I have verified both stages independently. The power stage works perfectly when driven by the function generator. The amplifier stage also works when disconnected. The oscillation only appears when you hook them together.

Does it still happen when the output of the OPA is not connected to the 25uF's, but just the power rails connected together.??

 

Does it still happen when the output of the OPA is not connected to the 25uF's, but just the power rails connected together.??

The output is clean with the power rails connected together.

 

I tested for noise on the power rails by separating the two stages and powering them from different supplies.

 

My comment about the phase shift on the collectors of the 2N2222 and PN2907 could be restated as those nodes could have poor high frequency response that might result in enough delay all the output stage to oscillate. But apparently, that's not the problem, as evidenced that the output stage works perfectly when driven by a function generator, and the oscillation only occurs when its connected to the opamps.

You might want to bypass that voltage follower stage just in case its the one that's oscillating. It is more prone to oscillation than the gain stages. Indeed if you reduced the number of gain stages and increased the gain of the remaining stages would also reduce those stages of oscillation.

Next guess: With a total gain of 200x, ground loops may be the source of the undesirable feedback. You might want to move the ground connections for the non-inverting inputs of the opamps to another place. You can do that by bending the corresponding pins so they don't go into the socket so you can jumper them elsewhere.

 

That does look like output stage oscillation to me.

I'm not sure the two 25 uF electrolytics need to be there. They bootstrap the drive current for Q5 and Q6, but also contribute frequency-dependent phase shift. Try the circuit without them. Also, I think the two 30K resistors are too high in value. You want a fairly low dynamic impedance driving the output stage. Try eliminating the 25 uF caps, connecting the output of U1D to the junction of D3 and D4, and reducing R9 and R11 to 4.7K or 4.99K.

ak

 

Okay, I solved the issue.

As DickCappels suggested, it was a ground loop issue. I moved the ground clip from the function generator close to the op amp ground and all of the oscillations disappeared.

Thanks for everyone's input!