Input attenuator help

I have to put my hand out for some unashamed spoon feeding, my learning curve on this is brutal, I know this must be simple for someone out there-

I have a few things Im racking right now and instead of using the standard -20dB pad I think I'd like to try a balanced input attenuator before the mic input transformer.

I'd like to build a suitable series attenuator to do 4 dB steps down to -20dB into 150 ohms.

Im trying not to get too spendy right now, I looked at the alpha switches available from mouser and cant at all understand how the F they have them categorized, which shorting switch will work for what I want? I want to do a six position switch, 5 pad sections and one bypassed.

This seems like a standard thing we could probably all use on any mic preamp with a 150 ohm input transformer, right? Might be good to store this in a meta somewhere. There could be a racking meta, everything you need to rack a box, links to pahntom powering, swithcing, this kind of stuff.

thanks in advance!

dave

I built a mic pre front end utility box prototype a few years ago that had this very thing you describe. Phantom, variable pad, phase, line level attenuator. I called it "The Sonic Janitor" Very handy.

Here is a very simple way to do it. I used Elma 2 pole 6 position shorting switches but any decent 2p6p shorting will do the job.

http://www.ethanwiner.com/gadgets.html

At the bottom of the page is a simple windows calculator for the R values.

It uses 2 fixed resistors on the balanced legs with a switchable shunt resistor across these legs. This is essentially the same as the shure switchable in-line pad.

Helsing

thanks for the link, hadnt looked at that in a while. sadly that does little to help because Im a retard. Im also on a mac, grrr...

dave

In the "R*DD" threads, Mr. O'Boogie posted a nice design for a mic input attenuator, 6dB steps, 0 to -30dB.

It's not exactly what you're looking for, I know, but it would do the job if you can tolerate 6dB steps. Otherwise, it's time to break out the calculator... The basic configuration needed is a 2K:200 "U" pad (balanced "L").

You need a switch of at least three poles, make-before-break (shorting).

When the subject of mic pads comes up, I usually recommend a look at my friend Rick Chinn's page:
http://www.uneeda-audio.com/pads/

did anyone save the pdf for the input pad in the redd thread? I tried to dl it but its corrupt.

dave

Dave,

Send me the pad steps you want along with your email address and I can shoot you the values and a scheme. Its really easy.

Helsing

thanks, thats awesome. My greatest thing with elctronics will always be switches, getting my brain around them kills me sometimes. I sucked at geometry in school but was good at calculus, go figure...

Id like to build somethign that will work on a 2 deck switch, mouser has a 2 deck 7 position shorting switch (alpha 10wr027) for 1.57 each. big difference between that and the elma, right...

How about this, 1 step no atten, then 5dB per step for 5 steps to feed a mic input transformer.

would the resistor values need to change to use the same switch in a line level application? I imagine so, right?

dave

mouser alpha 10WR046 should work for the redd you only use 3 of the poles.

I would install the resistors to a piece of perf board and run wires to switch so when the switch fails you can change it without hurting the resistors with more heat also if you find a nice switch you can just move the wires.

The equations at the bottom of this page should help too:

http://www.tkk.fi/Misc/Electronics/circuits/line_to_mic.html

JC

thank you for all the help everyone!

dave

Dave;
Your Computer is a piece of fruit?

Wow!

well it helped Dave but not me, i have still to find any kind of information on a stepped attenuator. all links go to simple single attenuators but make no mention of any kind to if the shunt R needs to change or not, how to figure that out etc..

To maintain the input and output impedances at their design values, the answer is no. You must change the shunt resistor along with the series resistors.

The link I posted earlier gives the equations for a U-Pad.

www.uneeda-audio.com/pads

I'm afraid there are no shortcuts. :wink:

well I'm not in it for shortcuts, I just haven't seen the answers to my questions.. since I want it switchable I assume that the series Rs will be, well, in series.. so do you sum the series resistors before the resistor that you are working on into the equation also or do you repeat the same attenuator values over and over again?(or why not?) why do I see the shunt value different on each step of the attenuator? why don't we use just one shunt and add up series resistors to get what we are after? Why can't something like your 600R stepped atten. be adjusted for use with MIC inputs?

There are a lot more questions that I have that I simply cannot find the answers to. A lot of people are very helpful in handing out answers but this is not what I want, I would like, simply, an elementary starting point that progresses to detailed explainations, otherwise I have no idea why or what I am doing or if it is even correct for that matter.

Sorry for the rant, I am finally getting over the flu but still feeling strange from all this cold medicine!

[quote author="Svart"]"why do I see the shunt value different on each step of the attenuator? why don't we use just one shunt and add up series resistors to get what we are after?"[/quote]

Because the attenuator is supposed to simply attenuate without changing the source and load impedances of the mic/preamp interface. Using a fixed shunt and then adding up series resistors will result in changes in both input and output impedances.

Unless we want to custom-design attenuators for each mic/preamp combination, and include reactive elements in the attenuators--in other words, deal with a massive pain in the ass--we have to make a couple of pretty big simplifications. First, we assume that mics and preamps truly possess their "nominal" source and load impedances (they don't), and we also assume that these impedances are purely resistive (they aren't).

