Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics 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.

Simple AM modulator for my power oscillator

Status
Not open for further replies.

neazoi

Advanced Member level 6
Joined
Jan 5, 2008
Messages
4,119
Helped
13
Reputation
26
Reaction score
15
Trophy points
1,318
Location
Greece
Activity points
36,918
Here is my power oscillator along with a simple AM modulator I have found in an old book (uses Be BJTs for tube cathode modulation).
But I wonder, will a power mosfet suit better as a modulator as far as concern fewer components count? Will I be able to modulate it on a single stage (line source) with something like this https://diyaudioprojects.com/Solid/ZCA/ZCA.htm and instead of the 40W resistor to be the BJT emitter?

Note that the multi-bjt modulator has the same number of components than that of the mosfet and the bjt will allow for mike input (desired) not just line.
 

Attachments

  • tmp1.PNG
    tmp1.PNG
    66.3 KB · Views: 233

Hi,

This is for your website, right? Who is Mike? (Mike for Michael and mic for microphone)
 

Both designs are horrible!

The quad darlington arrangement will be VERY unstable and require constant adjustment of the bias pot, the transistors are germanium and went out of production 50 years ago!

The MOSFET will work after a fashion if you substitute a power resistor for the key and feed the audio across it but do bear in mind it draws around 1A quiescent current and has very low audio gain.

If you want a simple modulated supply, use an LM317 in the supply line, use a 240 Ohm resistor between its output and 'adj' pin and from there to ground use a 910 Ohm resistor. Apply the audio across the 910 Ohm resistor so it modulates the output voltage. The impedance at the input point will be about 150 Ohms so driving it will be easy.

Do note though that modulating a crystal oscillator like that is prone to all kinds of problems, keep the modulation level low or it will stop running on negative cycles and clip hard on positive ones.

Brian.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
Both designs are horrible!
If you want a simple modulated supply, use an LM317 in the supply line, use a 240 Ohm resistor between its output and 'adj' pin and from there to ground use a 910 Ohm resistor. Apply the audio across the 910 Ohm resistor so it modulates the output voltage. The impedance at the input point will be about 150 Ohms so driving it will be easy.

Do note though that modulating a crystal oscillator like that is prone to all kinds of problems, keep the modulation level low or it will stop running on negative cycles and clip hard on positive ones.

Brian.

I built the modulator with a bc549c and a 2sc2078 (the same transistor used for the oscillator currently. It worked but the modulation was low. Increasing the AF volume only increased distortion.
All right I will try the LM317 PSU. If I power this from 20v (~10W RF output) the components values you described are ok?
 

The 240 Ohm should stay the same, it satisfies the load current requirements for the LM317. The 910 Ohm is to set the output voltage to half the input so that value is to give 6V out based on the 12V in your schematic. Change the value so you get about half of whatever supply you choose. The idea of half voltage is so you get most chance of the voltage swinging above and below that point, in other words the carrier with no modulation is at 50% of the peak which is where it should be for AM. You should place a decoupling capacitor at the input of the LM317 as normal but at the output use something smaller than usual, maybe 100nF, so it decouples RF but not the audio frequencies.

Brian.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
The 240 Ohm should stay the same, it satisfies the load current requirements for the LM317. The 910 Ohm is to set the output voltage to half the input so that value is to give 6V out based on the 12V in your schematic. Change the value so you get about half of whatever supply you choose. The idea of half voltage is so you get most chance of the voltage swinging above and below that point, in other words the carrier with no modulation is at 50% of the peak which is where it should be for AM. You should place a decoupling capacitor at the input of the LM317 as normal but at the output use something smaller than usual, maybe 100nF, so it decouples RF but not the audio frequencies.

Brian.

I used a 470R instead of 1k. I got a voltage of around 10v which is about the half of the input. However the modulation is very low lowest than with the emitter modulator. I am feeding the AF through a 470nf to the adjust pin. Any ideas?

UPDATE: it is just more insensitive, I had to feed a much greater signal in than the output of the PC sound card. Shouldn't I see the power meter move a bit up and down on the music? Now I see only 2w (cause I set the voltage to half) whereas with the emitter modulator I saw about 5W

- - - Updated - - -

Both designs are horrible!
Do note though that modulating a crystal oscillator like that is prone to all kinds of problems, keep the modulation level low or it will stop running on negative cycles and clip hard on positive ones.

