• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Simple transistor headphone amp

Status
Not open for further replies.

carbonzit

Active Member
I don't understand your simulation.

Where does "negative feedback" come from in the 2nd sim? Is there an off-circuit connection we can't see? It looks like you just removed the emitter capacitor. Please explain the feedback mechanism (in your simulation, not in general theory).

And while it looks reasonable, how did you come up with those distortion figures (40% and 3%)? Did you run a FFT analysis on both sims? or did you just look at the waveforms and guesstimate them?

Also, the voltage sources are different (25mV vs 460mV) by more than a factor of 10; why is that? shouldn't we be comparing apples to apples here?
 

audioguru

Well-Known Member
Most Helpful Member
I don't understand your simulation.

Where does "negative feedback" come from in the 2nd sim? Is there an off-circuit connection we can't see? It looks like you just removed the emitter capacitor. Please explain the feedback mechanism (in your simulation, not in general theory).
The unbypassed emitter resistor provides lots of negative feedback and it reduces the gain from about 180 with no negative feedsback to 10 with negative feedback.

And while it looks reasonable, how did you come up with those distortion figures (40% and 3%)? Did you run a FFT analysis on both sims? or did you just look at the waveforms and guesstimate them?
An experienced guess. Most half-decent amplifiers have a distortion of 0.05% or less.

Also, the voltage sources are different (25mV vs 460mV) by more than a factor of 10; why is that? shouldn't we be comparing apples to apples here?
Negative feedback reduces distortion by cancelling most of the signal which reduces gain and reduces distortion. A TL07x audio opamp has a distortion of 0.003% because the negative feedback reduces its open-loop gain of 200,000 times to 10 times.
 
Last edited:

Mr RB

Well-Known Member
Hi Mr RB,
Have you seen how much distortion is caused by a single transistor without any negative feedback and at high levels? The distortion compresses the top of the waveform so much that you cannot see how much voltage gain there is. That is one reason that opamps have so many transistors.
I understand you tweaked the simulations to make a point, but I make common emitter amps like your second sim. Keeping Rc/Re to 10:1 or less, with no emitter bypass cap and 10:1 is plenty enough gain from a common emitter amp.

So when I said before "common emitter amp with no negative feedback" I was referring to something almost identical to your second sim. At least I think we're on the same wavelength even if not using the exact same terms. :)

To me "negative feedback" in an amplifier context specifically means feeding back some of the output signal back around into the input of the amp.
 

audioguru

Well-Known Member
Most Helpful Member
An unbypassed emitter resistor provides negative feedback to reduce the gain and reduce the distortion.
My simulated transistor with horrible distortion has no negative feedback because it has no emitter resistor and its gain is about 180.
If an emitter resistor has a value that is 1/10th the collector resistor then the gain is reduced to about 10 and the total distortion is about 3%.
 

Attachments

Mr RB

Well-Known Member
Thanks AG I did understand that. :) I still would not call that "negative feedback" to me it's more accurately called an amp with no feedback that just has a more linear response (as there is no DIRECT coupling from the output (collector) to the input (base)).

This is taking me back to 1979 when I did common emitter amps in basic electronics class at the college, and I remember something about keeping the gain under 10 and the output amplitude under 50% of the supply voltage.

You said it has 3% distortion, and your sim shows an output of 7.5v? What distortion does it have with a 4.5v output?
 

crutschow

Well-Known Member
Most Helpful Member
To me "negative feedback" in an amplifier context specifically means feeding back some of the output signal back around into the input of the amp.
Negative feedback is, in general, a signal from the output of an amp (or gain stage) that affects the input in a manner as to reduce gain and distortion.

It's perhaps not readily apparent but an unbypassed emitter resistor does provide local negative feedback to the transistor stage. The emitter resistor, in effect, creates a signal voltage that subtracts from the base signal voltage reducing the stage gain and also the distortion, the same as a separate negative feedback path would do.

