Hi,
I don’t know how to phrase this question but maybe someone can help me. I have been doing some reading through Mr. D. Self’s books and it brought me to a point where I am confused when it comes to negative feedback.
Components C4, R10 and R11 from image2.pdf form the negative feedback. If we for a moment forget about C4 and the -3dB low pass roll-off point then the resistor components calculates to 30.4 dB. What do we call this value? Is this the negative feedback factor, ratio ...?
From what I have read it is best to choose an amount of xdB of NFB at 20 kHz to ensure stability. I have also learned that once open loop gain has reached P1 it falls with 6dB/octave when standard milliard compensation is used and the same goes for the NFB available.
For me to get the xdB NFB available at 20 kHz do I subtract the open loop gain at that specific frequency from the calculate component value as set above to get the available NFB at that frequency.
Image3.pdf is my calculations and image1.pdf is the ouput(top green), phase shift(bottom green), open loop gain(top red) and resistor calculation(top blue) for my design so far. When we talk about unity gain, when does this occur based on the simulation? Is this when open loop gain reaches the resistors calculated value?
If the questions are unclear please indicate.
Any assistance will be appreciated or reference to a threat or site.
Thanks CorrieB
I don’t know how to phrase this question but maybe someone can help me. I have been doing some reading through Mr. D. Self’s books and it brought me to a point where I am confused when it comes to negative feedback.
Components C4, R10 and R11 from image2.pdf form the negative feedback. If we for a moment forget about C4 and the -3dB low pass roll-off point then the resistor components calculates to 30.4 dB. What do we call this value? Is this the negative feedback factor, ratio ...?
From what I have read it is best to choose an amount of xdB of NFB at 20 kHz to ensure stability. I have also learned that once open loop gain has reached P1 it falls with 6dB/octave when standard milliard compensation is used and the same goes for the NFB available.
For me to get the xdB NFB available at 20 kHz do I subtract the open loop gain at that specific frequency from the calculate component value as set above to get the available NFB at that frequency.
Image3.pdf is my calculations and image1.pdf is the ouput(top green), phase shift(bottom green), open loop gain(top red) and resistor calculation(top blue) for my design so far. When we talk about unity gain, when does this occur based on the simulation? Is this when open loop gain reaches the resistors calculated value?
If the questions are unclear please indicate.
Any assistance will be appreciated or reference to a threat or site.
Thanks CorrieB
Hi,
The resistors set the closed loop gain.
As Self states there is no fixed level of feedback.
The unity gain criteria is confusing until you realise the closed loop
gain attenuates the output by that amount at the inverting input.
So the open loop gain has to be lower than the closed loop gain
before the phase hits 180 degrees. The higher the closed loop
gain the less feedback there is and higher open loop bandwidth
can be used.
For a unity gain buffer you really have to slug the amplifier, as the
inverting input is at the output. For fast op-amps to be used at
reasonable gains it is undesirable that they are unity gain stable,
because at higher gains they will be overcompensated.
Selfs level of feedback at 20KHz comes from equation 2 and is a rule
of thumb in essence for typical output stages. In reality varying the
feedback factor at 20Khz for a given closed loop gain is also moving
the "unity gain" point.
/sreten.
The resistors set the closed loop gain.
As Self states there is no fixed level of feedback.
The unity gain criteria is confusing until you realise the closed loop
gain attenuates the output by that amount at the inverting input.
So the open loop gain has to be lower than the closed loop gain
before the phase hits 180 degrees. The higher the closed loop
gain the less feedback there is and higher open loop bandwidth
can be used.
For a unity gain buffer you really have to slug the amplifier, as the
inverting input is at the output. For fast op-amps to be used at
reasonable gains it is undesirable that they are unity gain stable,
because at higher gains they will be overcompensated.
Selfs level of feedback at 20KHz comes from equation 2 and is a rule
of thumb in essence for typical output stages. In reality varying the
feedback factor at 20Khz for a given closed loop gain is also moving
the "unity gain" point.
/sreten.Hi,
I'm going to take you a little further than you asked.
Let's interpret your gain and phase plots.
The OLG becomes 0dB @ 10MHz.
The phase is about 170 to 175 degrees @ 10MHz.
the phase margin is 180 - 170 = 10degrees at best.
Adding a tiny bit of capacitance to this amp will make it into an oscillator.
