Removing coupling capacitors from console modules outputs - "for and against"?

[ln76d]

Another quick question from my sleepy mind (4AM here) 

In my console  most of modules have coupling capacitors on the inputs and the outputs - mostly electrolytics but there are also film caps (wiii!!! )
For example in a simple connection:
mic pre -> eq (switched in chain) -> active fader
Cable max. 30cm length

What do you Guys think about removing output  capacitors from the preamp and eq?
Eq switch is after it's input balancing stage so it shouldn't be audible at all.
Any disadvantages in my nasty plan? Any problems with dc offset running on the cables?
Or did i missed something?

 

[abbey road d enfer]

Another quick question from my sleepy mind (4AM here) 

In my console  most of modules have coupling capacitors on the inputs and the outputs - mostly electrolytics but there are also film caps (wiii!!! )
For example in a simple connection:
mic pre -> eq (switched in chain) -> active fader
Cable max. 30cm length

What do you Guys think about removing output  capacitors from the preamp and eq?
Eq switch is after it's input balancing stage so it shouldn't be audible at all.
Any disadvantages in my nasty plan? Any problems with dc offset running on the cables?
Or did i missed something?

You may do it if you accept switching clicks and noisy pots.
That is supposing the design is sane enough as not to have indecent offset that would push the operating point away from the comfort zone.

 

[gyraf]

You may do it if you accept switching clicks and noisy pots.

..and just to clear up the ambiguity: By that he means: DONT!

Jakob E.

 

[synthiaks]

Isn't most preamp circuits in consoles Transistor based? Most of the ones i know are and removing any Input or output caps from transistor stages doesn't make sense to me. But i might have misunderstood the question.

My take on this is that if there is capacitors anywhere in a circuit,
filtering DC they are there for a reason, removing them most probably wont improve anything, quite the opposite. 

But im also quite Conservative

S

 

[JohnRoberts]

You can not remove the phantom DC blocking capacitors from the front end of mic preamps, unless you give up ever using phantom power after that.

For outputs one concern is interfacing with transformers, that don't not like DC offsets.

In reality customers judge books (consoles) by their cover (clicks, pops, scratchy pots, etc). The vast majority of coupling caps in consoles are for such cosmetic concerns.

The majority of customers do not like the incidental sounds from DC coupled paths.

Some hot rodders use expensive modern op amps with very good DC offset specifications to forgo at least some DC coupling caps.

When properly used DC coupling caps are innocuous.  My general advice is to test the audio paths in question. If there are some tired/faulty caps they will show up. If they measure good, they probably are good. If they measure bad replace them.

JR

 

[ln76d]

Thanks guys for help.

I know what DC offset can cause in case of switches and potentiometers 
Here's a different story.
Worst thing is that i can't check this in the circuit at the moment, so i'm trying to "build" theoretical concept of the console first.

It's a broadcast console (even two merged to one) where the original routing was totally not usable for studio use.
Each module had inputs and outputs wired to the krone patchbay and there was many options for inserting and switching modules in broadcast use in both -  engineer and presenter console.  It was a  type of instalation, which schould be configured once and never touched Many logic functions, fader start, cue, tone and ident etc. - things which were cool 30 years ago in the radiostations
I'm assuming that because most of the preamps were switchable by external bank of switches so there was need of output coupling.

Simplifying routing PRE -> EQ -> Active fader. Removing output coupling from preamp - dc offset with an audio signal goes to the eq input (with coupling - balancing stage - coupling - BYPASS SWITCH).
Then it goes thru the balancing output stage - another coupling remove and then it go to the active fader input which has coupling too.
Each electronic balancing stage is IC based.
I'm assuming that the input coupling of each another stage will null the dc offset so there wouldn't be any audible effects on the switching or potentiometers using.

So here's my question Are there any other (than pots and switches scratching) disadventages of running dc offset thru the cables between modules?
Do the resistor termination on each stage can affect in this case on something? Each output have 100k resistor between + and - and each input have 43K termination. For the first look values are weird

 

[JohnRoberts]

When you have a multiple channel console, you can mod one or two channels and answer your own questions.

