I was just looking at the schematics that have a capacitor at the NFB, and the positions of the cap and resistor on the "ground leg" of the NFB. . . seem to be backwards (resistor on the wrong side of the cap)--not the alignment for passing a signal intact.
The effect of adding a resistor before input to a capacitor is usually a noise--the exact same noise from putting a cap at CI. Did anyone else notice it? Well, I don't know what the "right" way is, but that's surely not it.
What are some other options for AC coupled NFB?
The effect of adding a resistor before input to a capacitor is usually a noise--the exact same noise from putting a cap at CI. Did anyone else notice it? Well, I don't know what the "right" way is, but that's surely not it.
What are some other options for AC coupled NFB?
It makes absolutely no difference which order the two components are in for the same RC effect.
It can be C followed by R or just as effectively R followed by C.
The same applies to the Zobel RC at the output.
However, when a signal is tapped off from the junction of the RC then the order of component placement in the signal path is paramount. eg, See Dr. Cherry's NFB tapping point from the mid point of the Zobel.
It can be C followed by R or just as effectively R followed by C.
The same applies to the Zobel RC at the output.
However, when a signal is tapped off from the junction of the RC then the order of component placement in the signal path is paramount. eg, See Dr. Cherry's NFB tapping point from the mid point of the Zobel.
reversing shield and core on both sides of a cable, disables the shield.
Reversing it at one end only requires true floating devices.
On most audio equipment ground is connected to the case, which connects the amplifier input to an antenna, if the sources polarity is reversed.
Also in most devices case and ground are connected to mains live through a capacitor.
This works fine, if the equipment is connected together in traditional ways, i.e with coaxial cables in correct polarity.
Reversing polarity would introduce all kind of RF and hum issues.
regards
AndrewT said:
Yes! Any AC on the 0v line goes through the zobel's capacitor and into the negative feedback loop--"a noise-powered dynamic tone control."
And probably. . . bog standard op amp tone control value of approximately 15nF (or smaller) polyester capacitor that's now used for the speaker zobel?
I like this a lot! It operates only in the presence of a noise--the worst noise there is in the amplifier, which is all that racket (AC) coming back from the speaker ground line. And, this works on power noise present on the 0v line too.
The more noise you make, the less noise you get? Brilliant!! Thanks man!
danielwritesbac said:
There is a "right and wrong" way. The cap is more prone to pick up noise due to it's physical size and so should be at the "low impedance" end as a generalisation.
Polystyrene caps have a red mark on them indicating the outer part of the foil. Again use this to "shield" the cap by making sure this is again connected to the low impedance point.
Worth doing if you can.
Re: Re: CI, is that RC backwards?
I think that's what I was saying?
#1. I like this: source, C, R, load (flat response)
#2. Not this: source, R, C, load (peakish)
But, I was asking for help identifying which direction the signal is going at the low (ground-seeking) leg of the NFB.
I'm just having a hard time with the concept of negative feedback loop, relating to where the signal is coming from--either coming from the direction of the inverting input OR coming from the direction of the 0v line. But, there sure is a signal, and it came from somewhere!
Mooly said:
I think that's what I was saying?
#1. I like this: source, C, R, load (flat response)
#2. Not this: source, R, C, load (peakish)
But, I was asking for help identifying which direction the signal is going at the low (ground-seeking) leg of the NFB.
I'm just having a hard time with the concept of negative feedback loop, relating to where the signal is coming from--either coming from the direction of the inverting input OR coming from the direction of the 0v line. But, there sure is a signal, and it came from somewhere!
Mooly said:
I'd like to put the capacitor of the low (ground-seeking) leg of the NFB (of a non-inverting amplifier), as close as possible to the origin of the signal that will travel through that capacitor. Nat-Semi names that cap "CI" on the Overture series schematics.
Or, in some way, find an option that actually works.
I made a promise to AndrewT that I would update my beginner's amplifier project to fully AC coupled "when" I could do it without loss of quality. That has not yet come to pass.
The sound made by R and C in the wrong order in an audio path is baritone fallout and peakish upper mids. This is what happens in LM3886.pdf with a capacitor at CI. So, I'm hunting options, but it would be easier if I had a clue.
Testing, and doing so on a theme:
We often think of glowing tubes, huge coaxial drivers, real wood, strong metals and LP's as having the charms of hi-fi. When I look at it, there's a lot of multipass, some reverbenance (another form of multipass), a bit of distortion; and, in such equipment, there's the big carbon resistors that block capacitive noise like a drinking straw would peanut butter.
But, some of the charm could be here: http://www.youtube.com/watch?v=E_v468ptuXw
Even on the original recording LM3886 won't play it back correctly, and if Jonas Brothers' "Year 3000" is more to your taste, well try that on valves and you might find out they really can sing. Personally, I think that most of the "magic" is within the supporting circuits rather than inside the glass.
If the job of hi-fi is increasing the value of our music investment. . .
And, we don't want our selections limited to just a few tracks. . .
And, if we don't want to fry any tweeter hooked to an L-pad (that isn't also protected by an additional cap straight on the tweeter's + terminal). . .
Then we need to solve the caveats of the cap at CI.
Since that part of the schematic in LM3886.pdf is clearly wrong, we need to "quit doing what doesn't work." However. . .
There are many benefits to a fully AC coupled amplifier:
Increased usable power output (dynamics)
Speaker protection (less tweeter-fry issues)
Can use fuses on DC rails
Cooler amplifier operating temperatures
Fun parallel and bridged now much easier
Different amplifier input impedances available (impedance match)
and more. . .
EDIT: If the caveats of the "cap at CI" can be solved.
