Thanks! Yeah that makes complete sense. Thanks for the tip. I know his is a complex question, but with such a small layout at ~2mhz, is here a risk of ground loops using both sides as a ground plane? Obviously I dont expect a definitive answer, I just don wan to stress out thinking about it if its not really an issue. Of course, thats what makes engineering fun, but I want this thing to work first try, dammnit! :p
Not so much the frequency as the frequency and impedance. If this were a battery powered application, it wouldn't matter (~mA switching currents). Out of 10uH, I'm guessing it's a little punchier than that (100s mA?). So all the inductance around the traces adds up.
It's not that you're using it as a ground plane. Look at it, it's cut up to all hell! Think about the current flow through the negative space, it has to go along the edge, all the way around the traces and components.
You do get local ground loops, which is usually a good thing, having a much lower impedance between circuit areas. The problem notoriously associated with ground loops (voltage drops) is improved because your ground is better.
Also, a careful and comprehensive approach can deal with ground loop at high frequencies. If there's a difference in ground voltage where the voltage leaves the board (say, if you bring the voltages out with twisted pairs at C7 and C8, respectively), it's not something you can fix with additional CLC filtering (that doesn't filter ground). For that, you need a common mode choke. Which also helps deal with noise picked up on cables, and transmitted from other loads. (Whether that's a concern or problem, depends.)
I tried adding a via under C6, but I could see that creating problems with he solder mask so maybe Ill just leave it out....
Yeah, that's a short. Try moving C6 up so ground can go under, from D2 to the pour area on the other side.
Im planning on having this dc-dc module on its own ground plane, and then attaching it to the rest of the design through a .024 trace. Hopefully that will reduce any sor of ground loops or switching noise....
Ah, so ground difference between left and right sides might be a concern. Don't use weak traces, use proper inductors. To keep noise down, bring everything out at one point, and filter at that point (if there's a single point where ground comes together, there's no ground loop pushing noise into your underlying circuit!).
Regarding via placement: the ones by C7/C8 and 9V don't do anything -- the common pins already connect to both layers.
You can put them just anywhere though, besides the important places.
Tim
Home
Help
Search
Login
Register
