If one was to design a multi-channel amplifier for an active loudspeaker, what would be the desirable attributes of the amplifier devoted to the treble, the midrange, and the bass? Could the circuitry be more simple than that for a full-range amplifier? Are design compromises made in a full-range amp, that could be eliminated in a narrow range amp?
Very good question.
Much of the power in music is in the bass. You can make a powerful amplifier to drive the bass driver that doesn't need to slew nearly as fast as it would if it were driving the tweeter as well. That simplifies the drive section significantly.
For the tweeter, you often don't need much power (depending on the crossover frequency) so you can use smaller transistors with less capacitance, and hence faster slew rate for a given drive current.
An example of the philosophy is what I've got in mind for the two-way speakers in my study. These use a crossover frequency of 3KHz or so. I've got a design for the mid-bass (that I'm currently using to drive both bass and tweeter via a passive crossover) using large MOSFETs, that's capable of some 50W or so.
The driver for the tweeter is based on the TPA6120 headphone amp. This chip is designed for driving 32-64 Ohm headphones, to a watt or two. I worked out that if I put eight in parallel, I can get something like ten watts, at vanishingly low distortion levels, and with eye-popping slew rates.
A similar setup would be useful to replace the 100W amp that I use to drive my Infinity RS-5b's three way speakers, with crossover frequencies of 600Hz and 4KHz. I could use a 50W (single pair of FETs) amplifier for the bass driver, another for the mid, and the little 10W amp for the tweeter.
Much of the power in music is in the bass. You can make a powerful amplifier to drive the bass driver that doesn't need to slew nearly as fast as it would if it were driving the tweeter as well. That simplifies the drive section significantly.
For the tweeter, you often don't need much power (depending on the crossover frequency) so you can use smaller transistors with less capacitance, and hence faster slew rate for a given drive current.
An example of the philosophy is what I've got in mind for the two-way speakers in my study. These use a crossover frequency of 3KHz or so. I've got a design for the mid-bass (that I'm currently using to drive both bass and tweeter via a passive crossover) using large MOSFETs, that's capable of some 50W or so.
The driver for the tweeter is based on the TPA6120 headphone amp. This chip is designed for driving 32-64 Ohm headphones, to a watt or two. I worked out that if I put eight in parallel, I can get something like ten watts, at vanishingly low distortion levels, and with eye-popping slew rates.
A similar setup would be useful to replace the 100W amp that I use to drive my Infinity RS-5b's three way speakers, with crossover frequencies of 600Hz and 4KHz. I could use a 50W (single pair of FETs) amplifier for the bass driver, another for the mid, and the little 10W amp for the tweeter.
I've been wondering about this "myth". In full active system (3 amps), the best sound will come if all the amps are the same type from the same manufacturer.
If we use different amps, the sound will not blend nicely. (say for tweeter we use small classA amp, for mid we use chip amp, for bass we use classD amp)
Is this myth true?
If we use different amps, the sound will not blend nicely. (say for tweeter we use small classA amp, for mid we use chip amp, for bass we use classD amp)
Is this myth true?
Hi,lumanauw said:
I don't think the myths can be true , neither of them.
I'll go further, I am convinced the myth is blatantly untrue.
I accept that some amps are good in the midrange,
some are exceptional in the treble,
some are very good in the bass.
It is very rare that a single amplifier can be very good throughout the whole frequency range.
I think if amps were selected to do the job that they do best then the resulting amplifier/driver will sound better overall.
If the amplifier were designed to perform well in a narrower frequency range deliberately tailored to a particular driver (or frequency range) then the synergy may be even better.
By all means experiment with different amplifiers, there must be some combinations that will work better than others.
I think the area of biggest benefit will come from a no electrolytic treble amplifier. That target would be totally uneconomic for a bass amp and probably almost as unaffordable in a mid frequency range amp.
Slew rate requirement is related to frequency and to maximum drive voltage.
I believe that treble requires the same drive voltage as mid and bass when all the drivers are the same impedance and the same sensitivity.
Following from that the treble requires a higher slew rate, or more accurately the bass requires less.
However, many treble drivers are much more sensitive than the other drivers in the speaker, particularly the the low Fs bass drivers. This then brings about the need for different levels of drive voltage and different slew requirements.
It may be when you do the numbers, that the slew requirements become very similar for typical combinations of chosen drivers.
I believe that treble requires the same drive voltage as mid and bass when all the drivers are the same impedance and the same sensitivity.
