Power Scaling.

ICBM

jpfamps said:

There has to be a significant resistive element other wise you wouldn't dissipate much power in the attenuator.....

Ha, true .

What I meant was that most which claim to be 'reactive' in fact are still mostly resistive even though they have a reactive element, and don't actually have the same sort of impedance curve as a real speaker. Oversimplification comes back to bite me .

jpfamps

I've built / installed variable HT into several amps.

As a general rule with most forms of attenuation, they seem to work more naturally at lower levels of attenuation (of course there may well be Fletcher Munson effects working here as well).

My experiences with power scaling is that when you turn down the HT, you need to also reduce the drive to the power valves / power amp, otherwise you can experience a harsh octave doubled type distortion as the power valves are driven harder with respect to the HT as the power is turned down.

Essentially I think the "problem" is that the end user expects to have a single knob on the amp that will make the amp sound the same at all volumes. This is unrealistic. A good practical solution is having a range of controls that allow the use to adjust the sound at different power levels.

gringopig

.

ICBM

gringopig said:

Yes an attenuator is just a big resistor. It drops voltage across it and that voltage multiplied by the current flowing through it is the power in electrical Watts lost in heat dissipation.

No, it isn't. A proper attenuator is a lot more complicated than that - even a purely resistive one. It consists of a series and a parallel load combined with the speaker so the amplifier sees at least close to a constant impedance while the proportion of power sent to the speaker goes down.

gringopig said:

It's this coil of wire which causes simple resistive loads to sound poor as it has a characteristic depending on the frequency of the electric energy going into it.

Actually a well-designed resistive attenuator can sound very good - the Dr.Z Airbrake is a good example. They work especially well with no-negative-feedback amps, in fact.

I actually think the bigger problem is that at high levels of attenuation, the parallel resistance across the speaker is quite a low value, and this will heavily damp the speaker in comparison to it being connected directly to the amp, which is what makes the sound 'flat'. It may need a three-way variation with a series resistance to the amp, a parallel load and then another resistance in series to the speaker to eliminate the damping effect, but without custom-made resistive elements that can't be done with a single rotary control. It could be done with switching instead of a continuous control though.

gringopig

.

gringopig

.

martinw

tekbow said:

@martinw , thanks for that, much appreciated.

Just how invasive is it to install in an amp that it wasn't built in to? Something like a Soldano HR50+ with 2 channels, separate channel masters and shared EQ? Would that require 2 drive controls or one?

And if I may ask one more question, what's the purpose of the RBX bias control?, why wouldn't the on-board biasing be adequate?

It's moderately invasive, but it depends on how it's installed. In some cases you can use all existing holes to mount new controls, so no new holes are needed in the control panels. Other installations need 2 new holes so they are ore of a commitment. You may also need heatsinks and fans on larger amps. Internally, it's all reversible.

The RBX kit is an auxiliary raw bias supply for those amps which have too high an impedance in their own bias supply, i.e. they can't supply enough bias current. The Power Scaling tracking bias regulator draws more current than some amps' bias supplies can provide, usually because they tap off the HT primary via a high value resistor to drop the voltage.

tekbow said:

The way you're describing it's use is how I envisioned it, I.e. find a level of power amp effect I'm happy with (the point where the amp "opens up" perhaps) then use the scaling to bring that down to a manageable volume level.

That's what I do, and advise my customers to do. It's always best to use it to add some power amp colouration, not as a kind of rough volume control. @jpfamps makes a good point about user expectations, and it's a reason why it's not more prevalent IMO. One knob systems are limited and can sound poor at times, whereas two knob systems require a little finesse to set up. I'm lucky in that my customers are self selecting (and few!)

jpfamps

ICBM said:

gringopig said:

Yes an attenuator is just a big resistor. It drops voltage across it and that voltage multiplied by the current flowing through it is the power in electrical Watts lost in heat dissipation.

No, it isn't. A proper attenuator is a lot more complicated than that - even a purely resistive one. It consists of a series and a parallel load combined with the speaker so the amplifier sees at least close to a constant impedance while the proportion of power sent to the speaker goes down.

gringopig said:

It's this coil of wire which causes simple resistive loads to sound poor as it has a characteristic depending on the frequency of the electric energy going into it.

Actually a well-designed resistive attenuator can sound very good - the Dr.Z Airbrake is a good example. They work especially well with no-negative-feedback amps, in fact.