I still prefer the measure twice and cut once approach,

Do you hear the extra caps? If they are driving high impedance terminations, nonlinear artifacts will be minimal.

JR

 

[ln76d]

You can not remove the phantom DC blocking capacitors from the front end of mic preamps, unless you give up ever using phantom power after that.

For outputs one concern is interfacing with transformers, that don't not like DC offsets.

In reality customers judge books (consoles) by their cover (clicks, pops, scratchy pots, etc). The vast majority of coupling caps in consoles are for such cosmetic concerns.

The majority of customers do not like the incidental sounds from DC coupled paths.

Some hot rodders use expensive modern op amps with very good DC offset specifications to forgo at least some DC coupling caps.

When properly used DC coupling caps are innocuous.  My general advice is to test the audio paths in question. If there are some tired/faulty caps they will show up. If they measure good, they probably are good. If they measure bad replace them.

JR

Hi John!

In the topic above i described the situation  (hope it's understandable - language barrier )
I didn't considering removing coupling on any input stage In fact on microphone pre is transformer input and the switching doesn't affect the line input at all but since i'll be using line inputs for mixing, there's need of coupling dc from soundcard interface.
I'm not such purist so of course there will be tons of coupling capacitors in the chain but there wouldn't be bad to make it more "lightweight"

I've read a lot about low dc offset opamps since we've "talked" here about Jim Williams
Idea is even interesting but here i will have not many options to work with that.
Unfortunately most (these which i've found) are bipolars and they wouldn't be easliy interchangable with fet opamps do the impedances etc.
In fact in console in the preamp patch there are only bipolar NE5534/32 but i don't want to change them because i like the sound.  In other preamp simply swapping NE for lm4562 wasn't my favorite configuration so i will stick with NE here too Other thing is that i have no option to even simulate most of the parts of the circuits because they have single 24V supply with many reference points etc. It's really hard to make it in any software to get proper results.
Some electrolytics will be changed for MKT, since there are some points with high enough impedances to put 3.3uF and 4.7uF foil caps with result of minor cut below the 10Hz .

 

[ln76d]

When you have a multiple channel console, you can mod one or two channels and answer your own questions.

I still prefer the measure twice and cut once approach,

Do you hear the extra caps? If they are driving high impedance terminations, nonlinear artifacts will be minimal.

JR

Of course that will be first step which i will make, but for some time i have no option to do that
That's why i've asked - out of courisity 
Much more am interested (do my poor education ) are there any other issues are possible? Affecting on signall, hum, interferncies etc.

 

[abbey road d enfer]

Other thing is that i have no option to even simulate most of the parts of the circuits because they have single 24V supply with many reference points etc. It's really hard to make it in any software to get proper results.

Weird! It is not significantly more difficult to simulate a single-supply circuit than a dual-rail. Just a few more components and extra care with naming the supply rails. Not enoug to make it unfeasible.

Some electrolytics will be changed for MKT, since there are some points with high enough impedances to put 3.3uF and 4.7uF foil caps with result of minor cut below the 10Hz .

Beware of the fact that the parasitic capacitance of a larger film cap may be a nuisance; unwanted coupling between nodes due to increased stray capacitance may lead to unstability and interference susceptibility.

 

[ln76d]

Weird! It is not significantly more difficult to simulate a single-supply circuit than a dual-rail. Just a few more components and extra care with naming the supply rails. Not enoug to make it unfeasible.

I've tried several times, especially in multisim. Different substitutions for reference IC, naming rails etc.
Never worked as it should be and for example this is one of the simplest circuits:

http://groupdiy.com/index.php?action=dlattach;topic=56753.0;attach=28745;image

Daughter board have 3 IC more (two single for gain trim on each channel, one double for stereo line to mono mix and phase reverse in mic section)

I gave up

Later "i'll be back" for stray capacitance topic

 

[ln76d]

Beware of the fact that the parasitic capacitance of a larger film cap may be a nuisance; unwanted coupling between nodes due to increased stray capacitance may lead to unstability and interference susceptibility.

