2M FM oscillator w/ 150 Hz tone

[kc3ase]

In the spice model circuit and measurement shown attached I'm not sure of a few things.

1) What would be the best way to represent an electret mic in LT spice if possible?
2) Although the transistor drive current oscillates at 150 Hz I don't seem to have the expected near 146 Mhz carrier. I'm assuming it's because of my I1 AC current sign wave which does not properly respond like an electret mic's variable capacitance.

I'm also looking at the possibility of using a 555 to create the 150 Hz tone and then mixing the audio in from the mic later through a 741.

First I'm trying to understand how the basic RF circuit should work.

[moffy]

I cannot see how you expect it to oscillate at 146MHz, there is no feedback around the transistor, and even if there was C3 would likely short it out. There are a multitude of circuits on the web for what you want, why not use one of them?

[kc3ase]

I cannot see how you expect it to oscillate at 146MHz, there is no feedback around the transistor, and even if there was C3 would likely short it out. There are a multitude of circuits on the web for what you want, why not use one of them?

I actually picked this circuit off the web. Originally intended for FM radio band and changed the tank values.

[moffy]

http://www.talkingelectronics.com/projects/Spy%20Circuits/SpyCircuits-1.html "One transistor..."

You left off the CE capacitor of 4.7pf, or what would be appropriate for the changed values. Also Re is much smaller.

[TheMG]

I'm assuming the 150Hz you're trying to add is supposed to be a PL (CTCSS) tone to access a repeater?

While the circuit you suggest is simple and easy to build, it comes with significant disadvantages that make it impractical for anything more than simple experimentation and toying around with.

The most notable disadvantage is stability. The frequency will drift by a significant amount with changes in temperature and supply voltage. Acceptable for an experimental spy "bug" to play around with, but you'll quickly find this to be extremely frustrating if your objective is to get into a repeater and those you are trying to communicate with will be equally annoyed.

2) Although the transistor drive current oscillates at 150 Hz I don't seem to have the expected near 146 Mhz carrier. I'm assuming it's because of my I1 AC current sign wave which does not properly respond like an electret mic's variable capacitance.

With a circuit like this, since the components have wide tolerances and the construction of the circuit introduces significant stray inductances and capacitances, you need to have a variable capacitor or variable inductor to be able to adjust the circuit to the desired oscillating frequency. Usually this is done by spreading or compressing the inductor turns. You can add or remove a turn if you still can't get within the desired frequency range.

As for putting in an AC waveform instead of the electret microphone... that won't work. In the original circuit, an electret microphone is used because it behaves very similarly to a capacitor, the capacitance of which varies according to the applied acoustic waves (sound). The change in capacitance causes a slight change in the oscillation frequency of the circuit, which is what produces the FM modulation.

In order to apply audio modulation in the form of a voltage waveform, you'll need to use a varactor diode. Varactor diodes act like a variable capacitor which changes its capacitance based on the applied reverse bias voltage.

Most traditional FM transmitters use such varactor diodes in the modulation circuit. Usually, the transmitters employ either a crystal-controlled or PLL synthesized oscillator as a stable frequency source. The output of the oscillator is then phase-shifted by one or more varactor diode circuits to produce a phase-modulated (PM) signal, not much different from FM.

[Zero999]

As for putting in an AC waveform instead of the electret microphone... that won't work. In the original circuit, an electret microphone is used because it behaves very similarly to a capacitor, the capacitance of which varies according to the applied acoustic waves (sound). The change in capacitance causes a slight change in the oscillation frequency of the circuit, which is what produces the FM modulation.

It will work.

The type of electret microphones used in this kind of circuit, output an AC voltage and include a built-in JFET amplifier to reduce the output impedance.

http://www.openmusiclabs.com/learning/sensors/electret-microphones/

The output from the oscillator will be both FM and AM. As the base voltage is varied, the transistor's capacitance will vary too which will change the oscillator's frequency. The change in base voltage will alter the collector current and gm will alter the amplitude of the oscillator.

You are right about the poor stability of the circuit and that a varactor diode will give better results. Stability can be improved somewhat by: using a varactor diode, running the oscillator off a regulated power supply and adding a buffer to the output so changes in load impedance don't alter the frequency.

[kc3ase]

http://www.talkingelectronics.com/projects/Spy%20Circuits/SpyCircuits-1.html "One transistor..."

You left off the CE capacitor of 4.7pf, or what would be appropriate for the changed values. Also Re is much smaller.

Ah, the original drawing image I found called out 4.7u which I thought was wrong. Since that 4.7u cap had no effect I pulled it out. Let me try 4.7 pF.

[kc3ase]

I'm assuming the 150Hz you're trying to add is supposed to be a PL (CTCSS) tone to access a repeater?

While the circuit you suggest is simple and easy to build, it comes with significant disadvantages that make it impractical for anything more than simple experimentation and toying around with.

