Second edited version 3 june 2010
, thanks Peter!Some thoughts about the design of the power supply in the 606 till 909 series.
Quad introduced a new power-supply in the 606 with an active earth circuit.This basically means they removed the central earth wire from the transformer. As both power supply voltages now only depend on the value of the capacitors, an additional transistor-circuit was added to adjust the voltage across the two power supply capacitors. If you used ‘matched’ capacitors, the extra circuit would not be necessary.Looking at the simplified circuit diagram you will notice that it is impossible to have any DC current trough the loudspeaker. Quad introduced an elegant DC protection. The DC protection itself has nothing to do with the active earth, and without the transistor circuits, still there will be no DC current.If one of the dumpers would fail the voltage over the capacitor will be doubled, hopefully the fuse will be blown before the capacitor does, as it is only rated 63V. Anyway, this is a far better solution than the Clamp circuit in the 405. From the DC point of view…
But is there a “Free lunch”?
If you compare the return current from the loudspeaker in a conventional setup, like the 405 (see the diagram), and the 606 supply, it looks like there is an extra power supply capacitor, a diode and half the secondary transformer winding in series with the loudspeaker in the case of the 606. This is worse than the output circuit of a 303!
Remember, the capacitors are of large capacity and the diode and the transformer winding are “behind” the capacitor. In the real world the loudspeaker current is delivered by the capacitors, they act as power supplies. Even in the situation were the centre connection of the transformer is used the current is supplied by the capacitors. There is only a small current through this wire, this is due to the tolerance in capacitors value. If the capacitors were matched, there is zero current in that wire. Check this with two power supplies in series and two resistors, the principle is the same. So in both setups the speaker current goes ‘’ through’’ the capacitors. In both setups the quality of the capacitors and the value are important. In the case of the Quad virtual earth the stability of the centre voltage is determent by the value (capacity) of the capacitors and the “speed” of the regulating transistor circuits.
Why this post?
In this paper I described two situations, the conventional earth wire and the 606 situation where the wire has an infinite resistance. Based on a idea of Keith Snook, what will happen when the wire is replaced by a resistor? If the value is low enough, the voltage on the power supply capacitors will stabilise, without the need of the extra transistor circuits. If the value is high enough the DC current will be limited. The idea was to build a 405 PSU with the resistor and omit the DC protection. Until now I did not found a value that will match both the former criteria. The main reason for this is the fact that there is only very little time (5mS for the 50Hz situation, 60Hz is even worse in this aspect) for the capacitors to charge, the value of the resistor must be very low, but a low value wont protect your valuable speakers. So until someone comes with a better idea, our own DC protection or the Quad implementations will still be needed.