No feedback amp

[Rafael L]

Hi

I like to know how to use this simple topology without negative feedback.
In other circuits use complementary pair in the input to avoid the DC output, for the simple circuit below, what is the simplest way to avoid the DC output, when negative feedback is removed?

Thanks

[Corben]

How are you going to select the total gain if you eliminate the feedback?

[alexcp]

You can replace emitter degeneration resistors (R2 and R3) in the differential pair with a trimpot, connecting its wiper to the current source (Q3); this would allow you to adjust DC offset.

However, open loop gain will be high enough to make precise balancing difficult, and the result unstable. This amp without NFB is likely to be impractical, much as an op amp without NFB.

[Rafael L]

How are you going to select the total gain if you eliminate the feedback?

Placing a voltage divider, generally difference between the bases of differential pair is 5mV-30mV, depends on the gain of the VAS.

You can replace emitter degeneration resistors (R2 and R3) in the differential pair with a trimpot, connecting its wiper to the current source (Q3); this would allow you to adjust DC offset.

However, open loop gain will be high enough to make precise balancing difficult, and the result unstable. This amp without NFB is likely to be impractical, much as an op amp without NFB.

DC I have is half the source (15V), I need another transistor to compensate, but I do not know how is the circuit!

[Bonsai]

Is this for simulation purposes? If yes, make R21 a very low value (.001 Ohms) and C4 a very high value. You can then run your AC sims to get an idea of th eopen loop response.

If you want to build a low or ZGFB amp, then you can try heavy degeneration of the input devices and/or loading VAS to reduce the loop gain. A third method is to tap off the feedback signal from th e VAS rather than th eoutput, but this leaves you with no feedback top correct the cross over distoprtion, unless of course you are planning to run it heavily class A. Of course, you can use any combination of th e above as well (e.g. heavy degen of th e LTP and take f/back off the VAS).

[Rafael L]

[QUOTE=Bonsai;2103958]Is this for simulation purposes? If yes, make R21 a very low value (.001 Ohms) and C4 a very high value. You can then run your AC sims to get an idea of th eopen loop response.

Yes is for simulation, this will increase the gain and had to use two voltage dividers. It is used in opamps

Originally Posted by Bonsai

If you want to build a low or ZGFB amp, then you can try heavy degeneration of the input devices and/or loading VAS to reduce the loop gain. A third method is to tap off the feedback signal from th e VAS rather than th eoutput, but this leaves you with no feedback top correct the cross over distoprtion, unless of course you are planning to run it heavily class A. Of course, you can use any combination of th e above as well (e.g. heavy degen of th e LTP and take f/back off the VAS).

I'm trying to do is one pre-amp in class A, +- 4V RMS. I want to remove the negative feedback because not is necessary!

Distortion in 20KHz (with feedback)




Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 2.000e+04 5.500e+00 1.000e+00 -1.00° 0.00°
2 4.000e+04 8.877e-06 1.614e-06 -119.72° -118.73°
3 6.000e+04 1.169e-06 2.126e-07 -141.78° -140.79°
4 8.000e+04 1.533e-07 2.787e-08 176.90° 177.90°
5 1.000e+05 3.605e-08 6.555e-09 167.56° 168.55°
6 1.200e+05 3.098e-08 5.633e-09 -170.17° -169.18°
7 1.400e+05 1.221e-08 2.221e-09 -155.21° -154.21°
8 1.600e+05 2.014e-08 3.662e-09 166.43° 167.43°
9 1.800e+05 2.153e-08 3.916e-09 171.56° 172.56°
Total Harmonic Distortion: 0.000163%

[revintage]

No wonder you have low distortion-figures as NFB might be somewhere in the +40db region.

It will be easier if you do a Diamond output stage with a servo. Crossover-distortion will not be a problem as long as you stay in Class-A. Check this out:


http://www.andiha.no/articles/audio/cascfig/sch01.gif

Cascode Power Amplifier

[Rafael L]

Originally Posted by revintage

No wonder you have low distortion-figures as NFB might be somewhere in the +40db region.

It will be easier if you do a Diamond output stage with a servo. Crossover-distortion will not be a problem as long as you stay in Class-A. Check this out:


http://www.andiha.no/articles/audio/cascfig/sch01.gif

Cascode Power Amplifier

Negative feedback is 36 dB, so what the distortion in open-loop is <0,01%.
the circuit is the blameless without power TRs output.
With VAS balanced, I had already seen, here an example:

http://users.ece.gatech.edu/mleach/papers/tcir/tcir.pdf

But would like a solution to blameless.

[wahab]

[QUOTE=Rafael L;2104068]
the circuit is the blameless without power TRs output.

QUOTE]

I would call it a LIN...
Douglas Self didn t invent nothing in that matter, yet, a lot
of people name this topology with his "trade marks"..

[AKSA]

The original 1955 Harry C Lin amp used singleton input, SE VAS and PNP germanium quasi output stage with Sziklai npn driver.

If you want to build the amp without feedback, then to be strictly correct you must eschew interstage feedback, so only local feedback is acceptable..... this is really quite difficult.

As Bonsai suggests, it is best done with lashes of emitter degen on the LTP, lots of degen on the VAS, and some pretty stiff loading or feedback on the VAS collector. If the output stage is not operating in Class A, crossover could be an issue with no global feedback.

Interstage feedback between VAS collector and the LTP - the same RRC network transposed from output back to the VAS collector, and thence to fb node - is a better suggestion, however, as it will also reduce source impedance driving the output stage, always a good thing.

I have done a suitable Class AB zero global fb amp in the attachment, with simulation. It uses lashes of local feedback, and then adds interstage feedback from VAS back to LTP. It is possible to cut OLG to about 30dB with even more local and no interstage, but the results are not as good. As you can see, H2 is around -48dB, not very impressive. However, if you build this circuit you will be quite surprised......

Hugh

Lin1.gif