Speaker attenuator AKA Power Soak

Thread Starter

#12

Joined Nov 30, 2010
18,224
A musician wants to play his amplifier at rather high volume to get the sound of distortion as he likes it, but not annoy the neighbors. In math, that resembles changing 106db to less than 70db for a minimum range of 36db SPL. Retail power attenuators are designed to smoothly attenuate from maximum power to some lower level. This approach requires a high wattage rheostat. I designed to avoid the High Watt rheostat. This means the minimum attenuation is not a tiny part of a db. It is 18db (measured) to use a 5 watt potentiometer and 25db (measured) for a 1.5 watt audio taper pot. The largest wattage of audio taper pot that I could find is 1.5 watts, so the 5K 5W solution is a linear taper pot. The resistor to help a 5 watt pot survive is 80 ohms (82 ohms = standard size) 18W and the resistor required to help a 1.5 watt pot is 267 ohms (270 actual) 6 watts. My design contains a bypass switch to avoid changing the wiring every time you want to not have the attenuator in the circuit.

I know the math says -20db and -31db. The measured response does not match the math. Sorry, that's the tools I have to work with.
1582441919782.png
 
Last edited:

Thread Starter

#12

Joined Nov 30, 2010
18,224
So, how does it sound? I mean, is the distortion retained, just quieter?
That's the theory, as if using only resistors does not change the shape of the audio wave. The musician came to me because he has used attenuators. That says this kind of circuit can be done well. This circuit isolates the speaker inductance from the output transformer of the vacuum tube amplifier and there is a common practice of adding a bypass capacitor to enhance high frequency reproduction. I finished assembling it this morning. It works at all, but I have not examined the tone with my ears.

I can use the SPL meter as an input to my scope and maybe I will see some differences. Maybe ears are better. Maybe the speakers have a different frequency response when driven 40 db lower than usual. One of my problems right now is that my tube amp burned up its power transformer. Another problem is that the musician has much better ears than I have. I have done 3 or 4 original designs for him previously and haven't gotten any of them perfect on the first try.
 
Last edited:

Thread Starter

#12

Joined Nov 30, 2010
18,224
WB,
How are you measuring the results (SPL meter?)?
Radio Shack SPL meter bought decades ago. Tests done in the shop with no attempt to control acoustic reflections or place the meter in any "valid" measuring position. Simple goals. Do not exceed 36db as the minimum attenuation and have a lot of room below that in the potentiometer.
 

DickCappels

Joined Aug 21, 2008
10,152
Just be careful to use the same acoustic environment each time you perform a measurement. Maybe you don't have to wear exactly the same color shirt each time, but the furnishings in the room, the state of doors and windows, and the placement of the meter can affect the reading.

Is that really a 5k pot with its wiper connected to the speaker? Or does it go to a speaker amp's input?

I suppose you have kicked around the idea of just making a fuzz box so he can get the distortion without driving his amp hard, or would it be too difficult to obtain similar sounding distortion?
 

Thread Starter

#12

Joined Nov 30, 2010
18,224
Just be careful to use the same acoustic environment each time you perform a measurement. Maybe you don't have to wear exactly the same color shirt each time, but the furnishings in the room, the state of doors and windows, and the placement of the meter can affect the reading.

Is that really a 5k pot with its wiper connected to the speaker? Or does it go to a speaker amp's input?

I suppose you have kicked around the idea of just making a fuzz box so he can get the distortion without driving his amp hard, or would it be too difficult to obtain similar sounding distortion?
The acoustic environment is not very important and the db meter is not very important. That's why I didn't try to do anything precise with them. I am not writing legally binding specifications for retail sales. I am helping a friend get a cheap version of what normally sells for $160 and up.

The schematic is true. The wiper feeds the speaker. You can't knock a guitar output down 35db (down to .016 V p-p) and then plug it into an antique vacuum tube amplifier without wrecking the signal to noise ratio. Besides, you can't drive the tubes into distortion by working 30 to 40 db below their maximum power capabilities.

My 'scope says the practical range is -21db to -53db and the frequency response is flat from zero to 200KHz. If that amplitude range is too high, a) I included a 270 ohm resistor he can use by moving one wire. and b) I can change the circuit to use a 1.5 W audio taper pot.

I have been working with this musician since 1973. You can be sure he has tried a dozen fuzz boxes in 47 years. His opinion is that vacuum tube amplifiers are the holy grail. Nobody has ever duplicated that sound with a fuzz box to his satisfaction. The only problem here is playing at over 100db in an apartment building. This is not two kids guessing about what to do and how to do it. It is two guys with nearly 100 years of combined experience fixing a small problem. The reason I posted this is so other people can see how to build a Power Soak with about 4 electronic parts and a whopping big heat sink.
 
Last edited:

PeteHL

Joined Dec 17, 2014
473
If you want to attenuate by -30 dB, then you would need to put 160 Ohm in series with the fixed 80 Ohm resistor, I would think. This would mean that your 5k Ohm pot is too high of resistance to get the range of attenuation that you want?

20 X log [8 Ohm/ (80 Ohm + 160 Ohm)] = -30 dB
 
Last edited:

Thread Starter

#12

Joined Nov 30, 2010
18,224
If you want to attenuate by -30 dB, then you would need to put 160 Ohm in series with the fixed 80 Ohm resistor, I would think. This would mean that your 5k Ohm pot is too high of resistance to get the range of attenuation that you want?

