Thanks Frank.
Would there any point in putting a 10uF cap to feed each output and driver transistor instead?
I've read that this is a "good idea" but I don't know if it really matters.
Really Frank? I think you mean impedance not ESR since ESR is a fixed value that does not vary with frequency. The much larger ESR of a film cap in parallel with the much smaller ESR of a large electrolytic will make no discernable difference to the ESR of the electolytic on its own. Do you have specific manufacturer's data to back up this commonly held fallacy about HF impedance?
Last edited:
Gopher, there is a graph that shows what you are looking for here:
http://www.intersil.com/data/an/an1325.pdf
http://www.intersil.com/data/an/an1325.pdf
Big E
That graph only shows the way to prevent resonances when using multiple small value LOCAL bypass caps. The impedance of a large (10,000uF) reservoir electrolytic cap is several orders of magnitude less than that of a small (0.47 uF) film bypass cap at frequencies up to several hundred kHz. Adding the bypass cap in parallel does NOTHING to improve HF impedance (just do the sums) and can actually introduce audible resonances in the audio band.
That graph only shows the way to prevent resonances when using multiple small value LOCAL bypass caps. The impedance of a large (10,000uF) reservoir electrolytic cap is several orders of magnitude less than that of a small (0.47 uF) film bypass cap at frequencies up to several hundred kHz. Adding the bypass cap in parallel does NOTHING to improve HF impedance (just do the sums) and can actually introduce audible resonances in the audio band.
Last edited:
Like you say, the bypass was removed. Maybe he listened instead of just believing the urban legend! Your Jensen 4 pole caps are specifically designed for very low inductance/HF impedance. If you add bypass caps you could screw that up.
What's far more important is having substantial LOCAL elecs at or near the output transistors in addition to the main reservoir caps. Anything from a few hundred to a couple of thousand uF should do nicely, depending on the power output. A small value resistor before the local caps would help too.
What's far more important is having substantial LOCAL elecs at or near the output transistors in addition to the main reservoir caps. Anything from a few hundred to a couple of thousand uF should do nicely, depending on the power output. A small value resistor before the local caps would help too.
Last edited:
If you refer to what they have done in that app note on pages 6 and 10: that is complete nonsense: they are adding magnitudes of complex functions (real and imaginary parts) over frequency. What happens in reality is that small capacitances together with stray inductances of large capacitances form a resonant circuit, creating very high impedances at certain frequencies. As a result, adding capacitance without clearly knowing all relevant parameters of all parts (for each capacitor at least besides C: stray inductance and ESR), things tend to get worse than before. The stray inductance is normally not specified but can be calculated from the resonance frequency when the impedance graph of the part is known. Then run a spice simulation to see what makes sense.
Furthermore: small bypass capacitors only make sense very very close to the load. If placed in parallel to the large supply capacitors, the long lead to the amplifier creates another inductance that renders all positive effects void.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.