Class X is for "across the line" voltage applications. Class Y is for "line to ground" applications. You should read up on the different variations before specifying caps for line voltage applications.
For this circuit, I'm not sure why a Class Y cap would have been specified at all. I would use Class X for both C9, and C11.
I found alot of info about the function of these caps, and various other linked articles, but nothing about which material to use in a given situation. A quick search on digikeys website produced a variety of materials like ceramic, film type etc. I just don't know which is appropriate for this application. Thanks
Sorry, I must have misunderstood.
The values required will dictate the composition of the caps you use. I would think you should be able to find film caps in both values, with the correct lead spacing.
Like what Stormrider said, X class for both C9 and C11 should be OK.
If you are into experimentation, C9 can be altered to make the delay shorter/longer. for 220VAC lines, values from 0.22uf to 1uf were tested to work with varying time delay. For 100VAC to 120VAC mains, 0.68uf to 1uf were also tested.
The build guide for the Soft Start has already been submitted but I guess it's still under scrutiny and editing.
I read that in another article I found, it also stated that the y caps could be substituted for x caps, but x caps could not be substituted for y. Interesting.
Transformer voltage and Power Output
30+30Vac gives ~ +-42Vdc & ~ 70 to 80W into 8r0 and suits 4ohms speaker
35+35Vac gives ~ +-50Vdc & ~ 100 to 110W into 8r0 and suits 4ohm speaker
40+40Vac gives ~ +-59Vdc & ~ 140 to 160W into 8r0 and possibly suits 6ohms speaker
45+45Vac gives ~ +-68Vdc & ~ 190 to 210W into 8r0, does not suit lower impedance speaker.
The smoothing cap voltage ratings are
>=63Vdc suit 30Vac & 35Vac
>=80Vdc suits 45Vac
63Vdc may suit 40Vac but must be checked for actual worst case voltage for your transformer and mains supply.
My 230:40+40Vac transformer fed with 240Vac gives ~ +-58.5Vdc on the smoothing caps with single bridge rectifiers, but 254Vac takes the smoothing caps to just below their safe and guaranteed limit of 63Vac if the rail fuses have blown, i.e zero quiescent current.
30+30Vac gives ~ +-42Vdc & ~ 70 to 80W into 8r0 and suits 4ohms speaker
35+35Vac gives ~ +-50Vdc & ~ 100 to 110W into 8r0 and suits 4ohm speaker
40+40Vac gives ~ +-59Vdc & ~ 140 to 160W into 8r0 and possibly suits 6ohms speaker
45+45Vac gives ~ +-68Vdc & ~ 190 to 210W into 8r0, does not suit lower impedance speaker.
The smoothing cap voltage ratings are
>=63Vdc suit 30Vac & 35Vac
>=80Vdc suits 45Vac
63Vdc may suit 40Vac but must be checked for actual worst case voltage for your transformer and mains supply.
My 230:40+40Vac transformer fed with 240Vac gives ~ +-58.5Vdc on the smoothing caps with single bridge rectifiers, but 254Vac takes the smoothing caps to just below their safe and guaranteed limit of 63Vac if the rail fuses have blown, i.e zero quiescent current.
smoothing capacitance and load impedance vs bass frequency response.
For F-3dB of 10Hz (no useful lower Bass response, but good upper Bass, Mid & Treble response) you can set the High Pass input filter to ~16ms. This requires the NFB filter to be set to >=23ms and the smoothing capacitance filter effect to be set to >=33ms
For 8ohms speaker that equates to >=+-4m2F and for 4ohms to 8m4F per channel.
If one requires the F-3dB to be one octave lower then all the capacitance values and the RC time constants must be doubled.
i.e. F-3dB=5Hz requires 4ohms capable PSU to have +-17mF/ch.
I find that setting F-3dB to 5Hz, giving F-1dB ~ 10Hz, cuts a little of the low Bass performance of music reproduction. I recommend you aim for F-3dB of between 1Hz and 2Hz for a truly wideband capable amplifier.
For F-3dB of 10Hz (no useful lower Bass response, but good upper Bass, Mid & Treble response) you can set the High Pass input filter to ~16ms. This requires the NFB filter to be set to >=23ms and the smoothing capacitance filter effect to be set to >=33ms
For 8ohms speaker that equates to >=+-4m2F and for 4ohms to 8m4F per channel.
If one requires the F-3dB to be one octave lower then all the capacitance values and the RC time constants must be doubled.
i.e. F-3dB=5Hz requires 4ohms capable PSU to have +-17mF/ch.
I find that setting F-3dB to 5Hz, giving F-1dB ~ 10Hz, cuts a little of the low Bass performance of music reproduction. I recommend you aim for F-3dB of between 1Hz and 2Hz for a truly wideband capable amplifier.
Last edited:
Really? This seems extreme because only very large, or extremely inefficient, speakers will be F-3db at 20hz, so the amplifier being F-3db at 5hz, or 10hz should be completely unnoticeable. By 10hz you are well outside the audible range. Are you sure any differences you perceived were not caused by other factors?
I think those are best "guesstimates" for output power in terms of power supply. Andrew is not too far off though because it is true that power in is (often less) equal to power out.
these are based only on the 2pr output stage driven by a single pr driver stage. This amplifier does not have a pre-driver stage and so cannot drive high transient currents into low impedance and reactive loadings. It can drive low resistance loading. It can also drive reactive speakers. It cannot drive low impedance AND reactive speakers.
Now back to output stage design. You need to copy exactly what is given, or learn how to correctly design for reliable temperature de-rated SOAR for your proposed output stage.