Following from that the treble requires a higher slew rate, or more accurately the bass requires less.
However, many treble drivers are much more sensitive than the other drivers in the speaker, particularly the the low Fs bass drivers. This then brings about the need for different levels of drive voltage and different slew requirements.
It may be when you do the numbers, that the slew requirements become very similar for typical combinations of chosen drivers.
djk said:
Hmmm, Maybe I should be re-thinking my project.
I would be bi-amping some two-way horns (after an electronic crossover) using an Adcom 555 for the lows. For the Highs, I had been thinking about some Class A, either: 1) Forte 1A, 2) Monarchy SM70Pro, 3)Threshold s150, 4) Sumo Nine. These have less than "typical" amounts of feedback.
Before this I have used the Adcom 555 on some Klipschorns and loved the bass (this is why I am set on the Adcom for the lows).
Food for thought,
-Tom
WithTarragon said:
Gee...., Any thoughts or advice?
The Cabinets will be a clone of a Klipsch Jubilee: hornloaded bass bin with dual 12 in drivers (impedance can be as low as 3 Ohm). The Top is crossed at about 800 Hz (electronic xover with steep slopes and time-alignment). This will be a tractrix horn (store bought from Klipsch) and will not present a very low impedance to the amp. The efficiency on both section is good (so huge amounts of power are not required), although I never want to even come close to clipping.
I like the sound (bass) of the Adcom 555ii on a Klipschorn, but I would like something very detailed and accurate sounding for the top section (NOT "warm, lush & musical"). That is why I was considering a Class A design for the top (Forte 1A which is affordable, Monarchy SM70, Sumo Nine, Threshold S150 , which may not be affordable) .
The only other constraint on the amp for the top section is that I would like the input impedance to relatively high since I am using a passive "pre-amp" (10k pot conventionally wired and coming after the electornic crossover).
This is a chance to test the notion of mixing/matching amps to their relative strengths when bi-amping.
Thanks, & Any thoughts?
-Tom
Andrew, Thanks for your response/question.
The configuration is the digital out to a Behringer DCX crossover (digital in). The only ADC conversion is at the output of the DCX crosssover. This signal then goes to a pair of stereo pots (mechanically ganged together) then to the two stero amps.
By putting the passive "preap" after the electronic crossover, I avoid the extra conversions (DAC & ADC) and I also do not have to worry about making sure the signal is hot enough so that all bits are being used (maximizing my SNR, although with delta sigma modulation this is less of a concern now days).
The sensitivity on the bass bin is probably about 105 dB or so and the top section is a bit better (within a handful of dB). I will not be using an L-pad or autoformer. So the sensitivity-matching between the drivers (or compensation for differences in gain between the two amps will be done digitally in the DCX, so there will be some bit-loss, but not more than a bit or two).
Obviously, using the Adcom GFA 555 (200 watt) is gross overkill given the relative efficiency of the bass bin. I am aware of this. However, not all amps are equal when it comes to producing a clean & deep bass from a horn loaded bass bin. The Adcom seems to do well in this regard, and others I have tried do not. The nature of this interaction is not obvious to me, since it is not simply a function of damping factor, watts, current handling etc. It really did require experimentation.
I am not sure what you mean by "inverting input" in this context.
Thanks,
-Tom
The configuration is the digital out to a Behringer DCX crossover (digital in). The only ADC conversion is at the output of the DCX crosssover. This signal then goes to a pair of stereo pots (mechanically ganged together) then to the two stero amps.
By putting the passive "preap" after the electronic crossover, I avoid the extra conversions (DAC & ADC) and I also do not have to worry about making sure the signal is hot enough so that all bits are being used (maximizing my SNR, although with delta sigma modulation this is less of a concern now days).
The sensitivity on the bass bin is probably about 105 dB or so and the top section is a bit better (within a handful of dB). I will not be using an L-pad or autoformer. So the sensitivity-matching between the drivers (or compensation for differences in gain between the two amps will be done digitally in the DCX, so there will be some bit-loss, but not more than a bit or two).
Obviously, using the Adcom GFA 555 (200 watt) is gross overkill given the relative efficiency of the bass bin. I am aware of this. However, not all amps are equal when it comes to producing a clean & deep bass from a horn loaded bass bin. The Adcom seems to do well in this regard, and others I have tried do not. The nature of this interaction is not obvious to me, since it is not simply a function of damping factor, watts, current handling etc. It really did require experimentation.
