DIY Attenuator's speaker out sounds perfect, line out loses high end

[nashville]

Hello! I built a simple passive attenuator box and I'm running into an issue.

Goal: -20dB speaker output and a direct line output

Method: high wattage 3.6 ohm resistance in series with speaker out and 0.44 ohm resistance in parallel (to keep the load the amp sees at 4 ohms when connected to the attenuator into a 4 ohm cab), then tap the input through 11kOhm resistance to a 10kOhm pot, then to the line output

Result: when recorded and level-matched in Pro Tools, the speaker output works perfectly, sounds just like the full power clips. But when the Line Out is recorded and put through a custom IR of the same cab/mic/position as was used in the full power clips, the line out+IR is much darker sounding and has less dynamic range.

Question: How am I losing high frequencies in my line out? I've done the same test with a Two Notes Torpedo Captor 4 Ohm and the Line Out+IR sounds identical to the full power amp+cab, so I know it can be done.

Picture of the circuit:
https://i.imgur.com/iJus9tg.jpg

Thanks!

[GibsonGM]

Hi Nash, is there a build document somewhere about this particular attenuator we could 'peruse'?  Or did you home-brew it?  First thought is that the line out is presenting too high of an output impedance to the interface, causing tone sucking.   


I've been up all night dealing with -28F/40mph winds and keeping my house from freezing, so my head isn't sharp enough to see the issue from the build pic...a schematic might tell a lot more!

[Rob Strand]

At first glance the pot wiring looks wrong to me.

[FSFX]

Hello! I built a simple passive attenuator box and I'm running into an issue.

Try putting a 270p or 330p capacitor across the 11k resistor. That should help compensate for any capacitive loading on the output due to cables.

[nashville]

Hi Nash, is there a build document somewhere about this particular attenuator we could 'peruse'?  Or did you home-brew it?  First thought is that the line out is presenting too high of an output impedance to the interface, causing tone sucking.   


I've been up all night dealing with -28F/40mph winds and keeping my house from freezing, so my head isn't sharp enough to see the issue from the build pic...a schematic might tell a lot more!

Thanks for the input. It is homebrew, the idea came from trying out an L-Pad and hearing it sound more accurate than my Dr. Z Airbrake but the L-Pad started melting after 20 minutes of playing. It kept the impedance at 4 ohms, so I thought I'd get fixed resistors that do that at -20dB.

I thought it might be an impedance thing with the interface, so I put a Boss pedal between the line out and interface to see if that changed anything, and it did not. I tried low Z and hi Z inputs as well.

[nashville]

At first glance the pot wiring looks wrong to me.

Yeah, I accidentally wired the pot in reverse but it still functioned as a pot. I also tried the circuit with only fixed resistors on the line out and it still had the high end loss.

[Rob Strand]

Yeah, I accidentally wired the pot in reverse but it still functioned as a pot. I also tried the circuit with only fixed resistors on the line out and it still had the high end loss.

When you use a resistive dummy load it's normal to lose some highs.   On a tube amp (and most SS guitar amps) the signal at the amp output gets a "bright" boost and bass boost from amp output impedance interacting with the speaker voice-coil impedance.

You can see the effect in this patent.  Don't worry about the circuit in the patent, just look at the response plots.
https://patents.google.com/patent/US5467400A/en

If you use same pot connection you have now with a speaker load the highs should return.  If not the problem is something else.

You can add an RC network before the pot that *approximates* the bright boost.   

Another method is to use a more complex dummy load.  The components R1, R2 and L1 (1mH) give the bright boost from the voice-coil inductance.   At the end of the day this is also an approximation.

https://www.aikenamps.com/index.php/designing-a-reactive-speaker-load-emulator

The components L1 (50mH), C1, R3 give the peaking bass boost from the speaker resonance.