New attempt to bring order to attenuators

[wkbdgeorge]

So I was thinking and reading and dong more reading and what seems to lack in the attenuator info world is real world applications to attenuator type. I think this might clear the air.

Let's create a list of the different types of attenuators and say what you successfully used it for.

 

[shabtek]

Really just a series of pads on a multi-gang switch--or a custom pot rough-equivalent.
Use 600 ohm pads on line-level ins or outs of gear that is happy with 600 ohm world ...
keep reading

http://en.wikipedia.org/wiki/Attenuator_(electronics)
http://www.m0ukd.com/Calculators/Attenuator_Calculator/T_Attenuator_Resistor_Calculator.php
edit:  http://homepages.tcp.co.uk/~nroberts/atten.html

 

[YLab]

Usefull link for different types of pad: http://www.uneeda-audio.com/pads/

Use 600 ohm pads on line-level ins or outs of gear that is happy with 600 ohm world ...

Now thats the point that is not clear for me. My thought is that 600 ohm may be too low for some modern gear,,i really hear a change of tone when the output of some of my preamp are loaded with 600 ohm. At this time im using a u-pad made of 680-150-680 (the preamp see the total of this.. 1.5K) So my question would be: how to determine the best ''impedance'' for a pad. I've read somewhere that it should kinda match the input impedance of the sound card? Does that make any sense? 

Hubert

 

[shabtek]

apple C apple V apple C apple V apple C apple V apple C apple V apple C apple V apple C apple V apple C apple V apple C apple V apple C
http://www.rane.com/note124.html
matching

(from wiki)
Comes from:Telephone systems
Telephone systems also use matched impedances to minimise echoes on long distance lines. This is related to transmission lines theory. Matching also enables the telephone hybrid coil (2 to 4 wire conversion) to operate correctly. As the signals are sent and received on the same two-wire circuit to the central office (or exchange), cancellation is necessary at the telephone earpiece so that excessive sidetone is not heard. All devices used in telephone signal paths are generally dependent on using matched cable, source and load impedances. In the local loop, the impedance chosen is 600 ohm (nominal). Terminating networks are installed at the exchange to try to offer the best match to their subscriber lines. Each country has its own standard for these networks but they are all designed to approximate to about 600 ohms over the voice frequency band.

and is how old gear was designed

bridging
(from wiki)
In electronics, especially audio and sound recording, a high impedance bridging, voltage bridging, or simply bridging connection is one which maximizes transfer of a voltage signal to the load. The other typical configuration is an impedance matching connection, which maximizes power delivered to the load.

When the output (output impedance) Zout of a device (the source, ZS in illustration) is connected to the input (input impedance) Zin of another device (the load, ZL in the illustration), it is a bridging connection if the second device appreciably loads the previous device.

Essentially no power is transferred. The second device is sensitive to the output voltage of the first device, and this is maximized when loading is minimized. All good audio connections are bridging, even the connection between the power amplifier and the loudspeaker, since it is more important to accurately control the loudspeaker than to drive it with the maximum amount of power available. (Maximizing power transfer is more important in applications such as driving electric motors.)
In systems involving very long lines (such as telephone or cable TV systems), the source, line, and load impedances must be matched to prevent reflections of the signal at the ends of the line from causing reflections and echoes. This is known as terminating the line.
A connection is commonly said to be bridged if the load impedance is at least ten times the source impedance.

 

[YLab]

Thanks shabtek! great answer and links. not sure i understand everything perfectly but i've learn sommething for sure.

Still questionning myself on how to match those though ;D

let say that the the imp of the input of my sound card is 24k. without the pad the preamp see 24k.
now put the 600ohm pad in paralele with that.  24k/600 in paralele give an imp of 585. now thats a big difference.
hmm dont know what to think about that! But my hears still ear a drop in the hi end with the pad!

Hubert

 

[shabtek]

what is your preamp?
trust your ears
600 ohm pad is 'matching' the (lo-z) line level output of the pre
24k ohm card input is 'bridging' the pad (or the lo-z out with no pad)

 

[lassoharp]

Usefull link for different types of pad: http://www.uneeda-audio.com/pads/

Use 600 ohm pads on line-level ins or outs of gear that is happy with 600 ohm world ...

Now thats the point that is not clear for me. My thought is that 600 ohm may be too low for some modern gear,,i really hear a change of tone when the output of some of my preamp are loaded with 600 ohm. At this time im using a u-pad made of 680-150-680 (the preamp see the total of this.. 1.5K) So my question would be: how to determine the best ''impedance'' for a pad. I've read somewhere that it should kinda match the input impedance of the sound card? Does that make any sense? 

Hubert

Hubert,

I don't think it's exactly clear how you are loading the output of your preamp with a 600ohm load. With another piece of gear that has 600ohm input or resistively with a pad?  The U-pad you mentioned is most commonly used as a mic pad.  It matches - approximately- the mics output impedance of ~150 ohms so that the mic pre's input sees a ~150ohm source. The mics output sees the preamps input impedance which is usually on the order of 10 times the mics output imp. or ~ 1.5K or higher.

