Bourns T-pad as output attenuator (Hairball, ClassicAPI, etc)

[soapfoot]

Hello.

Bourns makes a 600 ohm T-attenuator that is sold for about $30 by Hairball Audio, Classic API, and perhaps others.

The Classic API website says

This potentiometer can be used to attenuate the input or output of a balanced line level device.

However, it has been my understanding that a T-pad is an unbalanced attenuator, and that I'd need a (larger, more complex) H-pad to maintain both constant impedance and balanced output.

An older post from PRR on this site (from 2005) states:

If you keep the wiring tight and inside a metal box, you CAN use an "unbalanced" pad in balanced 600Ω lines. We old-timers did it all the old-time. It is "floating", the only un-balance is stray capacitance, and with a tight layout it is not a problem in 600Ω lines.

(Reply #11 here: http://www.groupdiy.com/index.php?topic=4825.msg71831#msg71831)

This sounds appealing, and like it would be more than adequate for my purposes. I was just wondering if anyone could assist me with understanding how the T-attenuator would be connected to use it in this manner. 

Hairball and Classic API both supply little PCB daughter boards that have "input +" "input -" "output +" and "output -" marked.  It's nice of them to attempt to make it convenient.

Is it just as simple as "input +" and "input -" connected to the output transformer (on a transformer-balanced device), and "output +" and "output -" connected to pins 2 and 3 of the XLR connector on the output?  Would this maintain the pseudo-balance (aside from stray capacitance) described by PRR above?

Thanks for any clarity anyone can provide!

 

[abbey road d enfer]

Is it just as simple as "input +" and "input -" connected to the output transformer (on a transformer-balanced device), and "output +" and "output -" connected to pins 2 and 3 of the XLR connector on the output?  Would this maintain the pseudo-balance (aside from stray capacitance) described by PRR above?

Yes and yes. It's true that an H-pad would have better CMRR at 1MHz, but who cares?

 

[soapfoot]

Thank you so much!

 

[weiss]

does anyone know if mouser sells them?

 

[qmp audio]

does anyone know if mouser sells them?

i could be wrong but i believe these are custom made for hairball/CAPI.

 

[weiss]

i could be wrong but i believe these are custom made for hairball/CAPI.

yeah seems like

 

[jsteiger]

does anyone know if mouser sells them?

i could be wrong but i believe these are custom made for hairball/CAPI.

Yes we have them custom made 

 

[squib]

My limited experience of these us that they have crappy gain vs rotation.  I found them annoying.

 

[alexc]

What do you mean  by 'crappy Gain vs - rotation'.

I've used a good few of these t-pads from Classic API and I haven't noted any excessive crappiness.

Do you mean unit to unit variations or are you saying they have a lumpy, uneven response ?

My last one was on the secondary of a hefty tube line amp stage.

Worked a treat. I fully appreciated having one of these capi units at hand.

Much less lumpy than the higher powered mallory ones.
(although they too are great when the higher power is needed)

 

[abbey road d enfer]

My limited experience of these us that they have crappy gain vs rotation.  I found them annoying.

Remember that the proper operation of a constant-Z pad relies upon proper input and output loading. Failure to adhere to this basic predicament leads to unproper operation.

 

[mjrippe]

Anyone tried the 600 ohm variable H pad shown below?  Or any knowledgeable criticism of the design?

 

[abbey road d enfer]

Anyone tried the 600 ohm variable H pad shown below?  Or any knowledgeable criticism of the design?

It is not constant-impedance, both at input and output, which may be a problem with some load-sensitive circuitry, or at the contrary  a non-issue.

 

[mjrippe]

Thanks, Abbey.  It seemed too simple to be the perfect solution, just couldn't put my finger on why.

 

[buildafriend]

There are many variations of pads. L,T, Pi, H and I think one or two more. The hairball and CAPI T ones both worked awesome for me.

A circuit where you might need an H pad might be something like a sta  level where the signal runs balanced all the way through. I think?

 

[abbey road d enfer]

There are many variations of pads. L,T, Pi, H and I think one or two more. The hairball and CAPI T ones both worked awesome for me.

A circuit where you might need an H pad might be something like a sta  level where the signal runs balanced all the way through. I think?

An H-pad is needed when the input and output impedances must be strictly constant. Typically all analog telephone ; since many audio and broadcast products were designed by telephone engineers, they continued to apply the same principles. Since the 70's, I don't think one piece of audio equipment has been designed with this constraint.

 

[PRR]

> H-pad is needed when the input and output impedances must be strictly constant.

There are several forms of "H".

The one cited above image is not constant impedance. Assuming double termination it runs from 1920 Ohms at minimum loss (which is around 10dB!) to 660 Ohms at maximum loss.

(Also even with audio-taper pot most of the action is below half-way.)

A true balanced and constant-Z "H" pad needs *5* pot elements. Which is why you never see them. There are 4-pot "O" pads which do as good for less cost.

True balanced may be needed on *long* lines with large capacitive balance to consider. In the studio you can (or should) always be able to use an un-balanced pad. If the pad is in-the-box it *really* should be fine un-balanced (no large capacitances to bad places). The "unbalanced H" is the "T". T-pads were THE way to go when amplification was expensive and all mixing and levels were done passively.

It is very hard to beat the single 1K pot ("L-pad") with 500 Ohms from wiper to load. Unbalanced but who is perfectly balanced? The impedance varies but well within the range that classic gear tolerates (and double-terminated interface is very tolerant).