LC Circuit at Amp's output

[ericgibbs]

Hi Eric,
Thanks.

I understand that summing all odd harmonics results with a square wave.

I'm trying to figure out how the below circuit sums them up - in this case the 1st, 3rd and 5th harmonics.

hi,
This LTspice sim of your circuit shows how the f3 and f5 harmonics of the a 100KHZ square wave are attenuated by the circuit.

A sum of 100K/300K/500K are shown for Ref only.

EDIT:
This pdf may be helpful.

 

[MrAl]

Hi Eric,

Nice simulation there. Could you try raising the value of C3 to seven times the other two (that would make it 7nf)?
That might provide for a much better output as the 100kHz center will be more accurate that way.
The other two should be accurate the way they are though.

 

[ericgibbs]

Hi Eric,

Nice simulation there. Could you try raising the value of C3 to seven times the other two (that would make it 7nf)?
That might provide for a much better output as the 100kHz center will be more accurate that way.
The other two should be accurate the way they are though.

hi,
Initially changed C3 to 7n, kept the Inductor the same, nice exp build up.!

Then adjusted the inductor L3 to bring the tuned circuit back to 100KHz.
The plot is unstable for the first 100uSec, then it settles down OK.

 

[MrAl]

Hi Eric,

I was asking to see the expanded plot with the cap set to 7nf because that is the optimal value. The last 'tuned' LC is best 'detuned'. If we could see the waves more closely we would see that setting the cap to 7nf and then 'adjusting' the inductor to get that LC 'tuned' back to '100kHz', we actually end up tuning the circuit to a higher frequency, which of course means it doesnt work as well to remove unwanted harmonics. Setting the cap value to 7nf should do it though.
Notice the difference in amplitude with the 7nf and original inductor, and the amplitude with the 7nf and the new inductor. With the old inductor the amplitude is way high, while with the new inductor it's not nearly as high. That's because 'detuning' the last LC actually causes the circuit to get better tuned.

Sound strange?
For this circuit we want to block two frequencies and pass 100kHz. 'detuning' the last LC gets this done pretty well i think.
In other words, we should see the best 100kHz sine wave out of the network with the last LC detuned with 7nf and the original inductor.