It will probably ring quite badly into a resistive load with a 5 to 10 degree phase margin.
Most recommend at least 45degrees of phase margin for good characteristics. However, even with an adequate phase margin, this will change when a reactive load or cables are added to the amp. You now need to compensate the amplifier to retain the near 45degree phase margin into a wide variety of reactive loads.
Some recommend 60degrees of phase margin but this can cut off transients and make the amp sound too smooth. This often happens when an inductive load is applied to the output of a poorly compensated amplifier.
So firstly. change the amp gain and phase to obtain an adequate phase margin at an open loop 0dB gain.
Then, do the difficult part. Compensate the amp to be tolerant to reactive loading.
I'm going to take you a little further than you asked.
Let's interpret your gain and phase plots.
The OLG becomes 0dB @ 10MHz.
The phase is about 170 to 175 degrees @ 10MHz.
the phase margin is 180 - 170 = 10degrees at best.
Adding a tiny bit of capacitance to this amp will make it into an oscillator.
It will probably ring quite badly into a resistive load with a 5 to 10 degree phase margin.
Most recommend at least 45degrees of phase margin for good characteristics. However, even with an adequate phase margin, this will change when a reactive load or cables are added to the amp. You now need to compensate the amplifier to retain the near 45degree phase margin into a wide variety of reactive loads.
Some recommend 60degrees of phase margin but this can cut off transients and make the amp sound too smooth. This often happens when an inductive load is applied to the output of a poorly compensated amplifier.
So firstly. change the amp gain and phase to obtain an adequate phase margin at an open loop 0dB gain.
Then, do the difficult part. Compensate the amp to be tolerant to reactive loading.
Hi Sreten,
just to extend the lesson a bit.
If Corri's original amp has the gain reduced to +10dB (similar to that 5534 min value).
The phase margin looks like it is around 60degrees @ ~3MHz.
Is that correct?
Will this amp sound different at different gains since the margin changes with the chosen gain?
If we decide to adopt the +30dB gain value should the Cdom (=Miller comp cap) be reduced to bring back the more correct phase margin rather than the overcompensated resulting from the 100pF?
When the followers are added does the phase margin change again?
just to extend the lesson a bit.
If Corri's original amp has the gain reduced to +10dB (similar to that 5534 min value).
The phase margin looks like it is around 60degrees @ ~3MHz.
Is that correct?
Will this amp sound different at different gains since the margin changes with the chosen gain?
If we decide to adopt the +30dB gain value should the Cdom (=Miller comp cap) be reduced to bring back the more correct phase margin rather than the overcompensated resulting from the 100pF?
When the followers are added does the phase margin change again?
Guys,
Thank you for the feedback. Correct me if I am wrong according to my graphs my OLG at 25 dB happens at +- 800 kHz and my phase shift is +- 125 degrees.
For the HF Gain equation how do I calculate gm? Which one of the follow two are correct.
1. gm=Ic/25mV
2 re=25/Ic (Ic in mA)
gm=1/(Re+re)
I have been looking around and this was the best that I could find.
CorrieB
Thank you for the feedback. Correct me if I am wrong according to my graphs my OLG at 25 dB happens at +- 800 kHz and my phase shift is +- 125 degrees.
For the HF Gain equation how do I calculate gm? Which one of the follow two are correct.
1. gm=Ic/25mV
2 re=25/Ic (Ic in mA)
gm=1/(Re+re)
I have been looking around and this was the best that I could find.
CorrieB
AndrewT said:
Hi,
Q1 - Yes.
You can safely increase the gain of an amplifier to increase
input sensitivity as stability is increased, at the expense of
lower reduction of any distortion by feedback.
Reducing amplifier gain you have to be more careful, distortion
reduces but so do your stability margins, proceed with caution.
The Miller cap is the dominant pole, in relation to the second pole,
the second will obviously lower when you add the output stage.
100pF is in the right ballpark for a complete amplifier.
/sreten.
For a nice introduction into Bode-analysis I can heartly recommend Ron Mancinis app notes, especially
http://focus.ti.com/lit/an/sloa017a/sloa017a.pdf
He gives a very readable and at the same time sophisticated overview on the method.
All the best, Hannes
http://focus.ti.com/lit/an/sloa017a/sloa017a.pdf
He gives a very readable and at the same time sophisticated overview on the method.
