Custom PCB for a Split Flap Display

[ecreative]

Hey,

My first post, which I've likely put in the wrong category, but I wasn't sure which one is most relevant to my project. I'm building this awesome Split Flap display project by Morgan Manley https://www.instructables.com/Split-Flap-Display-3D-Printed-Modular-Compact-Encl/. It involves an ESP-32, connected to a PCF8575 Expander Board connected to a ULN2003 driving a small stepper motor - it's very cool to see it running.

However, my current build is hand soldered together. Ever the tinkerer, I wanted to see if I could combine the two required boards. My first time making a schematic and PCB. I've only ever used breadboards with simple circuits before, so I wondered if we had any experts who might be able to look over my design and look for anything obviously wrong, as I couldn't find a schematic for the generic PCF8575 and even if my schematic is correct. I'm sure I'll make some people wince at my PCB design but I gave it a good go based on visuals and probing boards I have to hand..

• I've attached someone else's schematic for the ULN2003 - Unlike their schematic I went for a 0.1uf capacitor as that's what my boards have on them.
  An image of the PCF8575
• Two screenshots of my KiCad schematic and PCB

FYI: The reason for the two lanes for PWR and GND is due to seeing a voltage drop when daisy chaining 7 JST PH cables across 8 modules. I'm hoping doubling up will solve the issue.

Any help would be appreciated, even if it's a grilling

[Benta]

First: very, very cool! Reminds me of Frankfurt Airport, the big arrival/departure signs always fascinate me. 
Second: very good first schematic. You didn't succumb to the sad fashion of drawing little boxes with one or two parts in each. 
I need a little time to look at the actual circuit.

[thm_w]

This is the correct section.

FYI: The reason for the two lanes for PWR and GND is due to seeing a voltage drop when daisy chaining 7 JST PH cables across 8 modules. I'm hoping doubling up will solve the issue.

How much current is one stepper drawing? Might be a good idea to measure it.
Does the stepper motor have to stay on or can the power be shut off once its flipped?

In terms of the PCB, get rid of the ground traces and add a big ground fill on the bottom of the board.

[ecreative]

First: very, very cool! Reminds me of Frankfurt Airport, the big arrival/departure signs always fascinate me. 
Second: very good first schematic. You didn't succumb to the sad fashion of drawing little boxes with one or two parts in each. 
I need a little time to look at the actual circuit.

Yes I remember them at airports when I was younger and wanted to make my own. Thank you very much, I truly appreciate the help.

FYI: The reason for the two lanes for PWR and GND is due to seeing a voltage drop when daisy chaining 7 JST PH cables across 8 modules. I'm hoping doubling up will solve the issue.

How much current is one stepper drawing? Might be a good idea to measure it.
Does the stepper motor have to stay on or can the power be shut off once its flipped?

In terms of the PCB, get rid of the ground traces and add a big ground fill on the bottom of the board.
[/quote]

It draws around 0.3-4amps while running.

The stepper motor is de-engerised in the firmware after the flip.

Good shout on the big GND pour. I wasn't sure what was the best approach.

[ecreative]

As mentioned, I've placed the GND with a larger pour. Also thought this might help with heat dissipation from U2.

If anyone is able to check my schematic, I'd really appreciate it

[thm_w]

I can't comment on if schematic is all good, but to make it more readable I would add more Ground symbols instead of having wiring going across the page.

New PCB layout definitely looks better.

[ecreative]

I can't comment on if schematic is all good, but to make it more readable I would add more Ground symbols instead of having wiring going across the page.

New PCB layout definitely looks better.

Thank you for the input. Is something like this a better approach?

[thm_w]

Looks way better

[xvr]

The RN2 connection looks a bit mysterious. If you untangle the wires around it, you'll find that these are pull-up resistors for both sides of the SCL and SDA lines, while the corresponding power lines for the pull-ups have been forgotten.

[ecreative]

Ok after speaking with a few people, I've made some changes to my schematic:

Changes made:
Removed the Mosfets
Added an AP2112K 3.3V regulator to drop VIN down to 3.3 V for the logic.
Powered the PCF8575 at 3.3V so its I2C pins match what it says for the ESP32.
Moved the SDA/SCL I2C pull-ups to the 3.3V rail.
Left the ULN2003, stepper motors and previosuly incorectly named A3144 on the 5V rail.

I'd still love someone to check over my AP2112K and capacitor/resistor layout to make sure I've not gone off the rails.

[ecreative]

Forgot to include the updated PCB. See attached.

[Benta]

I've looked at your schematic and I think you're going the wrong way.
The first approach was OK, but the MOSFETs looked strange. I suppose they comprise some kind of level shifting between 3.3 V and 5 V?

First, I'd stay with running your circuit on 5 V and leave it at that and remove the 3.3 V LDO.
Second, I'd solve the 3.3/5 V incompatibility with bidirectional level shifters. The TI TXS0102 comes to mind, but there are probably others. It's a two-channel device, and that's all you need.
Third, the A0, A1, A2 jumper circuit won't work, you can't leave any of those inputs open. You'll either have to add pullup/pulldown resistors, or use a two-way jumper setup. The PCF8575 data sheet is terrible on this point (and on others).

Cheers.