input chassis sockets / grounding
- Thread starter jimbo1968
- Start date
grounding is a mysterious art mastered by few. i don't understand it too well myself (i have no real background in electronics though), even after attentively following many posts about it here. if you really want a thorough and complete theoretical explanation, you're better off finding a book, i had a hard time finding good info myself on the internet.
a star ground i think is not the optimal way of grounding, but it is by far the simplest and easiest way of obtaining a functional, safe hum-free setup. Because of the simplicity of a gainclone, star ground is the obvious complement to it. i imagine in more complex systems, a star ground isnt even possible.
the idea of a star ground is that all your grounds join up into a final point. this inherently eliminates all possible ground loops. why? well because there is only one possible path to ground for every component. if you had uninsulated RCA's, you would create a second path to ground, and probably a nasty hum. It's a simple idea and it works.
other options are bus grounds for example. but i don't know much about that except they are harder to get right.
if it makes you feel better, most commerical equipment out there violates many grounding rules and is far from optimal.
a star ground i think is not the optimal way of grounding, but it is by far the simplest and easiest way of obtaining a functional, safe hum-free setup. Because of the simplicity of a gainclone, star ground is the obvious complement to it. i imagine in more complex systems, a star ground isnt even possible.
the idea of a star ground is that all your grounds join up into a final point. this inherently eliminates all possible ground loops. why? well because there is only one possible path to ground for every component. if you had uninsulated RCA's, you would create a second path to ground, and probably a nasty hum. It's a simple idea and it works.
other options are bus grounds for example. but i don't know much about that except they are harder to get right.
if it makes you feel better, most commerical equipment out there violates many grounding rules and is far from optimal.
Hi,
I disagree with Homer about a star ground not being optimal. But then we are tolerant and allowed to express differing views, so here's mine.
As stated by Homer the only connection to chassis (we agree) is the star ground. But there are three differing philosophies.
1. Connect ALL grounds and returns directly to the main central ground. Personally not my favourite.
2. Connect all grounds and returns to the main central ground except the Chassis. Connect the chassis via a coupling network using a selection or all of the following. 10nF to 100nF cap, 10r to 100r resistor, heavy duty two way diode.
3. Connect all the signal, power, decoupling, Thiel network, speaker returns, and the chassis to central star ground. Interconnect the signal and power ground with a resistor 10r to 47r. Connect input RCA ground via screen to input signal ground & then to central ground.
I believe that 2. & 3. are most likely to work with incorrectly wired ancillaries. You may need to experimnet to suit your equipment. 1. is intolerant of incorrectly wired ancillaries and may also cause buzzing an/or hum even when everything does comply.
Do not combine the central star ground with the PSU common to rectifier. Instead connect a wire of any length from PSU common to central ground.
SAFETY. Always connect the mains safety earth to all exposed metalwork forming the chassis. NO exceptions.
I disagree with Homer about a star ground not being optimal. But then we are tolerant and allowed to express differing views, so here's mine.
As stated by Homer the only connection to chassis (we agree) is the star ground. But there are three differing philosophies.
1. Connect ALL grounds and returns directly to the main central ground. Personally not my favourite.
2. Connect all grounds and returns to the main central ground except the Chassis. Connect the chassis via a coupling network using a selection or all of the following. 10nF to 100nF cap, 10r to 100r resistor, heavy duty two way diode.
3. Connect all the signal, power, decoupling, Thiel network, speaker returns, and the chassis to central star ground. Interconnect the signal and power ground with a resistor 10r to 47r. Connect input RCA ground via screen to input signal ground & then to central ground.
I believe that 2. & 3. are most likely to work with incorrectly wired ancillaries. You may need to experimnet to suit your equipment. 1. is intolerant of incorrectly wired ancillaries and may also cause buzzing an/or hum even when everything does comply.
Do not combine the central star ground with the PSU common to rectifier. Instead connect a wire of any length from PSU common to central ground.
SAFETY. Always connect the mains safety earth to all exposed metalwork forming the chassis. NO exceptions.
Hi,
a physical separation of the PSU common from the star ground is essential. The length and/or resistance of this link is not important but for good regulation smaller is better.
The reason for separation is to ensure the hi current charging spikes that circulate between the rectifiers and the smoothing caps do not generate a voltage across the central ground.
The spikes charge for a small part of the 50/60Hz cycle and provide all the power to drive the amp in this short time so the charge current can be 20 or more times the steady DC current. These currents will generate larger voltages than you might expect and you do not want to contaminate your ground reference with this pulsing buzz. Microvolts here get amplified by your gain to be audible at the output transducer. Interaction with other frequencies will create a mush in the noise floor that is avoidable.
a physical separation of the PSU common from the star ground is essential. The length and/or resistance of this link is not important but for good regulation smaller is better.
The reason for separation is to ensure the hi current charging spikes that circulate between the rectifiers and the smoothing caps do not generate a voltage across the central ground.
The spikes charge for a small part of the 50/60Hz cycle and provide all the power to drive the amp in this short time so the charge current can be 20 or more times the steady DC current. These currents will generate larger voltages than you might expect and you do not want to contaminate your ground reference with this pulsing buzz. Microvolts here get amplified by your gain to be audible at the output transducer. Interaction with other frequencies will create a mush in the noise floor that is avoidable.