Control RF preselector frequency by injecting same freq?

[Circlotron]

Is there some kind of circuit arrangement that will allow you to tune to a specific frequency the same as you would with a parallel LC but by injecting the same frequency into the circuit? That is to say, to tune to and receive 1MHz I poke a 1 meg wave into it. What I want to play with is using my function generator to select a particular frequency in the AM broadcast band and also generate a local oscillator signal exactly 455kHz above. The idea being that the two will track perfectly across the tuning range. The idea is not to listen to anything in particular but just to see how well the idea works.

[voltz]

Not completely sure i understand what you want but i think you are referring to a frequency controlled filter, or tracking filter arrangement.
In other words, a filter whose center frequency is adjusted by another frequency.

I cant think of any easy designs off hand that will do this, but various patents exist which describe methods for doing it.
Sounds complicated for a simple receiver design to me.

Maybe by using the tracking input of a function generator (DC controlled frequency input) , you could then use the DC control line to also adjust an LC filter with varicaps as C. That would then effectively produce a tracking filter controlled by the actual frequency in use. Although instead of turning a frequency dial to adjust frequency, you would change a voltage.

Or could you just implement a Direct Conversion Receiver design? Which uses the injected frequency as the actual  receiver frequency.
So you could demodulate signals just by injecting your function generator. No 455Khz IF needed.

 

[rs20]

+1, might as well just do direct-to-baseband conversion.

[T3sl4co1l]

Isn't that just describing a hetrodyne receiver?

Tim

[Circlotron]

Yep.
The point is, it is easy to generated two signals 455kHz apart and make them track perfectly. The question is, can the lower of these be used to control what incoming frequency the receiver is tuned to? The best I can think of is use a switched capacitor bandpass filter that is driven by a variable frequency clock but I expect that would be too noisy for the low signal levels involved.

[T3sl4co1l]

I don't get what's significant about the lower frequency.  Or the upper for that matter, or both being present.  Why would you want to?

If you just mix it with RF, you get all the sums and differences, a mess.  If you use quadrature mixers to produce the frequency pair and another to perform the RF mixing, you might be able to null one or the other, or their mixing products, but that's a lot of bother, and you're better off starting with just one frequency in the first place.

A switched capacitor filter exhibits sampling, so any frequencies higher than Fs are aliased down.  This could be used for conversion, in the same way that RF sampling ADCs can be used for conversion (in part).  Of course, image filtering is still required.

Would this be related to a Wadley loop, by any chance?

Tim

[David Hess]

That sounds like a Wadley Loop superheterodyne receiver.

[Circlotron]

Just a different way of doing a superhet function, but instead of using a 2-gang variable capacitor with one side controlling a local oscillator that is 455kHz above the desired received frequency and the other side determining what received frequency enters the mixer, just generate two adjustable frequencies a fixed 455kHz apart and use the higher for the local oscillator and the lower one to *somehow* control what received frequency enters the mixer.

The *somehow* is what needs to be solved.

[T3sl4co1l]

Ah, well that'd be a double superhet, then!

Usually done with selectable or independent LOs, for band and tuning.

If you take a step back and use a VCO, then you can use the CV to set a varactor as well.  Same idea, and same problems, but it's purely electric.  Or if LO is synthesized, then a DAC can tune the varactor.

The reason an AM radio needs gang tuning is, the RF stage needs low enough bandwidth to prevent images.  Images can show up because f_IF < RF_max - RF_min, and all the various mixing orders and LO harmonics.  It's also a necessity of the electrically small antenna, which necessarily also has low bandwidth.  You cannot receive the entire AM BCB with a small resonant loopstick.  (You can with a nonresonant loop, but then you lose a lot of gain, and noise factor becomes critical.)

AM doesn't need 455kHz IF, that just happens to be what they settled on.  If wideband RF were practical, an IF above the band (at maybe 2 or 3MHz) might've been chosen.

When you can receive the entire band, like the FM BCB, RF can be fixed tuned, and only one mixer is needed.  Note that the conventional IF, 10.7MHz, is more than half the band width (20MHz), which is just enough to keep images out.

When you may have a wide RF bandwidth available (like an array of HF antennas for a SW receiver), tuned RF is impractical, and a double hetrodyne system (or more), like the Wadley loop, or like many radios through history, may be desirable.

Spectrum analyzers also usually have a lot of IF stages; probably because of image rejection.  The input BW might be 2GHz, so you need IF of 3GHz or more, and then that has to be converted down to a few MHz for filtering.  The conversion may be done in stages because the 3GHz 1st IF filter has a very wide bandwidth (10s or 100s MHz), which therefore can't be converted straight down to final IF or baseband.

Tim

[radiogeek381]

Is there some kind of circuit arrangement that will allow you to tune to a specific frequency the same as you would with a parallel LC but by injecting the same frequency into the circuit?

That describes a direct conversion receiver.  (Some might also say it describes a Regenerative receiver as well.)

That is to say, to tune to and receive 1MHz I poke a 1 meg wave into it. What I want to play with is using my function generator to select a particular frequency in the AM broadcast band and also generate a local oscillator signal exactly 455kHz above. The idea being that the two will track perfectly across the tuning range. The idea is not to listen to anything in particular but just to see how well the idea works.