So the pad's impedance should match the source impedance - in this case a mic - not the soundcard's 24K input impedance. If the source impedance is a 600ohm output then your pad should usually be built to match 600 ohms.  It may help to think of this situation loosely as - the mics output imp is 'matched' to the soundcard's input imp via bridging condition.  The pad's impedance is matching the mic's(or other source) imp via true matching condition - For all the reasons you read in shabteks posts.

For the case of the tone changing with 600ohm loading, it would be helpful if you list the exact pieces of gear that you observed this in as there are a number of things that could be happening.

 

[wkbdgeorge]

Hubert,  What kind of resistors are you using?  If they were an inductive type, correct me if im wrong, but wouldnt they affect frequency response?  -George

 

[YLab]

So the pad's impedance should match the source impedance - in this case a mic - not the soundcard's 24K input impedance. If the source impedance is a 600ohm output then your pad should usually be built to match 600 ohms.  It may help to think of this situation loosely as - the mics output imp is 'matched' to the soundcard's input imp via bridging condition.  The pad's impedance is matching the mic's(or other source) imp via true matching condition - For all the reasons you read in shabteks posts.

Ok, now i understand, i was kind of seeing it the other way arround. :

Hubert,  What kind of resistors are you using?  If they were an inductive type, correct me if im wrong, but wouldnt they affect frequency response?  -George

Some 1/2w carbon comp resistors. Is there a big difference with metal film in that application?
Sorry by the way for the  ''thread hijacking''.

Here's is a schemo from the rane web site. So input on the right and output on the left. Let say that i want a 600ohm 20 db pad. How do we caculate the right values? I think that it will help me understand!

 

[abbey road d enfer]

Hubert,  What kind of resistors are you using?  If they were an inductive type, correct me if im wrong, but wouldnt they affect frequency response?  -George

Doesn't matter. Even wire-wound resistors, which are an actual coil of wire around a core, do not exhibit enough parasitic inductance to create a noticeable error in the calculation.

 

[abbey road d enfer]

Here's is a schemo from the rane web site. So input on the right and output on the left. Let say that i want a 600ohm 20 db pad. How do we caculate the right values?

You have to understand that there's nothing like the right values. There is a range of answers which will all be true. That depends on what interpretation you give to "input impedance shall be higher than source impedance". The answers will be different for 5 times or 10 times or 100 times higher. Same with "output impedance shall be significantly lower than load impedance". Depends also on the precision you want. Do you want 20 +/-1 dB or 20 +/-0 .01 dB. You will find that in many cases, there is no solution if you ask for too much adherence to your design specs.
I can give an answer to your question without pencil, paper, calculator: 2x 900 ohms resistors, 1x 200 ohms. The actual results will depend on the source and load impedance, that are not specified. But I can guarantee you that they will be within 10% (provided the source is ca. 200R and load is about 2k): input Z 2k, output Z 180 ohms, attenuation 20dB.

 

[YLab]

Thanks for your answers abbey,
the pad will go between a 3124+ or a neve 1290 (reading 40ohm on my multimeter) and my lynx sound card (24k input imp) if that help!
Thanks
Hubert

 

[lassoharp]

So the pad's impedance should match the source impedance - in this case a mic - not the soundcard's 24K input impedance. If the source impedance is a 600ohm output then your pad should usually be built to match 600 ohms.  It may help to think of this situation loosely as - the mics output imp is 'matched' to the soundcard's input imp via bridging condition.  The pad's impedance is matching the mic's(or other source) imp via true matching condition - For all the reasons you read in shabteks posts.

Ok, now i understand, i was kind of seeing it the other way arround. :

Hubert,  What kind of resistors are you using?  If they were an inductive type, correct me if im wrong, but wouldnt they affect frequency response?  -George

Some 1/2w carbon comp resistors. Is there a big difference with metal film in that application?
Sorry by the way for the  ''thread hijacking''.

Here's is a schemo from the rane web site. So input on the right and output on the left. Let say that i want a 600ohm 20 db pad. How do we caculate the right values? I think that it will help me understand!

Here's some more resource material.  For T and H types:

http://www.nu9n.com/tpad-calculator.html

Note that you have to assign both input and output impedances - so 2 conditions: constant impedance and/or network matching. Use the one that fits your needs best.

For L and U type:

http://www.uneeda-audio.com/pads/      (gives steps for calculating the long way)

http://www.sengpielaudio.com/calculator-Lpad.htm  

 

[keefaz]

Thanks for your answers abbey,
the pad will go between a 3124+ or a neve 1290 (reading 40ohm on my multimeter) and my lynx sound card (24k input imp) if that help!
Thanks
Hubert

I think you are measuring DC resistance of both unit output transformers secondaries with your multimeter, this is not the output impedance.
For a line attenuator pad, I would use a 10k impedance in and out or a 600R impedance in and out, if you want to load the units outputs more heavily