|
Update - Managed to measure the efficiency, it's around 79%, but EC gets too unstable with the proper input filter. A good filter would be an inductor of 22uH with internal resistance 80mOhms and a capacitor as large as you can spare (maybe even a few in parallel) with a total value of no less than 4.7mF. Even then I recommend a robust power supply, maybe from an old laptop, a really old one, because their psu's have a higher current rating than modern ones. They must be able to handle at least 10 times their nominal current rating for 500us or so.
Most Power supplies of this kind work well up to 5-10mA, basically being able to power no more than a preamp tube or something. This switching supply has the capacity to power a push pull EL84 power amplifier. Moreover other supplies often use no galvanic separation between the switching circuit and the output. This often makes the circuit very noisy. RF noise typically finds its way through the imperfect ground rail and to the amplifier itself. This may cause the amp to oscillate or misbehave in some other way. A good workaround is to use an RF transformer to separate galvanically the switching circuit from the output. Galvanic separation, though, sometimes causes problems with the servo loop, controlling the output voltage, and this is why the VCVS is used (or an optotransistor). The 500Mohm resistor of course needs to be somewhat lower, but EC doesn't work well with hysteresis, and shows no solution. There are still things to be considered, like input filtering, maybe some output filtering as well, although Tube amps are already well enough filtered against frequencies thousands of times lower. Input filtering is absolutely necessary, because initially, untill the stabilisation circuit kicks in, the circuit draws peak amounts of current of up to 100A. A properly designed LC input filter, should hold the ''front,, long enough until things settle down. I can't measure the efficiency, because EC is too unstable with such complex circuits, but it looks very good after stabilisation.
P.S. The hysteresis is way too low, but I've noticed EC doesn't cope well with hysteresis, and I've just brought it to the brink of stability just for the example. In reality the output needs better filtering and the hysteresis should be way higher. The 500Mohm resistor should be more like 10Mohm or less perhaps.
|