The reason the emitter feedback reduces the output distortion is that the emitter current also flows through the collector load resistor (with the slight difference of the base current) and thus the collector and emitter signals are essentially equal in shape except for the gain caused by the difference in resistor values. Thus a reduction in the distortion of the emitter voltage (and current) reduces the collector distortion the same amount.

Does that make sense?
 
Last edited:

carbonzit

Active Member
Progress

With my new-found ability to compute THD from LTspice's FFT plots, I've improved the performance of this li'l guy from a pretty bad 6+% to just over 2%.

Two changes: I reduced the Q2's resistor (to 100Ω), and changed Q2 to a 2N4401 (the NPN complement to Q3).

Further improvement seems to be possible ...
 

carbonzit

Active Member
The reason the emitter feedback reduces the output distortion is that the emitter current also flows through the collector load resistor (with the slight difference of the base current) and thus the collector and emitter signals are essentially equal in shape except for the gain caused by the difference in resistor values. Thus a reduction in the distortion of the emitter voltage (and current) reduces the collector distortion the same amount.

Does that make sense?
Not really, but I don't know if that says more about your explanation or my density ...

As I understand it, negative feedback reduces distortion by feeding back a signal consisting of the original signal (after being amplified) plus any distortion, but inverted from the original signal. The sum of these two signals would be the distortion, which is then (at least partially) cancelled or removed from the previous amplifier stage. Is that more or less correct? In other words, it acts somewhat like the error signal in a voltage regulator.

So how does your emitter resistor behave in this regard? is it somewhat the same mechanism at work?

And how would this be applied to my li'l amp?
 
Last edited:

audioguru

Well-Known Member
Most Helpful Member
Your Li'l amp already has some negative feedback with the resistor from collector to base on the first transistor.
But for very low distortion you need lots of gain provided by many transistors then the negative feedback reduces the very high gain a lot and also reduces the distortion a lot.
 

Mr RB

Well-Known Member
...
The reason the emitter feedback reduces the output distortion is that the emitter current also flows through the collector load resistor ... (snip)

Does that make sense?
Yes I already knew and fully understood that, even bfore AGs example. I think it's a communication preference, I personally would not describe that as "a common emitter amp with negative feedback" I would have described it as "a standard common emitter amp (with no feedback)".

I would have reserved the term "amplifier feedback" for specifically taking signal from the output and routing that back to the input.

Sorry to sound argumentative but what would you say for a 2 stage or 3 stage (standard) common emitter amp? To me I would only describe that as "a 3-stage common emitter amp with feedback" if it had signal from the output of Q3 feeding back to the input of Q1.

If each of the stages has some mechanism that makes that stage more stable or more linear or mimics the properties of negative feedback, to me would not mean the amp is described as "a 2 stage amp with negative feedback" which brings us back to my orignal comment and meaning of what I said to AG about his mods to Carbonzit's amp.
 

audioguru

Well-Known Member
Most Helpful Member
Most audio amps and all opamps have a differential input so that the non-inverting input is used for the input signal and the inverting input is used for the negative feedback. Then the input impedance is very high.
If the amp has only a single input then negative feedback to it will result in a very low input impedance.
 

crutschow

Well-Known Member
Most Helpful Member
Yes I already knew and fully understood that, even before AGs example. I think it's a communication preference, I personally would not describe that as "a common emitter amp with negative feedback" I would have described it as "a standard common emitter amp (with no feedback)"......
Well yes, I believe there have been previous threads where you have stated your preference for using terms in a manner different from the usual engineer interpretation of the terms. You are, of course, perfectly free to use a term in any manner you like. Just don't expect the rest of us to agree with your definitions. ;)

But your definition of a common emitter amp with an un-bypassed emitter resistor as a "a standard common emitter amp (with no feedback)" is incorrect. A "no feedback" stage would either have the emitter connected to ground, or a capacitor across the emitter resistor to put the emitter at AC ground over the frequencies of interest. An un-bypassed emitter resistor does indeed provide negative feedback in the engineering sense of the word. Edit: The evidence of this is that it stabilizes and reduces the stage gain, and reduces the distortion, all characteristics of negative feedback.
 