20 X log (8 Ohm/ (80 Ohm + 160 Ohm) = -30 dB
I had a 5K 5W Wire Wound pot laying around for free. It accomplishes a range of -21db to -40db in real world operation. Instinctively, I would think 500 ohms to 1K audio taper would work, but 5K gives it a nice adjustment range and pretty good knob feel. Using available audio taper pots would place one limit at -31db. I think that's too close to the request of -36db. One important consideration is the ignorance of a musician. I have to talk to him about a tad of difference and a doubling of loudness because he simply can not understand the logarithmic nature of human senses. When he describes a problem, I have to use my best interpretation, include adjustability, and hope I have his specs surrounded on both sides. Math only goes so far, then the human factors get important.

You did give me an idea. I can add a 500 ohm resistor from the wiper of the pot to one end (I forget which end and will have to math it out). That will make a pseudo audio taper, improve the knob feel, and reduce the impedance of the attenuator circuit...I hope.

Edit: I added 470 ohms from the wiper to ground and got 9 more db of attenuation and 9 more decibels in the lower numbers of the dial. Wider total range and better linearity in terms of db per number on the knob.
 
Last edited:

Thread Starter

#12

Joined Nov 30, 2010
18,224
If you want to attenuate by -30 dB, then you would need to put 160 Ohm in series with the fixed 80 Ohm resistor, I would think. This would mean that your 5k Ohm pot is too high of resistance to get the range of attenuation that you want?

20 X log (8 Ohm/ (80 Ohm + 160 Ohm) = -30 dB
Thanks for the idea. I got 9 more db or adjustment and it was all in the bottom, previously useless, aread of the pot rotation. Part= installed
 

Alec_t

Joined Sep 17, 2013
14,280
Welcome back to AAC.
Couldn't he just use a dummy load to keep the amp happy distorting at full power, plus headphones instead of the loudspeaker to keep the neighbours happy?
 

AnalogKid

Joined Aug 1, 2013
10,987
In post #1 you said the measured response does not agree with your math, but you don't say what the measured response is and you did not post your math.

ak
 

Audioguru again

Joined Oct 21, 2019
6,673
A speaker resonates. It relies on the very low output impedance of the amplifier (0.04 ohms or less for a modern amplifier) to damp the resonance. Your attenuator messes up damping.

The 8 ohm speaker's impedance is about 40 ohms or more at the resonant frequency which messes up your attenuation of the resonant frequency. The result is sounds like a bongo drum.

BUT, The "musician" likes distortion, fuzz and vacuum tube sound so maybe he will also like his geetar to sound like a bongo drum.

Why not load the 100W amplifier with a huge, hot 8 ohm resistor then attenuate the output with two high resistance resistors in series. Then it feeds a little amplifier or a headphones amplifier?
 

Attachments

PeteHL

Joined Dec 17, 2014
473
The issue that "Audioguru again" is referring to in post # 14 could be an issue, but not necessarily. It depends on whether or not the guitar plays at frequencies in the resonance magnification range of the speaker.
 

PeteHL

Joined Dec 17, 2014
473
This is how I would go about making the attenuator, see the diagram below. Including a single pole, 5 position rotary switch, this would give you about +/- 6 dB of level difference for each change of the position of the switch. The volume control of the amp could be used to get an output sound level between what the circuit can provide, without substantially reducing distortion, I would think.

-Pete

LOW-Z-ATTENUATOR.png
 

Thread Starter

#12

Joined Nov 30, 2010
18,224
Welcome back to AAC.
Couldn't he just use a dummy load to keep the amp happy distorting at full power, plus headphones instead of the loudspeaker to keep the neighbours happy?
Sure. All he has to do is plug the headphones into the Power Soak output instead of the usual 10 or 12 inch speaker.
 

Thread Starter

#12

Joined Nov 30, 2010
18,224
In post #1 you said the measured response does not agree with your math, but you don't say what the measured response is and you did not post your math.

ak
-18db (measured) and -25db (measured). First post
The math is: db=20log e2/e1
 

Thread Starter

#12

Joined Nov 30, 2010
18,224
A speaker resonates. It relies on the very low output impedance of the amplifier (0.04 ohms or less for a modern amplifier) to damp the resonance. Your attenuator messes up damping.

The 8 ohm speaker's impedance is about 40 ohms or more at the resonant frequency which messes up your attenuation of the resonant frequency. The result is sounds like a bongo drum.

BUT, The "musician" likes distortion, fuzz and vacuum tube sound so maybe he will also like his geetar to sound like a bongo drum.

Why not load the 100W amplifier with a huge, hot 8 ohm resistor then attenuate the output with two high resistance resistors in series. Then it feeds a little amplifier or a headphones amplifier?
Yes, this messes up the damping factor. I hope you will let me see your method for how to cut 35db to 45db off the output of a 100 watt amplifier without changing the speaker interaction with the amplifier. If you can do it for under $60 in parts, I will build your design instead of mine.

What the musician plugs this in to is up to him. Headphones, a small amplifier, his cat's patooty. I don't care. I just build what he asks for.

Speaking of cats, you let one out of the bag. There is a multi-million dollar industry supplying various effects for musicians. If they find out their geetar sounds like a beatnik doing a bongo solo in a dark, smoky bar, they will crash the whole industry, and it's all your fault! Then there is the psychological damage you have done. Musicians all over the world will suddenly realize that the didn't hear what they heard and they don't like what they believed they liked when they test drove these devices. It all sounds like bongo drums and they didn't know. Oh, the humanity!
 

Thread Starter

#12

Joined Nov 30, 2010
18,224
The issue that "Audioguru again" is referring to in post # 14 could be an issue, but not necessarily. It depends on whether or not the guitar plays at frequencies in the resonance magnification range of the speaker.
Guitars and speaker resonances are very much interacting in the same range of frequencies. Wrecking the damping factor is going to change the results, a lot. But that's what the customer asked for, and paid for. He has heard plenty of other Power Soaks, and he says he likes them for this one job. There's no accounting for taste. In this case, it isn't my taste we are talking about.
 
Top