Posted by Wayne Parham [ 66.137.58.85 ] on April 06, 2005 at 06:05:40:
In Reply to: Got under the hood.... posted by bernieclub on April 05, 2005 at 21:47:28:
I'm with you on this one. I'd replace the selector switch and the tube sockets, because that's a no-brainer. Just find a suitable switch replacement, some tube sockets and get the iron hot.
As for the rest of the upgrades, do you have a scope and a signal generator? You could verify your amplifier's frequency response by testing output voltage at several frequencies. This would allow you to try different components in the circuit and be sure of what your response is.
For the noise, you will want to measure power supply ripple and buzz and try to track it down. You will also probably want to see if any of the gain stages increase it, telling you it is coupled and amplified as opposed to being from B+.
What I found is not really new, and that is the snubbers are effective at reducing the switching buzz and that RC or LC filters are responsible for hum reduction. The large value components used to reduce hum don't have much effect on the switching artifacts, even if small value caps are used to bypass them. The best results come from installing 0.022uF caps right across the diodes. I'm not sure why any amp builder or kit maker would consider these parts optional - I think they should be installed on every amp with solid state rectifiers.
Another thing that's an issue is grounding. There's always the issue of ground loop verses ground path conductivity. You want a central single point ground and a big conductor out from it to every other component in the system. Ideally, the voltage drop across each conductor is exactly the same, and is very, very small. Ideally, it's zero, but I mean realistically ideal. 
The problem is that you aren't going to have this kind of balance between conductor length, cross-section area and current flow. It's all pretty much just a matter of getting a good conductor to each component, trying to make the voltage drop as low as possible and keep the grounding essentially free of noise producing loops.
But there's the rub. Single point grounding sometimes lets the ends float too much. Isolation doesn't work as well then, so you sometimes find that straping the grounds together works better. I've run into this on several occasions, both in digital electronics in noisy environments and on audio circuits too. I think in general, I try to make a single point ground and if it doesn't work, then I start tying grounds together. If that doesn't work either, then I have to completely isolate, like using opto-isolators.
So if I were you, I'd probably do some experimentation with grounding and with power supply filters. See what is most troublesome and eliminate it. You'll probably find several components, some ripple, some switching artifacts, some grounding and some inductively or capacitively coupled noise. If you have a scope, it would be helpful first to look at the noise and describe exactly what you see in terms of amplitude and frequency. Do it with inputs shorted to see power supply noise and with inputs open to see the amount of amplified noise coupled into the amp.
[ FSAudio Forum ] [ Help ]