Impedance Matching Attenuators

[YouAre]

To follow up on this thread:
http://www.diystompboxes.com/smfforum/index.php?topic=114446.msg1063583#msg1063583

I wanted to make sure I'm buying the correct resistors before I start building my attenuator! To recap, I'm building a resistive attenuator for a mic preamp. Links for preamp schematic and attenuator design are below.

Schematic:
https://seventhcircleaudio.com/assets/N72/N72R33/docs/n72_sch-938c7a738d8d4a951255670cc4d2a9b3.pdf

Attenuator Designs:
http://www.uneeda-audio.com/pads/
http://diystrat.blogspot.com/2013/08/diy-attenuators.html

The last link indicates when using the U-pad (pictured below), we should match the Shunt Resistor (R2) to the impedance that the output is expecting to see.



In the case of the attenuator designed in the blog linked above, he's using Mic with a 250ohm output impedance. So he uses a 250ohm resistor for R2. The mic preamp/interface he's using is expecting to see a 250ohm device connected to the input.

By that logic, if the interface I'm plugging the mic preamp interface into is expecting to see a 600ohm device (typical output impedance of analog Mic Pre's), we should use a 600ohm (or something close to it) shunt resistor, right?

Something doesn't sit with me on that though...
Let's say we're plugging preamp (600ohm output impedance) into the attenuator into the interface (expecting to see a 600ohm output impedance on the gear plugging into it) and we want a 10dB cut.

By my calculations, we want a 680ohm resistor for R1/2 if we're using a 604ohm for R2 (closest standard value to 600ohms). So what's that doing to our overall circuit?  We get:

600R (mic pre ouput) in parallel with (680R + 604R +680R) = ~459 ohms total for our attenuator tacked onto our mic pre output.

So does it still make sense to use 604 for the shunt resistor? To maintain the output impedance of the preamp, would we want to tack on a higher resistance? For example:

600R in parallel with 100K = 596R.

Based on these numbers, I'm inclined to think that it makes more sense to use higher value resistors.

What are your thoughts?

Thank you ahead of time for the help!

[PRR]

PLAGIARIZE PLAGIARIZE PLAGIARIZE ! ! !

A billion mike pads have been done and many of them published. Look on the front-end of mike preamps and console inputs.

The mike wants to see over 1K, better 2K.

Lower changes mike sound. A few mikes sound "better" when loaded, but the most common convention is light loading.

The mike preamp input really needs to see around 200 Ohms. Much higher and current hiss dominates. Much lower turns out to be unlikely for reasonable attenuation.

Using the appropriate U-pad, and avoiding hard math, this leads to a minimum loss pad of 1K-200r-1K. Nearest Handy Value to 1K-220-1K. Assume the mike preamp is 2K actual impedance. So "220" acts more like 200 (220||2K). Then 1K-200-1K makes 200/2200 loss or 0.091 or -20.8dB loss.

Getting lower loss forces a compromise on impedance somewhere. 500-200-500 approaches 15dB loss but gets close to 1K to the mike.

Higher loss is trivial. Lower loss happiness depends on how the mike and preamp feel about unconventional impedances.

[YouAre]

PLAGIARIZE PLAGIARIZE PLAGIARIZE ! ! !

I taught my daughters, Sasha and Malia, better than that!

I'll see myself out....

A billion mike pads have been done and many of them published. Look on the front-end of mike preamps and console inputs.

Of course. That's why I included the website that introduced the concept to me, and also the blog of someone that used/referenced that design. To the point of what you mentioned later, there are compromises with this design and I really want to understand what those compromises are so I can decide which ones I can live with.

Apologies ahead of time if I get a little too granular in my analysis of your post, but I really want to make sure I'm understanding everything correctly.

The mike wants to see over 1K, better 2K.

Lower changes mike sound. A few mikes sound "better" when loaded, but the most common convention is light loading.

Ok, so in you're example we're using a Mic into a preamp while my case is using a mic preamp into the interface. That's totally fine, the concepts should still apply.

So here we're seeing that the typical input impedance for a mic preamp is ~2K since the mic preamps are presumably designed to make the Mic happy.

The mike preamp input really needs to see around 200 Ohms. Much higher and current hiss dominates. Much lower turns out to be unlikely for reasonable attenuation.

So our mic's output impedance is 200R in this case, right?

Using the appropriate U-pad, and avoiding hard math, this leads to a minimum loss pad of 1K-200r-1K. Nearest Handy Value to 1K-220-1K.

I'd like to differentiate the values we're using as a given based on the gear we're using and which we're calculating. I think it's an important distinction because it will dictate how and where we can deviate for other attenuation values. Again, thank you for bearing with my attempts to be meticulous in my understanding. 

I'm guessing the 1K comes from taking the 2k impedance that the Mic is expecting to see (and not the 2K actual input impedance of the mic preamp.), and dividing that by 2 since it's a balanced signal.

