Schematic / PCB review

[MT_266]

Hi, everyone.

I'm working on a small and battery powered CO2 Sensor. As I'm new to making PCBs and have no experience I'd highly appreciate it if some of you could take a look at what I have done so far and maybe answer a few questions.
I've included everything in the attachments - even the zipped KiCad Project if anyone wants to take a more detailed look than just from the included images.

The project consists of a MCU for low power application, a 2.8V LDO (to be able to run from 3V battery), the CO2 Sensor and an E-Paper Display (The v2 of the Waveshare 1.54'' Display).
The CO2 Sensor uses I2C to communicate with the MCU and the E-Paper uses Half-Duplex SPI. (There also is a temperature Sensor included in the CO2 Sensor which I will also use for the E-Paper).
There is a green, yellow and red LED which I will turn on dependent on the level of CO2 measured (Those will be connected through wires to the PCB).

For some parts like the decoupling of the LDO the datasheets told me really precisely what values to use.
For the E-Paper there is a reference circuit which I included in the attachments and tried to reproduce in the schematic / PCB. If I understood correctly it is a charge pump to produce the voltage required to update the display.
Though I'm not sure if I read the reference correctly. Their labeling confused me bit (like the 3.3V on VDDIO. Is it just a notice or should I connect it?) and there are no notes regarding the layout on the PCB.
The diodes they used weren't available anymore so I replaced them with the closest available part I found.

So they first questions:

• Is my schematic correct? (I mainly asking this for the E-Paper part but of course I'm happy with any kind of feedback I can get)
• What is the inductor for? (I think it is used for blocking noise from the supply rail but I don't really know)
• Should I place the decoupling capacitors or the charge pump closer to the connector on the PCB?
• Do I even use the right connector?
• Do I have to somehow shield the charge pump so the Sensor doesn't pick up noise?
• Will the charge pump introduce noise to the rest of the supply rail?

For the Half-Duplex SPI I found official documents which told me to use a resistor (to eliminate the risk of shorts) between the MOSI-Pin of the MCU and the Data-Pin from the device to talk to. Though I found no information for the value I should use for the resistor. I picked a 1k so I don't risk a too large voltage drop across it. How do I estimate the value needed?

Next to the main questions above there are a few more things that I'd like to ask if anyone has the time for answering:

• I used a lot of vias in the Pads of the components. I heard others saying one should "fill" those. Do I have to care about this? How would I do that in KiCad?
• Under the MCU there is an almost isolated GND-Fill which I had connected with a trace to the main GND-Fill. Will this impact the performance of the PCB? Is there a better way to solve this?
• I included connectors for SPI, I2C and all unused Pins of the MCU to be able to connect more things if I have to later on. Is this okay or does it occupy to much space one the PCB?
• The datasheet of the CO2 Sensor states that while soldering it should not be heated above 235 degree celsius. As I will have to solder it with hot air: Will I be better of preheating (a few minutes long?) to about 150 degree and then increase to 235 degree for soldering or should I increase to like 400 degree and therefore solder it in a shorter amount of time?

I know some of the questions are probably a bit too specific to ask or are a bit wrong for this part of the forum but I don't know where else I should post this.
Thank you all for taking the time to read this!

This is the schematic:
[ Specified attachment is not available ]

This the PCB:
[ Specified attachment is not available ]
[ Specified attachment is not available ]

And this the reference circuit for the E-Paper:
[ Specified attachment is not available ]

[mariush]

Damn stupid schematic ... why the hell do you have to separate every chip and not show the whole connections.  Was looking for R5 for 5 minutes.

From right to left sort of..

I wouldn't cut the corner of that chip on the bottom right with the traces. You have plenty of room to angle the traces and go with them higher and then to the left. That chip can also be shifted a bit to the left to make room, if needed.

Align R1 and R2 near the header ... shift the chip a few mm higher if you need to have a trace on the back side.
I don't know what's happening with that 2.8v on the header? Is the whole top fill 2.8v?  If not, I'd think a decently thick trace would be a good idea.
C3 and C6 closer to the IC pads.

J5 header : looks to me like the resistors are too spaced, yet not spaced enough to have the silkscreen proper. I'd get the resistors a notch closer and move the printed text below them or somewhere else..
Traces should come out the center of the resistor pads, and then go 45 degree to where you want them to be.

center chip : I'd rotate it left 45 degrees, now the traces to J1 are shorter and straight forward. Traces from J4 don't have to go down that much and traces on the bottom can come out through vias under the chip and then straight to the pads,  without coming out under the chip
Oh there's a via between the two top right pads of the chip, that looks like a bad layout on your picture.

Thicker traces for 2.8v and GND coming from regulator, and I'd route the traces so that you don't have those slivers of ground fill

If you move Q1 to the left of the power input and the IC2 regulator to the right then you could also move shift the center chip a bit higher

More to the left I don't like that narrow bit that connects C15 pad to the ground fill between the two capacitors. I'd try to have more space above C15 to get a thicker fill there going to the C15 pad.

[MT_266]

First of all thank you for the detailed advice!

The whole top-Fill is 2.8V. Both of the top right pins from the center chip go to GND so I connected them and placed the via in the middle.

For the schematic:
I thought spacing everything out like that and working with labels would make it more readable. What should I connect directly and when should I use the labels?

Edit:
I tried to do what you said. The attachments in the main post are now updated.

[ataradov]

I'm fine with MCU mostly having labels, but the sty with sections is pointless and disorienting. Just place the decoupling capacitors near (under) the chip they are related to.

Connections that are simple to do without overloading everything with wires, should be direct. Like in case of LDO. You can just connect all the capacitors, and that part will be instantly more readable.

I2C pull-up resistors would be more obvious if they were located close to the I2C pins of the MCU, even if you leave them with labels. But in this case you can just connect them directly on the MCU side and keep the label for the other chip.

In case of those ePaper passives (capacitors and resistors) - just remove the labels and connect all of them directly. Half of the labels will go away, leaving only functional labels.

Even if you want to label everything, simple rearrangement of all of those things to be closer to the place they are actually connected would make this schematic more readable. Those compartments are not necessary.

[MT_266]

I changed the schematic according to your tips and updated the attachments (the numeration of a few parts changed with this). It really looks a lot cleaner!

[ataradov]

Yep. This is way better.