ESP8266 Community Forum

- Thu Jun 06, 2019 10:44 am #82678 So you have ignored my suggestion about preferred directions
About U1 I was talking about the schematics - good schematics has to be drawn with signal flow from left to right, and from top to bottom. I you rotate U1 CW the schematics can be read easily.
It seems like you are not very careful... Did you try to run some DRC tests? GND from ESP is unconnected because your moved junction. Also GND path between C7,C6 and ESP pad 15 would be too long. R15 is placed wrong - why you didn't put it near R17? Your ground areas (copper pours) are not correctly connected (do you see the thin blue lines?), GND pads of SW3 and SW4 aren't connected... I don't want to waste my time here looking at your design and check for stupid mistakes like unconnected pins/pads.

- Thu Jun 06, 2019 11:19 am #82680 You are completely right, however I do run DRC tests after every change but there is no error.

The problem was, don't ask me why, but GND was in skipped nets list.

This is the reason why there was no blue copper plate under ESP.

I first did it with auto router but now I try to follow your advices.

I absolutely do not ignore what you suggest but I want to reply fast, because manually doing the tracks will take time so I will do at the end.

- Thu Jun 06, 2019 11:31 am #82681 You have diodes in parallel with the coils. Your first schematic/layout did not have that. Are you sure this will work? The diodes will short out the flyback pulse when the mosfets are turned off. Isn't this counter to what you want to happen? Maybe you want to test this before you get a board made.

It is very difficult to tell you what to do to fix the issues with your circuit board. First, it is hard to see what is happening with the bottom layer. There is little contrast between the blue and the black. There are the top and the silk screen layer on top of that. The resolution is low so it is hard to make out detail. And the clearances you have between traces and parts still are a problem. Maybe not as far as getting a board made, but seeing what is going on in your layout. All of this and the layout issues make it difficult to suggest improvements. Also I don't have enough time to go over much.

When I start a layout I first decide how I want power to be run to the higher current consuming parts. The ones that have changes in current. Steady state is not usually an issue.

Parts that do significant switching of current, like the ESP8266 and the mosfet driver, need good bypass capacitance from their power terminal to ground. Ideally the paths must be as short as possible.

Looking at your layout, you have a short path from the ESP12 +ve power pad to the C7 (good). Then I look at the path from the negative lead of C7 and try see how short, how direct, the path is to the ESP12 ground pad is. And I can not find the path from C7 to the GND on the ESP12. Not only is it not a direct short path. There is no connection.

Where is the ground connections for the ESP12, SW3, SW4, P3, P7. Your grounds are all chopped up. They are worse than bad, they are nonexistent.

I suggest you spend some time reading, learning, how to design a good layout. I have no intention of redesigning this board for you through these comments. It is just too difficult and frustrating explaining all I think you should know.

- Thu Jun 06, 2019 11:33 am #82682 I see that you both have posted as I was writing my above comment, with some of the issues I raised. At this point I am done with this thread. I have better ways to spend my time.