Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

FM demodulator in a ratio detector

DavidYebadlo

Junior Member level 1
Joined
Apr 6, 2020
Messages
17
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
171
Hi All,
The aim of the task is to design a FM demodulator in a ratio detector. I've been trying to do this for several hours but I have a feeling that still isn't it. I am asking for help and suggestions on what should be improved in the system. I also send the layout diagram in LTspice and characteristics
ltspice-scheme.png
plot-ratio-detector-fm.png
Blue plot is output voltage and yellow is input voltage.
 

dr pepper

Full Member level 5
Joined
Mar 15, 2010
Messages
318
Helped
34
Reputation
68
Reaction score
37
Trophy points
1,308
Location
lancs
Activity points
3,594
You have the transformer coupling co-efficient set to 1, try adding some losses to L1 L2 and/or L4.
 

DavidYebadlo

Junior Member level 1
Joined
Apr 6, 2020
Messages
17
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
171
I would assume the yellow trace is from V1 but where is the blue trace measured?
Oh sorry for my mistake. V(n008) is voltage input and signal V(n004) is voltage output.

You have the transformer coupling co-efficient set to 1, try adding some losses to L1 L2 and/or L4.
I'll check it.
 

betwixt

Super Moderator
Staff member
Joined
Jul 4, 2009
Messages
14,697
Helped
4,804
Reputation
9,625
Reaction score
4,585
Trophy points
1,393
Location
Aberdyfi, West Wales, UK
Activity points
125,215
Still not sure - where is V(n004) on the schematic.

Normally, modulation recovery would be across a load resistance in the ground connection (below L3 on the schematic).

Brian.
 

diggermail

Newbie level 4
Joined
Apr 24, 2018
Messages
6
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
31
DIODES D1 D2 should be replaced with Schottky 1n5711 the sensitivity will increase and the sound will improve.
 

DavidYebadlo

Junior Member level 1
Joined
Apr 6, 2020
Messages
17
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
171
I corrected the circuit diagram 1609684579930.png
Input signal waveform:
1609684685286.png

Output signal waveform
1609684712990.png

Is this system working properly? What else should I improve?
 

diggermail

Newbie level 4
Joined
Apr 24, 2018
Messages
6
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
31
C2 C3 10n in my opinion a very large capacity, you need 100-330p
R1 you can safely throw out
С7 discard, diode 1N 5817 is not suitable, replace with 1N5711
 
Last edited:

betwixt

Super Moderator
Staff member
Joined
Jul 4, 2009
Messages
14,697
Helped
4,804
Reputation
9,625
Reaction score
4,585
Trophy points
1,393
Location
Aberdyfi, West Wales, UK
Activity points
125,215
A frequency/amplitude sweep rather than time/amplitude sweep would be more useful.

You may note that in the three schematics in post #7 the upper and lower (L1 and L2 in your schematics) tuned circuits are set to slightly lower and slightly high than center frequency. They are effectively two slope filters working in opposite directions and producing opposite polarities. Your schematic doesn't do that and additionally, if L4 and L5 are part of the same transformer, their fields will cancel.

Brian.
 

diggermail

Newbie level 4
Joined
Apr 24, 2018
Messages
6
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
31
L1 L2 is of course a midpoint coil
 

DavidYebadlo

Junior Member level 1
Joined
Apr 6, 2020
Messages
17
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
171
diode 1N 5817 is not suitable, replace with 1N5711
This diode model is not in LTspice. Is there any replacement?



You may note that in the three schematics in post #7 the upper and lower (L1 and L2 in your schematics) tuned circuits are set to slightly lower and slightly high than center frequency. They are effectively two slope filters working in opposite directions and producing opposite polarities. Your schematic doesn't do that and additionally, if L4 and L5 are part of the same transformer, their fields will cancel.
It may be. I noticed that by changing the induction value of the coils, the amplitude of the signal at the output changes.
 

diggermail

Newbie level 4
Joined
Apr 24, 2018
Messages
6
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
31
This diode model is not in LTspice. Is there any replacement?
not

I know that there is no such model, but I am not talking about models but about a real circuit. Will not work with a diode 1N5817
 

betwixt

Super Moderator
Staff member
Joined
Jul 4, 2009
Messages
14,697
Helped
4,804
Reputation
9,625
Reaction score
4,585
Trophy points
1,393
Location
Aberdyfi, West Wales, UK
Activity points
125,215
BAT54 should be close enough. The 1N5817 is a power rectifier!
It may be. I noticed that by changing the induction value of the coils, the amplitude of the signal at the output changes.
Its a frequency discriminator, to see how it responds you have to sweep the input frequency across its passband and look at the resulting amplitude change. The point I made about the inductors is that unless the simulation understands the construction of the transformer, it will not give the correct result. If treated as simple inductors and all being the same value, it becomes nothing more than an RF rectifier circuit.

