Hi @Drbulj
Here is what I would do.
There are benefits that let the signal ground floating.
1. It breaks the ground loop.
2. The ripple on the remote signal ground helps the amp cancel out the same ripple on the +input. If you shunt signal ground to earth, you only got ripple on the +input. That will cause the ripple got amplifiered.
Here is what I would do.
There are benefits that let the signal ground floating.
1. It breaks the ground loop.
2. The ripple on the remote signal ground helps the amp cancel out the same ripple on the +input. If you shunt signal ground to earth, you only got ripple on the +input. That will cause the ripple got amplifiered.
Last edited:
Amp that I showed previously had diagram like this, with 10 ohm "lifter" and it hummed like fridge.
10 ohm is also effectively added to resistor Rg. Now SGND, a feedback correction point, sees more of dirty signal from source ground than from speaker return (what it should measure)
Dirty ground signals and ground difference potentials now must pass through SGND and will only cancel if the same dirty signal exists in signal wire, but signal wire is shielded, and shield is not... Than better use twisted wires on RCA cable, no shielded cable.
Ground loops we cannot avoid unless we use class2 devices without safety earth (and open bigger can of worms). Differential topology is right way to get rid of it. But if single ended RCA must be used (whomever invented it should be charged for waste of time of millions 🙂) only way is to avoid these dirty loops going through 2 nodes of amp input, signal and SGND. Better to channel them through chassis.
Anyway, do it as you propose, but I believe with some sources you will still have hum, with others not (like battery powered, or class 2 to lesser extend). Plus 10 ohm lifter will slightly degrade THD.
10 ohm is also effectively added to resistor Rg. Now SGND, a feedback correction point, sees more of dirty signal from source ground than from speaker return (what it should measure)
Dirty ground signals and ground difference potentials now must pass through SGND and will only cancel if the same dirty signal exists in signal wire, but signal wire is shielded, and shield is not... Than better use twisted wires on RCA cable, no shielded cable.
It is not always floating because in single ended source (I assume also source device is single ended) this ground is attached to safety earth and is making ground loop via power cords (this if source is Class1 device , grounded one), just this time via sensitive SGND and not through chassis as I drafted. This diagram will be good if source device is equipped with transformer or transformer - like circuit. Than we are back to differential topology. Could also be not to bad if source device is not grounded through power cord (class2 device) . Many are like that so it could work. But too many if-s.
Ground loops we cannot avoid unless we use class2 devices without safety earth (and open bigger can of worms). Differential topology is right way to get rid of it. But if single ended RCA must be used (whomever invented it should be charged for waste of time of millions 🙂) only way is to avoid these dirty loops going through 2 nodes of amp input, signal and SGND. Better to channel them through chassis.
Anyway, do it as you propose, but I believe with some sources you will still have hum, with others not (like battery powered, or class 2 to lesser extend). Plus 10 ohm lifter will slightly degrade THD.
I have a question about the ground lift resistor. Maybe this is the right place.
•When both input and feedback ground are lifted together through a 10 ohm resistor like shown in post #41
•and the preamp is earthed.
Now isnt the lowest impedance path from the feedback resistors to the negative speaker terminal going through the earth wires?
Should this path from the feedback resistors to the speaker terminal not be as direct as possible?
Or am I thinking about this the wrong way?😂
•When both input and feedback ground are lifted together through a 10 ohm resistor like shown in post #41
•and the preamp is earthed.
Now isnt the lowest impedance path from the feedback resistors to the negative speaker terminal going through the earth wires?
Should this path from the feedback resistors to the speaker terminal not be as direct as possible?
Or am I thinking about this the wrong way?😂
Yes it does, even worst it goes through RCA shield, via signal device, its power cable , wall socket and back to this amp in question via its power cord.... Nice route to connect speaker return and amp feedback, but this path still has less impedance than 10R
🤣 🤣 🤣 🤣
Yes, and you are thinking right!
Do the benefits of a ground lift resistor remain if only the input is lifted and Not the feedback ground path?
I suppose is would still help breaking earth loops
I suppose is would still help breaking earth loops
Some people put 10 ohm and small capacitor between RCA and chassis, at place where jx put red X in post #41.
