I have been working on a GPS data logger design. Just finished the and am asking anyone with experience in the design and having boards manufactured to have a look at my design and point out anything that's wrong or will stop the device from working. It's my first PCB design, so there must be things I can improve. Thanks in advance for any help.
Design description.
Aim:configure a GPS receiver,receive and filter the GPS msgs and send the filtered data out. Menu system to select the filter to apply, selectable output logic levels 5v or 3.3v.
The mcu will setup and receive signals from the GPS module Via a uart port filter the GPS data and send it out on another uart port. The board will also save the GPS data to an on board sd card. It will also have a bypass mode to let the GPS signals go out without filtering.View attachment Design.zip

 

Running your layout package's ERC and DRC utilities will ensure that the PCB reflects the circuit and should pick up clearance errors, unconnected traces etc.

You can submit your Gerbers to freedfm.com and get a report on clearances, holes etc. from a manufacturers point of views.

You can also verify that the Gerbers plotted OK using the free GCprevue.

 

I manually check my designs with the aid of several highlighters.

Working from the schematic, with one color I mark all ground connections. I do the same thing with another color for power. Finally, with one or more colors, I mark the remainder of the connections/traces. Sometimes, I will break the PCB and schematic into sub-circuits for this step.

Any unmarked traces or any connection that cannot be found or is connected incorrectly is identified with this method.

Printing oversized PCB layouts and schematics are useful. Also, the oversized PCB layout can aid in identifying sharp corners in traces, awkward routing, and other design in efficiencies. After printing the layout, I may tune it several times for better component placement or routing. I've even gone back after manufacture and redesigned a layout I was not satisfied with.

Hope these suggestions help.

 

On first glance the layout looks very good. I'm not sure what all the features are (is that a huge ground pad under the SD socket or the socket itself?) but the twists and turns seem well laid out.

One proofing method I've never seen mentioned but have seen the necessity is making the netlist in readable form (several variants are just ASCII files) and scanning the net names.

It will save you a board spin if you catch one net named "3.3V" and another named "3.3_V"

I prefer to print the layout 1:1 and place the components I intend to use. A good magnifying glass or microscope helps here too.

 

On first glance the layout looks very good. I'm not sure what all the features are (is that a huge ground pad under the SD socket or the socket itself?) but the twists and turns seem well laid out.

Thank you for all the advice, Now to answer some of the questions asked. the pad under the SD socket is Ground in fact the whole top layer is a ground plane.

I planned on using a Service called Seeedstudio over in china to do the manufacturing, so I downloaded their DRC file and CAM processor to Check the board design the whole way through the design process, I didn't get any errors. I am in Zambia by the way.

I routed 2 power lines +3.3V and +5V lines on the bottom layer of the board both 50mil am worried about stray capacitance in that the top is ground separated by board and then power traces, is their a way of checking that?.

 

I routed 2 power lines +3.3V and +5V lines on the bottom layer of the board both 50mil am worried about stray capacitance in that the top is ground separated by board and then power traces, is their a way of checking that?.

I don't follow all that. The board looks like a 2 layer board to me (red top, blue bottom). That is just metal, I do not see any solder masking, and also cannot tell if you have a silk screen layer either.

Checking all that stuff is importaint. There is a free program called GC Preview that does an awesome job of displaying your Gerber file data which the board house should eb working with (it;s the industry standard). Get a copy and give a close look how things line up.

I would be concerned with the SD socket: you have a huge layer of metal and a huge piece of metal (the socket) that need to heat up and flow together. That at minimun needs proper solder mask to keep solder on the socket tabs and not flowing away under the socket.

Generally you don't have to worry about power lines cross talking to the signals above.