Brian.

I also used a tip122 (darlington) at the emitter. It worked in keeping the output power to high levels but on some bias settings I achieved very high modulation levels but highly distorted. I guess these are the kinds of problems you mention?

- - - Updated - - -

Do note though that modulating a crystal oscillator like that is prone to all kinds of problems, keep the modulation level low or it will stop running on negative cycles and clip hard on positive ones.

Brian.

All right problem solved Brian. The best was to use an LM317, but instead of feeding the AF directly to the adjust through a capacitor, I fed this to the base of a BJT, which is biased by a resistor. the emitter is connected to the gnd and the collector to the adjust pin. This creates a variable resistor, which provide high amounts of modulation without distortion!
I am amazed, but my idea worked great! The lamp (RF indicator) and the meter needle go up and down as they should. No audible distortion and the modulation is by far the loudest I have heard from all my other little AM modulators, believe it or not.
 
Last edited:

I fooled with sinewave oscillators (including phase-shift type and twin-tee). I applied a signal at various nodes in search of an easy way to change carrier amplitude, depending on signal voltage.

This twin-tee looks as though it might do the job:

amplitude modulation applying 10kHz sinewave withing twin-tee oscillator.png

The above appears to be a better way to modulate compared to my previous idea of using a PNP transistor to attenuate (or more likely clip) the carrier waveform.
 

Twin-tee is a good and stable configuration but doesn't lend itself to high frequency operation very well. The values become impractical beyond maybe a MHz or so because they are so small that keeping the DC conditions in range is no longer possible and stray capacitance overtakes the wanted one.

Amplitude modulating any oscillator has to be done with extreme care. Neazoi is using a quartz crystal to stabilize the frequency but 'overcooking' it massively, I would expect a twin-tee arrangement to shift frequency far too much as the bias point changes. It is still far better to produce a clean carrier and modulate it afterwards.

Brian.
 
Neazoi is using a quartz crystal to stabilize the frequency but 'overcooking' it massively.
Brian.
Hihi! I like the overcooking part :)
Seriously now, using old large crystals OR a few modern crystals in parallel helps keeping the crystals cool even at high powers. One reason is that I am not feeding the whole signal back to the crystal, like done on pierce oscillators, but through a capacitive divider, so it helps a bit. See my construction here **broken link removed**

- - - Updated - - -

Here is the modulator. Not currently optimized, this circuit achieved the highest modulation and trouble-free operation of the TX on AM.

Note that the VCC could be anything from a few volts to 20v or so. But I have only tested it at 20v.

Shall I better connect the base resistor to the adjust pin instead of the VCC for that reason (or any other benefit)?
Shall I try to remove the 10k at the C-E at all and leave only the transistor there?
If I wanted the circuit to operate on different voltages, which resistor should I make variable (to adjust modulation at different voltages). Right now I adjust the 470k, but adjusting the 10k could be better?

- - - Updated - - -

Hihi! I like the overcooking part :)
Seriously now, using old large crystals OR a few modern crystals in parallel helps keeping the crystals cool even at high powers. One reason is that I am not feeding the whole signal back to the crystal, like done on pierce oscillators, but through a capacitive divider, so it helps a bit. See my construction here **broken link removed**

- - - Updated - - -

Here is the modulator. Not currently optimized, this circuit achieved the highest modulation and trouble-free operation of the TX on AM.

Note that the VCC could be anything from a few volts to 20v or so. But I have only tested it at 20v.

Shall I better connect the base resistor to the adjust pin instead of the VCC for that reason (or any other benefit)?
Shall I try to remove the 10k at the C-E at all and leave only the transistor there?
If I wanted the circuit to operate on different voltages, which resistor should I make variable (to adjust modulation at different voltages). Right now I adjust the 470k, but adjusting the 10k could be better?




UPDATE2 I have just realized that I had make a mistake. I originally connected the base to the output of the regulator (modulated) instead of the input.
I have found it works both ways, but the modulation with the "mistake" circuit was cleaner and the carrier was kept at higher levels! Why is that?
 

