Ground pplane

[gootee]

Originally Posted by epilot

Hi, and thanks.

What do you really mean by "conductor pairs"? I though that the enclosed loop problem is just for ground or each trace itself, now you seem to say that it happens between pairs too (for instance between VCC trace and ground)!
Furthermore plz how it acts as an antenna!?

It would be ok if you could plz show me what you are meaning by a pic...

Thanks again

Google Faraday's Law.

In practical terms, for the layout you gave, just make the distance between the power rails and the large copper ground area as small as possible, everywhere. And twist tightly together any wire pairs that go from or to the board, and everywhere else in your power supply and other circuits.

It will also be very important for things like input signal and signal ground pairs. Either twist tightly together, all the way from input jack to input resistor, or, even better, use shielded twisted pair with shield connected to chassis (NOT signal ground to chassis!) at input end and not connected to anything at other end. Signal ground conductor connects ONLY at the input and at the ground end of the input resistor of the first amplification device (which then has a conductor of its own that goes to the main star ground point. i.e. the input jack (both signal and signal ground portions), must be completely isolated (electrically insulated) from the chassis.

[gootee]

Originally Posted by epilot

But plz google the schematics for the 78/79xx regulators. MOST of them have big caps at their outputs....

And you would believe "most schematics" instead of believing the manufacturer's datasheet?!

You should probably read several different linear regulator datasheets from each of several different manufacturers. ALSO, search their websites for Application Notes about linear regulators. Sometimes, they might say that you could use a somewhat-larger capacitor at the regulator's output, but they almost always mention, in some way, that it shouldn't be "too large".

Unless you have a proper mathematical justification, just follow the recommendation on the manufacturer's datasheet for the exact device that you are using.

If you are going to use a capacitor on the load side of a regulator, then you would want to place it as close as possible to the point of LOAD, where it can do the most good. And often or usually you would want to have both large and small capacitors at the point of load, e.g. some electrolytic and some small ceramic such as 0.1 uF. ("What C values?", you ask. Ideally, the capacitors' self-resonant frequencies will cover the frequencies of the currents demanded by the load. Caps' self-resonant freqs depend on C values and caps' parasitic inductances and the trace's parasitic inductance. If in doubt, first try 10 uF electrolytic in parallel with 0.1 uF X7R ceramic, at each chip power pin, to ground. Larger current demands would need larger than 10 uF. Higher frequencies would need lower than 0.1 uF.)

But still you have to be careful that none of those capacitances will form an unwanted high-frequency resonance with either the trace inductance or the inductance of any of the capacitors. Lossy capacitors are sometimes best.

[marce]

http://www.x2y.com/filters/TechDay09...%20_JohnWu.pdf

http://www.elmac.co.uk/pdfs/Lord_of_the_board.pdf

[gootee]

Originally Posted by marce

http://www.x2y.com/filters/TechDay09...%20_JohnWu.pdf

http://www.elmac.co.uk/pdfs/Lord_of_the_board.pdf

Perfect! Thanks for these, marce!

[gootee]

Originally Posted by epilot

Hi, and thanks.

What do you really mean by "conductor pairs"? I though that the enclosed loop problem is just for ground or each trace itself, now you seem to say that it happens between pairs too (for instance between VCC trace and ground)!
Furthermore plz how it acts as an antenna!?

It would be ok if you could plz show me what you are meaning by a pic...

Thanks again

You PCB would be much better if you would use a two-sided board and use the entire bottom side as a ground plane. You would only need to etch small areas on the bottom ground-plane side that were below a through-hole pad that would be soldered on the top side. (OK, you might need a couple of very short traces on the bottom side, for the large capacitors, unless you can angle them enough, or lay them on their sides or something, to enable soldering two of their leads on the top side of the board. And the jacks might need some short bottom-side traces to get to vias next to them, unless you used right-angle jacks.)

[magnoman]

Originally Posted by marce

http://www.x2y.com/filters/TechDay09...%20_JohnWu.pdf
http://www.elmac.co.uk/pdfs/Lord_of_the_board.pdf

marce

Thanks for the links.
Briefly went through the first, one area where I dont typically comform is the placing bypass caps on the same layer as the active parts and avoiding via's.
For me the added via inductance is still near the capacitors ideal self inductance and being able to place them elsewhere clears up the critical io routing. This can be done with ground and power planes, then I am able to place the connecting via's right at the active components power leads. Even without power planes I still find that I am able to better able to keep the supply impedance lower out to higher frequencies with large traces on the bottom side (especially since the by-pass capacitors spacing can then more easily straddle power leads, in the absence of planes).
This is just a practical situation that I seem to run into all the time.
Hope this helps
-Antonio

[AndrewT]

Good links.
Thanks.

[marce]

Henry Ott, page 458
Most of my data is in PDF form with my CAD libaries so I can access the info readily,
at the moment I have to use a DVD to back this lot up! So I am trying to do some housekeeping and checking documents for updates etc, as I have been collecting this stuff for years (including magazines such as Printed Circuit Design and Fabrication and SMT) and as I am rather disorganised its a bit of a long term project... But here is a zip of a collection of PCB related links including soem good Signal Integrity sites that you may find useful.
One day I will get organised...

[gootee]

I ordered the latest version of Henry Ott's book.

[tvrgeek]

78/79 regulators are not the best by today's standards, but that was not your question.
Let me be very contrary: There is no such thing as ground, just a place arbitrarily designated as a circuit reference point. There is no magic in shape. You should treat the source and sink with the same diligence. My habit is to use double side board and make the top all ground. This gives me more space to deal with V+ and V- on the bottom side and a cleaner, more straight, and more compact layout. ref: comment by gootee.