Board layout and single-point grounding

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

zapnspark

Well-known member
Joined
Feb 17, 2005
Messages
107
Location
Midwest
From EDN. A good perspective on why single point grounding isn't necessarily the best design approach.

http://tinyurl.com/bnwhlq

Cheers.

ZAP
 
zapnspark said:
From EDN. A good perspective on why single point grounding isn't necessarily the best design approach.

http://tinyurl.com/bnwhlq

Cheers.

A continuous ground plane with no cuts is your friend.

-a
 
Thanks. Good read.

Is he saying narrow traces are better for both hi and low gain signals?  It sounds contradictory.


"The smaller the signal, the smaller you want the node (Microphone inputs are very small signal)
The higher the node impedance (Z), the smaller you want the node.
The higher the gain from the node, the smaller you want the node. (By small node I mean narrow, short trace)"


 
Area.. more likely short than narrow.

High impedance and high gain nodes are like antennas to pick up noise. Reduce the size of the antenna, reduce the noise pickup.

JR
 
What makes sense to me is that before gain is applied to signal you would want traces narrow and short.  So if there are traces like antennae then RF pickup will be minimized.  right?

But, say the output of a Tube preamp has much current and Voltage gain.  The signal is clearly less susceptible to noise so I would think fat traces would be fine.  But the problem with fatter traces seems like it would be crosstalk radiation, and that would be an argument for thin traces.

-I used to think bigger as better but It's finally making sense to me why Mic pre Input XFMRS have such thin wires.
 
This article caught my attention because I just finished a battery powered mic pre-amp.
Designed for phantom powered mic. field recordings with various SD card recorders.
Everything was designed "rule-of-thumb", "by the book", single point grounding.
The measurements looked excellent until -- my cell phone went off.
ZZZZTTT, Bzzzfft.
Back to the drawing board.

Cheers.

ZAP
 
zapnspark said:
The measurements looked excellent until -- my cell phone went off.
ZZZZTTT, Bzzzfft.
Back to the drawing board.

Update.

The solution turned out to be simple -- for a change.
I used metal standoffs for the mic. pre-amp board.
One of the screws holding this board was off center and contacting the pre-amp ground plane.
That created a ground loop through the chassis.
I went to nylon standoffs + screws and the problem has been solved.

ZAP
 
I see no apparent benefit to skinny traces, they should be sized based on current requirements. Length and where you put them OTOH impacts their antenna behavior. Just like antennas, where they are relative to any ground plane also matters.

Skinny wire in transformers is more likely to fit a lot of turns on a given core, and testament to the small currents involved with low level signals.

It is not just what the voltage or current is, but what the node impedance is that should suggest care and handling to avoid noise pickup. A high voltage high impedance node can still be susceptible to noise pickup and crosstalk.  This is pretty apparent in some guitar amp designs.

JR

PS: FWIW I recall one amp designer arguing that he intentionally used narrow traces to force discipline on his PCB fab house. I dismissed his explanation as a rationalization. If serious he could have printed a tapered pattern to determine fab trace resolution. I have seen such patterns but not very common or useful these days.


 
JohnRoberts said:
PS: FWIW I recall one amp designer arguing that he intentionally used narrow traces to force discipline on his PCB fab house. I dismissed his explanation as a rationalization. If serious he could have printed a tapered pattern to determine fab trace resolution. I have seen such patterns but not very common or useful these days.

...maybe time to get a new fab house???  ::)

With respect to node impedance, I sometimes have trouble identifying high impedance nodes. Some seem straightforward, FET gates (when bias Rs are large), opamp non-inverting inputs (depending on parallel R on input)... Are there any general rules or techniques? I've definately had to troubleshoot oscillating circuits @ board level and discovered sensitive nodes through trial/error and experimentation. There's gotta be a smarter way to do it...

ian
 
Petty straightforward for some things.. opamp outputs are low impedance, both inputs are high impedance, while NF may make - inputs look like low impedance, any noise there, ends up in your output.

With active devices it can depend on how used.  a collector or drain is always high Z, base or gate high Z, emitter-source in a follower can be low Z, but in common base designs an input.

You'll figure it out, just look at the circuit and what everything is doing..

JR
 
Back
Top