Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

How to understand RF circuits.

Fluffyboii

Active Member
Hi, I kind of don't understand circuits that operate at frequencies like FM range or more. I look at radio circuits, transmitters and see LC tank circuits instead of resistors to make an class A amplifier for example. When I try to replicate on LTSpice I get inconsistent results with these types of circuits. It is probably because I never really understood how to calculate AC currents and voltages and memorized my way through it. Also stuff like regenerative receivers with feedback and such seem to be impossible to understand with rigorous way.

For example Class C amplifiers doesn't make sense to me since it will short to ground at DC (I get that this one at least has an resistor to prevent it). And there is no bias for the transistor itself. Maybe it is assumed that it is supposed to be used with regular AM signal so that it will have some DC bias but that would make it very unreliable. Testing it in LTSpice confirms my suspicion since it doesn't seem to work. When I add feedback from collector to base with a resistor to possibly turn it into Class A amplifier it outputs an distorted amplified signal. But it is impossible to set the Q point so that it is not distorted even with adding an emitter resistor it is always very bad at the output. Maybe the filtering at output supposed to fix it.
1672147831586.png

There are very weird circuits like this FM transmitter for example:
1672148328609.png

C9 and C10 should act like an voltage divider. Why is it decreasing the input signal before amplifying it. How bias voltages are determined here. How does the oscillator change its frequency exactly according to input signal.
1672148397233.png

This one looks similar as well. I understand that first stage is just an amplifier and second is an oscillator but how exactly that oscillator changes its frequency according to the voltage at base. I would expect it to amplitude modulate it instead of changing its frequency.

I read that it is possible to demodulate FM by taking input signal's derivative first than using an envelope detector to approximate message signal. I know the basic op amp circuit to take derivation but idk how to do something similar with an BJT. I saw a tuned RC filter can approximate derivation. I couldn't find a circuit that uses this unreliable way of demodulating FM but I am sure it can be done with 2 transistors or something like that since people seem to be able to use a single transistor for like 4 things at the same time. How can that be possible anyway? Like how can it both amplify audio frequencies and demodulate AM in a Reflex receiver. The simple analogy of thinking it as an valve that is controlled with voltage current doesn't apply there.
It is just so bizarre. Just when I think I understand something another black box of an circuit appears.

I learned that RF course in my collage just teaches drawing RF integrated circuits in Cadence. I know I can't get anything more than C of a class of that caliber but should I give it a try. It is probably only using mosfets in ways it is not applicable in discerete designs though.
 
C9/C10 are for impedance matching, rather than using a tapped coil for L2.

Nothing particularly weird about it, it's a pretty standard 'toy' FM transmitter, basically a high powered bug.

The second one is the same, without the two amplifier stages, so just a low power bug.

The modulation is crude and nasty, and produces both AM and FM, of fairly poor quality.
 
C9/C10 are for impedance matching, rather than using a tapped coil for L2.

Nothing particularly weird about it, it's a pretty standard 'toy' FM transmitter, basically a high powered bug.

The second one is the same, without the two amplifier stages, so just a low power bug.

The modulation is crude and nasty, and produces both AM and FM, of fairly poor quality.
I see. Would a better approach be just buying high speed/slew rate op amps for like 5-10USD a piece and making a VCO to feed the input signal directly into it? I am not sure the VCO I made in the past with CD4XXX series inverters used as amplifiers would work much more than 100Khz. Also analog VCOs always drift in frequency.

Since condenser microphones are just capacitors I wonder if one can be used to change the frequency of an oscillator if it is used in parallel with the RC tank circuit.
 
A proper transmitter (as opposed to a toy bug) uses either a crystal (fixed frequency) or a PLL (variable frequency), and NBFM is done using a varactor diode across the crystal, or a phase modulator on the crystal oscillator. The low(ish) frequency is then amplified and multiplied in a number of amplifier stages, this increases power, and also the amount of frequency modulation.

A common crystal would be around 12MHz, then multiplied 12 times to reach the 2m amateur band.

For a PLL, you can apply audio to the varactor in the loop to provide frequency modulation.

You don't really use opamps for RF purposes.
 
A proper transmitter (as opposed to a toy bug) uses either a crystal (fixed frequency) or a PLL (variable frequency), and NBFM is done using a varactor diode across the crystal, or a phase modulator on the crystal oscillator. The low(ish) frequency is then amplified and multiplied in a number of amplifier stages, this increases power, and also the amount of frequency modulation.

A common crystal would be around 12MHz, then multiplied 12 times to reach the 2m amateur band.

For a PLL, you can apply audio to the varactor in the loop to provide frequency modulation.