Last edited:

ericgibbs

Well-Known Member
Most Helpful Member
I would agree with Carl, I was always taught the definition of a common emitters operation and configuration in the way he described it.
 

carbonzit

Active Member
So to bring this discussion back to earth a bit, would one add some measure of negative feedback by simply adding an emitter resistor to Q1 in my amp? If one did so, what other changes would be necessary in order to correctly bias that stage? a collector resistor?
 

audioguru

Well-Known Member
Most Helpful Member
So to bring this discussion back to earth a bit, would one add some measure of negative feedback by simply adding an emitter resistor to Q1 in my amp? If one did so, what other changes would be necessary in order to correctly bias that stage? a collector resistor?
Q1 in your circuit already has plenty of negative feedback through the collector-base resistor.
It has so much negative feedback that the distortion is fairly low at 3% (low for a simple and cheap amplifier) and the gain of the amplifier is less than 1 at 0.6.

This collector to base resistor provides Q1 with too much base current so its average collector voltage is low which reduces the max ouput swing. I showed a SIM of it clipping the bottom of the waveform in my post #9.

I showed the negative feedback coming from the output instead of from the collector of Q1 and I showed the collector resistor with bootstrapping which increases its active value for more open-loop voltage gain and it gives the NPN output transistor a base voltage that rises above the supply voltage for a higher output level.
 

carbonzit

Active Member
My amp beats AG's amp!

Well, you're wrong. I know I've said that a lot here, but this time I can prove it.

I went ahead and simulated your circuit in LTspice, as you gave it in post #16 here (.asc file attached below: notice the "AG" suffix):



I checked it very carefully against your posted picture, and I don't think I made any mistakes. (If I did, please point them out.)

When I run it, I get a distortion figure (using LTspice's .four function) of 3.45%. Not too bad.

My circuit, however (minimally altered from the original one I posted and also attached below) gives a THD of 2.16%.

Therefore, at least by the measure of THD, my amp is better than your amp. In other words, your so-called "improvements" didn't improve anything.

If you don't believe me, you can run it for yourself.

Next?
 

Attachments

audioguru

Well-Known Member
Most Helpful Member
Of course my output distortion is higher than yours. My output level is 20 times more than yours.
Try it with the same output levels.

The gain of your modified amp is 0.29. Mine is 0.6 like your original circuit. Try it with the same gains.

But your amp simulates pretty well for a cheap and simple circuit.
 

carbonzit

Active Member
In order to compare apples to apples, I changed the input to your amp to 200mV, same as I used for the sim for my circuit.

The THD value I got was actually a little bit higher: 3.54%. (I noticed the same thing with my circuit, where smaller input voltages resulted in higher distortion figures. Counterintuitive, but there must be a simple explanation for this.)

audioguru said:
But your amp simulates pretty well for a cheap and simple circuit.
Heh. I'm going to print this out and frame it on my wall. Grudging praise is the best. Made my day. (That, and a good stiff martini.)
 
Last edited:

alec_t

Well-Known Member
Most Helpful Member
I don't think I made any mistakes. (If I did, please point them out.)
In the first .meas directive 'R6' should read 'R1'
 

Mr RB

Well-Known Member
...
But your definition of a common emitter amp with an un-bypassed emitter resistor as a "a standard common emitter amp (with no feedback)" is incorrect. A "no feedback" stage would either have the emitter connected to ground, or a capacitor across the emitter resistor to put the emitter at AC ground over the frequencies of interest. An un-bypassed emitter resistor does indeed provide negative feedback in the engineering sense of the word. Edit: The evidence of this is that it stabilizes and reduces the stage gain, and reduces the distortion, all characteristics of negative feedback.
Thank you Crutschow (and Eric and AG too). I stand corrected. :)

I will expand my use of the word "feedback" to the engineering definition to include things that are not actually feedback but which mimic the effect of actual feedback.
 
Status
Not open for further replies.

EE World Online Articles

Loading
Top