Brian.
 

DavidYebadlo

Junior Member level 1
Joined
Apr 6, 2020
Messages
17
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
171
AC analysis of ratio detector in output:
1609747473323.png
 

DavidYebadlo

Junior Member level 1
Joined
Apr 6, 2020
Messages
17
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
171
I changed the SFFM source to a sine source and added a modulator.
1609862654378.png
This is how the waveforms look like:
1609862731972.png

Probably something with the transformer (coils) is wrong.
 

betwixt

Super Moderator
Staff member
Joined
Jul 4, 2009
Messages
14,697
Helped
4,804
Reputation
9,625
Reaction score
4,585
Trophy points
1,393
Location
Aberdyfi, West Wales, UK
Activity points
125,215
Now C5 makes no sense, C5/L4 resonate at about 795KHz which I guess is not what you want.
You should get an output voltage that swings between negative and positive as you sweep the input signal through the resonant frequency.

What frequency is it supposed to work at?

Brian.
 

DavidYebadlo

Junior Member level 1
Joined
Apr 6, 2020
Messages
17
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
171
What frequency is it supposed to work at?
Initially, I assumed that the demodulator would demodulate a signal with the following parameters: freq signal 10MHz and carrier frequency 90MHz. But it was only preliminary data, then I modified the layout by changing parameter values etc.

At the beginning, I implemented a circuit with an SFFM source, now a normal sinusoidal source with an FM modulator.

Recently updated layout scheme:
1610122699427.png

1610122712061.png

I changed the frequency of the signal to the one for which there is resonance in the LC circuit. Did I understand correctly?

As a curiosity, the AC analysis of this circuit:
1610122876488.png
Regardless of the frequency, the voltage at the input and output does not change.
1610122998270.png
 

DavidYebadlo

Junior Member level 1
Joined
Apr 6, 2020
Messages
17
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
171
I corrected my scheme:
1610577794080.png
and characteristics:
1610577866017.png

1610578419291.png

Is the system is ratio detector FM demodulator? It seems to me that I am getting closer to the end of task, but I am asking for good advice and tips.
 

betwixt

Super Moderator
Staff member
Joined
Jul 4, 2009
Messages
14,697
Helped
4,804
Reputation
9,625
Reaction score
4,585
Trophy points
1,393
Location
Aberdyfi, West Wales, UK
Activity points
125,215
You seem to be missing the entire point about what a ratio detector is for.
It converts a frequency to a voltage within a narrow passband. At center of the passband the voltage is normally zero and within the upper and lower end of the passband it swings negative and positive. Imagine the center frequency is 10MHz and the bandwidth is +/- 200KHz. Now imagine sweeping the source frequency from 9MHz to 11MHz. At 9MHz the output will be zero because the signal simply doesn't pass the resonant frequency of the detector, as it tunes beyond 9.8MHz (entering the bandwidth) the output voltage swings sharply negative. As it tunes through 10MHz, the center frequency, the output drops to zero and as it tunes even further towards 10.2MHz is swings increasingly positive. After that it again falls outside the bandwidth and the output drops to zero. So frequency deviation around the center frequency is converted to a voltage proportional to how far away from the center it is and in which direction.

In classic FM systems, the source would be a carrier modulated by a signal that shifted it around the central carrier frequency, the ratio detector would do the reverse, it woould see the shifting frequency and convert it back to the original modulating voltage.

In order for it to work, the coupling between the coils has to be correct, you can't treat them as individual inductors. Modulation at 10MHz when the carrier is 90MHz isn't really practical for this kind of demodulator, it works best with narrow deviation. Consider that a 10MHz modulating signal will have sidebands extending close to the carrier frequency itself. You will find it used frequently for recovering audio frequencies (< 15KHz) where the center is 10.7MHz in domestic receivers.

http://www.tscm.com/NEETS-v12-Modulation.pdf page 3-15 will give you more information.

Brian.
 

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top