My designs are always noisy :(

Hardware questions and issues with the FV-1

Moderator: frank

iampoor
Posts: 27
Joined: Sat Nov 15, 2014 4:44 pm

Post by iampoor »

Okay, so Im convinced the problem was my layout. Nothing else would seem to make sense at this point. I have added some additional pads for bypass caps (just in case) on the plugin board, and increased the number of via's between the ground planes. Some of the pins on my original PCB were not connected very well to the bottom ground plane, an I have tried to keep the bottom ground plane as solid as possible.

Here is my layout and schematic. Anything else wonky or suggestions? I will be ordering these in a few days. :D


http://imgur.com/a/TKN6M
frank
Posts: 1244
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank »

Would be easier to look at if each layer were posted separately, can't realy see if there are any issues on the bottom layer.
Frank Thomson
Experimental Noize
iampoor
Posts: 27
Joined: Sat Nov 15, 2014 4:44 pm

Post by iampoor »

frank wrote:Would be easier to look at if each layer were posted separately, can't realy see if there are any issues on the bottom layer.
Here you go. Let me know if you would like anything clarified!

Thank you for looking this over and being so helpful!

http://imgur.com/a/nxfef
frank
Posts: 1244
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank »

Well, I would get rid of the ground pours on the top layer, better to have one good ground plane.

It looks like you may have some broken current return paths. See the pdf at
www.ti.com/lit/ml/sloa089/sloa089.pdf on page 17-15 it appears you have things like "Version B" and you reall want to look like "Version C" where the current return path is directly under the signal trace. Also see http://www.ultracad.com/articles/loop.pdf for more info.
Frank Thomson
Experimental Noize
iampoor
Posts: 27
Joined: Sat Nov 15, 2014 4:44 pm

Post by iampoor »

frank wrote:Well, I would get rid of the ground pours on the top layer, better to have one good ground plane.

It looks like you may have some broken current return paths. See the pdf at
www.ti.com/lit/ml/sloa089/sloa089.pdf on page 17-15 it appears you have things like "Version B" and you reall want to look like "Version C" where the current return path is directly under the signal trace. Also see http://www.ultracad.com/articles/loop.pdf for more info.
Great! Thanks for the info. How do these look?

There are a few breaks in the plane on the bigger board should I try through hole jumpers? Or does it look sensible? I tried to minimize the length of any traces on the bottom side, and not put them in super obvious current return paths.

http://imgur.com/a/1jEIM
frank
Posts: 1244
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank »

That looks much cleaner. You will almost always have to put in some break in the ground plane to route signals but as long as you try to avoid the current return paths, avoid running signals in parallel to each other (watch for that between layers), etc. you can usually get a good clean PCB.
Frank Thomson
Experimental Noize
iampoor
Posts: 27
Joined: Sat Nov 15, 2014 4:44 pm

Post by iampoor »

frank wrote:That looks much cleaner. You will almost always have to put in some break in the ground plane to route signals but as long as you try to avoid the current return paths, avoid running signals in parallel to each other (watch for that between layers), etc. you can usually get a good clean PCB.
Great! I will order it and report back. I hope my layout was the problem, it would be nice to finally figure this out. The other pedals I have built havent been particularly sensitive layout wise, so this has been a good experience! :D
dientulke
Posts: 8
Joined: Thu Oct 16, 2014 9:34 pm
Location: Vietnam
Contact:

remove noise

Post by dientulke »

note at red color circle. insert a 100 ohm/1W resistor between VCC and IC4 ( 3.3V ).
noise can be removed.
Image

http://imgur.com/0IJlQk2
Post Reply