I'm building my own version of a home theatre system with bi-amplified speakers. I will be using several amplifier modules to power the various drivers. I want to power the amp modules off a single central power supply. This supply will provide +/- 35 volts with some basic filtering. Some amp modules will be in the same case as the power supply while others will be remote mounted in speakers. For these remote amps I will run a DC power cable with the +/- 35 volts & GND.
However, I want each local amp module to have its own reserve of filter capacitors even though it gets its power from the main supply. Some modules will have a large reserve (bass amps) and some will have smaller reserves (tweeter amps).
What I don't want is to have one amplifier draw power from the capacitors in another amp module during music peaks. What I want to know is the easiest way to isolate the local power supply filter caps in each amp from other amps so they don't rob each other of power. I know how to do this with a single supply (diodes), but how do I do this with a split supply?
I read in sci.electronics.design that Curtis wrote (in ) about 'Power Supply Isolation Question...', on Sun, 12 Dec 2004:
In one word, DON'T. You are extremely liable to run into utterly insoluble crosstalk and distortion problems. You are bound to compromise the grounding, because you need one common ground for the input signals and with a common power supply you must have a common point at the other end, where all the large distorted currents lurk. Guess how I know.
Use a separate power supply for each amplifier. It may cost a bit, but it will be well worth it.
You can use a central power supply for line-level stuff but NOT, without a LOT of hassle, for power amplifiers.
--
Regards, John Woodgate, OOO - Own Opinions Only.
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
If this is what you are trying, there is likely to be trouble. Some of your return current will flow down the signal's GND line. This will create offset voltages in the signal that depend on output circuit current. This is going to be trouble.
If this what you intend:
You really should send the signal differentially and have the amplitude as large as practical. The CMRR of the differential input section in the remote box has to be quite good. Beware of the high frequency (~20KHz) CMRR it has to be good too. This signal should go on a "bell foil" shielded twisted pair.
The output section in the remote box should be designed as to pass current between the +35V and -35V and not return much (if any) current on the ground. Using a full bridge or half bridge output does this. One of the really troubling parts of the design is the fact that you will have a lot of distortion products in the power wires. The power really needs to be a twisted pair also and all other anti-crosstalk measures observed in this area.
This filter needs to include a series impedance to reduce the high frequency AC currents flowing in the cable. The power amplifier will have some power supply rejection ratio. If the high frequency PSRR of the amplifier isn't high enough and the ripple on the capacitors isn't low enough, the harmonics on the power supply will show up as harmonics in the output signal.
[...]
Each remote box should have rectifiers in its power input connections. This prevents the current from returning down the wires. It does not completely protect you though.
Assume an amplifier is drawing, lets say, 1A through the diode, If another amplifier suddenly draws enough power to reduce the main supplies voltage, the current through the diode stops. This means that the tweeter boxes have to contain extra capacitance to carry through the power dips caused by the bass amplifiers.
Having the inductors at the inputs more or less forces you to also include rectifiers across the supply lines. Imagine that someone (you would never) unplugs the connector with the power on. The inductors prevent the current from stopping and can make quite a large voltage spike. If you don't take care in the design, this spike could break the electronics or arc weld the end off connector pins.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.