Pre-amplifier grounding compendium?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Member
Joined 2007
Paid Member
Hey Diy'ers,

I'm in the process of making a preamplifier and wonder if any of you knows of a thread here in DIYaudio - or elsewhere - that can be considered a "compendium" on the best routing & grounding strategies for pre-amplifiers?

I know of this article by Guido Tent of Tentlabs:

http://tentlabs.com/Components/Shuntcomp/assets/Supply_decoupling.pdf

- which is about supply decoupling and tracing in that respect, but is it e.g. wisest to place the common ground close to the input or close to the output? And what would be the reasoning for either choice?

So if any of you knows of a short, easy-to-read, yet valid "compendium" on grounding I'd appreciate hearing about it.

Spring greetings from Denmark,

Jesper
 
Administrator
Joined 2007
Paid Member
I don't know of or indeed haven't seen a compendium on this.

Every design is unique and every design MUST follow basic principles. It helps to think of it all as a DC problem sometimes. Ask yourself, "if any current were to flow in this line or that track etc, then will it alter the output of the circuit in way".

To just follow the "star ground" and connect everything to that approach doesn't work in practice... there are just to many ifs and buts.

If you can read all of this then it might give you a little insight into the problems.
http://www.diyaudio.com/forums/solid-state/101321-3-stage-lin-topology-nfb-tappings.html
 
If working on a preamp with a 2 layer pcb why not dedicate one layer for power/decoupling traces and grounds and the other for signal routing and signal ground.

Use a ground plane on each layer and then have them connect at one point/via on the board for connction to the PSU.

I have seen Borbely use this approach, although he did not use ground planes but simple traces due to the single layer.
 
I won't even try to do 16+ bit resolution instrumentation circuit designs for anybody not willing to spring for 4 layer PCB - at low production volumes it is a waste of time and money trying to squeeze the performance into 2 -layer layouts - requires more design time, often more pcb layout/prototype passes

for pre-amp level signals a single gnd plane is likely OK, possible slots/necking down to constrain power supply/return current paths

the bigger issue can be deciding what to do with source signal gnd, chassis, safety gnd, gnd loops from PS leakage currents

Whitlock's papers at the Jensen site is good for the system interconnect issues - even if he keeps ending up recommending signal isolation transformers http://www.jensen-transformers.com/apps_wp.html


there is also an article here at diyAudio: http://www.diyaudio.com/forums/diya...udio-component-grounding-interconnection.html
 
Last edited:
I won't even try to do 16+ bit resolution instrumentation circuit designs for anybody not willing to spring for 4 layer PCB - at low production volumes it is a waste of time and money trying to squeeze the performance into 2 -layer layouts - requires more design time, often more pcb layout/prototype passes

for pre-amp level signals a single gnd plane is likely OK, possible slots/necking down to constrain power supply/return current paths

the bigger issue can be deciding what to do with source signal gnd, chassis, safety gnd, gnd loops from PS leakage currents

Whitlock's papers at the Jensen site is good for the system interconnect issues - even if he keeps ending up recommending signal isolation transformers JENSEN TRANSFORMERS, INC. - APPLICATION PAPERS AND SCHEMATICS


there is also an article here at diyAudio: http://www.diyaudio.com/forums/diya...udio-component-grounding-interconnection.html

Right, so it's more difficult, but possible, but at the same time not important b/c it's likely ok. :confused:

At low, DIY volume, say one or two boards for personal use, 4 layer boards are cost prohibitive for a hobbyist such as myself.

I don't see anything wrong with a 2 layer, 2 ground plane approach for diy. No iso. transformers please.
 
if you don't insist on "professional" look - all signals routed on board, no jumpers, no bus bars, no aux boards then you don't have to go 4-layer if you don't have mixed signal MHz DSP, uC, ADC, DAC on board

imaginative use of jumpers, twisted pair pwr distribution, all Class A, low current circuitry you can do OK on a nominally 2-layer board - too many diyers try to make their board layout look "pro"/mass producable not understanding the effort, compromises that went into those layouts

in fact you can do really well on gnd plane perf board for one-off circuits if you let actual electrical performance demands guide your build instead of tyring for a "pro look"

read everything Jim Williams has written on prototyping, layout - Pease was OK for audio frequency signals
 
Last edited:
if you don't insist on "professional" look - all signals routed on board, no jumpers, no bus bars, no aux boards then you don't have to go 4-layer if you don't have mixed signal MHz DSP, uC, ADC, DAC on board

imaginative use of jumpers, twisted pair pwr distribution, all Class A, low current circuitry you can do OK on a nominally 2-layer board - too many diyers try to make their board layout look "pro"/mass producable not understanding the effort, compromises that went into those layouts

in fact you can do really well on gnd plane perf board for one-off circuits if you let actual electrical performance demands guide your build instead of tyring for a "pro look"

read everything Jim Williams has written on prototyping, layout - Pease was OK for audio frequency signals

I'll have to look up your recommendations.

Ha, no more perf-board please! I've used them for years with great success but have recently started using layout software, Eagle and ExpressPCB. Much nicer IMO.
 
Member
Joined 2007
Paid Member
Hey All,

& thanks so much for your replies ;)

I've taken a look at your comments and your links and I can see that there's a bit of reading to do - and things to ponder - so will say thanks for now, and start reading/pondering ;)

'Best for your day (evening it is here in Denmark),

Jesper
 
Member
Joined 2007
Paid Member
Hi again,

And thanks once more for sharing your knowledge and insights. I've read up on what you've suggested and have decided on star-grounding in two different places (there are two units) - similar to what is suggested in djoffe's article.

Also, depending on how noisy things become I'll put a copper shield around the interconnect wires and not distance the wires too much from eachother - inspired by the small loop area recommendation in the "HAM" article on RFI shielding.

Also, inspired by one of Bonsai's posts in one of the links you gave, I'll experiment with having the ground and supply wires, e.g. to filter caps lie close to eachother.

I have a question, though: One part of this "preamplifier" really is a microphone with adjacent circuitry, the other part is an A/D converter. All is battery powered, i.e. one battery pack near the microphone(s), another at the A/D converter. If I place a shield around the wires connecting the two parts where would then be the most feasible place to put the connecting point for this shield?

I reckon the shield would pick up high frequency noise e.g. from mobile phones etc., and my own impulse is to connect it directly - through an efficient ferrite bead - to the star ground on the A/D converter's supply and then make sure that this supply is very well filtered high-frequency wise. This already is HF area, and the microphones are very low and volatile signal levels. Does this make sense?

Insights appreciated :)

Jesper
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.