I'm back

Did you mean "parasitic capacitance" or parasitic inductance due to "larger" size of capacitor?
If the "larger film cap" is about its physical dimensions and longer leads (as in many axial film caps from uF range) it shouldn't be any problem because i want to use small size radial MKT  (3.3uF and 4.7uF) caps, which aren't much bigger than, let say 100uF/63V Panasonic FC electrolytic. For any bigger sized capacitor there's no place anyway.
If you mean something else, please expand the topic. I'll be very grateful 

 

[abbey road d enfer]

Beware of the fact that the parasitic capacitance of a larger film cap may be a nuisance; unwanted coupling between nodes due to increased stray capacitance may lead to unstability and interference susceptibility.

I'm back

Did you mean "parasitic capacitance" or parasitic inductance due to "larger" size of capacitor?
If the "larger film cap" is about its physical dimensions and longer leads (as in many axial film caps from uF range) it shouldn't be any problem because i want to use small size radial MKT  (3.3uF and 4.7uF) caps, which aren't much bigger than, let say 100uF/63V Panasonic FC electrolytic. For any bigger sized capacitor there's no place anyway.
If you mean something else, please expand the topic. I'll be very grateful 

I actually mean stray capacitance between the external armature of the cap and the other parts of the circuit.
The MKT series is polyester, which comes as the worst choice for film caps, only slightly better than electrolytics.
Check the articles by Walt Jung and also Cyril bateman.

 

[ln76d]

I actually mean stray capacitance between the external armature of the cap and the other parts of the circuit.
The MKT series is polyester, which comes as the worst choice for film caps, only slightly better than electrolytics.
Check the articles by Walt Jung and also Cyril bateman.

I've read both some time ago, i refreshed some today
According to the Bateman articles:

#5
"From these 1 volt tests the best 1 mF electrolytic, the Bi-polar type, was clearly beaten by the good metallised PET"

#6
"10mF choice
We have three possibilities. A double Bi-polar using two 22mF 50/63 volt Bi-polar electrolytics, a 10 mF metallised PET or an assembly of three 3.3mF PPS capacitors.
The lowest cost solution for use with signal voltages less than 1 volt and no significant bias, is a double Bi-polar series pair.
A 10 mF MMK metallised PET takes the same PCB area and distorts less with DC bias.
The PPS capacitor assembly ensures lower distortion, especially when used with increased AC signals or DC bias voltage.
However it occupies more board area and is expensive.
An assembly of Polypropylene capacitors, as used in the DC bias network, would provide the lowest possible distortion but
requires a five times larger board area and is most expensive.
For small AC signals and modest DC bias, I choose the 10mF MMK metallised PET capacitor"

Interesting article is also here:
http://sound.westhost.com/articles/capacitors.htm

Of course in my equipment are different conditions, change from much higher value electrolytic to the lower MKT etc.
Tests should give me some light
In microphone circuits for example: Audix CX112 and AKG C414, change of output tantalum for identical value MKT gave realy audible difference which i didn't suspect at all.
In cheap, moded,  four channel preamp also. 10uF electrolytic vs 4.7uF MKT.
My friend did blind test, with two channels with MKT vs two with electrolytics (which one have "premium" e-cap inside).
He choosed both with MKT inside as better and both with electrolytics as sounded worse and the same.
I did some other recapping on different gear, in some were audible effects in some completely not

Of course it would be better to use polypropylene, but here "size matters"
What do you think about parallelling polyester cap with lower value polypropylene?
I didn't  notice improvement in bypassing electrolytic with film cap, but maybe i haven't done too much configurations like that.

 

[ln76d]

Other thing is that i have no option to even simulate most of the parts of the circuits because they have single 24V supply with many reference points etc. It's really hard to make it in any software to get proper results.

Weird! It is not significantly more difficult to simulate a single-supply circuit than a dual-rail. Just a few more components and extra care with naming the supply rails. Not enoug to make it unfeasible.

Thanks abbey!

You gave me a "kick" to get back to circuit simulation
After another several battles with software i realized where was the fault.
None of the 5V reference voltage circuits wasn't appropriate for REF50Z.
I substituted it by 5V Zener and it works ok. Shows proper AC analization and the voltage gain
Maybe this info would useful for someone