My MC2500 amp, while indeed much
better after moving the 60Hz carrying wiring, was still not
right. Using my Tektronix scope, I measured almost exactly 3
times the residual noise from the left channel as I did from the
right. This had me perplexed for some time. I localized the
issue to the op-amp and it's associated components.
However, after moving all of the components including the opamp from
one board to the other, the issue remained. I was left with the
conclusion that the board itself was leaky. I have heard of this
sort of thing happening, but I had not experienced it until now.
When dealing with high impedance, high gain circuits, seemingly small
changes can make a huge difference and this was no exception.
After some deliberation, I took another approach to the problem. I knew
that the MC2600 was the last version of that series of amplifiers, so I
decided to look at the same circuit and see if McIntosh had done
anything different. While the basic concept was indeed the same,
there was a few small subtle additions that made a large
difference. It was the addition of a 22pf capacitor, a 68 pf cap
and the replacement of a 47pf cap to 22pf cap. This dropped the
noise levels on the left channel to lower than the right. Of
course this forced me to perform the same changes on the right channel.
The result was a lowering of the p-p noise at the output by about 33%
in the good channel. Now both channels measure the same noise level and
while the amp made spec before, it greatly exceeded them now.
Below is a partial schematic of the circuit. The
changes/additions are circled in red.
I performed measurements to determine if I had changed the frequency response or distortion, and I could not detect any change. I also confirmed my measurements by performing a circuit simulation. A listening test also indicated that the sound had not changed. The annoying noise was gone. I can now place my ear an inch or so away from the speaker and can hear little to nothing with the input to the amp disconnected. My C22 preamp is not as quiet as my amp so if I want to reduce the noise of my system even more, I will have to look at it and see if there is something I can do about the residual noise. I do not have RTA pictures documenting this modification. You will have to take my word for or do it yourself and measure the difference. For those that care, the cap across pins 2 and 3 was 47 pf. I decided to leave the 33pf cap in. and just add the 68 pf. I would appreciate some feedback from those that perform this change. To leave feedback, please post a message to www.audiokarma.org in the McIntosh section. Spammers have caused me to not publish an email address.
Reducing noise in analog circuits is like trying to attain absolute zero. For every advance toward zero noise, the task gets more difficult the closer you get.
After giving the entire design some thought, I concluded that the way to really get the noise down, one would have to place this portion of the circuit at the back panel and that is exactly what McIntosh did with the MC2600.
Back to main page
I performed measurements to determine if I had changed the frequency response or distortion, and I could not detect any change. I also confirmed my measurements by performing a circuit simulation. A listening test also indicated that the sound had not changed. The annoying noise was gone. I can now place my ear an inch or so away from the speaker and can hear little to nothing with the input to the amp disconnected. My C22 preamp is not as quiet as my amp so if I want to reduce the noise of my system even more, I will have to look at it and see if there is something I can do about the residual noise. I do not have RTA pictures documenting this modification. You will have to take my word for or do it yourself and measure the difference. For those that care, the cap across pins 2 and 3 was 47 pf. I decided to leave the 33pf cap in. and just add the 68 pf. I would appreciate some feedback from those that perform this change. To leave feedback, please post a message to www.audiokarma.org in the McIntosh section. Spammers have caused me to not publish an email address.
Reducing noise in analog circuits is like trying to attain absolute zero. For every advance toward zero noise, the task gets more difficult the closer you get.
After giving the entire design some thought, I concluded that the way to really get the noise down, one would have to place this portion of the circuit at the back panel and that is exactly what McIntosh did with the MC2600.
Back to main page