I know a person who once got into a debate with me about choosing the value of amplifier's power supply filter capacitors.
I told him that supply filter cap's value are chosen according to the amount of current it should filter. But, from his point of view, the more the capacity the better the sound quality from every aspect like low, mid and highs are touching the sky.
Anyway, I know that supply filter cap's value make a difference in sound quality. But according to logic/physics, is there a point where this theory becomes useless?
I told him that supply filter cap's value are chosen according to the amount of current it should filter. But, from his point of view, the more the capacity the better the sound quality from every aspect like low, mid and highs are touching the sky.
Anyway, I know that supply filter cap's value make a difference in sound quality. But according to logic/physics, is there a point where this theory becomes useless?
Not useless, by any means, but there is a point of diminishing returns, after which, say, doubling capacitance (so doubling cost andb real estate inside the chassis) provides a very small improvement.
Say you have +/-35V supplies, 1000uF per side give you inacceptable 10V ripple, 2200uF=5V (cheap design but just acceptable in a cheapish commercial product), 4700uF=2.5V which is "good" and 10000uF=1.25V which is very good.
Each time doubling cost and space.
It´s always a compromise and *you* draw the line.
Another factor: say it´s an "Audiophile" amp or you decide "I will spare no costs, it´s for my own use".
So you choose, say, 25000uF per side.
Fine, you will have very little ripple, ample reserve for strong Bass, etc. , all good things, but you may stress your power transformer, you will *definitely* stress the rectifier diodes, you will need some kind of soft start to avoid nuisance fuse blowing, turn on thumps might be very loud, etc.
As you see, every problem has a range of solutions, none is "perfect", you choose.
Say you have +/-35V supplies, 1000uF per side give you inacceptable 10V ripple, 2200uF=5V (cheap design but just acceptable in a cheapish commercial product), 4700uF=2.5V which is "good" and 10000uF=1.25V which is very good.
Each time doubling cost and space.
It´s always a compromise and *you* draw the line.
Another factor: say it´s an "Audiophile" amp or you decide "I will spare no costs, it´s for my own use".
So you choose, say, 25000uF per side.
Fine, you will have very little ripple, ample reserve for strong Bass, etc. , all good things, but you may stress your power transformer, you will *definitely* stress the rectifier diodes, you will need some kind of soft start to avoid nuisance fuse blowing, turn on thumps might be very loud, etc.
As you see, every problem has a range of solutions, none is "perfect", you choose.
Hello,
When beginning with this hifi hobby i started with a Yamaha ca410 then some Nad ones before jumping into diy.
All my first amplifiers coming from the shop were modified by usually doubling the capacitors by putting a bigger pair in parallel. Usually it also meant a bigger rectifier. Those were very nice improvements for the money!!!
Later with diy audio i sometimes went into more extreme caps. lol
Of course just adding caps will eventually put a lot of stress on transformer and rectifier and can turn out into something sounding not as good.
Now i will usually use smaller caps and create a CRC or CLC filter keeping the first cap big enough but not to big and the second cap add some extra.
Greetings, Eduard
The value of power supply caps will determine the lower limit of amplifier power bandwidth. The rule of thumb for 20 Hz up is about 4700 µF for 2x8 ohm loads and 8200 µF for 2x4 ohm loads. Given that there are relatively few true 8 ohm speakers around, I'd consider 6800 µF parts about the minimum for a stereo amp.
Beyond 10-12000 µF you're generally entering diminishing returns territory, though if you want to be truly 2 ohm proof (requiring a correspondingly beefy circuit), 15000-18000 µF would be advised. 22000 may give some more margin for aging if it's not at the cost of voltage handling.
Beyond 10-12000 µF you're generally entering diminishing returns territory, though if you want to be truly 2 ohm proof (requiring a correspondingly beefy circuit), 15000-18000 µF would be advised. 22000 may give some more margin for aging if it's not at the cost of voltage handling.