Ideal impedance of in-line attenuator
Sep 28, 2005 at 1:39 PM Thread Starter Post #1 of 5

chillysalsa

Headphoneus Supremus
Joined
Sep 26, 2002
Posts
1,989
Likes
187
Just trying to make some interconnects that will drop my CD player output by 6 or 10 dB, since the amp gain (on the integrated amp for speakers) is pretty high and my speakers are sensitive.

So I figure I will just use a simple voltage divider:

v_divide2.gif


in which the interconnect will have a resistor R(in) in series with the hot/signal connection, and R(g) from the signal wire to the ground/shield.

So, R(in) + R(g) is the resistance of the interconnect (without knowing the capacitance we can't calculate the true impedance of it, right?). But still, what is this value ideally supposed to be???

My source is 2V output, and amp had input impedance of 100K.
 
Sep 28, 2005 at 5:45 PM Post #3 of 5
Thanks Rick...

But wouldn't that give me 15.56 dB of attenuation?

I was thinking of using values of R(in) = 47K ohms and R(g) = 27.5K ohms. Output would drop to 0.63V, and attenuation is 20 log (0.63/2) = 10 dB.

But, is that better than using lower values like 47 ohms and 22 ohms, which would present the CD player with an effective load of 69 ohms.

I mean, in terms of 'impedance matching' what I am I trying to match with what, if anything at all?

Found a quick little calculator for this:

http://www.sengpielaudio.com/calcula...agedivider.htm
 
Sep 28, 2005 at 6:36 PM Post #4 of 5
5:1 attenuation Rin/Rg with the RG setting load to the previous stage with the 10K still having a good 10-to-1 relationship to the following 100K intput stage.
Think of attenuators as "reverse gain" with the same basic formulas but also in reverse
wink.gif




BTW-you can scale this "resistor pair" downwards (lower RG allowing a lower Rin for the 5:1 ratio) for a lower impedance driving the next stage input but there are limits to just how low which is entirely dependant on the driving stage output impedance,also requiring at 5-to-1 at the minimum impedance ratio.
you could also use a variable pad (potentiometer) which would make the Rin "static" (no change) while the Rg element would be entirely dependant on postion.a better side of the pair for changes..
another option would be a DPDT switch that would have both a -5dB and -10dB setting simply by switching in a second set of identical resistors for the lower (-5) setting.
Parallel resistors are "Divide by 2" while series is "Multiply by 2" (or however many resistors in series/parallel and why digital control attenuators can work with only a few resistors with value determined by what combo is switched in
cool.gif
)
 

Users who are viewing this thread

Back
Top