Can I confirm why there is the resistor lifting the signal ground from chaasis/power ground?
I've read a lot of this stuff about grounding to remove hum, and found myself creating the star by natural instinct. But what I believe is happening in that star is similar to a balanced antenna or resonant cavity which oscillates against itself, thus canceling a good deal of the 60 cycles.
Another take from all this is that there is little I can do about a computer audio card whose connector goes to a computer power supply with a 3-prong plug. Indeed, I have no intention of cutting the chassis safety PE ground on the computer. So the net resuslt is use a pull-up resistor from the ground of the chassis to the signal ground input from the RCA plugs coming in.
Do I sound like an amateur, or what?
Another take from all this is that there is little I can do about a computer audio card whose connector goes to a computer power supply with a 3-prong plug. Indeed, I have no intention of cutting the chassis safety PE ground on the computer. So the net resuslt is use a pull-up resistor from the ground of the chassis to the signal ground input from the RCA plugs coming in.
Do I sound like an amateur, or what?
It's much simpler than that
You have to think in terms of "volt drops across conductors" and whether your method of connection will cause a portion of that "volt drop" to find it's way into an audio circuit. It's nothing more than ohms law really.
Second to that is when a "loop" exists that can have current induced into it from say a transformer magnetic field.
It helps to just think in terms of DC currents. If you imagine supplying 10 amps into 10 ohms as a load for an amplifier, then put in values for all the wiring to and from the PSU (Use 1 ohm say) and then "make the wiring "disappear" through good layout and routing.
there are three 'ground' in a decent designed amp.number one is safe ground from the chassis to the outlet protected field (E),number two is power ground from the regulated power ground to the chassis.number three is signal ground.if connect the signal ground directly to the power ground,then it will share the safe ground too.the protected field is not clean.
the resistor was set up between the signal ground and power ground to avoid safe ground noise.
zang
Sound Card & AMP Ground Feedback
I got a pair of these TK2050 Class D amplifiers of Ebay from Sure Electronics. They are connected to a 7.1 channel computer audio card - the Asus Xonar D2X.
I have a single switched power supply too. I am observing the hot neutral and earth ground on it using a 3-prong power plug. Its a 27volt 37A Meanwell PS.
I got most of the hum out, but why can't I get it all? I ran separate power lines - 16ga wire - from the PS terminals to each amp. For signal grounding, there is connector wire soldered to the the RCA shields of 8 recepticals. They are joined only to themselves.
One thing I did that may seem peculiar is mount each amp and the PS in separate cookie tins. I did this for cham, but I cannot believe that would be the cause of this humming problem.
Speaker out terminals are not allowed to connect to ground on this amp. Speakers are Koss and are magnetically shielded. Maybe I need shielded speaker wire?
Pictures:
I got a pair of these TK2050 Class D amplifiers of Ebay from Sure Electronics. They are connected to a 7.1 channel computer audio card - the Asus Xonar D2X.
I have a single switched power supply too. I am observing the hot neutral and earth ground on it using a 3-prong power plug. Its a 27volt 37A Meanwell PS.
I got most of the hum out, but why can't I get it all? I ran separate power lines - 16ga wire - from the PS terminals to each amp. For signal grounding, there is connector wire soldered to the the RCA shields of 8 recepticals. They are joined only to themselves.
One thing I did that may seem peculiar is mount each amp and the PS in separate cookie tins. I did this for cham, but I cannot believe that would be the cause of this humming problem.
Speaker out terminals are not allowed to connect to ground on this amp. Speakers are Koss and are magnetically shielded. Maybe I need shielded speaker wire?
Pictures:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
When you say higher level I assume you mean because there are two separate grounding stages for the outputs and the star ground?
Is the diagram correct for a safe and ground loop free design or do I have to move all the output ground points to the star ground too?
Sorry the diagram is so poor I went through a few revisions myself as I was sketching it.
Thanks
Is the diagram correct for a safe and ground loop free design or do I have to move all the output ground points to the star ground too?
Sorry the diagram is so poor I went through a few revisions myself as I was sketching it.
Thanks
Hi Matt
Since you have mono-blocks each PA will have its own star at its individual PS ie at the output caps common point. Note for most designs this means the speaker return will tie at this star as well, instead of the PA.
Then each PA star will connect to chassis forming another star at the disconnect network.
The preamp should connect to its own power supply and input/out depending on the circuit/PCB layout. Usually this means PRE outs to PA signal gnd/ separate from the PA gnd star. PRE ins should be direct to isolated connectors. Use twisted wires or shielded cable for signals.
Since you have mono-blocks each PA will have its own star at its individual PS ie at the output caps common point. Note for most designs this means the speaker return will tie at this star as well, instead of the PA.
Then each PA star will connect to chassis forming another star at the disconnect network.
The preamp should connect to its own power supply and input/out depending on the circuit/PCB layout. Usually this means PRE outs to PA signal gnd/ separate from the PA gnd star. PRE ins should be direct to isolated connectors. Use twisted wires or shielded cable for signals.
Thanks for the reply, after looking through every stage carefully it looks as if I don't have any choice as the power supply for the regulator effectively grounds all four power amps together anyway.
Should I make a 4 link ground path on the board otherwise they will be grounded down the wires of the regulator?
I also have the same problem with the B1 buffer unless I went to extremes of making totally separate supplies for every part to ground stuff separately.
I am using two large transformers for each channel pair which are centre tapped so this already grounds two boards also.
Can anyone suggest what should be done?
Thanks
Should I make a 4 link ground path on the board otherwise they will be grounded down the wires of the regulator?
I also have the same problem with the B1 buffer unless I went to extremes of making totally separate supplies for every part to ground stuff separately.
