ESP8266 drops wifi connection when Display HD44780 is connected

ESPHome
1

Hi everyone,

I’m on my first little esphome project:
I bought a Wemos D1 mini ESP8266 and connected it to a small display HD44780.
I used the esphome wizard, flashed it from the command line and connected it to esphome. It nicely connects to my wifi as is see from the pings.

Now I connected the display with an I2C adapter to GND, 5V and D1, D2 for SDA and SCL, modyfied the yaml (see below) and flashed it succesfully.

This is what happens (found out after playing around with it):

  • Connect D1 mini Micro-USB to power
  • D1 boots successfully and display shows desired output
  • No gwifi connection (IP does not reply on ping)
  • Now, I disconnect the 5V pin, display turns off, IP appears back, when I ping the device continously after 3-5 seconds.
  • Now I reconnect the the 5V pin, wifi signal stays, but display is showing garbage

This is what I tried:

  • As it seems like a power issue, tried different power supplies. Not only PC usb and phone charger, but also supposedly better Raspberry power supply (HNP10I-microUSB, 5V, 2A)
  • Another wemos d1 mini and a nodemcu
  • Another display (also HD44780)
  • Connecting display to 3.3 V (display seems to not work)

Any ideas, what I could possibly try to get the display output and wifi running at the same time?

esphome:
  name: wemos
  friendly_name: wemos

esp8266:
  board: d1_mini

# Enable logging
logger:

# Enable Home Assistant API
api:
  encryption:
    key: "abcd"

ota:
  - platform: esphome
    password: "abcd"

wifi:
  ssid: !secret wifi_ssid
  password: !secret wifi_password

  # Enable fallback hotspot (captive portal) in case wifi connection fails
  ap:
    ssid: "Wemos Fallback Hotspot"
    password: "abcd

captive_portal:

i2c:
  sda: GPIO4
  scl: GPIO5

display:
  - platform: lcd_pcf8574
    dimensions: 20x4
    address: 0x27
    lambda: |-
      it.print("Hello World!");
2

I would suspect it’s still a power issue. Try powering the display from an external 5V source - say your spare D1 mini and see how it goes.

3

Yeah a lot (most) of these D1 mini clones are known for not having robust or adequate power supplies.

4

Since you power it at 5V, it’s not going through d1 mini voltage regulator.
That points to your USB power source.
Try to power your setup with some reliable 2A phone charger and decent USB cable.

5

Ammm… not exactly. He said he connected lcd to pins 5V and GND, which means that it indeed doesn’t use onboard 3.3V psu, but it uses board as “transfer” from usb to 5V pin. PCB traces are quite small on these boards, so voltage drop can be an issue (there were such cases in the past).
I’d suggest to power lcd from 5v, but separately, not from ESP board,s 5v pin.

6

VUSB is usually going directly to 5V pin through diode. That can cause issues powering some 5V device that can’t tolerate diode voltage drop.(Not the case here)
If the trace from VUSB to 3.3V regulator of this d1 mini is so thin that it cant handle the current, the board is really trash.

In any case, separate power supply is best option, remember common GND.

7

Sadly i only have D1 mini pro at home, so i can’t check, but i searched for pics online and if this one is true there’s a kind of resistor in between usbV and 5V pin onboard…??? It might be a diode, but it would be strange, since pin is marked 5V, which is definitely not true if diode is there… but i did see small value resistors for protection in the past.

image
image898×992 92.8 KB

8

It’s diode. That on the image is one of the most common design. It’s there to protect USB if someone powers the board on 5V pin while USB is connected.
There seems to be fuse as well, the first component from USB connector.
I have one that doesn’t have protection diode, aliexpress $1 one. I guess diode was costing too much… :sweat_smile:

9

I found better image to see the trace:

image
image893×618 76.1 KB

1 Like
10

Thanks, resistor is from 5V to GND as i see, and it’s 100kohm. Since diode is schottky (only 0.2V drop) it shouldn’t be a problem, but some parts of pcb trace are still quite narrow. But distances are small, so, again, i guess it shouldn’t be a problem…

11

You need to measure the voltage drop of schottky. It can really be anything between 0.1 and 0.6V depending the specific diode used and current draw through it.

By the way, I’m wondering what regulator that 4A2D is… Anybody knows?
Thought of Torex, mut it would be rated only for 150mA…

12

0.6V is quite a lot for a schottky, in that case it will be significantly overloaded - at the way to a second (smoked) life… especially these small diodes.
Nominal drop is around 0.2V, while 0.6V is for clasic diode.

13

I agree that with normal current draw 100-200mA it’s probably much lower.
TCB5819WTR for example has rating 0.6V at 1A.

Any idea for the voltage regulator above?

14

Interesting… that’s very high for a schottky… i’d consider this diode useless for my designs. 0.6V has standard diode, not schottky:

A Schottky diode is a type of diode that uses a metal-semiconductor junction instead of the traditional PN junction. This design gives it a much lower forward voltage drop, typically between 0.2V and 0.45V, compared to 0.6V to 0.7V for a silicon diode.

15

If voltage drop is the factor I agree.
But low-Vf ones usually leak more.
And high-Vf ones are still much faster and slightly more efficient than regular diodes.

I don’t know the factors d1-mini designers were looking for, if any…

16

I doubt they were looking anything else than price, but if they did i hope they looked to minimíze voltage drop.

1 Like
17

original d1-mini uses that diode

image
image427×405 10.8 KB

18

Hm…good to know, since in that case pin “5V” is in fact around 4.5V, which makes it pretty useless for many applications (like lcd for example).
But, i guess that this pin is not primarily intended to power external circuits, but rather to power esp from external source.

19

Could be. Also, at low current the voltage drop is likely not so high.
I can’t measure it since the only d1 mini clone I have ever bought, doesn’t have that protection diode.

20

Since original topic is about HD44780 i can say that 5V or 4.5V can cause at least that segments visibility drops to almost none. These displays have trimmer for contrast setting and contrast is very sensitive to powr issues. I used to work quite a lot with these lcd’s back at old AVR days… but to cause esp to drop wifi… well, i guess if lcd is backlit… backlight can be quite current demanding, depends on lcd type.