Opamps feedback component ratios and frequency response

[Winston OBoogie]

I was approaching it from a line input perspective, it's possible that was incorrect.

Not really.    I  would say it's the same exercise, just as Abbey and John R. have explained above. 

Of course, the exact source impedance that would feed the filter is unknown and out of your control, but you can look at worst case (higher source Z) and use that as your lowest possible dominant pole.  These days most any source is low impedance - within the low to mid end of the range PRR said above - and the series R you use in the filter doesn't have to be so high as to degrade the noise performance to any great degree.

Edit, forgot:  HPF - for better or worse, I'm still attracted to using low noise J-Fets as input devices which, besides offering good immunity against gremlins, allow the use of a nice film cap to block any incoming DC and define the dominant low end.    John R's technique of using a later in the topology higher Z point sounds good for when the input is BJT with a lower input impedance. 

 

[Winston OBoogie]

OK, this might sound a bit Kooky but, I used to define my low and high pass band based on the product of the two.  This was not an original idea and was proposed many decades ago for achieving a 'balanced' response from receivers and such.  It might have been in RDH that I first read about it but who knows?

For instance, for a long while, I used 600,000 as the approx. product.  So, with a top end  at  20,000 Hz, my low end would be set to about 30Hz.  If the top went to 30KHz, low was  20Hz.  A miserable 8K top would have been 75Hz bottom and chucked in the bin, 44KHz and 13Hz,  200K +3Hz, is getting there, and so on...   
I  thought it had some merit at the time.  I can't hear to the same degree at the top end now so have no idea if I'd still feel the same.
I could probably find that miserably crap 8K top end thing I threw away and be done.


Hope I didn't  just admit all that in public 




Edit:  corrected faulty memory.

 

[dogears]

It’s ok, some of us (me) have junk for ears from the get-go!

 

 

[Winston OBoogie]

It’s ok, some of us (me) have junk for ears from the get-go!

 

Haha    Nah, I doubt it

 

[gridcurrent]

OK, this might sound a bit Kooky but, I used to define my low and high pass band based on the product of the two.  This was not an original idea and was proposed many decades ago for achieving a 'balanced' response from receivers and such. 

arguably, a concept with great merit.
discussed in the loudspeaker book by Badmaieff and Davis (page 25), who stated a multiplication number of 500,000.
perhaps worth mentioning that time honored electronics, tube and solid state, typically have limited bandwidth.

 

[JohnRoberts]

arguably, a concept with great merit.
discussed in the loudspeaker book by Badmaieff and Davis (page 25), who stated a multiplication number of 500,000.
perhaps worth mentioning that time honored electronics, tube and solid state, typically have limited bandwidth.

I was not familiar with a formula, but indeed it is an old concept where HF and LF skirts need to be complimentary... too much of one or the other does not sound natural.

JR

 

[abbey road d enfer]

I'm a little confused by the topic.  Are we now talking about mic inputs as relates to op-amps?

indeed the circuit in the OP is a line input, but there's no law that prevents using the same concept to any kind of circuit, is there?

 

[Winston OBoogie]

indeed the circuit in the OP is a line input, but there's no law that prevents using the same concept to any kind of circuit, is there?

No.  I said as much above.    Forget it, I  lost the plot 

 

[Winston OBoogie]

..  an old concept where HF and LF skirts need to be complimentary... too much of one or the other does not sound natural. 

Yes, exactly.  That's really what I meant to emphasise but didn't.    A formula doesn't matter, I might even have remembered the number I used to use incorrectly, it's been ages. 

But when working on some of the old  valve mic amps such as V72, V76 etc.,  I'd extend the top end beyond the old German 15Khz broadcast filter, and the whole thing would loose its naturalness.  I'd have to lower the dominant LF skirt to get that naturalness back.  I did it by ear, playing high quality orchestral recordings I knew.  Juggled the skirts around and then looked at the product of those skirts.
If I then used the same product number but increased or decreased the LF and HF for sh*ts and giggles, it would for the most part still sound balanced and natural.

