questions on power supply for audio amp

A

alan

Guest
I built a simple unregulated power supply for an audio amp. The usual:
bridge rectifier with snubber caps and resistors on the diodes, goes
into electrolytic caps with a resistor+capacitor snubber, ceramic caps
at the power pin of the amplifier, etc. The problem is that there is
some high frequency noise on the supplies that is getting thru the amp.
Looking on a scope, I can see that there is a broad peak at ~10MHz and
another at ~40 MHZ. I don't think these are due to capacitor ringing
since they go away when I turn the power off while the caps are still
charged.

What's the best way to get rid of this noise? Should I do any filtering
on the primary or secondary of the transformer? Or do I need some
better bypassing at the filter caps? (e.g. finding a capacitor that has
a minimal impedance at these frequencies)
 
"alan" <no-longer-valid@yahoo.com> wrote in message
news:cshjur$41k$1@news.Stanford.EDU...
I built a simple unregulated power supply for an audio amp. The usual:
bridge rectifier with snubber caps and resistors on the diodes, goes
into electrolytic caps with a resistor+capacitor snubber, ceramic caps
at the power pin of the amplifier, etc. The problem is that there is
some high frequency noise on the supplies that is getting thru the amp.
Looking on a scope, I can see that there is a broad peak at ~10MHz and
another at ~40 MHZ. I don't think these are due to capacitor ringing
since they go away when I turn the power off while the caps are still
charged.

What's the best way to get rid of this noise? Should I do any filtering
on the primary or secondary of the transformer? Or do I need some
better bypassing at the filter caps? (e.g. finding a capacitor that has
a minimal impedance at these frequencies)
Click to expand...

Millions of audio amps operate with a simple rectifier and filter cap
without "RF on the powwer supply". Yes, you can get RF breakthrough if the
interference is strong enough. My initial response is "is this actually RF
interference?".

An oscilloscope on the power rails shows the rectifier sawtooth *plus* some
of the output signal - so RF on the power supply can mean either the RF is
coming in OR the amplifier is generating RF.

First off, I would look for amplifier instability. You need star
grounding - if you have a trace resistance common to power or loudspeaker
current and the input circuit, you will get instability. Does your CRO show
bursts of high frequency on peaks of the signal waveform ?

How large is the RF signal on the output with no signal ? How large on the
supply rail with no audio signal ?

If you live in the city and crank up your CRO, you will see RF in all sorts
of places, so 10mV of RF on the CRO may not mean much.

Is the amplifier quiet with no audio signal ? What does it sound like ?
What kind of diodes did you use for the power supply ?

An RF filter on the mains and speaker and audio input leads can be added -
but lets establish what is going on first.

Roger










If you see a signal other than your rectifier sawtooth on the supply rails,
there are TWO possibilities.





Unless you are running RF transmitters nearby, it is unlikely that RF noise
is being picked up


1) Your oscilloscope is picking up RF. and you don't really have a
problem. How big is the signal on the
 
Subject: questions on power supply for audio amp
From: alan no-longer-valid@yahoo.com
Date: 18/01/2005 00:06 GMT Standard Time
Message-id: <cshjur$41k$1@news.Stanford.EDU

I built a simple unregulated power supply for an audio amp. The usual:
bridge rectifier with snubber caps and resistors on the diodes, goes
into electrolytic caps with a resistor+capacitor snubber, ceramic caps
at the power pin of the amplifier, etc. The problem is that there is
some high frequency noise on the supplies that is getting thru the amp.
Looking on a scope, I can see that there is a broad peak at ~10MHz and
another at ~40 MHZ. I don't think these are due to capacitor ringing
since they go away when I turn the power off while the caps are still
charged.

What's the best way to get rid of this noise? Should I do any filtering
on the primary or secondary of the transformer? Or do I need some
better bypassing at the filter caps? (e.g. finding a capacitor that has
a minimal impedance at these frequencies)






Use a mains line filter.
Click to expand...
 
I read in sci.electronics.design that alan <no-longer-valid@yahoo.com>
wrote (in <csic0i$mo4$1@news.Stanford.EDU>) about 'questions on power
supply for audio amp', on Mon, 17 Jan 2005:

I think approx 5mV p-p.
Click to expand...
It is of absolutely no significance. Just forget it.
--
Regards, John Woodgate, OOO - Own Opinions Only.
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
 
"John Woodgate" <jmw@jmwa.demon.contraspam.yuk> wrote in message
news:puC2d3AzEM7BFwlh@jmwa.demon.co.uk...
It is of absolutely no significance. Just forget it.
--
Click to expand...
The mains wiring is an antenna. Also speaker wires. RF levels are high in
city areas. You have computers and other switching power supplies. So you
have low level RF crawling all over your circuit, especially visible when
you use a CRO with its own ground. Perhaps breaking the mains connection
disconnects your antenna.

