1

In a superheterodyne receiver, let us assume, I have a signal in the RF range of 88MHz to 108MHz.

From what I have read, my understanding is that by tuning the local oscillator frequency, we can make the incoming RF signal to fall within the IF frequency range which we have designed.

My questions :

  1. Suppose I have a frequency of 95MHz, and my IF filter is designed at 98MHz (Am I correct in assuming this value for the IF filter?) The signal, after going through the filter, will come as a signal having 98MHz, right? But my actual signal is 95MHz, and all my modulation is done at 95MHz. Not after the filter, the signal is converted to 98MHz. Won't my actual signal present in the 95MHz, be lost or tampered with filter?

In other words, how is the modulation retained even after passing through the IF filter?

  1. Does high carrier frequency imply that more information can be carried and low carrier frequency imply that less information can be carried per unit time? Can someone explain.
JRE
  • 67,678
  • 8
  • 104
  • 179
  • Your example doesn't make much sense, both numerically and in that it sounds more like the description of a transmitter ("all of my modulation is done") than a receiver. Typical IF for an FM broadcast receiver is 10.7 MHz. The entire occupied bandwith of the chosen signal is translated there. – Chris Stratton May 18 '20 at 03:25
  • Ok. The thing I am trying to understand is that if I convert my 95MHz to 10.7MHz, won't my original signal get lost. How will the modulation be retained after the IF Filter as the IF filter selects only a particular frequency –  May 18 '20 at 04:15
  • There's no such thing as a filter that selects a single frequency, even if you wanted one. Filters have *width*. You use a filter that has a reasonably flat width of passband sufficient to pass all of the needed signal, and a hopefully steep enough skirt to sufficiently reduce adjacent channels to each side. The filter for FM broadcast would be wider than that for an AM broadcast, which is wider than that for single sideband, which is wider than that for morse code, because these different types of modulations result in different signal bandwidths. – Chris Stratton May 18 '20 at 04:17
  • Ok. Could you please explain on how this works. If my input signal is 95MHz, and the IF Filter range is 10.7MHz, what is the output of the IF Filter and how does my original modulation of the 95MHz, carrier remain unaffected even after coming out of IF Filter? –  May 18 '20 at 04:20
  • 1
    The IF filter's center frequency is 10.7 MHz, not its bandwidth. You mix the RF signal with a higher or lower local oscillator so that RF-LO = 10.7 or LO-RF = 10.7 ie, you mix the signal down to 10.7 MHz. The entire modulated signal goes through that translation, as does a lot adjacent that you probably don't want - though real receivers tend to have a tunable pre-selector filter that tracks the tuning adjustment and helps a little in diminishing further out interference before the mixer. The output frequency of the IF is the same as the input, because filters are (supposed to be) linear. – Chris Stratton May 18 '20 at 04:24
  • OK thank you. Two questions. 1. If 10.7MHz is the center frequency of the filter, what's the bandwidth of the filter? 2. So, the modulation is retained even after the IF Filter, as filters are supposed to be linear devices, which make them maintain the modulation even after IF Filter? –  May 18 '20 at 04:28
  • No, linearity means that the output frequency is the same as the input, only weaker for components of the signal further from the filter center. As for the bandwidth, how about you look up the specs of FM broadcast? – Chris Stratton May 18 '20 at 04:37
  • Ok, I think you can understand my misunderstandings in this concept. Would you be kind to provide an answer after you know where I am lacking in the understanding of this? –  May 18 '20 at 04:40
  • And can we change the center frequency of the IF Filter for FM bands? Because, the demodulation circuit can be design to demodulate whatever signal we get from IF filter? –  May 18 '20 at 05:12
  • Have you read: https://en.wikipedia.org/wiki/Superheterodyne_receiver ? Under "Principle of operation" there's a picture showing how the filters behave. *can we change the center frequency of the IF Filter* The IF frequency is a **design choice**, you cannot change it. If you want a different IF frequency: build a new radio. – Bimpelrekkie May 18 '20 at 07:17

1 Answers1

3

Take a Frequency Modulated (FM) signal going into a superheterodyne receiver which has a tuning range over 88 to 108 MHz.

Any given FM signal will occupy a much narrower bandwidth than that. Typically an audio signal will deviate the carrier, that is modulate the frequency, by around +/- 100 kHz. If we take the carrier to be 95 MHz, then the complete signal is represented by the frequency changing from 94.9 MHz to 95.1 MHz.

Let's say we have an Intermediate Frequency (IF) of 10.7 MHz, this is the frequency almost invariably used in commercial FM receivers. The IF could be another frequency in that ballpark, but 10.7 MHz has been settled on as a de facto standard, and there are a lot of cheap components made and used for that frequency.

