Future plans for Dev. Board v1.2

[ElEctric_EyE]

GARTHWILSON wrote:

EE, this is four layers, right?... Then I remembered you have parts on both sides...

Yes 4 layers. The only IC's on the bottom are Voltage Regulators.

[ElEctric_EyE]

I was looking again at the app note where it talks about using ferrite beads connecting the videoDAC power plane to the main VCC power plane. There's no page numbers, but it's about half way down under "Separate Power Plane". So, instead of the ferrite bead, I thought another regulator would an even better filter. Pretty simple thinking, but hopefully I am on the right track?

[Arlet]

Regulators are too slow to filter out high frequency (> 10MHz) noise. A bead should work better in those cases.

[ElEctric_EyE]

Shouldn't be a problem fitting one on the board, that is if I can use the one I think is well suited. It's a 0603 metric surface mount type BLM03AG241SN1 that's rated @200mA, .8ohm DC resistance, and 240ohm impedance @100MHz. The videoDAC will be using close to 100mA max.

I was not expecting to find a surface mount type like this. The appnote must be old because the type they recommended is no longer sold and it's an axial thru-hole type.

[ElEctric_EyE]

I'm thinking that DC resistance is too high. I found another one rated @750mA, .09ohm DC resistance, and 33ohm impedance @100MHz. Good news is they're both 0603 surface mount types. So I'll go ahead with it in the design with the freedom to experiment.

[GARTHWILSON]

GARTHWILSON wrote:

EE, this is four layers, right? Since you have chainsaw lines all over the power plane, and lots of traces crossing those lines, I was going to say be sure to specify the order of layers as signal-ground-signal-power, rather than signal-ground-power-signal which would give you signal lines that aren't against a legitimate plane layer. Then I remembered you have parts on both sides, so you can't do that! I don't know what to suggest then, unless you want to add another pair of layers so that both signal layers can have their ground planes without a chopped-up power plane interfering. I don't remember Dr. Howard Johnson having any articles about that. I'll think about it and see if I can come up with anything better. I wouldn't rush to get boards made yet.

I think the better thing to do might be to eliminate the power planes and just branch the power out to each point it's needed in a given area where you had a power plane, and bypass it to ground, as close as possible to each pin using it, with .01uF (or even a little less). Then the signal lines on both sides can get their real transmission lines for better high-speed performance since there won't be anything substancial between them and the uninterrupted ground plane.

[GARTHWILSON]

ElEctric_EyE wrote:

I'm thinking that DC resistance is too high. I found another one rated @750mA, .09ohm DC resistance, and 33ohm impedance @100MHz. Good news is they're both 0603 surface mount types. So I'll go ahead with it in the design with the freedom to experiment.

One I've been using in a product is the Laird Technologies HI1206T161R-10, which is in a 1206 package, is rated for 6A, has a DC resistance of .018Ω, and has an impedance of about 150Ω @100MHz @100mA. When you look at ferrite beads, be sure to look at the curves in the specifications, because increasing DC current reduces their high-frequency impedance a lot. I was disappointed the muRata didn't put that info in the data sheet you linked to.

[ElEctric_EyE]

GARTHWILSON wrote:

I think the better thing to do might be to eliminate the power planes and just branch the power out to each point it's needed in a given area where you had a power plane...

Unofrtunately there's no room to route the extra power signals through the top or bottom layers. It's why I' doing this through the power plane.

GARTHWILSON wrote:

...When you look at ferrite beads, be sure to look at the curves in the specifications, because increasing DC current reduces their high-frequency impedance a lot...

Thanks for that heads up! 1206 can fit into my design and that part you point out is about 5x cheaper than the one I was going to use. Being bigger gives these ferrite beads better ratings for this application it seems.