Looking around online I couldn’t help notice the number of audio amp manufactures offering fully balanced amplifies. Here is a short list:
Pass Labs
Mark Levinson
Krell
Balanced Audio Technology
Moon/simAudio
Theta Digital
What is behind this trend? One factor may be that they can lower there supply rails and they by get bigger p.s. caps for less money but I don’t think that’s the hold reason. I wonder what the most popular method of splitting the signal phase is?
Pass Labs
Mark Levinson
Krell
Balanced Audio Technology
Moon/simAudio
Theta Digital
What is behind this trend? One factor may be that they can lower there supply rails and they by get bigger p.s. caps for less money but I don’t think that’s the hold reason. I wonder what the most popular method of splitting the signal phase is?
Actually, balanced amplifiers have been around for a long time. That said, they are slowly gaining in acceptance, in part because they can lower background noise.
Lower supply rails don't figure into the equation. One--transformers are usually priced by VA, regardless of the ratio between voltage and current. Two--balanced signal doesn't imply a bridged amplifier circuit; see the Alephs as examples. Prices actually sometimes increase due to the fact that there are more parts in some balanced topologies; you need amplification for the other phase.
Grey
Lower supply rails don't figure into the equation. One--transformers are usually priced by VA, regardless of the ratio between voltage and current. Two--balanced signal doesn't imply a bridged amplifier circuit; see the Alephs as examples. Prices actually sometimes increase due to the fact that there are more parts in some balanced topologies; you need amplification for the other phase.
Grey
From my point of view, quite a lot of advantages:
- Lower rails means less dissipation in a single active device, so in some cases no need to cascode them
- you can use active devices with lower voltage ratings
- as you said, lower rated caps
- only half gain per side necessary, could lead to less amplifying stages
- Less modulated rails, as current draw is much more constant (at least in non class A amps
)
- In some cases, you can avoid coupling caps (absolute offset doesn't matter for the speaker if both outputs are "hot")
- Distortion (k2) cancellation
Tino
- Lower rails means less dissipation in a single active device, so in some cases no need to cascode them
- you can use active devices with lower voltage ratings
- as you said, lower rated caps
- only half gain per side necessary, could lead to less amplifying stages
- Less modulated rails, as current draw is much more constant (at least in non class A amps
)- In some cases, you can avoid coupling caps (absolute offset doesn't matter for the speaker if both outputs are "hot")
- Distortion (k2) cancellation
Tino
Hmmm...better call a doctor. This thread isn't doing so well.
--Okay, the rails are lower. True. But the bias current is twice what it was, so the Pd per device remains the same. Granted, this may allow you to use a different device, and that might work to your benefit, but it won't be because of lowered heat dissipation.
--Lower voltage caps: Unfortunately, a 1000uF 50V cap and a 1000uF 25V cap are not equivalent. Life isn't that simple. The proper way to compare caps is to look at the Joules of energy stored. Assuming that a 50V cap is actually charged to 50V, it contains 1.25J of energy. Cool. So how much energy does the 25V cap store (assuming 25V charge)? The same amount? Half that, since the voltage has been cut in half? Sadly, no. It only contains .3125J. The trick here is that the voltage is squared, then multiplied by half the capacitance. It's that squared that comes back to haunt you. This is the reason that tube guys get bragging rights for ridiculously small amounts of capacitance--their voltages are not only higher, they're squared in addition.
Like transformers, caps are usually priced more-or-less by storage, not strictly by voltage or capacitance. Granted, the pricing structure here is a little more complicated, but if you run through the math, you'll see that there's a definite trend.
--As I noted in my previous post--and as Jocko pointed out--there's a difference between balanced and bridged. The larger Alephs are good examples of non-bridged amplifiers that can accept a balanced signal (or unbalanced, for that matter). The Aleph-X is a bridged circuit that can accept an unbalanced signal (or balanced). If I recall correctly, the smaller X amps can accept either, but the larger ones require a balanced signal. This may have changed with the newer .5 series; I haven't looked.
There's been some discussion about Haflers recently. Even though they were unbalanced, the addition of a small board (basically a phase-splitter) allowed the amps to be used in bridged mode. They didn't really have enough current to do a good job, and you were limited to 8 ohm loads, but at least on paper they were bridged using a single-ended input.
Does that help any?
Grey
--Okay, the rails are lower. True. But the bias current is twice what it was, so the Pd per device remains the same. Granted, this may allow you to use a different device, and that might work to your benefit, but it won't be because of lowered heat dissipation.
