CMOY starts just fine, then gets VERY clippy in 3 or 4 seconds

[Pars]

The 3dB point is the half power point... 3dB down from nominal and decreasing at 6dB per octave for a 1st order filter. I was looking around for a Wikipedia on this, and this was about as good as I could find quickly.

 

[Mizerable]

Hey how does the 1/(2 pi R C ) formula work?


I was looking online for other people using output caps and I came across a bunch of people using something like .1uF 63V polyester caps for the output DC blocking or other caps in the low few uF to fractions of a uF caps.

Building a pre amp for a Gainclone chip amp.

Take a look at this one. This guy had a 670mV ( wow, just ..wow) offset and he is using 2uF to 4.7uF polyproplyene caps to block it. So according to that formula, this guy is losing frequencies of up to 665 Hz !?

EDIT: nevermind ! wikipedia'd high-pass filter for the formula explanation

Also, how badly would a 1000uF or a 2200uF electrolytic cap hurt sound quality?

 

[Ron.id]

Quote:

Originally Posted by Mizerable /img/forum/go_quote.gif This guy had a 670mV ( wow, just ..wow) offset and he is using 2uF to 4.7uF polyproplyene caps to block it. So according to that formula, this guy is losing frequencies of up to 665 Hz !?

It's a pre-amp. He will attach the output to input of power-amp, usually had 50K-100K input impedance. So, add that power-amp input impedance to 1/(2.pi.R.C) formula...

 

[tangent]

Quote:

Originally Posted by Mizerable /img/forum/go_quote.gif here are the pics

A few of those joints show signs of being "cold": the solder surface should always be shiny and smooth. Use desoldering braid to remove some of this solder (don't completely desolder it, just get rid of some of it) and put fresh solder back on the joint. Alternately, if you have some liquid flux around, dab a bit on the joint and reflow it with your iron.

Then, clean the flux off the board.

Finally, there are several places where I can see stray wire strands or overly long component leads on the bottom side of the board. Clip these away.

Quote:

how does the 1/(2 pi R C ) formula work?

A capacitor (C) in series with an AC signal (like music) followed by a resistor (R) to ground is called a high-pass filter. This means that it lets high frequencies pass through it [mostly] untouched. As frequency goes down, the filter attenuates the signal more and more.

There's a special frequency where attenuation changes from a very slow roll-off to a constant amount of significant roll-off, which is the -3 dB point: the point at which the signal coming out is attenuated by 3 dB relative to the input. (Negative dB is attenuation, or "down" from the input.) Below the -3 dB point, the signal is down an additional 6 dB with every octave, which is a frequency doubling or halving.

For example, if the -3 dB point is 12 Hz, the signal is 9 dB down at 6 Hz, 15 dB down at 3 Hz, etc. Because this is an exponential curve, at 0 Hz it's infinite attenuation, which is another way of saying that a capacitor blocks DC.

Quote:

how does the 1/(2 pi R C ) formula work?

R is the resistor value in ohms, C is the capacitor value in farads, and pi you should know. So to rework the calculation by Pars:

C = 0.00000022 farads (0.22 microfarad caps, in series with the signal)
R = 120 ohms (your headphones, a resistance between signal and ground after the C)

f = 1/(2 × pi × R × C)
f = 1/(2 × 3.14159 × 120 × 0.00000022)
f = 1/0.00016587609211
f = 6029 Hz

 

[Mizerable]

The problem is fixed. I wanted to see that formula's effect so I tried a large 1000uF 10v cap in series with the output and it sounded pretty much like it should.

I then tried a smaller .22uF film cap and only the very very high treble remained.

But I found the problem behind the distortion: the DC offset was pushing the weedy lil speakers past what they could do comfortably. That 1000uF cap on each channel fixed it right up. I'm gonna get a better opamp, but this will do till then

Thank you everybody ! I also cleaned up a lot of the joints. Found some small bridges and dirtiness that was covering it up.

 

[Mizerable]

I'm using this 16v IBM laptop powersupply from around 2003 and I assumed that since it is for a laptop, that it would be pretty stable.

The problem is that when I have the cmoy plugged into this power supply and I have the headphones in and there is no source plugged, there's this LOUD BUZZ. As soon as I connect a cd player or mp3 player, the buzzing stops immediately.

Also, when I dont have the cmoy connected to any source like the mp3 player, the right channel only always has a 180 mV offset. As soon as I connect it to the headphone and the buzzing goes away, the offset in both channels drops to below 1mV (the offset in the left channel is always very low).

This leads me to think that the buzzing is associated with the high DC offset in the right channel. When I have the thing powered with batteries, there is no buzzing ever and I assume there is never that high offset with or without a sound source connected.

I don't think this is a big problem anyways, since there is never an offset over 1mV in either channel when a source is connected. But I am just curious as to why connecting a source to the input seems to remove dc offset in the output.

Also, all these readings were taken without headphones plugged in.

 

[Pars]

This still has the output caps? Or did you remove them? Same TL082 opamp? Do you have a volume pot on this?

DC offset should be measured with no source or headphone connected, and the pot turned all the way down (lowest volume). If you don't have a pot, short the L&R inputs to ground. A Cmoy (or most other amps) are always going to make some noise with no source connected, unless they have a pot and its turned all the way down. Normal.

The laptop supply is a switching supply (not a linear) and is probably going to be noisier than hell. Can't help you there.