New Posts  All Forums:Forum Nav:

some cmoy mods

post #1 of 3
1/22/12 at 2:15pm
Thread Starter 

I did a few things that differ from the basic design. I don't use batteries but a 12v regulated walwart so Instead of 4.7k splitting resistors I used 1k which makes the rails more stable.

 

I found this info and incorporated it.

 

" the ideal value for the feedback resistors (to give minimum offset voltage) will be such that the two resistances in parallel roughly equal the resistance to ground at the non-inverting input pin."

http://en.wikipedia.org/wiki/Operational_amplifier

 

So I changed the feedback gain resistors to 150k to ground and 270k. This comes close to the 100k input resistor to ground. This also lowered the gain to 2.8 which is more than enough since I don't have nor want any picky cans and also negates any noise associated with large resistors of which I can't hear any. With my fluke meter I measured 1mv dc on one output channel and 0 on the other after this change.

 

I also noticed that with the standard .1ufd input cap the lows started rolling off early so I increased it to .18 (taken from the signal path of a dead denon receiver). After that there was very little rolloff until below 20hz.

 

 

Reply
post #2 of 3
1/23/12 at 9:01am

Nice work with selecting the gain resistors. Very few people take the time to balance the input offset currents. If nothing else its a step in the right direction if you want to use a BJT opamp. 

 

For the input cap, its best to shoot for a -3db frequency several octaves lower than the lowest frequency you really care about. If you get the chance you may want a larger cap still. Perhaps 0.47uF.

There are the -1.5db and -0.75db points to consider, phase distortion, and the fact that these roll-offs are cumulative. -3db at 10hz in your amp, and -3db at 10hz in your source is -3db at 20hz at the output of the system which puts us right back where we started... Its not hard or really expensive to get those frequencies so low that they never add up to anything anywhere near what you will hear. 

 

 

If you have not read it I would suggest Tangent's article on input caps:

http://tangentsoft.com/audio/input-cap.html


Edited by nikongod - 1/23/12 at 9:14am
Reply
post #3 of 3
1/23/12 at 9:49am
Thread Starter 

Thanks those are good suggestions. I forgot to mention that the higher gain resistors seemed to fix a loading problem also. I had been using a small value gain/feedback resistor ( the ungrounded one) to lower the gain but it caused the high frequencies to start to dip above around 8k. The high value resistors fixed this.

Reply
New Posts  All Forums:Forum Nav:
  Return Home