So i had an idea for a 4 layer board designating one layer as shield (or call it earth) layer that will fasten directly to the chassis, and the dc ground plane connecting to this layer at only single point. I can see how the shield layer would benefit pwr and signal traces but for a ground layer perspective wouldnt it just look like an antennae/oppportunity for a lot of parasitic capacitance? And is that actually still benefitial for audio?
Another way of asking this would be what happens to the gnd layer the closer its paralleled to the chassis?
Yes ground planes are used as a shield, also guard loops (ie with a voltage) too to adjust noise/capacitance. Etc
You could also use a mu-metal, copper clad and stainless steel boxes within each other to really reduce noise.
You could also use a mu-metal, copper clad and stainless steel boxes within each other to really reduce noise.
Thanks for that input but that doesnt really answer my question which is- when the chassis is brought as close to the signal ground as possible and a capacitance is formed, is that beneficial in any way?
well, yes in that signal to case can interact with resistances making a low pass filter - for example tube RF amps need to be aware of the distance to the sides of the case as that impacts the response (HV makes larger distances a capacitor).
You will also find that close proximity reduces noise. More layers of different metals (ie cable shield and case as shield too).
Typically the ground of the cable and the case are connected. So if there’s a an odd ground loop then you may get a capacitance to the chassis. So the schematic would define this.
You will also find that close proximity reduces noise. More layers of different metals (ie cable shield and case as shield too).
Typically the ground of the cable and the case are connected. So if there’s a an odd ground loop then you may get a capacitance to the chassis. So the schematic would define this.
Hi Donovas ... Can I ask you what the application is? Analog (up to ~100 kHz) or digital?
Cheers, Jesper
Cheers, Jesper
True, the old tube amp connect the input jack to the chassis then connect ground to it..
donovas, perhaps better to use and think of a pcb plane/pour as the dc ground for all the pcb circuitry - given that all signal voltages on that pcb would couple to the dc ground through part connections and local parasitic capacitance.
If parts on that pcb get close to chassis or chassis mounted parts, then stray capacitance may cause some signal current from a pcb part to flow to the chassis and then back to the pcb dc ground plane (given that the pcb ground plane has a link to chassis as would be anticipated). That's not ideal, so should be avoided. There effectively won't be signal currents flowing between the pcb dc ground and the chassis, because they are at the same dc potential (and ac potential for audio related frequencies) even when the pcb is separated by just a small distance, because capacitive current flow requires ac voltage. So the pcb dc ground plane/pour should be on the 'chassis' side of a multilayer pcb, and not buried inside the pcb.
If you have a setup where noise or signal leakage/coupling is occurring then you need to work at how that is occurring and find a root cause first, rather than apply a knee-jerk remedy which is not based on a good understanding.
Within a pcb there may be a need to minimise parasitic coupling to the dc ground plane for certain reasons. For example, additional stray capacitance of some nodes like valve input grids can lower high frequency roll off, which can be a stability concern in some audio circuits that need a 0dB loop gain out well past 100k to 200kHz. That can mean that coupling caps need to be purposefully raised away from the dc ground plane, and the part shouldn't sit touching the pcb, and such node traces shouldn't overlay the dc ground plane.
If parts on that pcb get close to chassis or chassis mounted parts, then stray capacitance may cause some signal current from a pcb part to flow to the chassis and then back to the pcb dc ground plane (given that the pcb ground plane has a link to chassis as would be anticipated). That's not ideal, so should be avoided. There effectively won't be signal currents flowing between the pcb dc ground and the chassis, because they are at the same dc potential (and ac potential for audio related frequencies) even when the pcb is separated by just a small distance, because capacitive current flow requires ac voltage. So the pcb dc ground plane/pour should be on the 'chassis' side of a multilayer pcb, and not buried inside the pcb.
If you have a setup where noise or signal leakage/coupling is occurring then you need to work at how that is occurring and find a root cause first, rather than apply a knee-jerk remedy which is not based on a good understanding.
Within a pcb there may be a need to minimise parasitic coupling to the dc ground plane for certain reasons. For example, additional stray capacitance of some nodes like valve input grids can lower high frequency roll off, which can be a stability concern in some audio circuits that need a 0dB loop gain out well past 100k to 200kHz. That can mean that coupling caps need to be purposefully raised away from the dc ground plane, and the part shouldn't sit touching the pcb, and such node traces shouldn't overlay the dc ground plane.