Banifi,
Here is my opportunity to explain the term "delay" before you get more confused. The term is used and misused in this forum quite often.
There are two types of delay that people talk about:
a) inertial delay
This is what people usually mean when they talk about phase delay or group delay. It is what is usually the cause of phase shift. You can imagine the circuit to have electrical inertia (like a mechanical system has mass inertia). The higher the inertia the more sluggish the response to the input signal. For a sinewave input, the output will track the input more closely at low frequencies than at high frequencies. For a sinewave input the output will be a sinewave who's amplitude will decvreas with frequency and who's "phase delay" or "phase shift" relative to the input will increase with frequency. However, there is no actual delay involved - there is the illusion of a delay when you look at the sinewave response - the output responds INSTANTANEOUSLY to the input.
b) propagation delay
This is where the output does not respond to the input at all for a certain amount of time. A springline is a mechanical version, the old fashioned way of creating echo effects. The delay between seeing a lightening flash and hearing the thunder is a propagation delay due to the speed of sound in air. Propagation delays are inherent in digital systems. Electricity has propagation delay due to the speed at which electrons travel through copper or semiconductor junctions. Transistors exhibit some delays due to the physics of the depletion regions. However, the latter are extremely small when transistors are operating in their "active" region. Transistor turn-on, turn-off and storage time delays that you see in databooks refer to changing transistor state between active, full on and full off states and should not be confused with the minute delays when amplifying in the active region.
For the purposes of designing a competent audio amplifier you can ignore the subject of propagation delay entirely. It is a very low priority indeed. When you read someone talking about delay make sure you unserstand what delay they are refering to and ask them if you are unsure. I prefer to use the term "delay" for propagation delay and use the term "phase shift" for inertial effects.
Here is my opportunity to explain the term "delay" before you get more confused. The term is used and misused in this forum quite often.
There are two types of delay that people talk about:
a) inertial delay
This is what people usually mean when they talk about phase delay or group delay. It is what is usually the cause of phase shift. You can imagine the circuit to have electrical inertia (like a mechanical system has mass inertia). The higher the inertia the more sluggish the response to the input signal. For a sinewave input, the output will track the input more closely at low frequencies than at high frequencies. For a sinewave input the output will be a sinewave who's amplitude will decvreas with frequency and who's "phase delay" or "phase shift" relative to the input will increase with frequency. However, there is no actual delay involved - there is the illusion of a delay when you look at the sinewave response - the output responds INSTANTANEOUSLY to the input.
b) propagation delay
This is where the output does not respond to the input at all for a certain amount of time. A springline is a mechanical version, the old fashioned way of creating echo effects. The delay between seeing a lightening flash and hearing the thunder is a propagation delay due to the speed of sound in air. Propagation delays are inherent in digital systems. Electricity has propagation delay due to the speed at which electrons travel through copper or semiconductor junctions. Transistors exhibit some delays due to the physics of the depletion regions. However, the latter are extremely small when transistors are operating in their "active" region. Transistor turn-on, turn-off and storage time delays that you see in databooks refer to changing transistor state between active, full on and full off states and should not be confused with the minute delays when amplifying in the active region.
For the purposes of designing a competent audio amplifier you can ignore the subject of propagation delay entirely. It is a very low priority indeed. When you read someone talking about delay make sure you unserstand what delay they are refering to and ask them if you are unsure. I prefer to use the term "delay" for propagation delay and use the term "phase shift" for inertial effects.
Jan is creating an example of propagation delay here to explain how feedback will not work correctly if the delay is too big. But remember that in a linear audio amp the propagation delay is negligible for your analysis. Instead, you need to consider the affect of inertia on the feedback effectiveness.
traderbam said:
Well, the point was to illustrate the effects of feedback related to the time it takes to go full swing from input back to the input. It really doesn't matter for this analysis whether you call it phase shift of propagation delay or quantum tunneling effects, thats why I carefully talked about the travel time around the loop. I agree that most of the delay is caused by capacitive phase shift effects in the amp. That is why the output does NOT respond instantaneously to the inpout, as I am sure you are aware.
And please don't try to second-guess me, by the time you get my point I have already changed my mind.
. Low loop delay, you know.
Jan Didden
Jan,
The most unpleasant form of delay is the time it takes between saying or writing the wrong thing and realising you've done it!
