Gents, I'm laying out a PCB for one of my mic pre projects. I always made a full ground plane on the top layer, but recently I thought that there may be better ways.
One thought was to make a separate ground plane for PSU and the rest of the circuit. The obvious question is where do you connect them together (i. e. where do you place the star point)? I suppose there should be some difference when you make a star connection near the cap following the bridge rectifier (where ripple current is high) or at the cap on the regulator output.
Maybe I'm better off making a discrete grounding for PSU?
What you think?
PCB layout - ground plane for PSU/star point
Re: PCB layout - ground plane for PSU/star point
This is a fairly old subject but I would look at PS layout not as a shielding (ground plane) issue, but a managing current paths and IxR voltage drops.
For a regulated supply the obvious common ground node is at the ground terminals of the regulators,
Care should be spent on reservoir cap grounds and transformer center tap.
JR
For a regulated supply the obvious common ground node is at the ground terminals of the regulators,
Care should be spent on reservoir cap grounds and transformer center tap.
JR
Cancel the "cancel culture", do not support mob hatred.
- mediatechnology
- Posts: 5466
- Joined: Sat Aug 11, 2007 2:34 pm
- Location: Oak Cliff, Texas
- Contact:
Re: PCB layout - ground plane for PSU/star point
+1 on what JR said particularly the above statement.Care should be spent on reservoir cap grounds and transformer center tap.
A large ground plane in the filter cap/transformer center tap area may actually work against you since it is more difficult to define the ripple current paths.
Re: PCB layout - ground plane for PSU/star point
Thanks! That's what I tought as well.
My current plan is to use a separate PSU board that includes bridge/reservoir caps for audio and regulation for +48V. Final regulation of power rails will be done on local preamp boards (for each channel individually). I'm going to lay out a star ground point at the output of VRegs. Each individual star ground will be taken to chassis star point. Does that sound sensible?
My current plan is to use a separate PSU board that includes bridge/reservoir caps for audio and regulation for +48V. Final regulation of power rails will be done on local preamp boards (for each channel individually). I'm going to lay out a star ground point at the output of VRegs. Each individual star ground will be taken to chassis star point. Does that sound sensible?
Re: PCB layout - ground plane for PSU/star point
Sounds sensible from a distance...
If each preamp board has a separate ground reference, audio signals shared between them need to handled differentially (like balanced inputs and outputs).
JR
If each preamp board has a separate ground reference, audio signals shared between them need to handled differentially (like balanced inputs and outputs).
JR
Cancel the "cancel culture", do not support mob hatred.
-
- Posts: 17
- Joined: Mon Jan 05, 2015 7:49 am
Re: PCB layout - ground plane for PSU/star point
I've always preferred ground plane on the opposite side of the component layer to have much greater coverage with less irregularities caused by components eating up the ground plane real estate. Running signal and power on the component side with traces, no power planes. Does a star return on a PCB make any difference over a ground plane?
Don't get me wrong, if I had RF and or high speed switching logic on the same board as low noise analogue I would run a split through the ground plane and join at the ground entry connection on the PCB or two ground wires back to the ground reference if it's the chassis or power supply. Physically one PCB but electrically two PCBs with a common ground connection.
But I haven't done much probing and noise analysis, just observed that from other mixed signal and RF designers results.
Don't get me wrong, if I had RF and or high speed switching logic on the same board as low noise analogue I would run a split through the ground plane and join at the ground entry connection on the PCB or two ground wires back to the ground reference if it's the chassis or power supply. Physically one PCB but electrically two PCBs with a common ground connection.
But I haven't done much probing and noise analysis, just observed that from other mixed signal and RF designers results.
Re: PCB layout - ground plane for PSU/star point
There's loadsa issues about using ground planes for high quality audio but I don't think it can be taught in a forum like this. But do listen to what the gurus here suggest. eg JR about ripple & signal current.
My $0.02 is to make sure you have Cutouts in the Ground Plane under any tracks connected directly to OPA input pins .. including any tracks connected to the input pins via capacitors. The extra capacitance on the inputs caused by ground planes will play havoc with your stability.
I hate Ground Planes .. especially in very low noise circuits cos they muck up my careful routing of signal currents ... but there's more than one way to skin a cat .. and ultra low noise may not be on your list of priorities.
It's more important you UNDERSTAND what you are trying to achieve than to adopt some prophet's gospel.
My $0.02 is to make sure you have Cutouts in the Ground Plane under any tracks connected directly to OPA input pins .. including any tracks connected to the input pins via capacitors. The extra capacitance on the inputs caused by ground planes will play havoc with your stability.
I hate Ground Planes .. especially in very low noise circuits cos they muck up my careful routing of signal currents ... but there's more than one way to skin a cat .. and ultra low noise may not be on your list of priorities.
It's more important you UNDERSTAND what you are trying to achieve than to adopt some prophet's gospel.
- mediatechnology
- Posts: 5466
- Joined: Sat Aug 11, 2007 2:34 pm
- Location: Oak Cliff, Texas
- Contact:
Re: PCB layout - ground plane for PSU/star point
+1.My $0.02 is to make sure you have Cutouts in the Ground Plane under any tracks connected directly to OPA input pins .. including any tracks connected to the input pins via capacitors. The extra capacitance on the inputs caused by ground planes will play havoc with your stability.
The higher the feedback impedance the greater the need for those cutouts.