Fader/attenuator circuit

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

KasperNyhus

Well-known member
Joined
Jan 18, 2010
Messages
85
Location
Denmark
Hi!

I'm about to build myself a sort of custom mixer that fulfill all my needs...  ;D but I need some advise for the fader implementation.

Here is how I designed the project so far:

It will be based around the 51x format so various types of preamps and eq's can be used in any combination.

I call it The 51x Tracking Rack

51x%20tracking%20rack%20v4.jpg


There will also be an option to patch e.g. a compressor in from the two slots in the upper right corner (fx1 & fx2) and an external fx insert for each channel.

The routing panel could look something like this:

channel%20500%20rack%20v4.png


There is 4 aux outputs so two or more channels can be summed together before they hit the soundcard. I used the following circuit:
https://dl.dropboxusercontent.com/u/14107936/Tracking%20Rack%20tr%C3%A5d/SUM-12_schematic.pdf

The idea is that the signal can either pass the tracking rack without being affected by any IC's and stuff like that or it can be eq'ed, compressed, attenuated and summed together...

But the attenuation is the problem right now.
The whole thing is balanced - between every 500 module and through the entire routing. So I thought I could make a balanced attenuator ala this one:
http://classicapi.com/catalog/product_info.php?cPath=48&products_id=254 but Im someone told me that it fu**s with the signal in a bad way.
And what would be really cool is to have some real faders. But with the option of bypassing them.

The way I see it I need to "debalance?" the signal, attenuate it, and balance it again? Or?

Here is a diagram of the circuit components:
51x%20tracking%20rack%20sch.jpg


How do I attenuate the signal correctly? can I make a circuit based on the NE5532?

Does anybody have a fader circuit in mind?

Best,
Kasper
 
KasperNyhus said:
So I thought I could make a balanced attenuator ala this one:
http://classicapi.com/catalog/product_info.php?cPath=48&products_id=254 but Im someone told me that it fu**s with the signal in a bad way.
What does exactly this "someone" hold against a balanced passive attenuator? The only way this type of attenuator can f..k with the signal is when it's not correctly connected (possible causes: wrong impedance termination, unbal connection)
And what would be really cool is to have some real faders. But with the option of bypassing them.
The way I see it I need to "debalance?" the signal, attenuate it, and balance it again? Or?
Unless you get stereo faders, connected as balanced mono. P&G makes them
http://www.pennyandgiles.com/Linear-Manual-Fader-pd-58,3,27,.php
How do I attenuate the signal correctly?
An attenuator is an attenuator is an attenuator. Now you want to make sure the input and output impedances are adequate for the surrounding circuit. Using a buffer (0dB on top), or a non-inverting output amp (10-12dB in hand) is a good bet, but yes, you have to either make it balanced, or debal-attenuate-rebal.
  can I make a circuit based on the NE5532?
Or just about any decent audio opamp.
 
Just use a regular 1k fader. Connect input hot to the top of the fader and input cold to the bottom. Output cold is from the bottom of the fader and output hot is from the slider.

Cheers

Ian
 
abbey road d enfer said:
What does exactly this "someone" hold against a balanced passive attenuator? The only way this type of attenuator can f..k with the signal is when it's not correctly connected (possible causes: wrong impedance termination, unbal connection)

My local pusher :) he doesn't think a three deck pot can ever be precise enough and he recommends making a stepped attenuator.

Unless you get stereo faders, connected as balanced mono. P&G makes them
http://www.pennyandgiles.com/Linear-Manual-Fader-pd-58,3,27,.php
those look expensive - but very cool

An attenuator is an attenuator is an attenuator. Now you want to make sure the input and output impedances are adequate for the surrounding circuit. Using a buffer (0dB on top), or a non-inverting output amp (10-12dB in hand) is a good bet, but yes, you have to either make it balanced, or debal-attenuate-rebal.
I asume that the impedance for the circuit is 600ohm?

Or just about any decent audio opamp.

Is this correct?
fadercircuit.png
 
ruffrecords said:
Just use a regular 1k fader. Connect input hot to the top of the fader and input cold to the bottom. Output cold is from the bottom of the fader and output hot is from the slider.

Cheers

Ian

Will that keep the correct impedance?
 
KasperNyhus said:
abbey road d enfer said:
What does exactly this "someone" hold against a balanced passive attenuator? The only way this type of attenuator can f..k with the signal is when it's not correctly connected (possible causes: wrong impedance termination, unbal connection)

My local pusher :) he doesn't think a three deck pot can ever be precise enough and he recommends making a stepped attenuator.
Who cares about precision in a mixing system? Resolution is the word. What's the resolution of a stepped attenuator? 1dB in a restricted range, then 2, 3 or more dB. With a pot, you have almost infinite resolution, at least as good as your hand and better than your ears.
Unless you get stereo faders, connected as balanced mono. P&G makes them
http://www.pennyandgiles.com/Linear-Manual-Fader-pd-58,3,27,.php
those look expensive - but very cool

An attenuator is an attenuator is an attenuator. Now you want to make sure the input and output impedances are adequate for the surrounding circuit. Using a buffer (0dB on top), or a non-inverting output amp (10-12dB in hand) is a good bet, but yes, you have to either make it balanced, or debal-attenuate-rebal.
I asume that the impedance for the circuit is 600ohm?
No. The impedance is variable but again, who cares?
Or just about any decent audio opamp.

Is this correct?
fadercircuit.png
It should work.
 
