# How exactly to track oscillator in superhet receiver

The IF is 490 kHz. I have learnt that the local oscillator and the tuning block before the IF stage must "be aware of each other" or track. So in my case, the tuning of both the circuits should differ by 490 kHz. See this article:

Specifically, it is mentioned that:

Since we have to tune the RF amplifier section throughout the entire broadcast band, the frequency of the local oscillator must also vary in a manner that it always maintains a gap of 455 kHz. To achieve this condition, the Local Oscillator and RF Amplifier section are 'ganged', i.e. their tuning condensers are connected/ganged mechanically in such a way that when we tune the variable capacitor in the RF section, the variable capacitor in the local oscillator also changes its value, it 'tracks' the frequency to which the 'Aerial Circuit' is tuned and remain separated from the tuned frequency by 455 kHz up.

I want to know how exactly is this tracking done. As far as I understand, the variable capacitor is made common to both the oscillator as well as the front end tuner. How will this lead to a difference of 455 kHz (or 490 kHz in my case) in the tuning of the tank as I vary the frequency of the oscillator?

• As the quote says, there are two mechanically connected variable capacitors, one to tune the preselector filter and the other to tune the oscillator. Typically, there is one shaft with two groupings of plates on it, each of which intermeshes with its own set of stator plates. google.com/search?q=dual+section+variable+capacitor – Chris Stratton Aug 2 '14 at 13:52

The problem with using identical capacitor sections is that the LO has a proportionally narrower tuning range than the front end, so tracking will be poor. It can be improved by adding a padding capacitor in series with the LO tuning capacitor. In a broadcast AM radio the value of the padding capacitor is usually is a little larger than the tuning capacitor's maximum value.

To fine tune the tracking a trimmer capacitor can be added in parallel with the padding capacitor. Some tuning mechanisms have this trimmer built in, along with smaller trimmers in parallel with the tuning caps to adjust tracking at the high end.

• Ok. I'll try this. Actually I thought both sections will work on frequncies exactly IF apart. I'll try the padding scheme you mentioned. – Plutonium smuggler Aug 3 '14 at 0:50

There is not just one variable capacitor. There may be one unit, but this unit contains at least two variable capacitors. These are all mechanically activated by the same motion so that they always move together. One capacitor adjusts the local oscillator frequency, and the other the center point of the RF bandpass filter.

Usually the front end filter is somewhat broad so that small errors in the tracking don't make much difference. The main purpose of the front end filter is to eliminate image frequencies. A product mixer produces both the sum and difference of the frequencies. Working backwards, there are therefore two different RF frequencies that result in the IF frequency after mixing with the LO. The job of the front end filter is to squash the other frequency, called the image frequency.

• you mean to say I can use exactly the same values of capacitor and inductor for the front end tank circuit as in the oscillator ? If that is the case, then wont the mixer generate a bunch of frequencies instead of the desired frequency, because its input itself would be a bunch of frequencies ? At least this is what I infer from this : ham.stackexchange.com/questions/1953/ssb-demodulation – Plutonium smuggler Aug 2 '14 at 14:29
• @Plutoniumsmuggler: No, that's exactly what you cannot do, since the resonant frequencies of the two circuits need to be different (by the IF frequency), not identical. – Dave Tweed Aug 2 '14 at 14:43
• Yes exactly @Dave Tweed. So that means at the front end tank, I need to tune the frequency which is 455 ( or in my case 490 KHZ) above the LO frequency. I have a gang capacitor with both capacitances identical. So how do I go about it ? – Plutonium smuggler Aug 2 '14 at 14:57
• Design the LO for a different frequency range. – Leon Heller Aug 2 '14 at 15:24
• @Plutoniumsmuggler, So I'm not a radio guy. But for tuning the dual identical variable caps, could you put another cap in series with one, (or parallel?, both?) – George Herold Aug 2 '14 at 23:10

First of all, there are two basic heterodyning schemes, one where the local oscillator frequency is above the RF, and one where the LO is below the RF. In either case, the goal is to produce an intermediate frequency - IF - which peaks in amplitude as radio stations are "tuned in" across the frequency band of interest.

Using, as an example, the US AM broadcast band, which covers a frequency range of from 540 to 1610 kHz with 10 kHz spacing between stations, The IF used is nearly universally 455 kHz and is produced by mixing the RF with an LO 455 kHz higher than the RF, which is called "superheterodyning". If the LO were below the RF it would be called "subheterodyning".

In any case, the point is that if there's a station at 590 kHz, the LO must be 455kHz above that, at 1045 kHz, for the mixer to generate the 455 kHz IF and, likewise, if there's a station at 1410 kHz, the LO must be at 1865 kHz to get the 455 kHz IF.

Therefore, as the radio tunes across the RF band, the LO must track that tuning, but must always be 455 kHz higher than the RF.

That's accomplished by using two variable capacitors, mechanically linked, so that as the capacitance of one is varied, the capacitance of the other is too, the purpose being to use one to tune to the RF, and the other to tune the LO. The capacitances of the capacitors are usually different in order to allow them to track, and there are static adjustments built in in order to make the tracking more accurate.

If you can read in russian - you can find good guide for RF and LO tank padding in this book (pp. 26-43, A. G. Sobolevsky, "I'm want to build a superhet."). Not too sofisticated and not too complicated. Described some theory and prctical schematics with "start values" for common wave ranges.