Yes, it's just for experimentation and learning. With no filtering, modulation limits, temperature compensation, etc etc I'm quite sure this will be all over the place. C1 is adjustable. I just can't find the symbol/model for varcap in LT spice.

[kc3ase]

After a little bit more modeling and just placing in a varactor a ~100 MHz carrier appears.

Questions:
1) Why does Ib for Q1 show negative base current flowing? Is this bad?
2) How can one calculate current flowing through the tank circuit into RF power (assuming a dipole antenna with perfect resonance in between C1/C2, and at ground)?

[AF6LJ]

After a little bit more modeling and just placing in a varactor a ~100 MHz carrier appears.

Questions:
1) Why does Ib for Q1 show negative base current flowing? Is this bad?
2) How can one calculate current flowing through the tank circuit into RF power (assuming a dipole antenna with perfect resonance in between C1/C2, and at ground)?

Your varicap might be conducting on part of the RF cycle.
I have seen this hundreds of times before in circuits where the RF voltage exceeds the DC bias voltage on the varicap. 

You might want to modify the circuit a bit by feeding the varicap bias with a separate voltage divider and using a DC blocking cap to couple the diode to the rest of the oscillator.
That will also clean up the RF output of the circuit is rectification is taking place in the varicap.

[kc3ase]

>Your varicap might be conducting on part of the RF cycle.
Not according to the model. But I did check it, thanks.

I'm also trying this other circuit I've found from Forrest Mims shown in the attachment from page 46 of his Engineer's Mini Notebook - Communications Projects (C) 1987    He says this is an FM transmitter with an adjustable audio tone.

He actually has L1,L2 as a tapped homemade inductor. I used a calculator online to try to estimate the values of what each side would be. He says L1 is 5 turns of bare solid hookup wire wound around 3/8" diameter with a total length of 1/2". Tap is wire soldered at 1-1/2 turns point. I assumed by solid hookup wire he was referring to about 22 AWG (typical breadboard stuff) and calculated based off that. The antenna at 7" is supposed to be attached on the collector of Q1.

There's also another change to this circuit I'm trying. He runs a second 555 to control output pulses at 10 second intervals to meet FCC requirements. The second 555 pulls the voltage divider R2-R3 to 0V during the intervals disabling the RF output.

He goes on to explain that Q1 oscillates at a frequency controlled by C3 and L1,L2 and that you can temporarily disconnect the disabling timer for the RF adjustment.

I can, depending on some experimentation with L1 and L2's values get the circuit to oscillate in Spice but not at anything near 100 Mhz. Also, when the tone is generated on the timer the RF output is simply turned off. I don't get the frequency change during the tone output I would have expected.

Questions:
1) Are my L1 and L2 values wildly wrong?
2) Am I asking too much of LT SpiceIV?
3) This circuit in spice also shows Q1's base going into negative current. All the RF circuits I seem to model show this issue in SpiceIV. Has anyone seen this also?
4) If I want to convert this to 9V it looks like the only consideration I have to make is some more R1 resistance besides the cap voltage ratings. Correct?
5) Am I missing something in SPICE by not modeling the antenna's influence on the circuits?

TIA

[AF6LJ]

A spice model is only as good as the date you feed it, this includes all the strays that add up to meaningful amounts of parasitic capacitance and inductance. This is good for getting your DC values all in order and most of the simulators out there on the net are not suited for VHF work. I seriously doubt one can find a model for a 2N2222 that includes reliable data for how it will behave at 144-148MHZ.


He doesn't specify what size wire was used, I am willing to guess you would get better results with #18 wire, you will end up with a higher inductance with smaller wire.

Trying to make that circuit work on two meters with C-2 as large is it is might be problematic.
The base of that transistor is loading down the tank circuit quite a bit and may be another contributing factor to why the frequency is low.

What you really need to do is build the circuit dead bug style on a piece of copper clad board.
(you need a ground plane which will aid in stabilizing the circuit.)
Play with it on the breadboard and see how it works.

[kc3ase]

A spice model is only as good as the date you feed it, this includes all the strays that add up to meaningful amounts of parasitic capacitance and inductance. This is good for getting your DC values all in order and most of the simulators out there on the net are not suited for VHF work. I seriously doubt one can find a model for a 2N2222 that includes reliable data for how it will behave at 144-148MHZ.


He doesn't specify what size wire was used, I am willing to guess you would get better results with #18 wire, you will end up with a higher inductance with smaller wire.

Trying to make that circuit work on two meters with C-2 as large is it is might be problematic.
The base of that transistor is loading down the tank circuit quite a bit and may be another contributing factor to why the frequency is low.

What you really need to do is build the circuit dead bug style on a piece of copper clad board.
(you need a ground plane which will aid in stabilizing the circuit.)
Play with it on the breadboard and see how it works.

Yes, I was figuring the parasitics of components would be having an effect. The model assumes all components are ideal.