You don't really use opamps for RF purposes.
1672152076083.png

Perhaps the thing you mention is this. Yeah I have no idea to construct a circuit like this. PLLs like CD4046 generate square waves. I can think of an way to use digital ICs to do these functions but I don't know how it can be done for a sine wave. Multiple analog multipliers like AD633 would be expensive. I assume cheap stuff just uses microprocessors to approximate this stuff. I guess an Arduino can do it with the fraction of the cost if it can operate at high frequencies.

This is a full package for FM transmitter.
 
View attachment 139747
Perhaps the thing you mention is this. Yeah I have no idea to construct a circuit like this. PLLs like CD4046 generate square waves. I can think of an way to use digital ICs to do these functions but I don't know how it can be done for a sine wave. Multiple analog multipliers like AD633 would be expensive. I assume cheap stuff just uses microprocessors to approximate this stuff. I guess an Arduino can do it with the fraction of the cost if it can operate at high frequencies.

The CD4046 generates square waves, but can also be used as part of a sine wave PLL design, however, most PLL chips use sine waves, the 4046 is fairly unusual as it doesn't, but it's not intended for RF use (as most PLL's are).

You seem to have a strange fascination for analogue multipliers?, which are fairly uncommonly used in RF (and pretty uncommon generally really - I've never used one, or indeed ever owned one).

Cheap stuff generally is either crappy (like the bugs above) or uses custom IC's for much of it's design, and (like everything else) it will almost certainly all be controlled by a micro-controller (or more than one).
 
You seem to have a strange fascination for analogue multipliers?, which are fairly uncommonly used in RF (and pretty uncommon generally really - I've never used one, or indeed ever owned one
It just seems like everything comes to multiplication in mathematical textbook stuff. Multiply sine wave with another at another frequench then filter it and you can get a frequency mixer from my understanding. I just don't know another way of doing it.

I want to buy like 10 AD633s and make the worlds most useless, analog calculator. It just seems too fun to solve matematical equations without using any processing power.

It all started because one guy at reddit synth group told me that I could build an ring modulator with LEDs
 
It just seems like everything comes to multiplication in mathematical textbook stuff. Multiply sine wave with another at another frequench then filter it and you can get a frequency mixer from my understanding. I just don't know another way of doing it.

I want to buy like 10 AD633s and make the worlds most useless, analog calculator. It just seems too fun to solve matematical equations without using any processing power.

It all started because one guy at reddit synth group told me that I could build an ring modulator with LEDs

You're not understanding what the textbooks are saying - a frequency mixer can be as simple as one diode, or one transistor (as used in almost all domestic radios), 4 diodes is a common double balanced one.
 
You're not understanding what the textbooks are saying - a frequency mixer can be as simple as one diode, or one transistor (as used in almost all domestic radios), 4 diodes is a common double balanced one.
Yes I can not comprehend how that could work to multiply the frequency of a sine wave. Ok I now remember the circuit you probably mention. But no textbook of mine explains the circuit diagrams behind those operations in detail. Do you know one that specially explains the operation of these types of circuits instead of the theory.
 
Last edited:
Class C amplifiers doesn't make sense to me since it will short to ground at DC
Yes, that is true, BUT not when the amplifier is driven with the correct input signal.

there is no bias for the transistor itself.
Correct, that is what makes it a class C amplifier.

Maybe it is assumed that it is supposed to be used with regular AM signal so that it will have some DC bias
A class C amplifier will not work correctly with an AM signal.
An FM signal - Yes
An unmodulated RF carrier - Yes
An AM signal - No. The signal will be very distorted.

What make you think that an AM signal has "DC bias" ? It does not.


How a class C amplifier works:

Consider this simplified circuit:
1672155755337.png

Transistor Q1 will not conduct while there is no input signal.

When an RF signal at (say) 30MHz is applied to the input of the amplifier, Q1 will start to conduct when the peak value of the RF signal is greater than Vbe of the transistor (about 0.6 V).

As the transistor conducts, the voltage at the collector will pulse down towards 0 V, the signal will look like the negative half of a 30MHz sinewave.
This signal will contain many harmonics of 30MHz.

If we want to use this signal to connect to an antenna and make a transmitter, we must use a lowpass filter to remove the harmonics, or we will not be popular with other users of the radio spectrum.

A practical lowpass filter circuit applied to this amplifier would look like this:
1672156867748.png

This would give a good level of attenuation to harmonics generated in the class C amplifier.

If we were to replace the lowpass filter with tuned circuits resonant at 60 MHz,
then the 30 MHz signal and higher harmonics (90, 120, 150 MHz etc) would be attenuated and we would be left with a 60 MHz signal.

So our original 30 MHz has become 60 MHz - We have created a Frequency Multiplier.

1672152076083.png

Perhaps the thing you mention is this
Wow, that is a real "tour de force".

OK, that is enough for now.