I actually think the bigger problem is that at high levels of attenuation, the parallel resistance across the speaker is quite a low value, and this will heavily damp the speaker in comparison to it being connected directly to the amp, which is what makes the sound 'flat'. It may need a three-way variation with a series resistance to the amp, a parallel load and then another resistance in series to the speaker to eliminate the damping effect, but without custom-made resistive elements that can't be done with a single rotary control. It could be done with switching instead of a continuous control though.

There are two issues to consider with attenuators.

1) The attenutor / speaker combination, should "look" like a speaker to the amplifier. With a properly designed reactive load this is relatively easy to achieve.

2) The amp / attenuator combination should "look" like an amplifier to the speaker. This, is much more problematic, and most (if any) attenuators don't achieve this.

What the speaker "sees" will vary very much with the amount of attenuation, eg ICBM says above the very low parallel resistance with the speaker causes the speaker to be damped more than if it were connected directly to the amp.

ICBM

jpfamps said:

There are two issues to consider with attenuators.

1) The attenutor / speaker combination, should "look" like a speaker to the amplifier. With a properly designed reactive load this is relatively easy to achieve.

2) The amp / attenuator combination should "look" like an amplifier to the speaker. This, is much more problematic, and most (if any) attenuators don't achieve this.

What the speaker "sees" will vary very much with the amount of attenuation, eg ICBM says above the very low parallel resistance with the speaker causes the speaker to be damped more than if it were connected directly to the amp.

At the most basic, adding a third resistance in series between the attenuator and the speaker which increases as the parallel resistance goes down should help fix that I think.

I wonder if a stereo L-Pad could be adapted to try this out...

olafgarten

ICBM said:

jpfamps said:

There are two issues to consider with attenuators.

1) The attenutor / speaker combination, should "look" like a speaker to the amplifier. With a properly designed reactive load this is relatively easy to achieve.

2) The amp / attenuator combination should "look" like an amplifier to the speaker. This, is much more problematic, and most (if any) attenuators don't achieve this.

What the speaker "sees" will vary very much with the amount of attenuation, eg ICBM says above the very low parallel resistance with the speaker causes the speaker to be damped more than if it were connected directly to the amp.

At the most basic, adding a third resistance in series between the attenuator and the speaker which increases as the parallel resistance goes down should help fix that I think.

I wonder if a stereo L-Pad could be adapted to try this out...

I'm sure you could design a load that emulates a speakers impedance curve using resistors, inductors and capacitors. You just need to create a peak at the resonant frequency and then an increase in impedance at the high end. 

I have some free time later today so I'll have a go and test it using LT SPICE

ICBM

olafgarten said:

I'm sure you could design a load that emulates a speakers impedance curve using resistors, inductors and capacitors. You just need to create a peak at the resonant frequency and then an increase in impedance at the high end.

As jpfamps said that's the relatively easy bit. In fact, apart from the resonant peak, the Marshall Powerbrake does the rising impedance curve pretty well, because it's a very unusual (I think unique) attenuator - it uses a multi-tap auto-transformer to split the power between the speaker and a load resistor.

It's a shame they stopped making it, but apparently it's because the transformer was too expensive - which also most likely explains why no-one else has taken up the concept. If you added something which produced the resonant peak to that, it should be pretty accurate.

Whether that would make it sound better or not, I don't know - the remarkable thing is that a well-designed purely resistive one can sound really good as well, and in theory it shouldn't...

MentalSharps

jpfamps said:

There are two issues to consider with attenuators.

1) The attenutor / speaker combination, should "look" like a speaker to the amplifier. With a properly designed reactive load this is relatively easy to achieve.

2) The amp / attenuator combination should "look" like an amplifier to the speaker. This, is much more problematic, and most (if any) attenuators don't achieve this.

What the speaker "sees" will vary very much with the amount of attenuation, eg ICBM says above the very low parallel resistance with the speaker causes the speaker to be damped more than if it were connected directly to the amp.

Would this mean that using a properly designed reactive load with an impedence curve similar to a guitar speaker, going into IR's and monitors/headphones/FRFR is the ideal sounding of all options? But using the reactive loadbox to attenuate the output to a guitar cab will be compromised in some way.

Could there be any benefit in using Power Scaling in combination with a reactive load for the sound/feel, or just to reduce wear on valves?