new to pcb design - Page 2

Definitely need to move your jacks to the edge of the board. Also, you have a lot of wasted space towards the top. That's not exactly a problem, but why waste space when you're paying for it? You also may want to consider how you place resistors around the opamp - some of the feedback loops look a little long.

R3P R4P need to be lower or your virtual ground will waste some power. I wouldn't use larger than 47 ohm load sharing resistors. R1P and R2P would better be about 200K , draws less current. Increase C1A to 1 UF. That is for JFETs op amps only.

Originally Posted by tamu 

So I started learning how to use eagle. It was a bit hard first but then all became very easy.

 

Ive done some cmoy designs so far. here is the last one I made. I would like to know what you guys think and is there anything that I should be improving?

 

thanks for the inputs guys!

 

Try doing it without the autorouter. This is a simple design

and you should be able to route it "by hand". The autorouter

is sometime more of a pain in the butt than helpful.

I also notice that your power supply cap footprint is too small.

It should be 8 to 10 mm. If you want to use a cap that is tall,

don't be afraid to lay it "on its side" like an axial cap.

You do not need to run the power traces "through" the TLE2426.

You can treat it as a separate element and just run power to

it and connect it's output to the ground plain.

You can connect components directly to the ground plain

by using "thermal relief" pads.

Good luck and keep trying!

That looks much better!

Don't be shy to use nice, wide traces.

Remember to set "Miter" the same width

as your trace. 

At low frequency, like we use in audio, the trace size is

dependent on the load it is expected to carry. Having

a larger trace than is needed just means less resistance.

The ground plain is used for noise reduction, it act to

capture noise to ground. Ground plains are also used

for heat sinks by some components like SMD voltage

regulators.

If you get into designing higher frequency boards like

RF or computer stuff then trace width and spacing

can be critical.

the series R inside the op amp feedback loops is always questionable - yes they can be useful current limiters but you need to justify how much you're limiting the current and they can cause oscillation with C load - like cable capacitance

the 1K in the active gnd buffer are way too big

when correctly using the series R it is good to have a local feedback C from op amp output to -in, and a feedback R from the output - the feedback R,C values ideally need to be coordinated with the expected C load to insure stability

don't be afraid of jumpers - especially if you're doing a single sided board - "old school" pcb draftsmen took pride in minimal jumpers, vias, feedthrus - but I say do what's necessary to get good electrical performance - sometimes more jumpers are better than squeezing traces between pins and looping around other parts

High speed op amps benefit greatly from all SMD build.

Also input impedances difference should be close to 0 , so that you can use bipolar input op amps.

Then you could use for example AD8397 and you'd get stronger cmoy.