--Lower voltage caps: Unfortunately, a 1000uF 50V cap and a 1000uF 25V cap are not equivalent. Life isn't that simple. The proper way to compare caps is to look at the Joules of energy stored. Assuming that a 50V cap is actually charged to 50V, it contains 1.25J of energy. Cool. So how much energy does the 25V cap store (assuming 25V charge)? The same amount? Half that, since the voltage has been cut in half? Sadly, no. It only contains .3125J. The trick here is that the voltage is squared, then multiplied by half the capacitance. It's that squared that comes back to haunt you. This is the reason that tube guys get bragging rights for ridiculously small amounts of capacitance--their voltages are not only higher, they're squared in addition.
Like transformers, caps are usually priced more-or-less by storage, not strictly by voltage or capacitance. Granted, the pricing structure here is a little more complicated, but if you run through the math, you'll see that there's a definite trend.
--As I noted in my previous post--and as Jocko pointed out--there's a difference between balanced and bridged. The larger Alephs are good examples of non-bridged amplifiers that can accept a balanced signal (or unbalanced, for that matter). The Aleph-X is a bridged circuit that can accept an unbalanced signal (or balanced). If I recall correctly, the smaller X amps can accept either, but the larger ones require a balanced signal. This may have changed with the newer .5 series; I haven't looked.
There's been some discussion about Haflers recently. Even though they were unbalanced, the addition of a small board (basically a phase-splitter) allowed the amps to be used in bridged mode. They didn't really have enough current to do a good job, and you were limited to 8 ohm loads, but at least on paper they were bridged using a single-ended input.
Does that help any?
Grey
Hello Grey,
the calculation with the joules is very nice but in reality it´s the ripple that counts. If you´ve got 20V power supply and want a 0,1V ripple when using X Ampère bias then you need a certain amount of microFarads. The voltage of the caps doesn`t come into play in this calculation (except that it needs to be higher than the supply voltage)
As for the balanced topic:
There is a difference between accepting a balanced input signal or putting out a balanced signal.
So for me an Aleph or an ONO (with inverting output stage) is not a balanced amp, an XA is as are the various BOSOZ.
William
the calculation with the joules is very nice but in reality it´s the ripple that counts. If you´ve got 20V power supply and want a 0,1V ripple when using X Ampère bias then you need a certain amount of microFarads. The voltage of the caps doesn`t come into play in this calculation (except that it needs to be higher than the supply voltage)
As for the balanced topic:
There is a difference between accepting a balanced input signal or putting out a balanced signal.
So for me an Aleph or an ONO (with inverting output stage) is not a balanced amp, an XA is as are the various BOSOZ.
William
That was the base of my post too. When I talk balanced, I mean balanced from input to output, two phases. Like Aleph X, Son of Zen, X-Series, the picture on the first page of the SuSy patent, ....wuffwaff said:
So the better expression would be bridged? No problem.....
What would you like to cure??GRollins said:
Why the bias current per device should be twice?? For a given output power in class A the bias current doesn't change. The total dissipation of a power amp will remain more or less the same (a little more in fact for the bridged). But as we are doubling the devices (we have two phases) the dissipation per device is less.GRollins said:
Moreover, in fact I was not talking dissipation or voltage rating of the output devices. Pd can be overcome with paralleling output devices. Think of JFets at input. Each stage has to perform the half output swing. So less need for cascoding, etc etc.
No, I do not think we need a doctor. Just a cool beer maybe.
Tino
zinsula said:
That is Fully balanced, the confusion probably is because of the difference between balanced and fully balanced.
Balanced designs, both bridged and full separation for the phases was a trend some 20 years ago, and probably halted because of the added costs because of having to double all parts.
My guess would be that because of the difference in harmonic dominance, both bridged and fully balanced symmetrical designs have become popular again.
As with everything else, audio development seems to follow a wavy pattern.
Don Camillo and Gino Cervi is nice, what sentimental times.
If only it were so...
I like the concept of a bridged amp that doesn't use any more current, but I'm afraid that it kinda runs counter to the laws of thermodynamics. For everything to remain the same output-wise, the current must increase in order to offset the decrease in rail voltage.
Or in the time-honored phrase: There ain't no such thing as a free lunch.
Grey
I like the concept of a bridged amp that doesn't use any more current, but I'm afraid that it kinda runs counter to the laws of thermodynamics. For everything to remain the same output-wise, the current must increase in order to offset the decrease in rail voltage.
Or in the time-honored phrase: There ain't no such thing as a free lunch.
Grey
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.