I am not sure what you mean by "inverting input" in this context.
Thanks,
-Tom
As soon as you said DCX, I knew your set up is the exception. Yes, the volume adjustment goes AFTER the DCX.
Only a bit or two lost when gain matching in the digital domain? I'm not sure but I suspect a lot more.
DCX supercedes my comment on inverting opamp input. Forget it.
Greater than 105db/W/m. That seems to indicate a power requirement for the LOUD bits of between 10W to 20W and normal average level of just a few hundred mW or less.
This clearly puts you into ClassA territory. SS (BJT or FET) or tube (valve).
It might be worth doing that in analogue after DCX as well.
Only a bit or two lost when gain matching in the digital domain? I'm not sure but I suspect a lot more.
DCX supercedes my comment on inverting opamp input. Forget it.
Greater than 105db/W/m. That seems to indicate a power requirement for the LOUD bits of between 10W to 20W and normal average level of just a few hundred mW or less.
This clearly puts you into ClassA territory. SS (BJT or FET) or tube (valve).
Andrew, Thanks for your comments, it looks like we are the same page.
Now, the question of which amp. I am sold on the idea of the Adcom GFA-555 for the bass bin of a high efficiency, horn loaded system. So the issue is what would be good for the top section.
We are in agreement that a Class A would be a candidate. However, I am being cautious given the comment by DJK about "mixing" amps since they could sound quite different. His example concerned amps with & without feedback (he used a SET as an example). While in my case I am wondering about mixing an amp with a "typical" of of feedback and one with a minimal amount of feedback (Class A amps I have listed previously)
Unfortunately, experimenting with Class A amps can be expensive.
Thanks,
-Tom
Now, the question of which amp. I am sold on the idea of the Adcom GFA-555 for the bass bin of a high efficiency, horn loaded system. So the issue is what would be good for the top section.
We are in agreement that a Class A would be a candidate. However, I am being cautious given the comment by DJK about "mixing" amps since they could sound quite different. His example concerned amps with & without feedback (he used a SET as an example). While in my case I am wondering about mixing an amp with a "typical" of of feedback and one with a minimal amount of feedback (Class A amps I have listed previously)
Unfortunately, experimenting with Class A amps can be expensive.
Thanks,
-Tom
Jubilees, eh? They're something I'd like to hear one day.
I have a pair of KHorns (sadly, now for sale) that have been through many experiments with drivers, crossovers (original, ALK and digital) and were one of the monitors I used to assess many different tube amps I built. The best, for different reasons, were versions of Lynn's Amity and Wrights PP2C. The 813 Super Amity shook the neighbourhood, but 50W of class A, even in a huge room, and a predisposition to annoying the old bag next door, it was too much. I used corner loaded LABhorns for <60Hz actively crossed and fed by a big Perreaux 6000B amp.
I agree with djk's comment re not mixing zero NFB tube and conventional SS, unless it's under about 80Hz. The differences will be really obvious.
I also agree with Andrew about doing the gain matching passively and not giving up any resolution in the xover. I tried Sony and dbx and hated both of them, though I mainly used analogue and thought I lost as much as I gained with the digital xover.
I'd either make two tube amps or roughly similar topology or two like a 20W version of Nelson's F4 using only one pair of outputs.
I have a pair of KHorns (sadly, now for sale) that have been through many experiments with drivers, crossovers (original, ALK and digital) and were one of the monitors I used to assess many different tube amps I built. The best, for different reasons, were versions of Lynn's Amity and Wrights PP2C. The 813 Super Amity shook the neighbourhood, but 50W of class A, even in a huge room, and a predisposition to annoying the old bag next door, it was too much. I used corner loaded LABhorns for <60Hz actively crossed and fed by a big Perreaux 6000B amp.
I agree with djk's comment re not mixing zero NFB tube and conventional SS, unless it's under about 80Hz. The differences will be really obvious.
I also agree with Andrew about doing the gain matching passively and not giving up any resolution in the xover. I tried Sony and dbx and hated both of them, though I mainly used analogue and thought I lost as much as I gained with the digital xover.
I'd either make two tube amps or roughly similar topology or two like a 20W version of Nelson's F4 using only one pair of outputs.
not necessarily.
An alternative would be to match gain to driver sensitivity.
Adjustable gain and/or attenuation is the much easier option.
Passive bi-amping does require amplifier gains to be set to match each other. The passive crossover and drivers already have matching sensitivities, that's the way the manufacturer sells them to the user.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.