Attachments

  • tmp1.jpg
    tmp1.jpg
    7.4 KB · Views: 226
  • tmp1.jpg
    tmp1.jpg
    25 KB · Views: 271
Last edited:

The 10K isn't necessary but does no harm, all it will do is keep the maximum output voltage a little lower than it could be. You should be able to increase the 100 Ohm to 220 or 240 Ohms, it is there so the minimum output current requirement is satisfied and to form the top of the 'adj' divider chain. Changing to 240 Ohms will have no effect on the modulation but it will drop the overall current consumption by a few mA.
UPDATE2 I have just realized that I had make a mistake. I originally connected the base to the output of the regulator (modulated) instead of the input.
I have found it works both ways, but the modulation with the "mistake" circuit was cleaner and the carrier was kept at higher levels! Why is that?
It is because the circuit is being used as a current booster, the current being the supply to your oscillator and dependent on the voltage it produces. The LM317 always tries to keep 1.25V between it's output pin and 'adj' pin but where you benefit here is the current at the 'adj' pin is very low so it is easier to apply audio to it. You are using the error amplifier inside the voltage regulator as an additional amplifier stage. The output pin voltage varies according to the voltage on the 'adj' pin so if you use it to bias the transistor you get negative feedback which improves the quality. More output voltage = more bias current = lower output voltage.

Ideally, you should use a potentiometer across VCCin and ground with the wiper connected to the transistor base through a suitable current limiting resistor (~10K) so you can set the output voltage to about half VCC with no modulation input. You can try it across the output pin instead of input pin an compare the results.

Brian.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
It is because the circuit is being used as a current booster, the current being the supply to your oscillator and dependent on the voltage it produces. The LM317 always tries to keep 1.25V between it's output pin and 'adj' pin but where you benefit here is the current at the 'adj' pin is very low so it is easier to apply audio to it. You are using the error amplifier inside the voltage regulator as an additional amplifier stage. The output pin voltage varies according to the voltage on the 'adj' pin so if you use it to bias the transistor you get negative feedback which improves the quality. More output voltage = more bias current = lower output voltage.
Brian.

Thanks so much about the explanation!

I experimented more. I removed the transistor circuit and I connected an LDR to the ADJ to the gnd (vt43n1).
By covering the LDR with my fand I could achieve RF output variations of a few 100s of mW and up to the full output power! This is a huge "modulation" range.
I wonder if I can exploit this.
I tried a led at the collector of a transistor and tried to drive the led brightness by the music in the base.. It worked but the led did not switch completely on or off, so the modulation was low.
It would be interesting if you have any ideas about this to try, just as an experement, and you never know
 

I tried a led at the collector of a transistor and tried to drive the led brightness by the music in the base.. It worked but the led did not switch completely on or off, so the modulation was low.
If you mean in the LM317 circuit, that would limit the lower output voltage to VCEsat + Vf so you would lose around 2V of modulation audio. Remember that the idea of modulating the 'ADJ' pin is it is a high impedance point, the current flowing there is quite low, a few uA from the LM317 and some current through the 240 Ohm resistor.

There is an even simpler implementation of this concept but it only works up to about 2MHz and produces REALLY bad spectral output. You can generate about 5W of RF using only 4 components. I leave it up to you to work out the details.....

Brian.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
There is an even simpler implementation of this concept but it only works up to about 2MHz and produces REALLY bad spectral output. You can generate about 5W of RF using only 4 components. I leave it up to you to work out the details.....

Brian.

What's that???
 

Thought that might get your attention. :grin:

It does work but maybe not with every sample of parts so you may need to experiment. Remember voltage regulators need decoupling capacitors at their input and output to keep them stable - well try placing a parallel tuned circuit in the ground pin of a voltage regulator and see what comes out of it. You could try it with the LM317, I have no idea if it will work on that device but if it does you might get even more output. It certainly works on fixed 5V regulators. Try connecting a resistive 'dummy' load of say 50 Ohms and see what voltage appears across it. Obviously don't connect a capacitor from the output to ground but you probably still need one at the input.

Brian.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top