From a quick glance it looks like you have the + and - inputs backwards on several of the opamps.

JR
Ahh yes, you are right. This is correct:
balfig12.gif


although all this back and forth debal - bal seams a lot of trouble for simple attenuation... hmm

For the resolution thing, I thought about making it 0,5db steps, and as I will never need to fully attenuate the signal in the summing it could work with a 24step.
 
KasperNyhus said:
ruffrecords said:
Just use a regular 1k fader. Connect input hot to the top of the fader and input cold to the bottom. Output cold is from the bottom of the fader and output hot is from the slider.

Cheers

Ian

Will that keep the correct impedance?

There is no 'correct' impedance.. Assuming the driving source is low (say 75 ohms or less) the 1K pot will look like an impedance of between 75 ohms and about 300 ohms. Provided you feed it to a 10K bridging input it will be fine. We used a very similar circuit a lot for monitor pots at Neve.

The probable fly in the ointment is driving buses which this circuit is really not suitable for.

Cheers

Ian
 
KasperNyhus said:
although all this back and forth debal - bal seams a lot of trouble for simple attenuation... hmm

For the resolution thing, I thought about making it 0,5db steps, and as I will never need to fully attenuate the signal in the summing it could work with a 24step.

I once designed a balanced input attenuator for a power amp, by using the pot wired as a rheostat in the middle leg of an H-pad, so 4 precision resistors and the pot. The only downside was that max kill was limited by the CMRR of the following differential stage.

JR
 
ruffrecords said:
Just use a regular 1k fader. Connect input hot to the top of the fader and input cold to the bottom. Output cold is from the bottom of the fader and output hot is from the slider.

Cheers

Ian

Like this?
pot.jpg
 
JohnRoberts said:
KasperNyhus said:
although all this back and forth debal - bal seams a lot of trouble for simple attenuation... hmm

For the resolution thing, I thought about making it 0,5db steps, and as I will never need to fully attenuate the signal in the summing it could work with a 24step.

I once designed a balanced input attenuator for a power amp, by using the pot wired as a rheostat in the middle leg of an H-pad, so 4 precision resistors and the pot. The only downside was that max kill was limited by the CMRR of the following differential stage.

JR
Like this?
HPAD.jpg
 
abbey road d enfer said:
KasperNyhus said:
I asume that the impedance for the circuit is 600ohm?
No. The impedance is variable but again, who cares?

So there is actually no need for the fancy 3-deck pot? Great! I can save some money there and use it for some actual good sounding electronics...
 
ruffrecords said:
KasperNyhus said:
ruffrecords said:
Just use a regular 1k fader. Connect input hot to the top of the fader and input cold to the bottom. Output cold is from the bottom of the fader and output hot is from the slider.

Cheers

Ian

Will that keep the correct impedance?

There is no 'correct' impedance.. Assuming the driving source is low (say 75 ohms or less) the 1K pot will look like an impedance of between 75 ohms and about 300 ohms. Provided you feed it to a 10K bridging input it will be fine. We used a very similar circuit a lot for monitor pots at Neve.

The probable fly in the ointment is driving buses which this circuit is really not suitable for.

Cheers

Ian

... but that is exactly what Im going to do? feed it to this circuit:
SUM-12_schematic.jpg
 
another thing.... In your EZ Tube Mixer thread, Ian, you point out that the fader needs to be placed before the eq? Why is that?

(sorry for all my noob questions – but I'm really learning a lot here thanks to you guys!)
 
KasperNyhus said:
ruffrecords said:
KasperNyhus said:
ruffrecords said:
Just use a regular 1k fader. Connect input hot to the top of the fader and input cold to the bottom. Output cold is from the bottom of the fader and output hot is from the slider.

Cheers

Ian

Will that keep the correct impedance?

There is no 'correct' impedance.. Assuming the driving source is low (say 75 ohms or less) the 1K pot will look like an impedance of between 75 ohms and about 300 ohms. Provided you feed it to a 10K bridging input it will be fine. We used a very similar circuit a lot for monitor pots at Neve.

The probable fly in the ointment is driving buses which this circuit is really not suitable for.

Cheers

Ian

... but that is exactly what Im going to do?

Then it depends on how many buses a fader can possibly feed. In your circuit the load per bus is 20K so you could certainly feed two buses from one fader. Four buses would look like 5K which I would say is probably the limit.

Cheers

Ian
 
KasperNyhus said:
another thing.... In your EZ Tube Mixer thread, Ian, you point out that the fader needs to be placed before the eq? Why is that?

(sorry for all my noob questions – but I'm really learning a lot here thanks to you guys!)

That is specific to the EZTubeMixer passive EQ circuits which need a relatively high impedance load - the 10K of a regular fader is to low for them so you need to place the fader before the EQ rather than after.

If you have an active EQ with op amps in it then the chances are it has a low output impedance so it will be quite happy driving a fader.

Cheers

Ian
 
abbey road d enfer said:
KasperNyhus said:
Like this?
HPAD.jpg
The problem with this arrangement is that you lose about 10dB, which may impair the actual noise performance.
Yes that is what I meant, and no I do not suggest this is the optimal solution.

For a line level application the S/N should not be a serious limitation. As I already mentioned the amount of kill when full down (due to CMR of following stage and pot minimum resistance) is more likely to be a qualitative concern. Another concern is the fader law.

If you are willing to throw a bunch of opamps at it, no big deal to interface to and from a standard grounded fader. 

JR
 
Back
Top