Well, this is a horse of a different feather.  This is a classic superhet organization where the product after the first mixer is at 455kHz and then passes through a passive filter network.  It is a great organizational scheme and was the basis for most of the AM broadcast receivers from about 1940 until, probably, the 1980s or later.

But then we got into all kinds of other things.  This may not answer your question, but I have a feeling it might give you some more tools to navigate through the fascinating engineering tradeoffs and technical fashion show that is radio design.

Bottom line takeaway here is:

1. There are many different receiver architectures, and it is worth giving each of them a look.  There's nuggets of wisdom in technical archaeology.
The ones that come to mind:  Coherer, Tuned-Radio-Frequency, Regenerative, Heterodyne or super-heterodyne, and direct conversion.  In particular, there is much to be learned from the various exotic super-het schemes that used multiple levels of conversion to take advantage of one kind of filter widget or another.

2. The goal of the front-end in a receiver is to deliver energy near the signal of interest from the antenna to the some element that will eventually get us to the demodulator/detector.  As you read about RX designs, pay attention to the flow of the desired signal and the interfering or undesired signal through the path. Much of what motivates good design is as much about the stuff you don't hear in the speaker.

Ideally, the front end would deliver only the signal of interest, but narrow filters bring with them a host of problems.  (That's why you don't see many TRF receiver implementations (even historically) above one MHz or so.)  This part of the story is about adapting engineering styles and decisions to contemporary technology, supply, and economics.

3. Almost any radio component with gain will have some non-linearity.  Mixers also have non-linear behavior -- that's what makes them mixers.  As Mr. C01l has pointed out, the mixer will produce lots of products that beat your incoming signals (and its neighbors) against each other and against the carrier oscillator (we're assuming a heterodyne architecture at this point).  Well designed receivers take a great deal of care in balancing the gain before and after the mixer, the selectivity before the mixer, and the local oscillator injection at the mixer.  (Superhets are organized to put most of the spurious products well outside the passband of the next stage of amplification.) The good news is that many of the spurious products are related to the cube of some fraction which makes them shrink in amplitude relative to the signal of interest.

This last part is, perhaps, the hardest thing for people to internalize.  Nonlinearity is our friend.  It does very interesting things.  Try some pencil and paper experiments.  Get yourself a list of trigonometric identities and start mixing some A*sin(f1*t) and B*sin(f2*t) and C*sin(f3*t) by squaring the sum of two or three of them.  Then cube the sum of two or three.  Take a look at what the trigonometry tells you. What happens when A is much bigger than B or C, but when B is bigger than C? How about A much bigger than B and C, but B and C are equal? The drill will take an hour or so, and grind through a pencil, but it will help embed things like no simulation or youtube video could. Do this experiment, and you'll understand the discussion of IP3 that you'll eventually get to on the path to enlightenment.

4. Since electronic components like diodes, varactors,  FETs, and BJTs have non-linear responses to their inputs, it is often a good idea to put some filtering between the antenna and the first one of these that the signal sees. That means that the preselector, for most sizes of budget, should be passive, otherwise it contains a diode, varacter, FET, or BJT and so you need to put a preselector in front of it.  which should be passive. <rinse-lather-repeat>

Preselectors are a great idea.  Heterodyne mixers are also a great idea.  But neither is a good substitute for the other.

Because of all these tradeoffs, you'll see receiver organizations that appear to be quite odd.  Some have ton's of loss in the front end, having spent lots of incoming energy in filters before the first active element.  This is common in HF where receiver overload is a real problem. 

But then there are receivers that stick a transistor to the very first non-metalic element in the signal chain -- right out at the feed of the antenna.  These often look at hot dots in a cold sky.  When there are no other signals to beat with yours, preselectors aren't all that important.  (But watch out for spurious signals from the microwave oven...)

This isn't a direct answer to your question, but I hope it can guide you in your search for other questions to come.

[voltz]

Just a different way of doing a superhet function, but instead of using a 2-gang variable capacitor with one side controlling a local oscillator that is 455kHz above the desired received frequency and the other side determining what received frequency enters the mixer, just generate two adjustable frequencies a fixed 455kHz apart and use the higher for the local oscillator and the lower one to *somehow* control what received frequency enters the mixer.

The *somehow* is what needs to be solved.

Seems to be some confusion here. Why generate two frequencies 455Khz apart to create a single superhet receiver?. Makes no sense to me. What you really should be aiming for is a single local oscillator (VFO or VCO) with a preselector stage (that is a variable front end filter) which tracks the VFO.

You dont need the lower 455Khz frequency to do this - the mixer output will automatically produce your 455Khz IF. ie. 455Khz = Vlocal + RXwanted, of course you can also use the Vlocal - RXwated result too from the high sided VFO. Then your local VFO will read an offset of 455Khz. You only need one oscillator to do this.

The trick is to make a DC controlled preselector filter. As i described in my first post - this can be done with varicaps (varactors) in a tuned LC arrangement forming a tracking filter. If you use a VCO with a DC control line rather than a mechanical VFO capacitor, this same DC control can then tune the preselector. Thats standard practice in all modern transceiver equipment, you find this in HF, VHF and UHF designs.

Hope that helps.