IMHO what I drow is best, post #39, hard wire to chassis. If you leave resistance in, voltage will develop over it and it is directly at your negative amp imput (aka SGND). Capacitor will short RF, but not 50Hz. Read 2 papers posted in #3
But it is not end of story.
Most hum is coming from inside, powerful trafo, and less from ground loops.
"modern" layouts with trafo in middle and 2 RCA each in their opposite corner make my hair go vertical... Same with stereo cables being completely separated... Somehow many actually want to make antennas out of their wires 😅
This looks like a similar technique to the one Douglas Self discusses in one of his books. You still have to take care of loop area issues and common impedance coupling on the PCB's etc, but it seems it produces good noise results based on what he said. I have not tried this myself BTW.
Hi, this wiring I learned from pro gear, studied a lot while fighting dreadfull hum. All my early DIY amps had some hum. Then switched to differential and forgot it. Than started playing with some single ended rebuilds and rediscovered hot water.
Yes, Douglas Self is using it, and it does work, with lots of care how to rout cables and separate PSgnd (for decoupling caps and eventually spk return) and SGND on pcb.
SGND is just point of connection between Rin and feedback resistor (or capacitor) to input cable shield and spk return, should be easy to separate it from PSgnd
Another nice article is from Tom
https://neurochrome.com/pages/grounding
In chapter An Improved Grounding Scheme he nicely with measurement explains why is SGND last gnd and should be connected to SPK return terminal.
This was kind of skipped in D. Self book.
https://neurochrome.com/pages/grounding
In chapter An Improved Grounding Scheme he nicely with measurement explains why is SGND last gnd and should be connected to SPK return terminal.
This was kind of skipped in D. Self book.
This only works well when the signal source is floated. It could not mitigate ground noise if the signal source is also grounded.
Also, gain loss won’t be necessarily translated to THD. If you have output inductors, you’ve got gain loss anyway.
It pretty much what is happening, but I don’t think there is anything wrong with it. When you have ground loop, this the price to pay to get reasonable noise rejection. The first option is to remove the ground loop, such as removing cable tv connection. When I try to integrate the amp into my AV theater system, at that point, breaking ground loop is not possible.
Last edited:
Hi @jxdking, I will give up talking in your thread, thanks for accepting my comments so far.
For end I would like to point few facts about electrons flow, all related to Ohm's law:
1. Current flow must be closed circuit
2. Current will always find way to flow where there is less impedance, proportionally to flow where is more impedance.
2. Every cable or pcb trace, or junction has some impedance, sometimes very small but it's there
3. Any current flowing through impedance will develop voltage
4. Many different signals can flow in the same cable (issue with RCA termination where Gnd and signal return are in same wire) but they can be separated from each other as all signals need return paths. Each signal in multiple of signals introduced in one copper wire will follow its return path to close the circuit.
Think about it, introduce small impedance (resistor if it is simpler) in every trace or cable, and follow the currents that must return to the source.
For end I would like to point few facts about electrons flow, all related to Ohm's law:
1. Current flow must be closed circuit
2. Current will always find way to flow where there is less impedance, proportionally to flow where is more impedance.
2. Every cable or pcb trace, or junction has some impedance, sometimes very small but it's there
3. Any current flowing through impedance will develop voltage
4. Many different signals can flow in the same cable (issue with RCA termination where Gnd and signal return are in same wire) but they can be separated from each other as all signals need return paths. Each signal in multiple of signals introduced in one copper wire will follow its return path to close the circuit.
Think about it, introduce small impedance (resistor if it is simpler) in every trace or cable, and follow the currents that must return to the source.
the schemes presented in the link above are "star" and "buss"
with "buss" you can choose in which order to return to psu
the schemes that are shown here and elsewhere only show a single amplifier - the problems start when you wire up 2 or more channels
Attachments
This guy below knows what he is doing. This will be my new go-to guideline.
https://hifisonix.com/wp-content/uploads/2019/02/Ground-Loops.pdf
https://hifisonix.com/wp-content/uploads/2019/02/Ground-Loops.pdf