Hello.
What size main power supply capacitors do people use for ordinary analog solid-state amplifiers at about 20 or 25 watts?
This place being what it is, people are going to start spewing all kinds of stuff that I don't care about. I look forward to hearing from the one or two level-headed people among the crowd who can just give me an answer to the question that I asked.
Thank you for your kind attention.
What size main power supply capacitors do people use for ordinary analog solid-state amplifiers at about 20 or 25 watts?
This place being what it is, people are going to start spewing all kinds of stuff that I don't care about. I look forward to hearing from the one or two level-headed people among the crowd who can just give me an answer to the question that I asked.
Thank you for your kind attention.
For an 8ohms speaker I use +-20mF
For stereo 8+8ohms speakers I would use +-40mF, or +-45mF.
For 4ohms speakers, if I used any, I would double that requirement.
The above allows one to choose the NFB DC blocking capacitor to be set to an RC value of ~115ms and this in turn allows the input filter to be set to ~80ms (F-3dB=2Hz)
This gives the amplifier an F-1dB ~4Hz, typical of commercial HiFi amplifiers.
For stereo 8+8ohms speakers I would use +-40mF, or +-45mF.
For 4ohms speakers, if I used any, I would double that requirement.
The above allows one to choose the NFB DC blocking capacitor to be set to an RC value of ~115ms and this in turn allows the input filter to be set to ~80ms (F-3dB=2Hz)
This gives the amplifier an F-1dB ~4Hz, typical of commercial HiFi amplifiers.
Valve or solid-state? Speaker impedance? First you have to decide what the supply rail voltage and peak current will be.
Then you have to know what the PSRR of the amplifier is, and the speaker sensitivity, so you know how little hum you can hear so you can calculate how much ripple you can tolerate.
Then you calculate, using the definition of capacitance. Then you (probably) double the answer because of the poor tolerance of electrolytics.
Or just copy what someone else has done and hope they knew what they were doing and hope your system is sufficiently similar to theirs that their decisions are appropriate for your system too.
Then you have to know what the PSRR of the amplifier is, and the speaker sensitivity, so you know how little hum you can hear so you can calculate how much ripple you can tolerate.
Then you calculate, using the definition of capacitance. Then you (probably) double the answer because of the poor tolerance of electrolytics.
Or just copy what someone else has done and hope they knew what they were doing and hope your system is sufficiently similar to theirs that their decisions are appropriate for your system too.
if you like an answer to be precise try at least to ask an precise and detailed question , and you will see that many people here are able to answer , in that specific matter not "all roads lead to Rome " .
.
Last edited:
I recommend you size the caps to give whatever ripple voltage you desire, at maximum output power (just barely clipped sinewave @ 2 kHz, into 8ohm or 4 ohm load). Free PSUD software, or pencil+paper calculations, will give you the ripple voltage.
Edit- here's a happy YouTube (link)
Edit- here's a happy YouTube (link)
Last edited:
Thanks to those "snifflers and bitches " one (that is me very often) can evolved , or not , what a marvelous choice .......
.
The problem with random insults is that nobody knows who is being insulted or why. Wait a minute, one person knows: the person handing out the insults. As he is the only one who knows it must be that the insults were issued as a means of getting something off his chest i.e. a form of therapy.Bamalama said:
Um .... you don´t want to get it very Techie, do you?
OK: get a couple of those thingies which come in an aluminum can but are neither beer nor deodorant and solder them inside.
OMG!!!! I said "solder" !!!!!!!!
EXCUSE ME !!!!!!!!!
Define "level headed"
Ok, don´t worry about definitions, I *guess* what you mean
FWIW each and every one of the 125000 (give or take a few thousand) Forum members firmly believes he fits that description (including you and me) , and most (all?) others are brainless people, with the exception of the few who think exactly like him.
Won´t even mention acceptable ripple under load.
Ok, now back to the original question: in theory (and actually in practice) "better filtering is best", (although some people like ripple mixed with their audio for unknown reasons), in any case you soon reach a diminishing returns point, where doubling values do no carry *audible* improvements.
Yet there are certain seemingly favorite filter cap values, no doubt found reasonable by trial and error, I guess less than 0.00001% PSU designers around the World use "PSU Designer" software
In your amp class it´s common to find 4700uF supply caps, a few cheap guys use 2200; those who think they really care can use 6800uF or even 10000uF , such values are not expensive or take unduly space inside your amp, any larger is fine, it just takes money and space.
Brand? : any major brand sold by the large established distributors (Mouser, Digikey,Farnell, RS, etc.) is fine, don´t get carried away on Mil Spec and such , unless you want to play music inside a B52 bomber at 40000ft and -20 degrees C temperature.
For your own sanity (and financial health) avoid snake oil salesmen for the plague they are.
Good luck
There is a problem with gross over capacitance like 50,000uf or even 100,000uf. This is a healthy dead short on the rectifier and power transformer until the caps charge up. Even mundane things like power switches have to be considered due to the inrush current.
So to spite the lack of any improved audio performance with ridiculous filter cap values, there is also a reliability downside.
Why go there?
So to spite the lack of any improved audio performance with ridiculous filter cap values, there is also a reliability downside.
Why go there?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.