ESP32 DevKitC schematic questions

[DihydrogenMonoxide]

I'm trying to design a custom PCB for my ESP32 DevKitC project.
Fortunately, schematics for the DevKitC are available for download here.
There are just a few things I don't understand:

1. Why are there so many capacitors between VDD33 and GND? Can't these all be replaced by a single capacitor?
2. What is the purpose of the Schottky Diode (BAT760-7)? Why is Vbus on the left hand side of the diode? Datasheet for the BAT760-7 is here.
3. What is the purpose of the Transient Voltage Suppressors (LESD5D5.0CT1G)? Datasheet for the LESD5D5.0CT1G is here.
4. Why did they mark the DSR and CTS pins as "arrow in" pins but then not wire them to anything? Shouldn't these be marked No Connect?

I've attached an image of the schematic below:

 

[KlausST]

Hi,

It seems you are not an experienced PCB layouter .... if so, then good luck.

1) No! This is not a DC circuit. There are high speed signals.
--> On each pin of each IC you should place a fast ceramics capacitor with very low impedance to GND. .. plus at least one big bulk capacitor. For this you need a very solid GND plane.

2) Btw: This violates USB specifications.
The diode makes that the circuit is supplied by the USB...an it prevents current back to the PC in case the circuit is self powered.
Left side = USB side = bus side, thus V_bus is correct.

3) This is usual (maybe mandatory, read CP2102 datasheet) ESD protection. Read about ESD protection.
ESD may cause immediate damage. But - even worse - it may cause a damage that causes delayed fails ... even after months.
Fails may be: increased current consumption, increased heating, loss of function..

4) You may mark them as n.c. If you like. It makes no difference.

Consider to use a bigger value capacitor at EN, to improve programming function. Read informations in the internet about ESP32 programming fails.

Klaus

 

[DihydrogenMonoxide]

It seems you are not an experienced PCB layouter .... if so, then good luck.

Hah! I guess it's obvious My plan is to copy the design and physical layout from the dev boards I'm using as much as possible.

1) No! This is not a DC circuit. There are high speed signals.
--> On each pin of each IC you should place a fast ceramics capacitor with very low impedance to GND. .. plus at least one big bulk capacitor. For this you need a very solid GND plane.

I'm glad I asked about this. I found a lot of articles about "bypass capacitors" for IC pins. Am I on the right track?
In my design, I'm adding copper fill for GND connections. What else can I do to ensure a "very solid GND plane"?

2) Btw: This violates USB specifications.
The diode makes that the circuit is supplied by the USB...an it prevents current back to the PC in case the circuit is self powered.
Left side = USB side = bus side, thus V_bus is correct.

Can I eliminate this diode if my plan is to always supply power via USB?

 

[KlausST]

Hi,

Yes, decoupling capacitors. Usually 100nF X7R.

, I'm adding copper fill for GND connections.

No. Copper fill has nothing to do with GND plane.
Simple method: Use one layer just for GND, no othervtraces, no cuts.

Can I eliminate this diode if my plan is to always supply power via USB?

You may do whatever you like. As long as you don't think about mass production nobody will complain.
It's still violating USB specifications, thus I can't recommend it.

Klaus

 

[DihydrogenMonoxide]

You may do whatever you like. As long as you don't think about mass production nobody will complain.

OK, so just to reiterate:
If my design provides power via USB only, there is no purpose to having that diode. And the presence of the diode is a violation of USB specifications anyway.
Do I have that right?

 

[KlausST]

Right.




Klaus