DIY audio shunt attenuator?

[DrewTube]

My Eros -> EFP3 rig has too much gain.  Like ~40 dB too much gain (high output moving magnet cartridge).  Without additional attenuation, I have to put the preamp on the absolute lowest volume setting to even be a tolerable listening level.  I have a string of three -12dB Harrison Labs audio attenuators on the input jack of each channel of my power amp, which brings it down to much more manageable levels and lets me put the volume control in the "sweet spot" toward the middle of the dial.  However, it seems dumb to have 8 inches of attenuator barrels dangling off my power amp (and seems to be asking for me to break off the RCA jack with that giant lever).

I have an RCA connector barrel that will let me put my own resistors in there, so I figure I can make a single L-pad shunt attenuator with my needed ~40dB of attenuation.  There are many combos of series and shunt resistors that will accomplish this, so I guess the question is: what is the ideal impedance for my attenuator?  10K? 35K? 50K?  More?

From the EFP3 manual, it says the output impedance is 700 ohms and recommended output load is >10K ohms.  My solid state power amp has an input impedance of 18K ohms.  Connection is via unbalanced RCA interconnect.

[Paul Birkeland]

What input resistors do you have on the input on your EFP3 for the Eros?  Padding of the inputs is discussed in the manual on on page 24, and I'd be starting with that.

What is the gain of your power amp and the sensitivity of your speakers? (and the output of your cartridge)

Sometimes a little padding is extremely helpful, but you have -36dB, which is intense!

[DrewTube]

Solid state power amp is 29 dB voltage gain, 18K ohms input impedance.

Speaker sensitivity is 90 dB, 4 ohms impedance.

Phono cartridge output is rated at 6.8 mV - 7.2 mV.

EFP3 padding resistors are all at stock 33K.

I had originally planned on replacing those with much higher resistance values, but I vaguely recall seeing a paper from PJ suggesting that the best place to put the attenuation is as close to the power amp inputs as possible.  I guess that in my situation, with as much gain/voltage as I have, putting some of the padding at the preamp input would still leave plenty to drive the signal through the line stage and the interconnects - and might present a more ideal load to the Eros, to boot?

[Paul Joppa]

In your case, you should address the issue of signal level in the Foreplay as well as in the power amp. An unattenuated Eros, with a 7mV cartridge, will produce a nominal 2v output - but peaks on some recordings could be 5 times as great, which will overload a Foreplay. I'd suggest 10-20dB attenuation at the Foreplay input, and 20-30dB at the power amp input, to get the Foreplay signal in the sweet zone between noisy and overloaded.

You can try moving one or two of the Harrison Labs attenuators to the Foreplay inputs to see what sounds best, then implement the DIY attenuations to match.

[Paul Birkeland]

To translate PJ's recommendations into something with parts values:

Take out the pair of 33K resistors in the Foreplay on the Eros input and replace them with with 182K resistors (1/8W is fine).

To make a -20dB Harrison Labs style attenuator, you would want a 9.09K series element and a 1K shunt element.  If you get reasonably small resistors, it's possible to just build these into each plug on your interconnects that plug into the amp.  Some solid state amps will have RCA jacks that are wired with short wires to a PC board, and that's another place where it's pretty easy to add some padding resistors.

Moving the Harrison Labs attenuators around will give some idea of how things will sound, but they are 10K attenuators and will load down the Eros quite a bit more than the input of a Foreplay otherwise would.

[DrewTube]

Many thanks to both PB and PJ.  This gives me a much better idea of how to approach this with confidence.  I will go do some analysis and let you know how this turns out.

[DrewTube]

PB,

Any specific reason you recommend a 182K series pad resistor in the EFP3 input vs. 176K?  By my calculations, 176K would be an incremental -12dB, while 182K is -12.3 dB.  Obviously, that's really close and presumably an irrelevant differential, but it piqued my curiosity.

Thanks!

[Paul Birkeland]

I just went with what was in the manual.  If 176K is easier to obtain, that will work just fine. 

[DrewTube]

OK, so here's what I ended up implementing.  As a reminder, I needed ~40 dB of attenuation.

Per PB's suggestion, I swapped out the 33K input padding resistor in the EFP3 with a pair of 182K Vishay Dale CMF .1% 1/8W resistors.  This gives me ~12 dB of attenuation.

Next, I populated my RCA barrels with a 16.9K Vishay Dale CMF .1% 1/4W series resistor and a 698 ohm Vishay Dale CMF .1% 1/4W shunt resistor.  This achieves ~28 dB of attenuation, while still approximating the source impedance of 700 ohms (@ 698) and the load impedance of 18K (@ 17.6K) that would be there with the preamp directly connected to the power amp.

That -12 dB + -28 dB gives me my total of -40 dB and also reduces the chances of overloading the preamp.

So far, it's working great.  Volume controls on the EFP3 are right in the sweet spot and I only have one barrel per channel (vs. three) hanging off my power amp.  I might be imagining it, but I could swear the imaging is a little better and the mid-bass is punchier and better defined, too.

[Paul Joppa]

Excellent! The sweet spot is always the best balance between low noise and low distortion.