A U-Pad is simply a balanced case of an L-Pad. Personally, I find it easier to understand attenuators when I think of them in their unbalanced versions.



First, imagine the circuit without the L-Pad. We have a mic with a nominal Rsource of 200 ohms, and a preamp with a nominal Rload of 2Kohms. Suppose we want to attenuate by 20dB, but we want to ensure that the mic continues to see 2K, and the preamp continues to see 200 ohms, in order to avoid skewing the frequency response away from "normal conditions."

If we use Rseries=2K and Rshunt=220 ohms, the mic sees:
Rseries+(Rshunt||Rload) = 2K+(220||2000)=2K+198=2198.
...and the preamp sees:
Rshunt||(Rseries+Rsource)=220||(2000+200)=220||2200=200.
The output voltage is:
(Rshunt||Rload)/([Rshunt||Rload]+Rseries)=198/2198=0.09
...which is down 20.9dB from the input voltage.

All well and good. But then, what if you want to simply double the value of Rseries to add 6dB more attenuation, without changing Rshunt? The mic sees 4198 ohms--quite a bit higher than "normal"--while the preamp sees 209 ohms, which is "normal." The mic will not behave quite the same into a 4K load as it would into 2K. You may or may not hear a difference, but we're introducing another variable which we really don't want. The situation gets worse at the mic end as you further increase Rseries, for greater attenuation, and worse at both ends as you decrease Rseries for less attenuation.

The situation is even poorer when you're starting with a lower value of attenuation--say, 6 or 10dB--and trying to go up from there by simply increasing Rseries while Rshunt remains fixed. Crunch the numbers and you'll see that the impedances shift significantly.

Multi-value pads with a fixed shunt have been built, but it's just not the right way to do it--unless, perhaps, it's only a couple or three steps that are just a couple of dB apart. In that case, the "fudge factor" is likely within acceptable limits; but the attenuator wouldn't be very useful, anyway.

"Why can't something like your 600R stepped atten. be adjusted for use with MIC inputs?"

Click to expand...

Apples and oranges. The Bridged-T is an attenuator for symmetrical impedances (600 ohms in/out in this case). A mic attenuator needs to work between unequal impedances.

If you want something ready-made, this was contributed by Winston O'B before he disappeared. Luckily, I had the foresight to save a copy. This shows a proper implementation of a stepped mic attenuator.

PDF

I'm grateful for the info, this is more info than I could find though a lot of searching.

I assume that all this applies to output attenuators also, most likely for 600r outputs to sandstate inputs (10k+? or what would you assume the optimum inpedence for sandstate is?).

hmm I am looking at all this and finding one more question to be answered. i see a lot of different pads for the same impedences and db drop but they are drastically different values. so where do we come up with the initial resistor values to figure the rest? do we just pick a number and work from there? this doesn't seem very solid.

also looking at WOB's pad, which is exactly what i needed, I am trying to use this to figure all this out rather than copy directly, and I don't understand why the 3k3 is paralleled with the 220r.. and another question, do the shunt R actually "connect" between the series R? his diagram is a little confusing.

[quote author="NewYorkDave"]If you want something ready-made, this was contributed by Winston O'B before he disappeared. Luckily, I had the foresight to save a copy. This shows a proper implementation of a stepped mic attenuator.

PDF[/quote]

Must admit it's not completely clear to me either how this circuit keeps presenting about a constant source-impedance to the TX and about a constant load-impedance to the mic.

Stated otherwise: I can understand how it's done with such a switch (the topology), but the values seem odd
- unless the WoB-circuit wasn't meant for mics & keeping constant impedances for the various att-settings.

Thanks,

Peter

[quote author="soundguy"]thanks, thats awesome. My greatest thing with elctronics will always be switches, getting my brain around them kills me sometimes. I sucked at geometry in school but was good at calculus, go figure...
[/quote]

Heh.. don't feel bad. I can't even translate switch schematics into how they're wired on the switch. Like the schematic in this thread of W. O'boogies swith... I have nooooo idea how that relates to the round things that click. :shock: I only got the idea with a DPDT switch by looking at one built and the scematic at the same time. :sad:

Kiira

[quote author="kiira"]Heh.. don't feel bad. I can't even translate switch schematics into how they're wired on the switch. Like the schematic in this thread of W. O'boogies swith... I have nooooo idea how that relates to the round things that click. :shock: I only got the idea with a DPDT switch by looking at one built and the scematic at the same time. :sad:

Kiira[/quote]

FWIW, for IC-layouts as made with a computer we have tools that indicate which wire needs to go where - but I haven't heard of a workbench-equivalent of that yet :wink:

But don't panic, it's actually not that hard. Sewrch for & have a look at the gain-switch for the Green-pre for instance. It's a single pole, 12 throw and there were detailed pics. Could be a nice starting point for rotary switches.
Ah, here it is:
http://www.groupdiy.com/index.php?topic=6242&highlight=green+gain+switch
Two or three wafers is then just more of the same.

Bye,

Peter