Also just want to confirm that the 200R you've chosen is based on what the mic preamp is expecting to see and not the actual output impedance of the mic.

Assume the mike preamp is 2K actual impedance. So "220" acts more like 200 (220||2K). Then 1K-200-1K makes 200/2200 loss or 0.091 or -20.8dB loss.

Got it, parallel the actual input impedance with the shunt resistor. Didn't think about that before!

Getting lower loss forces a compromise on impedance somewhere. 500-200-500 approaches 15dB loss but gets close to 1K to the mike.

Here's where it gets interesting! So before we were using values based purely on the respective input and output impedances.

Now can we split the difference here? Instead of 500-200-500, dropping only the series resistance, can we lower the series resistors by a smaller amount, and raise up the shunt resistance to compensate?


Say we do 750-300-750. Same 15dB drop, but now the mic is seeing 1.5K and the preamp is seeing 300R.
This way we're spreading the pain a bit of who needs to compensate, the mic or the pre. Is there a benefit to this?

Higher loss is trivial. Lower loss happiness depends on how the mike and preamp feel about unconventional impedances.

Is it trivial because we can raise the series resistance without any ill effects on the mic?



Not to be tedious and I appreciate your tolerance if I am, if we're putting this attenuator between the output of a mic preamp (600ohm output impedance) and the input to a digital console (10K? Page 26 of this document, Line input, http://media.uaudio.com/support/downloads/Apollo_Twin_Hardware_Manual.pdf)...

Do we want to start with 5K 600R 5K (let's ignore standard values for a moment, we'll calculate that later and accept the consequences).

So really, the mic preamp will see and impedance of....5K (R Series 1) + 5K  (R Series 2) + 600R || 10K (R shunt in paralel with the interface input impedance), right? I calculate 10,566K.

The interface only ever sees that 600R shunt resistance, right? Or do we need to parallel 600R || 5k+5K?


Again, thank you for your time and your effort on this thread and the other thread. You've been a huge help, and I really appreciate your patience. Thank you.

[PRR]

> my case is using a mic preamp into the interface.

So I misunderstood. Sorry.

The interface is probably 10K, and not hiss-critical.

The preamp can drive 600 Ohms.

*Some* vintage-like preamps work best very-near 600 Ohm load. I now recall this was a question in the thread you have left behind.

O-pad. 1K across preamp output ("near" 600r). Two 1K or 2K series resistors. A 10K audio-taper pot across the interface input, wired as a rheostat so it gives 10K full-up, ~~1K half-dial, zero full-down.

Do math. The preamp never sees more than 933r or less than 667r. (Maybe the "1K" could be 10%-20% lower to get nearer 500-700r total.) Full-up attenuation is small, half-dial is around 10dB, going to infinite at full-down.

If you like switches, do a pot and measure "nice" resistance settings, do it on your switch.

[YouAre]

> my case is using a mic preamp into the interface.

So I misunderstood. Sorry.

No no, my fault for being unclear.

The interface is probably 10K, and not hiss-critical.

The preamp can drive 600 Ohms.

*Some* vintage-like preamps work best very-near 600 Ohm load. I now recall this was a question in the thread you have left behind.

Cool, so we've got some wiggle room as I'd anticipated.

O-pad. 1K across preamp output ("near" 600r). Two 1K or 2K series resistors. A 10K audio-taper pot across the interface input, wired as a rheostat so it gives 10K full-up, ~~1K half-dial, zero full-down.

Interesting! Good way to keep the preamp happy, and we don't have to worry as much about the resistance of the attenuating portion of the O-pad affecting the load that the preamp wants to see.

Also, changing the single shunt resistor as opposed to the two series resistors does make things a little easier!

Do math.

::busts out calculator::

The preamp never sees more than 933r or less than 667r.

ok, so 933R = 1K || (2K + 10K + 2K), right?

So dropping that 10K resistance of the output shunt resistance to 0, I get a value of 800R.

Can you please clarify how you got the 667?

(Maybe the "1K" could be 10%-20% lower to get nearer 500-700r total.) Full-up attenuation is small, half-dial is around 10dB, going to infinite at full-down.

If you like switches, do a pot and measure "nice" resistance settings, do it on your switch.

Definitely, this is going to go on a switch for certain.

Just to make sure that I'm doing the gain reduction calculations properly...is there a difference between calculating the attenuation of a U-pad vs an O-pad, or does the extra resistor on the O-pad only help to keep the input device happy?



As an aside...I'm debating switching in caps in parallel to the series resistor, to mimic the caps we install on guitar volume pots. I'm assuming that even with our approximate maintenance of the impedance, utilizing the O-pad in lieu of the U-pad, that we may still encounter some high end loss. Not sure if these is true...but It's easy enough to switch in some caps to roll off some bass response to make it feel like we're brightening things up.

Again, thank you for all the help! I really appreciate it!