Opamps feedback component ratios and frequency response

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

JAY X

Well-known member
Joined
Jan 9, 2009
Messages
683
Hi!

This is a question where i didn't found much info.

A friend told me that is important to keep the same component ratios in the circuits you build in order to maintain the same frequency response bandwidth across all circuits.

for example:

Circuit 1: inverting opamp: 10k input . 10k/22pf at opamp feedback path.
Circuit 2: VCA fader:              22k input .  22k/10pf at opamp feedback path.

It is not always possible to maintain the same ratios, but simmilary bandwidth for example  with 4k7/47pf
and so on....

¿What do you think? ¿are component ratios important?

Jay x





 
you need more friends...

no need to keep all ratios the same, the end result will be the combined response of all the channels combined.... 

i.e. 2 circuits that are each -3dB @ 20kHz in series will be -6dB at 20 kHz, etc.  It seems prudent to make most poles well above 20kHz so they don't combine to affect passband.

JR
 
Ok, i see...So, combining circuits in series, is like adding an order to a lowpass filter...N circuits is n filter order...

Jay x
 
JAY X said:
Ok, i see...So, combining circuits in series, is like adding an order to a lowpass filter...N circuits is n filter order...

Jay x
Yes if each circuit block has a one pole roll-off, the combined path will exhibit N poles.

JR
 
JAY X said:
A friend told me that is important to keep the same component ratios in the circuits you build in order to maintain the same frequency response bandwidth across all circuits.
I don't think it's really a valid concept.
Signal transmission theory says HF roll-off is better controlled by applying a passive pole right at the input, that should be the dominant pole (lowest frequency), then the subsequent stages should have higher poles, such that their combined response does not significantly alters the response.
 
abbey road d enfer said:
Signal transmission theory says HF roll-off is better controlled by applying a passive pole right at the input, that should be the dominant pole (lowest frequency), then the subsequent stages should have higher poles, such that their combined response does not significantly alters the response.

That's a pretty condensed-info sentence!

And thanks - a whole range of classic design decisions now makes a whole lot of sense to me...

/Jakob E.
 
abbey road d enfer said:
I don't think it's really a valid concept.
Signal transmission theory says HF roll-off is better controlled by applying a passive pole right at the input, that should be the dominant pole (lowest frequency), then the subsequent stages should have higher poles, such that their combined response does not significantly alters the response.
I've never seen that articulated that way but have practiced a variant on that for decades.  I generally incorporate a LPF at the very input to protect against input signals that are faster than the circuitry can track to prevent rectification and slew rate related distortions. Then I execute the dominant HPF pole with a quality film capacitor in a later high impedance stage (does not need to be passive).

====

Speaking of passive input poles, my last phono preamp (to end all phono preamps) several decades ago, incorporated the 75uSec RIAA pole passively before any closed NF loops. That preamp circuit happily accepted wide band square waves at the input. As if a phono cart could ever make those.  ::)

JR 
 
If you try to put a passive dominant high frequency pole at the input,  how do you control its frequency? Meaning it will be dependent on an unknown and variable source impedance.  Audio is not like RF or video with specific impedances.

It would seem you would to install a large series resistance at the input  to minimize the source impedance contribution.
 
> Audio is not like RF or video with specific impedances.

Audio is generally LOW impedance source. Zero to 470r. A 4.7k build-out may be ample.
 
john12ax7 said:
If you try to put a passive dominant high frequency pole at the input,  how do you control its frequency? Meaning it will be dependent on an unknown and variable source impedance.  Audio is not like RF or video with specific impedances.
True... in many cases we can assume a range of source impedances, but not a precision exercise to depend on that impedance to define a bandpass skirt frequency.  Passive input poles can be set way above and way below the passband, to protect the circuit path from DC and RF without affecting frequency response.  As I already posted I preferred to  place my dominant bandpass filters in later high impedance stages using higher quality, and more reliable filter values. 
It would seem you would to install a large series resistance at the input  to minimize the source impedance contribution.
Adding excessive series resistance to an input my better define source impedance but can add noise, undesirable for high performance audio paths.

JR
 
JohnRoberts said:
Passive input poles can be set way above and way below the passband, to protect the circuit path from DC and RF without affecting frequency response.

This is what I've usually done,  passive input pole to deal with RF (MHz range) . Then the active stage sets a reasonable audio BW  (100 kHz range).  Abbey seemed to be suggesting the reverse is preferred.
 
john12ax7 said:
This is what I've usually done,  passive input pole to deal with RF (MHz range) . Then the active stage sets a reasonable audio BW  (100 kHz range).  Abbey seemed to be suggesting the reverse is preferred.
Who's right or wrong is irrelevant. What matters is what's right or wrong.

JR
 
john12ax7 said:
If you try to put a passive dominant high frequency pole at the input,  how do you control its frequency? Meaning it will be dependent on an unknown and variable source impedance.  Audio is not like RF or video with specific impedances.

It would seem you would to install a large series resistance at the input  to minimize the source impedance contribution.
I agree that my answer was somewhat restrictive and needs to be expanded, but still the idea is to prevent the risk of slew-rate limiting, by making sure the signal that enters a stage is properly LPF'd, as JR suggests.
For example, loading a mic input with 3nF capacitance makes sure a 600r mic has less than 0.1dB attenuation at 20kHz; then a 100r mic will have a BW of about 500 kHz.  In practice I put a 10nF cap on each leg. And remeber there's usually a mic cable with significant capacitance.
Ideally, the active stage should have a slew rate capable of reproducing a 500kHz sinewave at nominal operating level. Then the subsequent stages should be LPF's passively.
Indeed, this kind of alignment requires knowledge of the source impedance possible range. Applying a source of impedance out of range has consequences, risk of intermodulation distortion if too low, restricted BW if too high.
Another factor is that most active microphones include LP filtering.
Pasive (Dynamic or ribbon) mics have significant inductance that reduces the BW more than the nominal impedance suggests.
Line input receiver stages are less difficult, because inserting resistors/inductors in the signal can be done without introducing too much noise.
 
I'm a little confused by the topic.  Are we now talking about mic inputs as relates to op-amps?  Is there even a decent way to use a naked op-amp as a mic pre other than the chips designed for that purpose? 
Anyway, I erased my earlier post as I seem to have grabbed the wrong end of a stick or lost the plot    :eek: 
I am interested in the subject though and would like to know the specifics  :)
 
> Are we now talking about mic inputs

The start post clearly does not.

Later posters seem to be following their own trails.
 
The start post clearly does not.

Hey P.   

Later posters seem to be following their own trails.

OK, well probably best I just follow along rather than look stupid  :)   

 
abbey road d enfer said:
 
Signal transmission theory says HF roll-off is better controlled by applying a passive pole right at the input, that should be the dominant pole (lowest frequency), then the subsequent stages should have higher poles, such that their combined response does not significantly alters the response.

That works for me.

For mic amps I like transformers, I'm not ashamed of that.    By default, they determine my dominant low and high pass.  So far, they've also successfully stopped any RF parasites getting in the  door.   
Beyond the transformers, many circuits I mess with would pass signal beyond what I can measure (500KHz) and some would pass DC if given half a chance. 

In something without transformers such as a power amp, my approach would be the same but using passives  at the entrance and exit.

 
yes on transformers!

noise rejection, galvic isolation, natural Bessell rolloff,  it's like putting whipped cream on a hot fudge sunday,  :D
 
Back
Top