Speaker Model for Attenuators

[mac]

I'm playing near eviction levels and I need an attenuator for my 30 watt amp.
I built a T-Bridge pad, but since it's a resistive load it does not sound good @ 1/4x

I thought about 8ohm, 500uH inductors, but they don't have the cone mass that restricts the coil's movement.

I was wondering if this speaker model can do the job better,

http://sound.westhost.com/tsp.htm

Any thoughts?

mac

[Brisance]

How about cascading 2 op-amps to get less than 1 gain?

[mac]

How about cascading 2 op-amps to get less than 1 gain?

A high power opamp.
Sounds like a voltage regulator

mac

[Morocotopo]

Mac, how about an L-pad? I´ve read that they sound better...

I´m currently thinking of making an Airbrake type attenuator, purely resistive, many people say it sounds OK, at least at the first atten. levels. As far as I can tell it´s an L-pad type.

[teemuk]

I was wondering if this speaker model can do the job better,

http://sound.westhost.com/tsp.htm

It is definitely more accurate model of an actual loudspeaker load than a purely resistive one.

If you can find the parts (they are not "generic", the cap must be non-polar and all those inductors must handle the high load currents) then it should be a fine choice. With given component values, however, resonant frequency will be somewhat low in comparison to typical guitar loudspeakers, but then again you can derive all those component values directly from Thiele/Small parameters if you want to tweak further. The site even seems to have a spreadsheet for that so you don't even have to know the math, All you need to know are the T/S parameteters of interest.

[sajy_ho]

I know it sounds stupid; but how about using a speaker crossover with ratherly high cut off frequency in combination with resistor devider?

[mac]

Mac, how about an L-pad? I´ve read that they sound better...

I built a T-bridge pad which sounds better than the L-pad IMHO.
The T-pad does a decent job with my Epi Valve Jr, but not with my 30watts Laney.

I read that some people put a 2.2uf non polar across a resistor to get some highs, like a crossover.

If you can find the parts (they are not "generic", the cap must be non-polar and all those inductors must handle the high load currents) then it should be a fine choice. With given component values, however, resonant frequency will be somewhat low in comparison to typical guitar loudspeakers, but then again you can derive all those component values directly from Thiele/Small parameters if you want to tweak further. The site even seems to have a spreadsheet for that so you don't even have to know the math, All you need to know are the T/S parameteters of interest.

I simulated the circuit in LTSpice. I used Eminence and Celestion speakers data and changed some parameters,
Lcoil=500uH   Rcoil=6.8ohm
Lcone=5mH  Rcone=47ohm  Ccone=680uF

With this values and voltage source of 8ohm @ 1v, the resonant frequency is about 86hz.
After the resonance peak the impedance is flat until 1Khz, then it begins to get inductive and starts to climb.

5mH inductors are easy to find in low consumption light bulbs. I guess they can handle 10-20watts. They can also be modified to make 500uH inductors.
Count the number of turns N1 for 5mH. Then N2 = N1*sqrt(500uH/5mH)
If you are unsure about the wire, just replace it with a thicker one. N1 and N2 are the same.

I want 1/4x of 30watts, that is 3 dummy loads and the speaker, to the 8ohm amp output,


     ---///---///---
-----               -----
     ---///---SPK---


mac