JimB
 
What make you think that an AM signal has "DC bias" ? It does not.
It got an DC offset for it to be detectable with an envelope generator right? I read that AM indeed needs to be amplified with an class A amplifier instead of C since it needs to be linearly amplified. Which make sense, also explains why I am not getting "meaningful" results in LTSpice.

So for multiplication of an sine wave you just band pass filter the harmonic you want then amplify, repeat until you get the frequency you want. So it can be done the way in the diagram without reaching insane part counts that way. It is still difficult for me to understand stuff in frequency domain but that clears the image in my head.
 
So for multiplication of an sine wave you just band pass filter the harmonic you want then amplify, repeat until you get the frequency you want.
A pure sine wave contains no harmonics!

With a Class C amp, the transistor only conducts for a small part of each cycle, producing a distorted and harmonic-rich waveform, and with a pretty high efficiency.

It can either be fed to a tuned circuit set to a harmonic for frequency multiplication, or to the same frequency circuit and low pass filtering to be used at that frequency for a "clean" signal.
 
What make you think that an AM signal has "DC bias" ? It does not.

It got an DC offset for it to be detectable with an envelope generator right?
Like I said already, no, it does not.

So for multiplication of an sine wave you just band pass filter the harmonic you want
Like rjenkins said: A pure sinewave does not have harmonics.
A distorted sinewave, ie the output of class C amplifier has lots of harmonics, select the one that you want (within reason) using tuned circuits.
2nd and 3rd harmonics are OK.
4th and 5th harmonics can be a bit of a stretch.
Higher order harmonics are not often selected this way.

JimB
 
Like I said already, no, it does not.


Like rjenkins said: A pure sinewave does not have harmonics.
A distorted sinewave, ie the output of class C amplifier has lots of harmonics, select the one that you want (within reason) using tuned circuits.
2nd and 3rd harmonics are OK.
4th and 5th harmonics can be a bit of a stretch.
Higher order harmonics are not often selected this way.

JimB
1672169295961.png

Doesn't this count as DC bias added to the signal. I must be mixing it up with something else. I mean it will be gone after diode rectifies it and it passes through the RC filter so it is pointless to try to bias a transistor with it.

Ok nevermind it doesn't count as DC bias because it is on both sides. It is just changing the shape of the signal.
 
Yes I can not comprehend how that could work to multiply the frequency of a sine wave. Ok I now remember the circuit you probably mention. But no textbook of mine explains the circuit diagrams behind those operations in detail. Do you know one that specially explains the operation of these types of circuits instead of the theory.

Books are generally really bad - they explain complicated theory, and only give partial (mostly non-working) circuits.

For a more practical book, if you can find it, I have "Solid-state Circuits Guidebook by Brice Ward", a TAB book - mine is dated 1974 :D

For RF work, you should get either (or all of them), the RAE Manual, or the RSGB Books (Radio Communication 1 & 2, and the VHF/UHF Manual). These are all VERY expensive, but old copies are fine, and could well be better than brand new ones.

As for class-C amplifiers, it's not really an 'amplifier' as such, it's just a switch which 'rings' the tuned circuit in the collector. Think of the tuned circuit as a church bell, and the transistor the little clapper that hits it.
 
Books are generally really bad - they explain complicated theory, and only give partial (mostly non-working) circuits.

For a more practical book, if you can find it, I have "Solid-state Circuits Guidebook by Brice Ward", a TAB book - mine is dated 1974 :D

For RF work, you should get either (or all of them), the RAE Manual, or the RSGB Books (Radio Communication 1 & 2, and the VHF/UHF Manual). These are all VERY expensive, but old copies are fine, and could well be better than brand new ones.

As for class-C amplifiers, it's not really an 'amplifier' as such, it's just a switch which 'rings' the tuned circuit in the collector. Think of the tuned circuit as a church bell, and the transistor the little clapper that hits it.
A quick search on online second hand book marked didn't bring many results as expected but I will ask second hand bookstores I come across.
 
Sorry, I was a bit confused try - "the radio amateur's handbook" for the American one (mines the 51th edition, again from 1974).
There are only 1954 ones for some reason. Also few mistery books with no info that just says radio comminication
 

Attachments

  • Kitap_20200622162544109940411.jpg
    Kitap_20200622162544109940411.jpg
    19.4 KB · Views: 252
  • Kitap_2022082310314015736115.jpg
    Kitap_2022082310314015736115.jpg
    14 KB · Views: 265
One on Amazon in the UK.


Presumably it's still in print?, so you could get a new one?.
Found one from 1971 for fraction of the cost. Is that modern enough. There is one from 74 as well. Seems like it sold well for an english electronics book here. Can get one for like 3USD

Ok those were all sold listings :/ Will find one for sure at some point.
 
Last edited:

Latest threads

New Articles From Microcontroller Tips

Back
Top