I do disagree with you. Inertial effects (or capacitance or inductance if you like) do not affect the instantaneous response of the output to the input. Provided there are no propagation delays (let's not split hairs about the speed of light) the output will always respond instantaneously to the input. No matter how much capacitance in the circuit. The response may be very sluggish and of very low amplitude but the response is INSTANT.
That's my opnion, anyhow.
The most unpleasant form of delay is the time it takes between saying or writing the wrong thing and realising you've done it!
I do disagree with you. Inertial effects (or capacitance or inductance if you like) do not affect the instantaneous response of the output to the input. Provided there are no propagation delays (let's not split hairs about the speed of light) the output will always respond instantaneously to the input. No matter how much capacitance in the circuit. The response may be very sluggish and of very low amplitude but the response is INSTANT.
That's my opnion, anyhow.
traderbam said:
If you state it this way, yes, I agree. I interprete what you say that at the moment you apply a signal to say, an R-C filter, the voltage on the capacitor changes instantly, meaning if the signal is positive, it starts to rise instantly. Do I interprete you correctly? If yes, I agree.
Of course, the complete (don't know a better word) signal appears on the cap with a certain delay due to the phase shift. And in the examples above, that delay ultimately determines how the feedback reacts.
So, I think we are in agreement?
Jan Didden
Yes, this is what I'm saying.
Well, the problem is that if you think about it the output signal appears with a delay of infinity. I don't think that looking at inertial effects as if they are delays is helpful. I think it can lead to a mis-interpretation of what is going on. In a system without propagation delay the feedback isn't really sensitive to a delay it is really sensitive to the rate of change of output for a given input change - the electrical inertia if you like. Thinking about phase shift is a useful rule of thumb for designing a feedback system and we all talk about making sure the phase shift is less than 180deg when the loop gain is above 1. But we aren't talking about a 180 deg phase shift due to any time delay, it is simply due to inertia and we talk about it out of convenience as a angular shift in the special case of a sinewave.
traderbam said:
Hey, you promised not to split hairs!
I'm not sure I understand this term inertia in this context, it's new for me anyway, but I am sure you know what you are talking about.
On the sine as special case, I would say that a delayed sinewave (my use of the term) remains a sine wave, so the delay conundrum (if that is what it is) seems to describe reality pretty well, as all theories and analogies do by necessity. So, I'm not sure if it is *wrong* to use it, although there may be other ways to talk about it like your inertia, which also is no more than an analogy for reality. Sorry, I'm moving into philosophy here, but ultimately that's where all conversations go if you go deep enough.
And I guess I don't need to mention that any signal can be decomposed into sinewaves, each with their own *delay* or *inertia* or *phase shift*.
Jan Didden
Let me try explaining it this way.
When you put a sinewave through a system with inertia the output will "look" on an oscilloscope like the same sinewave delayed in time and with a different amplitude.
But it isn't that at all. For example, if you could draw a red dot on the peak of the input signal, the dot would not appear on the peak of the output signal. It would be in a different place. Same if you had a second signal, smaller, added to the original sinewave. The output only "looks" like a shifted sinewave because a sinewave is a special case, visually.
In a mechanical system the word "inertia" refers to the resistance of a mass to a change in its velocity (the kinetic energy change). In an electrical system the mass is typically mapped to capacitance and springiness to inductance, and voltage to velocity. Alternatively, you can map voltage to force, current to velocity, inductance to mass and springiness to capacitance.
When you put a sinewave through a system with inertia the output will "look" on an oscilloscope like the same sinewave delayed in time and with a different amplitude.
But it isn't that at all. For example, if you could draw a red dot on the peak of the input signal, the dot would not appear on the peak of the output signal. It would be in a different place. Same if you had a second signal, smaller, added to the original sinewave. The output only "looks" like a shifted sinewave because a sinewave is a special case, visually.
In a mechanical system the word "inertia" refers to the resistance of a mass to a change in its velocity (the kinetic energy change). In an electrical system the mass is typically mapped to capacitance and springiness to inductance, and voltage to velocity. Alternatively, you can map voltage to force, current to velocity, inductance to mass and springiness to capacitance.
traderbam said:
..but we know there really isn't a wave, isn't there? It's just that a certain node exhibits changing voltage levels. We call it a wave only because we can see the change over time, plot it against time and say hey, this really looks like a sine wave.
So, I am sorry, but you lost me.
Jan Didden
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.