Class i

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
If we do take Re (internal as well external) emitter resistances into account we get:
Gain = output/input = RL / ( RL + 1/gm + RE' + RE) ,
where gm ~= Ie / 26mV, RE' is the internal and RE is the external emitter resistance.
Now let's define Rtop = 1/gmtop + RE'top + RE for the 'top' tranny
and Rbot = 1/gmbot + RE'bot + RE for the 'bottom' tranny,
then gain = RL / ( RL + Rtop||Rbot )
and the gm of the whole circuit is 1 / ( Rtop||Rbot ), okay?

Now, let's plot this stuff and look at the black curve, which depicts the combined gm (of top and bottom tranny). Contrary to my previous post of gm, the green curve, this one doesn't show the 'doubling' at large currents. Instead, only a small increase in the crossover zone. This is because the circuit is a bit over biased (which is another story).
Despite the fact we have taken into account the effect of Vbe, RE' and RE, this kind gm does not explain the excess of distortion caused by a sliding bias.
Please tell me what I did wrong.

Cheers,
E.


The point is that the composite gm of the bipolar output trannies plus emitter resistors did not reveal the gm doubling at large currents. That means you can't use it for explaining the distortion. Perhaps you forgot it, but we are talking about the real cause of distortion: is it bias voltage modulation or is it gm modulation. According to my last plot (black curve) it is not gm modulation.

edit: the astute reader would notice that the 'green' gm curve does show gm doubling. So, wtf are we talking about?

Hi Edmond,

Foremost , happy new year to you and to all members of the forum,


There s nothing wrong in the equations above as per se , only
the way they must be interpreted differ between people in this thread.

If we take your circuit as exemple and apply the said equations we can
easily deduct that gm variation of your output stage for 100mA bias
range from 5.26S with no output signal to an asymptotical limit of 8.33S
at very large signal , that is 1.6 variation ratio and this should translate
in the simulation , wich it doesnt if we look at your curves for the non
sliding bias one.

Yet , the sliding bias curve show about this ratio in gm variation ,
wich is the cause of its higher distorsion , i think.

What the equation above dont readily show , or rather interpret,
is that the emitter resistors will cause a local degeneration , that is ,
local negative feedback within the device from emitter to base and this
will translate in what you are interpreting as "base modulation".

Of course , negative feedback will modulate the input signal at the
relevant node, explicitly the devices bases.

Implementing a sliding bias is just reducing the feedback ratio
from emitter to base , hence negating the NFB provided by the emitter
resistors signal and this translate in higher gm variation in the sims , hence ,
higher distorsion as well..................;)

As obvious in the equations above , quasi linearization of the caracteristic
imply gm variation being mitigated by increasing the emitters resistors values,
wich will induce lower gm ------ > lower but also more linear gain caracteristic.

Is there something wrong in this view?....
 
To me the non linearity is low enough to be controlled by overall NFB.
What I really like about this circuit, is how by combining a n channel VMOS source follower on positive half and n channel common source on the negative half we get a low cost all n channel output stage without the bias stability problems.
 
Hi Edmond,

Foremost , happy new year to you and to all members of the forum,

Thank you, wahab, the same to you.

There s nothing wrong in the equations above as per se, only the way they must be interpreted differ between people in this thread.

I wish I had never dropped these equations. I have never used them. They were only meant as an illustration of what approximately happens inside the simulator. The graphs I've shown were based on simulation, thus NOT based on these f*king equations.

If we take your circuit as example and apply the said equations we can easily deduct that gm variation of your output stage for 100mA bias range from 5.26S with no output signal to an asymptotic limit of 8.33S at very large signal , that is 1.6 variation ratio and this should translate in the simulation , which it doesn't if we look at your curves for the non sliding bias one.
>a 1.6 variation ratio ? I disagree.
The idle bias was 150mA, RE' = 0.04R and RE = 0.12R. So gm ~= 1 / ( 26/150 + 0.04 + 0.12 ) = 3A/V per side; thus for both sides 6A/V, while the graph shows 6.2A/V.
At the extremes of the graph, where Ie = 5A, (the 'other' Ie is almost zero, so we may ignore this one) we get gm ~= 1/( 26/5000 + 0.04 + 0.12 ) = 6.053A/V, while the graph shows 5.68A/V. So I would say reasonable in accordance with my simplified equations. Calculated and simulated gm variation are 1:1008 resp. 1:109.

Yet , the sliding bias curve show about this ratio in gm variation ,
which is the cause of its higher distortion , i think.

The sliding bias curve shows a 1:2 variation, which has nothing to do with your (erroneous) 1.6 variation. It is caused by the effect you have just explained below: 'negating the NFB provided by the emitter resistors signal'. :)

What the equation above don't readily show , or rather interpret,
is that the emitter resistors will cause a local degeneration , that is, local negative feedback within the device from emitter to base and this will translate in what you are interpreting as "base modulation".

Sorry, I can't follow you. Calculation as well as simulation has taken into account the effect of the degeneration resistors.
>"base modulation" ? AFAIK, I've never used this terminology. Perhaps you mean 'bias modulation'?

Of course , negative feedback will modulate the input signal at the relevant node, explicitly the devices bases.

Implementing a sliding bias is just reducing the feedback ratio from emitter to base , hence negating the NFB provided by the emitter resistors signal and this translate in higher gm variation in the sims , hence, higher distortion as well..................;)

I fully agree, though I still insist that the nonlinear bias voltage modulation is the root cause of increased distortion. :D

As obvious in the equations above , quasi linearization of the characteristic imply gm variation being mitigated by increasing the emitters resistors values, which will induce lower gm ------ > lower but also more linear gain characteristic.

Is there something wrong in this view?....

Nope.

Cheers,
E.
 
To me the non linearity is low enough to be controlled by overall NFB.
What I really like about this circuit, is how by combining a n channel VMOS source follower on positive half and n channel common source on the negative half we get a low cost all n channel output stage without the bias stability problems.

That's exactly what Marcel van de Gevel did. :D
See: 'Audio power with a new loop', Electronics World, Feb. 1996, pp. 140..143.
Perhaps you did mean this circuit, right?

Cheers,
E.
 

Attachments

  • Gevel.png
    Gevel.png
    324.3 KB · Views: 333
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.