6-channel-amplifier-grounding

[Mark Whitney]

Something like this should work. Made with TINYCAD, it is easy to learn and best of all it is FREE.

 

[AndrewT]

I guess there would be nothing wrong not using the center tap at all...
And then the all signal GNDs could/should go directly to these "floating" PS GNDs in the middle of the PS caps.

NO !
the centre between the smoothing caps has very high current charging pulses. You must NOT connect any point that requires a clean voltage refernce to this charging location.

But please say (or rather show an evidence) if this layout has anything wrong because I don't see any...
On the contrary: this might be the best solution regarding Matt's given circumstances.

 

[AndrewT]

Hi All,

Thanks for the contributions.

Andrew T please can you confirm if I have understood you correctly, my diagram is based on your comments.

Also please could you confirm if capacitor busbar commoning should be employed across channel sets as in the diagram? (commoning is badly off centre so all wires can be seen, but I would common properly at the centre).

I am also unsure where on each T the separate CT connections from Toroid should be made? I assume it should be as "1"..?

Lastly is this the most suitable method of connecting the CT to PE, or should it be made from the capacitor bus bar block?

Thanks

Almost right.

The Zero volts link (cap bridging) between the two sets of smoothing caps is not needed.

If I understand your inset diagram, you are asking where to place the OUTPUT zero volts to take to the amplifier.

Put it at the output side of the capacitors. 3 or 4 seems to be the output side.

You are showing twisted pairs or triplets everywhere.
That covers the wiring interconnections.
The icing on your cake is where the twisted wiring splits to make connections to terminals. Keep the LOOP AREA at the terminals as small as possible.

 

[Cortez]

NO !
the centre between the smoothing caps has very high current charging pulses. You must NOT connect any point that requires a clean voltage refernce to this charging location.

Sorry Andrew, but as I see that's just a (baseless) fear but not an evidence...

1. With the proper "T-shape" GND layout this is not a problem at all.
2. What's the difference not using the center tap with the case attached?

I also use a double rectifier PS without any problem and with 0 hum in my amp.
If you check it, it's the same layout when not using the center tap but just
the common point of the PS caps as a GND reference...
(The snubbers aren't necessary of course...)

 

[Mark Whitney]

Cortez, you still have a 0V that goes to each amp and should also be connected to chassis.

 

[Cortez]

But each "0V" is independent from eachother (means no galvanic GND loop).
Chassis: why? The chassis should be connected to the PE and that's all...

 

[Mark Whitney]

What happens when the transformer insulation fails?


Plus all exposed parts must be connected to PE. Speaker and RCA are also connected to 0V and therefore should be connected to PE.

It won't make the amp work any better, but it will make it safe.

 

[Cortez]

Use quality trafos and then this never happens...
And what about the "+" speaker terminal?
In this basis you should also connect it to the PE.

 

[Mark Whitney]

Cortez, best to study the regulations before giving advice. You don't want anyone to get hurt.

 

[Cortez]

If you don't use the center tap, there isn't any galvanic connection to the trafo just via
the output devices and the rectifier bridge but that's the same case like with any amplifier
being the "+" speaker output obviously not connected to the PE. But correct me if I am wrong...

 

[AndrewT]

Sorry Andrew, but as I see that's just a (baseless) fear but not an evidence...

1. With the proper "T-shape" GND layout this is not a problem at all.
2. What's the difference not using the center tap with the case attached?

...............

Be "sorry" and continue to be wrong.

GNDs in the middle of the PS caps

Taking any/all of the voltage references to one of the noisiest locations in the whole amplifier is just plain wrong !
Do NOT use the smoothing capacitors as a voltage reference.

 

[Cortez]

Whatever you choose as your reference, it can't be noisy.
As I said I see the case with the double rectifier layout basically the same.
And as long as it's true and my amp is quite I think it's a good solution...

 

[AndrewT]

Whatever you choose as your reference, it can't be noisy..................

That's a corker !

 

[DF96]

Why not? The common point at the PS caps (without connecting it to the trafo center tap)
gives a GND reference point. What's wrong with using that as a normal GND point?

You can't use a point where capacitors join as a DC reference point. You can't draw DC from a capacitor. You need to do some more reading about PSUs before you advise others.

 

[Cortez]

Why not? Please explain...
What's the difference comparing to the double rectifier
PS layout (which works perfectly, of course)..?

 

[Cortez]

Ok, I simulated, I see now: if there is a significant DC current then it pulls just one side so
the floating GND (or the +/- rails) will be asymmetric. Sorry guys, I didn't get it for 1st...!

 

[Cortez]

Then this is the final recommendation, right?

Summary

• for every PS a wire triplet goes using the trafo center tap as GND
• this triplet should tightly twisted all the way long till the IPS PS
• on the amp PCB the signal GND should be connected to the PS GND via a 15R resistor
• the chassis should be connected to the PE directly
• the trafo center tap should be connected to the PE via a 15R + rectifier combo
• both the signal wires from RCA to amp PCB
  + loudspeaker wires from amp PCB to terminals should be also twisted tightly
• when grounding to the local cap bank PS GND use the "T" rules

(Poor Matt: sorry for offing your thread budy! )

 

[AndrewT]

That looks OK.
It works for mono and can be extended to two channel.

I have not built multi-channel, so have never tested using 3 or more channels.
But the philosophy should extend to many channels.

It comes from understanding why a mono must be built and wired to minimise interference.

DF makes the point regularly that unbalanced signal should use a coaxial cable instead of twisted pair. The Builder has the right to choose.

Build a mono and come back and tell us if you have become a convert, or give evidence of why that wiring style is wrong.

 

[DF96]

on the amp PCB the signal GND should be connected to the PS GND via a 15R resistor

I don't understand the apparent distinction you appear to be drawing between signal ground and PSU ground. They are the same, aren't they?

the trafo center tap should be connected to the PE via a 15R + rectifier combo

The secondary CT would normally be one of the worst places to connect to the chassis (note: chassis, not PE) but I suppose in this design it is the only common ground so there is no alternative.

If I were building this I would have one PSU with one set of reservoir capacitors (put six in parallel if necessary to get the required value), then take six DC feeds from this to local decoupling/smoothing caps on each amplifier board. This makes sure we don't kid ourselves that we have six separate PSUs.

 

[matt09]

Almost right.

The Zero volts link (cap bridging) between the two sets of smoothing caps is not needed.

If I understand your inset diagram, you are asking where to place the OUTPUT zero volts to take to the amplifier.

Put it at the output side of the capacitors. 3 or 4 seems to be the output side.

You are showing twisted pairs or triplets everywhere.
That covers the wiring interconnections.
The icing on your cake is where the twisted wiring splits to make connections to terminals. Keep the LOOP AREA at the terminals as small as possible.

Sorry for my lack of clarity, I actually meant where should I connect the 0V centre tap from the main transformer on the T?

I'm relieved the capacitor bridging isn't required at this would be very hard to implement across 6 sets!

Is the CT to PE correct?

Thanks