Symptoms of an oscillating amplifier?

[Antennae]

I've got the talking Mighty Mite that makeRF has:
http://makerf.com/posts/so_you_want_to_be_a_shortwave_pirate

I was wondering how to tell if my amplifier is oscillating?

[ff]

For your application, depending on the frequency of oscillation, you might hear whistling, whining, a sound like rushing air, or "motorboating", which is kinda like a slow "wub-wub-wub" noise.  If your amp is oscillating, you should immediately be able to tell by listening.  Hope that helps...

[Antennae]

Thanks, those are easy to look for. It doesn't exhibit those symptoms. 
Wub wub wub... 

[makeRF]

The best way to see what's happening is with an oscilloscope, otherwise you are shooting in the dark.  But, like ff said, you should be able to hear the circuit oscillating on an AM/shortwave radio.  Tune to the frequency of the crystal, and tune your variable capacitor.  You should hear a tone if you are in USB/LSB mode or silence if in AM once it is oscillating at the resonant frequency.

EDIT:  I just re-read the subject line and realized that you were talking about an oscillating amplifier, not oscillating transmitter.  So, disregard everything I said above about crystals and tuning with the variable capacitor, etc.  A scope would still help identify stray oscillations in an amplifier.  What type of amplifier are you using, and how is it hooked into your circuit?

[ka1iic]

If you are talking about the transmitter on the MarkRF page using the LM386 chip a cheap and dirty way to know is to test the heat from that audio chip 1) run the audio chip into an 8 ohm load resister and compare that with the same chip running into the oscillator (transmitter).  If you notice a large difference in heat produced... you have a problem.

As the others have said the best way is to use a scope but some folks don't have such things :-)...  If it is oscillating in the ultrasonic region then the audio signal from the transmitter might sound 'weird' (high pitched squeals  on the audio) or there may be strange spurs off the oscillating frequency.  Ultrasonic oscillations can be real dogs but I don't expect that from the circuit as shown.

Hope this doesn't muddy the waters for you ....

[Antennae]

makeRF: I'm using the LM386 chip as an amp.

kaliic: The audio sounded fine, just quiet. The LM386 chip would get warm when transmitting.  It would get really hot running a speaker without resistors. It wouldn't get warm when running a speaker with resistors. 

Its not assembled anymore.  It was bugging me why the sound didn't carry very far.  I thought maybe because it was oscillating? I'll reply here if I tweak with it.

thanx

[makeRF]

Ah, I see... I was thinking you were talking about amplifying the RF.

Yes, the LM386 works as an audio modulation source for the transmitter, but it's not ideal.  In my tests the audio was always a bit dampened.  It works, but it's not the best.  It's really simple to construct, so that's the tradeoff.

As for the LM386 getting hot, I'm not sure if mine did or not... It's been a while since I tinkered with that circuit.  What I presented in the schematic was the bare essential parts to get it working.  You might have better luck building the full LM386 circuit presented in the datasheet.  It has a few more parts that might help with stability.  Also, now that I think about it, one reason it might be heating up is the voltage input is too high.  You could try to step the voltage down to the chip, which might help.  Or, the current might be too high, and in such a case you might need to add in a resistor to limit the current at the input.

[ff]

the current might be too high, and in such a case you might need to add in a resistor to limit the current at the input.

There are several versions of the LM386.  Dave Martin's Corsair also used one to series modulate the driver stage of his Corsair.  The consensus in our builder's group was that the best modulation came from the LM386-N3 or N4 that was capable of more current output.  The typical '386 is the N1, which runs out of gas in these applications.  You might have better results with an N3.  Worth a try...

[Antennae]

Ah, I was using the N1. 

[makeRF]

There are several versions of the LM386.  Dave Martin's Corsair also used one to series modulate the driver stage of his Corsair.  The consensus in our builder's group was that the best modulation came from the LM386-N3 or N4 that was capable of more current output.  The typical '386 is the N1, which runs out of gas in these applications.  You might have better results with an N3.  Worth a try...

Good point ff!  I used an N3 in my build.  The audio is still muted a bit, but perhaps not as much as the other variants.  I believe I ordered a handful of the LM386 N3's from Jameco at some point.  I simply chose the N3 because it puts out more power.  IIRC, Radio Shack only sells the N1.

[Antennae]

So it probably would have worked better!  Now, if I hadn't already pilfered parts from it...
 I'll amend the hfublog post when the blog comes back up.

[ka1iic]

The audio problem is most likely from an impedance mis-match.

One must take into consideration that these chips are designed to operate with maximum power into a 4 to 8 ohm load and I must assume that the oscillator is showing a much higher impedance (27-30 ohms) than what the chip is designed for therefore a larger chip would be called for.

The higher impedance of the oscillator in reference to the audio chip, would cause the audio chip to produce much less current (watts)because of the load of the oscillator.

I just looked at the circuit again and... my gut tells me to place a very large value capacitor (250 uf or greater) in the line between the audio chip and the oscillator and put a resistor of say 16 ohms from the audio chips output to ground.  The Capacitor will pass the audio but will keep any DC current from entering the audio chip.  The chip might (I say might) not like having DC entering the output of the audio chip.

If I were to try this circuit I would use a 380 chip and not the 386 as shown.

Please take anything I say with a grain of salt... <heh> I just had my 3rd spinal operation and "they" have me all "doped up"... <heh> 

p.s.  Syringomylia just plan sux...

[Antennae]

Sorry to hear about syringomylia, I just looked it up. A neighbor walks around with a walker because he had a tumor on his spinal cord and it slowly destroyed his leg control before they found it.

I have lots of questions:
How did you get the 27-30ohm range of the oscillator circuit? Is this about the same value for all circuits like this?

And how did you come up with the 16ohm resistor for the audio chip's output to ground?

So that its clear: Would I put the 16ohm resistor on the LM386 number 2 pin?

THANKS! The fires are kindled to make this thing work more better. 

[ff]

All of what Vince said is true, Antennae.  If you are delving into this far enough to be concerned about impedance matching and fidelity, you should be forewarned that modulating an oscillator is very poor methodology and you probably will never be satisfied with the results.  MakeRF's circuit is more to demonstrate the principles involved, and to give the newbie a taste of generating a modulated signal.  If you are reaching for more than that, you should build a more ambitious design, like a Corsair:

 http://freeradiotx.blogspot.co.uk/2009/06/corsair-transmitter-by-dave-martin.html

[Antennae]

Thank you ff, if I go for something more powerful I'll consider the Corsair. But I think it is now a personal vendetta against this wee little punk of a transmitter I have.   It took me a lot of time and effort to make this and it let me down (won't hardly transmit sound across the street).  I want to twist its nose to make it louder so to speak.