All the best, Hannes
Hi,
@ 340kHz the OLG crosses the +30dB line.
at this frequency, phase = 50degrees. Phase margin = 1809-50 =130degrees <<45degree OK.
Although his comment that overcompensated should be noted.
However, the slower output stage followers have still to be added. He reckons that your 100pF is about right when this is taken into account.
Tell what the sim says.
@ 340kHz the OLG crosses the +30dB line.
at this frequency, phase = 50degrees. Phase margin = 1809-50 =130degrees <<45degree OK.
Although his comment that overcompensated should be noted.
However, the slower output stage followers have still to be added. He reckons that your 100pF is about right when this is taken into account.
Tell what the sim says.
AndrewT,
I see what you mean. That crossover point is the unity gain point right?
In the following paragraph Bob Cordell mentions a gain crossover point of between 200kHz and 2MHz, does he refer to this same point?
"Here is an example:
Assume a degenerated differential input pair with RE = 225 ohms per
emitter
Assume this pair provides single-ended drive to the base of a CE VAS
transistor
Assume 1 mA through each transistor
Net gm = 2 mmoh (single-ended output of diff pair); gm = 0.002
Maximum available current for slewing = +/- 1 mA (single-ended
output of diff pair)
Now set the compensation assuming a 500 kHz gain crossover
(most power amplifiers will have a gain crossover between 200 kHz
and 2 MHz)
Assume 26 dB closed loop gain = 20:1.
To achieve this gain crossover frequency, we need OLG of 20 at 500
kHz.
Xc = 20 / gm at 500 kHz = j10,000 ohms
Cm = 0.159/(10k * 500k) = 32 pf
Slew rate SR = I/C = 1e-3 / 0.032e-9 = 31 V/us
Setting value of Miller cap and determining slew rate:
Xc = CLG/gm at Fc where CLG is closed loop gain and gm is that of
input stage
Cm = 0.159/Xc * Fc where Fc is the NFB gain crossover frequency
Cm = 0.159 * gm/(CLG * Fc)
SR = I/C = I * CLG * Fc /(0.159 * gm) = 6.28 * (I/gm) * CLG * Fc"
CorrieB
I see what you mean. That crossover point is the unity gain point right?
In the following paragraph Bob Cordell mentions a gain crossover point of between 200kHz and 2MHz, does he refer to this same point?
"Here is an example:
Assume a degenerated differential input pair with RE = 225 ohms per
emitter
Assume this pair provides single-ended drive to the base of a CE VAS
transistor
Assume 1 mA through each transistor
Net gm = 2 mmoh (single-ended output of diff pair); gm = 0.002
Maximum available current for slewing = +/- 1 mA (single-ended
output of diff pair)
Now set the compensation assuming a 500 kHz gain crossover
(most power amplifiers will have a gain crossover between 200 kHz
and 2 MHz)
Assume 26 dB closed loop gain = 20:1.
To achieve this gain crossover frequency, we need OLG of 20 at 500
kHz.
Xc = 20 / gm at 500 kHz = j10,000 ohms
Cm = 0.159/(10k * 500k) = 32 pf
Slew rate SR = I/C = 1e-3 / 0.032e-9 = 31 V/us
Setting value of Miller cap and determining slew rate:
Xc = CLG/gm at Fc where CLG is closed loop gain and gm is that of
input stage
Cm = 0.159/Xc * Fc where Fc is the NFB gain crossover frequency
Cm = 0.159 * gm/(CLG * Fc)
SR = I/C = I * CLG * Fc /(0.159 * gm) = 6.28 * (I/gm) * CLG * Fc"
CorrieB
yes,
Bob is picking the OLG to match the CLG @ 500kHz. (the crossover frequency as against the unity gain frequency where I went wrong in that example).
Then finding what amp parameters are needed to obtain that set of conditions.
Note, I said OK to the phase margin.
It should read
phase margin = 130degrees >> (much greater than) 45degrees.
Bob is picking the OLG to match the CLG @ 500kHz. (the crossover frequency as against the unity gain frequency where I went wrong in that example).
Then finding what amp parameters are needed to obtain that set of conditions.
Note, I said OK to the phase margin.
It should read
phase margin = 130degrees >> (much greater than) 45degrees.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.