 

[Winston OBoogie]

arguably, a concept with great merit.
discussed in the loudspeaker book by Badmaieff and Davis (page 25), who stated a multiplication number of 500,000.
perhaps worth mentioning that time honored electronics, tube and solid state, typically have limited bandwidth.

OK great, thanks.
I don't have that book so, that's another one to add to the list. 
Yes, I've seen a couple of different multiplication numbers stated in a couple of old books,  and there obviously won't be some magical 'Golden Ratio' type number that suddenly polishes turds.  Just be sensible with keeping things complementary.

 

[moamps]

...but indeed it is an old concept where HF and LF skirts need to be complimentary... too much of one or the other does not sound natural. ..

Sorry for bothering you, guys, but is there any recent scientific research or explanation about this phenomena?

 

[JohnRoberts]

Sorry for bothering you, guys, but is there any recent scientific research or explanation about this phenomena?

I am not aware of any specific research, but that does not mean it doesn't exist.

JR

 

[abbey road d enfer]

OK, this might sound a bit Kooky but, I used to define my low and high pass band based on the product of the two.  This was not an original idea and was proposed many decades ago for achieving a 'balanced' response from receivers and such.  It might have been in RDH that I first read about it but who knows?

For instance, for a long while, I used 600,000 as the approx. product.  So, with a top end  at  20,000 Hz, my low end would be set to about 30Hz.  If the top went to 30KHz, low was  20Hz.  A miserable 8K top would have been 75Hz bottom and chucked in the bin, 44KHz and 13Hz,  200K +3Hz, is getting there, and so on...   
I  thought it had some merit at the time.  I can't hear to the same degree at the top end now so have no idea if I'd still feel the same.
I could probably find that miserably crap 8K top end thing I threw away and be done.


Hope I didn't  just admit all that in public 




Edit:  corrected faulty memory.

I used to know that as the rule of 400 000  ;D
Seems consistent with 20Hz-20kHz, though.

 

[PRR]

> I am not aware of any specific research

Ratiotron Designer Handbook, 4th:

 

[Winston OBoogie]

Sorry for bothering you, guys, but is there any recent scientific research or explanation about this phenomena?

Heya Moamps,
I  can't remember where I first read about it, but  I remember it  being a study by, possibly (probably?),  RCA in the 1940's. 
It might be in Radiotron Designer's Handbook that I read it but I'd have to scroll through to know for sure.

 

[Winston OBoogie]

I used to know that as the rule of 400 000  ;D
Seems consistent with 20Hz-20kHz, though.

Haha, yeah 
I just read the excerpt from RDH that PRR posted and they state 500 000 so:
You = 400K, RDH = 500K, Me = 600K.  ;D ;D ;D


 

 

[moamps]

I just read the excerpt from RDH that PRR posted and they state 500 000 so:
You = 400K, RDH = 500K, Me = 600K.

On the bottom of the page PRR posted is a small print saying
"Various authorities place this figure from 400000 to 640000" so you are in.

Those sources are IMO outdated, and FR, THD, IMD figures of speakers, amps, sources are much better nowadays,  so I'm wandering is this rule for audio bandwidth related only to psychoacoustics  or technical parameters too.

 

[Winston OBoogie]

On the bottom of the page PRR posted is a small print saying
"Various authorities place this figure from 400000 to 640000" so you are in.

Those sources are IMO outdated, and FR, THD, IMD figures of speakers, amps, sources are much better nowadays,  so I'm wandering is this rule for audio bandwidth related only to psychoacoustics  or technical parameters too.

Looks like I just scraped by with an "also ran" then.  Cool   
Yes, you are absolutely right,  the numbers are outdated when compared to modern technology .  I certainly don't recommend folks design stuff by those rules.  Indeed, I'm more than a little ashamed to admit that I once  ran the numbers and adjusted LF & HF corners based on my quite un-scientific conclusions. 
As others have said though, I do think it matters that LF & HF 'skirts' are complimentary.  This is probably (IMHO) a psychoacoustic issue but, I'd love to  know for sure or otherwise. 
As I said, it could just be 'Kooky".   
Cheers mate