To 'tune to' our 95 MHz signal, the first Local oscillator (LO) is set to 105.7 MHz. We mix the incoming signal and the first LO, and our IF filter picks out the difference frequency, and signals near to the difference. Typically the IF filter has a bandwidth of a few hundred kHz, to cope with the full width of the frequency deviation, and a bit more. 95 MHz translates to 10.7 MHz. 95.1 MHz translates to 10.6 MHz, and 94.9 MHz becomes 10.8 MHz. All of those signals pass through the IF filter.

So at the IF, just as at the carrier, we have a frequency modulation. It's the same rate and same deviation, and represents exactly the same information.

By setting the first LO to other frequencies, we can get other carrier frequencies to pass through the IF filter. Perhaps the most important part of the filter is that nearby RF frequencies do not pass through the IF filter.

Once down to a fixed frequency range, we can build a high quality FM demodulator, that maps the changes in frequency around 10.7 MHz into changes in voltage, to recover the modulation. Most demodulators are designed to only accept the specific frequencies passing through the IF filter, so all tuning is done by changing the LO frequency.

It's the signal bandwidth that determines how much information can be carried, not the signal centre frequency. The 95 MHz version and the 10.7 MHz version of a standard audio FM signal carry the same amount of information. However, with a higher centre frequency, it's possible to make a signal with a wider bandwidth. So for instance if you wanted to make a 25 MHz bandwidth signal, that's not physically possible at a 10 MHz centre frequency, though is perfectly possible at 95 MHz centre.

Neil_UK
  • 158,152
  • 3
  • 173
  • 387
  • Thank you for the detailed answer. Just one question. Why do we need to have an IF Filter of say 10MHz? Why can't it be any other value? Because, from what you say, I understand like, even after the IF Filter, the modulated information is present. So, why specifically 10MHz ? –  May 18 '20 at 08:19
  • 1
    Until you've actually designed some superhet receivers, there are some rules of thumb you'll need to take on trust. Roughly 10:1 is the magic ratio, of one IF frequency to the next. With less, it's difficult to design the IF filter. With more, it's difficult to design the RF filter that goes before the mixer to eliminate the image. So with RF frequencies of about 100 MHz, an IF of 10.7 MHz is 'about right'. Once several manufacturers had chosen that frequency, it became a *de facto* standard. Anything from 5 MHz to 15 MHz would have done almost as well. – Neil_UK May 18 '20 at 08:41
  • Thank you for the answer! –  May 18 '20 at 09:00
  • Could you provide an answer for my Second Question too? –  May 18 '20 at 09:00
  • @Neil_UK I'd be interested to know more about this 10:1 ratio vs IF filtering vs image rejection. Any recommended learning material, or does this just come from experience? – Lundin May 18 '20 at 09:20
  • 1
    @Lundin The wikipedia Superhet article rather glosses over the issue of image frequency, which is a pity, I might expand it. Consider a 95 MHz signal, and 10 MHz IF. The 105 MHz LO will also mix down 115 MHz. It's possible to make a tunable RF filter that passes 95 MHz +/- 500 kHz and rejects 115 MHz, and stays stable, cheaply. Now imaging a 3 MHz IF. The LO is 98 MHz and the image is at 101 MHz. An RF filter to pass 95 MHz +/- 500k and stop 101 MHz would be impractical at a commercial cost, too steep. Once you've wrestled with a few filter designs, it's obvious, but it's insider knowledge. – Neil_UK May 18 '20 at 09:31
  • @Neil_UK Ah I see, so it's mostly about getting the image sufficiently far away from center, so that the IF filter can block out the image frequency? I'll read up on wikipedia to start with, thanks. – Lundin May 18 '20 at 09:45
  • @Lundin I had the privilege to work with a 'one of a kind' designer who took that 10:1 as a challenge, and built a tunable image reject filter at 1 GHz for a 10 MHz IF. It was very unusual, and used a few mm of the outer of some UT85 coax cable as a very high Q low value inductor to implement simultaneously the passband and the stopband - the main problem with implementing them separately is their relative stability, a very small difference between two very large numbers. However, it was research only, impossible to build commercially. – Neil_UK May 18 '20 at 09:50
  • @Neil_UK Ok I'm mostly a rookie at this, I'm a software guy who often ends up writing filter tuning & image rejection algorithms, without having a clue how the chip designer ended up with a certain IF and image. Had a project in the past where we simply couldn't get image rejection working well - the application was very narrowband with the image just 200kHz off carrier. And IF was nowhere near this 10:1 ratio either, so I got curious about your answer. – Lundin May 18 '20 at 10:30