I'm going for the breadboard now. Already got the 555 up and running no problem. About 650 Hz on the output, not much surprise there. I'll post things as I get set up.

[AF6LJ]

Don't be surprised if you can't get the oscillator to work on that breadboard.
Remember all those long parallel rows of tied together metal contacts.
You can breadboard RF circuits on that board and get somewhat reliable results as long as...
Your circuits don't have a lot of gain.
Your circuits don't operate much above 30MHZ.
And as long as your RF circuits are not high impedance.

[kc3ase]

Don't be surprised if you can't get the oscillator to work on that breadboard.
Remember all those long parallel rows of tied together metal contacts.
You can breadboard RF circuits on that board and get somewhat reliable results as long as...
Your circuits don't have a lot of gain.
Your circuits don't operate much above 30MHZ.
And as long as your RF circuits are not high impedance.

Well I found that basically I get no oscillation. Spikes do show up when the 555 triggers but no meaningful constant oscillation continues after that point. Interestingly enough, that's pretty much exactly what SPICE said would happen.

I wish he would have written what the inductor value should be. I have an antenna analyzer but only good to 6M. [For measuring L even if I would make this coil in the unknown wire gauge]

Yeah I've heard there's a good amount of capacitance in a breadboard and this used to cause problems the other way as well as a design that worked on a breadboard doesn't work on a PCB.

Looks like the next step is move down to 10M and start there and work upwards as I get a better understanding.

[kc3ase]

We'll here we are at 10M. Lucky for me in SPICE there's no losses from an ideal design. I have disconnected the feedback cap C2 to see what the oscillator wants to do naturally letting it get kicked off by the 555 timer.

So then I replaced C2 with a value based on a reactance of ~12 ohms @ 28 Mhz or so. But then there's no more oscillation. If I'm learning correctly on RF and please correct me if I'm not, I want a cap which has a reactance much higher to not take a majority of the energy from the tank itself. Just enough to kick the tank back into action.

Which means I need to look at the current flowing in the tank and only take some portion of that to feed it back.

[AF6LJ]

We'll here we are at 10M. Lucky for me in SPICE there's no losses from an ideal design. I have disconnected the feedback cap C2 to see what the oscillator wants to do naturally letting it get kicked off by the 555 timer.

So then I replaced C2 with a value based on a reactance of ~12 ohms @ 28 Mhz or so. But then there's no more oscillation. If I'm learning correctly on RF and please correct me if I'm not, I want a cap which has a reactance much higher to not take a majority of the energy from the tank itself. Just enough to kick the tank back into action.

Which means I need to look at the current flowing in the tank and only take some portion of that to feed it back.

You are correct about the feedback you don't need a lot of it.
Two things to remember about building a successful oscillator.
You need an amplifier with enough gain to overcome circuit losses.
And you need positive (in phase) feedback.
Also keep in mind that breadboard has several pf of shunt capacitance built in. At that frequency that oscillator (if you can get it to work on the breadboard, I am doubtful of this) those same values for your L and C won't be close to what it will take to get the oscillator to work in your finished circuit.

[kc3ase]

Thanks Susan.

I'm going to try this more tomorrow. I got embedded into software and digital designs but I've always loved radio. It's probably because I don't understand it so well. I want to overcome that, and I'm going to have to start from scratch.

I was showing my son today how to create a flip flop to turn off and on two LEDs in staggered pattern he likes that but I'm totally lost when it comes to analog and RF.

[AF6LJ]

Radio is a lot of fun, learning RF is even more fun.

[G0HZU]

If you want to learn a few basics about oscillator design my advice would be to forget both of those crude and nasty FM bug oscillator circuits and start by drawing up a short list of performance requirements in terms of frequency stability over time, power level, tuning range and noise performance.

Then just build a classic oscillator circuit to try to meet the above requirements. Whilst it isn't realistic to expect to use a free running oscillator on the 2m band because of the relatively poor frequency stability wrt temperature and shock/vibration, you should be able to make something that is stable to maybe +/-2kHz over several minutes as long as it is screened and kept at a fairly constant temperature and not moved. So this would be OK for a bit of learning and fun.

I would look to use a JFET oscillator here. Do you have anything like a 2N3819 JFET to play with? You can make a pretty decent oscillator at 145MHz with just a 2N3819 JFET, a couple of capacitors and some wire to wind a resonator coil.

The main aim would be to try to get lots of group delay in the resonator at 145MHz as this will help with stability and phase noise. There are ways to simulate the oscillator and its resonator and from this info there are a few simple sums you can do to predict the phase noise performance. Generally speaking, having lots of group delay in the circuit means good frequency stability and low phase noise.

The circuits for those crude FM bug oscillators haven't evolved in over 35 years and the typical circuit will have high phase noise and poor stability. I could probably design, simulate and build a simple oscillator at >4GHz that would have similar or better stability and phase noise than some of those old FM bug circuits can manage at 100MHz.