Dual Mono - Anti Mono Power Supplies

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm about to embark upon building a dual monobloc amplifier, the AKSA, paradoxically within the same chassis....

As specified, each transformer's secondary would be configured as a center tapped, feeding a single bridge,' capacitorized', providing +&- rails. and fed to its respective channel. Alternatively, each of those secondaries could feed, its own bridge, one end on each to a star ground, feeding the amps as before.

As another altenative, could one take a humbucker approach, attempting to capitalize on the fact that within a typical stereo signal, much of it is mono in nature. i.e. -
Use each transformer's secondaries, feeding their own bridges, cofigured out of phase, to supply the same, but isolated, rail to both channels. Hopefully, this might reduce flux desity within each transformer and provide some intewinding, intercapacitive benefits (whatever the hell that means). While this certainly eliminates the 'ideal' of true monobloc separation, would there be some counter benefits in the reduction of eddy currents....? more effective use of the cores?

Similarly, one could take a more straightforward approach and configure the secondaries in the standard monobloc configuration, but wire the respective primaries out of phase to each other....

I'm only build one pair of 'monoblocs' initially, so comparisons would be difficult. I'd like to know if this is a really stupid idea?
 
Dual Mono Bipolar Power Supplies

Hi Paul,

I will try to answer the points as you make them, as much to keep me on track as anything else! You wrote:

<<As specified, each transformer's secondary would be configured as a center tapped, feeding a single bridge,' capacitorized', providing +&- rails. and fed to its respective channel.

HD: If using ultrafast, soft recovery diodes I do not suggest using caps across each diode. This tends to offset the efficiency gains of using such diodes as the switch on event is lengthened and the soft recovery hardly changed at all.

In stereo, or at least dual mono, my experience has been that the channel separation is crucial to imaging. While some recordings have more image information than others (this information is incorporated within the 500Hz to 4KHz band), any degradation of separation will quickly erode the image, and so power supplies should be kept utterly independent.

The audio amplifier is a power supply modulator; the energy for the speaker comes exclusively from the power supply, and so we must consider two circuit blocks; the modulating circuit (or amp proper as we know it), and the power supply - the powerhouse for the process. Clearly any modulation of one power supply by the other is seen by the speaker as modulation pure and simple. Since it is the subtle signal phase differences between the two channels which control the imaging, we must give them every opportunity to deliver the goods by presenting a clean, unmodulated power supply to both channels. Complete isolation one supply from the other is the most effective way of ensuring this.


<<As another altenative, could one take a humbucker approach, attempting to capitalize on the fact that within a typical stereo signal, much of it is mono in nature. i.e. -
Use each transformer's secondaries, feeding their own bridges, cofigured out of phase, to supply the same, but isolated, rail to both channels.


HD: The anti-phase approach is ingenious, but flawed in that once either phase enters the bridge rectifier, the DC output is always from the same terminal. That is to say that a 60Hz AC waveform will always produce a 120Hz output ripple, and since the output is from a positive and a negative terminal, the phase or antiphase input to the bridge is immaterial.

If you wish to improve the power supply purity, perhaps the best way I know is to use three phase power. This effectively yields 3 x 120 = 360Hz ripple, which is much easier to filter out and requires less filter capacitance. It is one of the reasons AC systems in aircraft use 400Hz AC; much smaller transformers, much less filtering - and so much lighter.


<<Hopefully, this might reduce flux desity within each transformer and provide some intewinding, intercapacitive benefits (whatever the hell that means). While this certainly eliminates the 'ideal' of true monobloc separation, would there be some counter benefits in the reduction of eddy currents....? more effective use of the cores?


HD: The same circulating current exists in one transformer as another when they are both operating in antiphase. That is to say, the energy requirement of a phase and an antiphase transformer operating side by side is identical, since energy is a scalar, not a vector, quantity. From the bridge forward, the waveforms look identical, so I do not believe this would offer any benefits.


<<Similarly, one could take a more straightforward approach and configure the secondaries in the standard monobloc configuration, but wire the respective primaries out of phase to each other....


See above argument; to me (and I am just one opinion!) there appears to be no discernible benefit.

Power supplies are tricky beasts. For an audio amp, we want a power supply which delivers as close to DC as we can get. This implies diodes in the bridge which switch on very quickly, as near as dammit to the 0 degree crossover point on the incoming waveform, and turn off without the creation of a spray of voltage spikes which inevitably tax the filter caps beyond their abilities and finish up affecting the output transistors, and thus, the speaker.

We also want two independent power supplies which can operate without any discernible or measureable influence upon each other, regardless of signal through the amps. This is quite difficult, and in the final analysis it would be great if we could power the two separate transformers from separate power circuits.

Cheers,

Hugh R. Dean
www.printedelectronics.com
 
Hugh,
Thanks for the response. Although when I referred to 'capacitorizing' the output I was referring to seperate smoothng caps, not snubbers, your point about each individual bridge switching at twice the input frequency is still valid, so there is no 'antiphase' cancellation or better use of the core. Oh well, dual mono it is... didn't think it through.

Dang!,
Paul
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.