Hi Daniel,
I don't know what say really. If you can reliably ( and that means double blind etc ) identify an audible effect from swapping places of R and C then something untoward is happening.
Could it be the effect of stray "pickup" physically by the cap as I mentioned earlier adding noise and hash to the feedback signal ? It is just a series circuit electrically. I do wonder if there is some form of "instability" ocuring under dynamic conditions and any "hash" tips it over the edge.
It's many years since I last tried anything with power OpAmps but at the time I found them very critical on layout etc. They always seemed to be running on the verge of instability to me.
Feedback connection points are another issue really, very important yes, but you seem to have something else going on.
I don't know what say really. If you can reliably ( and that means double blind etc ) identify an audible effect from swapping places of R and C then something untoward is happening.
Could it be the effect of stray "pickup" physically by the cap as I mentioned earlier adding noise and hash to the feedback signal ? It is just a series circuit electrically. I do wonder if there is some form of "instability" ocuring under dynamic conditions and any "hash" tips it over the edge.
It's many years since I last tried anything with power OpAmps but at the time I found them very critical on layout etc. They always seemed to be running on the verge of instability to me.
Feedback connection points are another issue really, very important yes, but you seem to have something else going on.
You can make this same abberation at the + input side too.
Its Case #2 below.
The hallmark sound is quite recognizable as the exact same consequences as "the cap at CI."
Example setup:
Any Overture chip amp, whether gainclone or traditional layout.
22k (or 15k, or 10k) input impedance. Its spec'd 680R to match typical chip amp kits. Its a non-inverting amplifier.
Case #1: (okay)
from a potentiometer
to a capacitor for DC blocking
to a 680R resistor
to the chipamp + input
Case #2:
from a potentiometer
to a 680R resistor
to a capacitor for DC blocking
to the chipamp + input
The noise in case #2 may help a cheap capacitor (upper mids boost), but defeats attempts at using quality parts. The umistakable effects are loose bass (or less bass), added midbass, baritone dropout, peakish mids, and a roll-off in the treble. Case#2 is "retail amplifier sound."
Its Case #2 below.
The hallmark sound is quite recognizable as the exact same consequences as "the cap at CI."
Example setup:
Any Overture chip amp, whether gainclone or traditional layout.
22k (or 15k, or 10k) input impedance. Its spec'd 680R to match typical chip amp kits. Its a non-inverting amplifier.
Case #1: (okay)
from a potentiometer
to a capacitor for DC blocking
to a 680R resistor
to the chipamp + input
Case #2:
from a potentiometer
to a 680R resistor
to a capacitor for DC blocking
to the chipamp + input
The noise in case #2 may help a cheap capacitor (upper mids boost), but defeats attempts at using quality parts. The umistakable effects are loose bass (or less bass), added midbass, baritone dropout, peakish mids, and a roll-off in the treble. Case#2 is "retail amplifier sound."
Re: Re: Re: CI, is that RC backwards?
If you draw your two circuits, you will see, where the sonic differences come from. Cin and Rb cannot be looked upon as a simple isolated RC circuit, like the questions in your first posts imply. You have to analyze the entire circuit formed by the potentiometer, Cin, Rb, the ICs input resistance and Rin, if there is one.
Being AC it goes either way.danielwritesbac said:
The signal comes from the output. Most of it goes to the speaker. Part of it goes to the inverting input. Another part of it returns to ground. The load resistance and the NFB resistors Rf and Ri determine how much of the signal goes where.danielwritesbac said:I'm just having a hard time with the concept of negative feedback loop, relating to where the signal is coming from--either coming from the direction of the inverting input OR coming from the direction of the 0v line. But, there sure is a signal, and it came from somewhere!
Do you use Rin to ground between Cin and Rb? Or is the potentiometer your only Rin?danielwritesbac said:
If you draw your two circuits, you will see, where the sonic differences come from. Cin and Rb cannot be looked upon as a simple isolated RC circuit, like the questions in your first posts imply. You have to analyze the entire circuit formed by the potentiometer, Cin, Rb, the ICs input resistance and Rin, if there is one.
Re: Re: Re: Re: CI, is that RC backwards?
YES!!!
Okay, first, the example does have a resistor (10k to 22k) as a load, directly from the input + pin to the 0v line.
I'm sure that what you just said applies exactly to the NFB as well; however, I don't have the means to tell you why I think so.
+input compared with -input. . .
I can only say that installing NatSemi's CI makes exactly the same awful noise as Case#2 above. And, so, NatSemi's PDF must be showing that NFB cap in the wrong place (bad design). Where is the right place for CI?
pacificblue said:
YES!!!
Okay, first, the example does have a resistor (10k to 22k) as a load, directly from the input + pin to the 0v line.
I'm sure that what you just said applies exactly to the NFB as well; however, I don't have the means to tell you why I think so.
+input compared with -input. . .
I can only say that installing NatSemi's CI makes exactly the same awful noise as Case#2 above. And, so, NatSemi's PDF must be showing that NFB cap in the wrong place (bad design). Where is the right place for CI?
Re: Re: Re: Re: Re: CI, is that RC backwards?
If your listening test reveals that it sounds better the other way round, your ground is probably not clean.
Not bad design, but solid design.danielwritesbac said:
That is the reason, why the National engineers do it that way round. The other way round should not affect the sound quality directly, but due to additional noise indirectly.Mooly said:
If your listening test reveals that it sounds better the other way round, your ground is probably not clean.
Does it actually produce noise? Or are you trying to say that the sonic quality leaves something to be desired with the capacitor connected?danielwritesbac said:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.