You get used to the normal background levels for each location - for
instance, here in the country I see no RF on my CRO - in the city I saw 5 mV
pp quite often. AM stations are easily discernable by the modulation of
the envelope.

It might be interesting to see if the RF persists when you disconnect the
speaker leads.

If your amplifier is unstable, you can expect a volt or more of supersonic
oscillation.

If your amp sounds good, you can enjoy it. Case closed.

Roger
 
Roger Lascelles wrote:
"alan" <no-longer-valid@yahoo.com> wrote in message
news:cshjur$41k$1@news.Stanford.EDU...

I built a simple unregulated power supply for an audio amp. The usual:
bridge rectifier with snubber caps and resistors on the diodes, goes
into electrolytic caps with a resistor+capacitor snubber, ceramic caps
at the power pin of the amplifier, etc. The problem is that there is
some high frequency noise on the supplies that is getting thru the amp.
Looking on a scope, I can see that there is a broad peak at ~10MHz and
another at ~40 MHZ. I don't think these are due to capacitor ringing
since they go away when I turn the power off while the caps are still
charged.

What's the best way to get rid of this noise? Should I do any filtering
on the primary or secondary of the transformer? Or do I need some
better bypassing at the filter caps? (e.g. finding a capacitor that has
a minimal impedance at these frequencies)



Millions of audio amps operate with a simple rectifier and filter cap
without "RF on the powwer supply". Yes, you can get RF breakthrough if the
interference is strong enough. My initial response is "is this actually RF
interference?".

An oscilloscope on the power rails shows the rectifier sawtooth *plus* some
of the output signal - so RF on the power supply can mean either the RF is
coming in OR the amplifier is generating RF.


First off, I would look for amplifier instability. You need star
grounding - if you have a trace resistance common to power or loudspeaker
current and the input circuit, you will get instability. Does your CRO show
bursts of high frequency on peaks of the signal waveform ?
Click to expand...
already got star grounding.
I don't think this noise is due to the amp coz that amp's (whatever you
call it when the open loop gain goes to 1) is at 3-ish MHz. The fft
also shows flat noise up to and past this point. I don't think it rings
on a square wave either. This high freq noise also goes away when I
turn off the power, i.e. the caps are still charged so the amp is still
running, but the mains power is disconnected.

How large is the RF signal on the output with no signal ? How large on the
supply rail with no audio signal ?
Click to expand...
I think approx 5mV p-p.

If you live in the city and crank up your CRO, you will see RF in all sorts
of places, so 10mV of RF on the CRO may not mean much.

Is the amplifier quiet with no audio signal ?
Click to expand...
There is only a slight hiss which I expect. I can also hear a quiet
60*n Hz buzz on the right channel, although I can't find that signal
with the scope.

What does it sound like ?
What kind of diodes did you use for the power supply ?
Click to expand...
One of those FWBR's that all come in one square package.

An RF filter on the mains and speaker and audio input leads can be added -
but lets establish what is going on first.


If you see a signal other than your rectifier sawtooth on the supply rails,
there are TWO possibilities.

Unless you are running RF transmitters nearby, it is unlikely that RF noise
is being picked up

1) Your oscilloscope is picking up RF. and you don't really have a
problem. How big is the signal on the
Click to expand...
I don't think the scope is picking up RF coz this signal goes away when
I turn the amp off.
 
"alan" <no-longer-valid@yahoo.com> wrote in message
news:csj5f6$d27$1@news.Stanford.EDU...

Anyway, I would like to get the RF fuzz out of my amp, if only so that I
can measure clean signals on it.


If your amp sounds good, you can enjoy it. Case closed.

There is good, and then there is better.
Click to expand...
Alan, let me put it this way.

You see RF on your CRO, because you introduce the CRO into the circuit of
the amplifier, thus making a new circuit. Thus your CRO is not showing you
what goes on inside your amplifier - it is showing the combined result of
your amplifier with a great big lump of CRO metalwork hanging off it, and
the CRO ground.

You would have to make a tiny battery powered CRO or a tiny light coupled
CRO probe to see the true voltage between the nodes of your amplifier.

You have a measurement problem. We all have to live with it. Small signals
on CROs are suspect. Short your probe to its ground and clip it onto
another ground - you may see "RF", hum, all sorts of rubbish. Especially at
RF, the CRO lead and CRO PCB inductance, mains lead inductance starts to
cause voltage differences between different parts of the CRO. You can clip
your probe to $10000 amplifiers and see the "RF".

Most likely the amplifier doesn't have RF between any points within its
circuit.

This means you don't "have RF". It ain't broke.

Roger
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top