimes even shorting it out, can aid in accurate peaking. The biasing battery voltage should be the same as the voltage which appears on the AVC line during normal operation. The idea is to prevent the AVC from affecting the output level as the circuits are peaked. Of course with the AVC system disabled you cannot use a meter connected to it for an indicator.
"High Fidelity" receivers sometimes require a better method of IF alignment. One these sets the IF band-pass (or response curve) should not be a sharp peak, but rather a curve with a relatively flat top of, say, ten kHz width. One might presume that the adjustment of the "fidelity" control or switch would result in this flat-top curve but in practice it doesn't always do so. One way to check the curve is to reset the signal generator a bit above and also below the IF frequency and noting the change on the output meter. Moving from, say, five kHz below to five kHz above the IF frequency should result in the meter reading a nearly constant voltage. If it does consider yourself fortunate. Usually one side of the curve will be higher than the other, or there may be a dip as the generator passes through the desired IF center frequency.
Without a sweep generator (frequency modulated oscillator) and an oscilloscope it is pretty difficult to correct such curve deficiencies. Guessing at which IF adjustment screw to turn can result in perhaps reducing the level on one side of the curve but increasing the other side, reducing the overall curve, or moving the IF center frequency itself. With a sweep generator and scope it is easy to see what affects what.
Typical IF response curve. Note that one side of the center frequency is higher than the other. Ideally both sides should be equal, with a minimum of dip in the center.
Most signal generators which have an FM feature contain a separate FM oscillator of some fixed frequency. This frequency-modulated oscillator is then coupled to the regular variable oscillator and the resulting beat frequency is used for the alignment. For instance, if the FM oscillator is 2000 kHz, the variable oscillator would be set at 2450 kHz (or 1550 kHz) to produce a 450 kHz beat signal for the IF. Easier said than done! As a rule neither the 2000 FM signal nor the 2450 variable signal can be relied on to be accurate, let alone being able to set the dial on the generator that accurately. To overcome this problem the IF amplifier is pre-aligned at its desired frequency and then the sweep signal is applied so that with the sweep width control set at its minimum, it is adjusted also for maximum output. Now advance the sweep width control and the center frequency should be correct. A second signal generator, set at the desired IF frequency and loosely coupled to the set, can also be used as a "marker generator" which will put a beat pattern on the scope which will indicate just where on the curve this frequency will be.
Ideally you would check the curve beginning with the last IF stage. When that stage has the proper curve the generator is connected to the next previous stage and its transformer adjusted for the desired response. Then to the next previous stage, etc. Failing to do the alignment in this way could result in one transformer being adjusted to correct the poor alignment of another which could result in some loss in overall gain.
Each stage will have its own typical response curve. For example, the last stage may have a double-hump curve with a dip in the center. This center dip would be compensated in one of the earlier stages which would have a peak at the center frequency. Or perhaps the last stage would have the peak and a previous stage would have the double hump. The manufacturer's alignment might give this information but it is not likely this data will be at hand. Alignment stage by stage will reveal the typical curve for that stage as you make the adjustments.
Another method of alignment, and perhaps much easier, is to use the signal generator set at the desired IF frequency and to use an audio modulation frequency of, say, ten kHz. Most signal generators use an internal audio generator of about 400 Hz, but also have provision for an external audio signal. Here you need an audio oscillator set at the desired frequency connected as the external source. Since the ten kHz modulation generates an IF signal with ten kHz sidebands, adjusting the IF transformers for maximum audio output should result in a somewhat flat-top response curve.
Remember, again, keep the signal generator output down, and the AVC disabled, so as to not overload the circuits. Ideally, the signal level should be the same as the normal signal level during regular use of the receiver.
There is no substitute for experience. Take your time when doing
alignment. Don't get discouraged if things don't come out right the first (or
second, or third) time. Anyone can do the job if they take care. Now that you
have gone to all this trouble, wouldn't it be nice if there were some programs
being broadcast that would make it worth while!