I am using two large transformers for each channel pair which are centre tapped so this already grounds two boards also.
Can anyone suggest what should be done?
Thanks
Last edited:
In a monoblock there is one channel and one audio ground. Easy. The rules allow the audio ground to be located on the chassis or on the amp PCB or at the speaker return terminal or floating in space between all the main components of the amplifier.
Going to multi-channel does not changes the rules.
Each amplifier must have an audio ground. All the parts of that one channel must be referenced to that audio ground.
The problem is that the implementation forces one to adopt a common Safety Earth and this compromises the audio ground.
I see two choices.
a.) keep a separate audio ground for each channel. This demands a pair of isolated transformer windings for each channel.
b.) share a common audio ground among all the channels.
This can be a master star ground serving the PSU and chassis and the 4 audio grounds, plus 4 audio star grounds serving their own amp
or
it can be a master shared audio ground. This requires all the return routes for all the individual currents to have a wire back to the master shared audio ground.
That makes up 3 alternatives.
A giant star ground for everything
or
A star with 4 star off it
or
4 isolated channels each with their own isolated PSU and transformer windings.
BTW.
I agree with Infinia.
A receiver does not need a buffer, particularly if it is competently designed.
A transmitter may benefit from a buffer if it cannot source sufficient current to drive the cables and receiver.
Going to multi-channel does not changes the rules.
Each amplifier must have an audio ground. All the parts of that one channel must be referenced to that audio ground.
The problem is that the implementation forces one to adopt a common Safety Earth and this compromises the audio ground.
I see two choices.
a.) keep a separate audio ground for each channel. This demands a pair of isolated transformer windings for each channel.
b.) share a common audio ground among all the channels.
This can be a master star ground serving the PSU and chassis and the 4 audio grounds, plus 4 audio star grounds serving their own amp
or
it can be a master shared audio ground. This requires all the return routes for all the individual currents to have a wire back to the master shared audio ground.
That makes up 3 alternatives.
A giant star ground for everything
or
A star with 4 star off it
or
4 isolated channels each with their own isolated PSU and transformer windings.
BTW.
I agree with Infinia.
A receiver does not need a buffer, particularly if it is competently designed.
A transmitter may benefit from a buffer if it cannot source sufficient current to drive the cables and receiver.
Last edited:
Thanks for the reply,
Are there any drawbacks to having one master star ground for all channels as oppose to separate grounds?
At the moment it looks like this is the only possible implementation from looking at the design and your list.
The first problem is I was originally building the old style gain clone so bought 2x25vac x2 transformers, this means to use it with the new design I have to use them as a two lots of centre tapped outputs forcing the grounds of two amps together.
I am using a single super regulator for all four opamps across all channels as the load is tiny, this again however uses a single ground effectively now joining the two pairs of amps together resulting in a single ground. Surely it isn't worth building 4 separate supplies and getting different transformers just for the purpose of separate grounds?
It is noted in the builders thread that a buffer is highly recommended as the amp favors a stable source of constant impedance regardless of the receiver. I also want it to control the gain anyway. The buffer commons all 4 input and output grounds together and is on a separate board.
Should I have the ground strip as I have implemented below or leave the buffer connection to common the channels, or have both?
Many thanks.
Are there any drawbacks to having one master star ground for all channels as oppose to separate grounds?
At the moment it looks like this is the only possible implementation from looking at the design and your list.
The first problem is I was originally building the old style gain clone so bought 2x25vac x2 transformers, this means to use it with the new design I have to use them as a two lots of centre tapped outputs forcing the grounds of two amps together.
I am using a single super regulator for all four opamps across all channels as the load is tiny, this again however uses a single ground effectively now joining the two pairs of amps together resulting in a single ground. Surely it isn't worth building 4 separate supplies and getting different transformers just for the purpose of separate grounds?
It is noted in the builders thread that a buffer is highly recommended as the amp favors a stable source of constant impedance regardless of the receiver. I also want it to control the gain anyway. The buffer commons all 4 input and output grounds together and is on a separate board.
Should I have the ground strip as I have implemented below or leave the buffer connection to common the channels, or have both?
Many thanks.
Attachments
How about posting your schematics since you are designing your PCB. Showing a layout w/o them is pointless.
FWIW Ifn it were me I would just buy the chip-amp raw PCBs from some online vendors. Also some on ebay look good, cost is about $2-3 USD each. That way you can buy the good parts and stuff them as needed. Maximum flexibility electrically and mechanically! If you join them all on a single plane it does not buy you anything but possibly neater wiring, but only if you get it right (grounds) the first time.
FWIW Ifn it were me I would just buy the chip-amp raw PCBs from some online vendors. Also some on ebay look good, cost is about $2-3 USD each. That way you can buy the good parts and stuff them as needed. Maximum flexibility electrically and mechanically! If you join them all on a single plane it does not buy you anything but possibly neater wiring, but only if you get it right (grounds) the first time.
F5 transformer hum
I have built F5 amplifier with two power supplies and a separate box for the 2 amplifier PCBs. My problem is a transformer hum if both power supply umbilical cords are inserted. So one channel is fine but if you insert second cord the hum starts. I have followed the wiring diagram in post 54 by digi01 in this thread. How should I change the cofiguration in my case?
regards
Iivari
I have built F5 amplifier with two power supplies and a separate box for the 2 amplifier PCBs. My problem is a transformer hum if both power supply umbilical cords are inserted. So one channel is fine but if you insert second cord the hum starts. I have followed the wiring diagram in post 54 by digi01 in this thread. How should I change the cofiguration in my case?
regards
Iivari