Why connect DHT22 directly to ESP8266 NodeMCU V3 3.3V pins without level shifter?

Question

Hi,
I found that the ESP8266 nodemcu v3 has a voltage on the data pins of 3.3V. In various places you can find info that the pins tolerate 5.8V, although I don't see this in the official materials.
Now the question is, why are virtually all devices even the DHT22 connected without a level converter when the ESP8266 has 3.3V pins?

piterek-23 · Answer

And doesn't the DHT sometimes run on a 3V3 supply?

Anonymous · Answer

. It works as most circuits do. It is more common to convert 5V from an AVR (e.g. Arduino UNO) to a 3.3V peripheral than 3.3V from an ESP etc to a 5V peripheral.

labo · Answer

. Depending on the specification it has a 3.3-5V, 3.3-6V supply. However, with a 5V supply people feed the output directly to the ESP8266 nodemcu pins. This doesn't just apply to the DHT22 but to other circuits with 5V output, hence my question about the tolerance of the ESP8266 nodemcu inputs and the possible need for a level converter.

Anonymous · Answer

. And these other chips, what bus do they communicate over? Like the DHT pseudo 1-Wire or another with OC(OD), or SPI, UART?

labo · Answer

Maybe lame, but what's the difference? An acceptable pin voltage is an acceptable pin voltage. It doesn't matter which one, or what signal is flying with it.

Anonymous · Answer

Lame because it is important. If you have a 1-Wire or I2C bus, NO chip forces a high level, it forces a pull-up resistor. Similarly, you can do UART communication tered with OC(OD) outputs. It's also not about allowable voltage but allowable current and that's a big difference! It controls 3.3V inputs with 24V without a problem! So write down which buses you are referring to.

labo · Answer

. I have a device giving TX, RX 3.3V so no such problem here. The thing is, I was looking through various examples on the net looking for an answer to my question about pin voltages.

Anonymous · Answer

So you would have to do a longer lecture, in one sentence it is impossible to answer. Most uC's accept 5V on their inputs, especially interface inputs like SPI, I2C, UART etc. Whether and which uC pins accept 5V can be read in the note. I will leave it at that and write briefly which interfaces and which currents are or are not a problem. * I2C, 1-Wire and others with OC(OD) you pull up to +3.3V and there is no problem. The maximum voltage and current is determined by the pull-up resistor. A circuit with 5V levels will correctly accept levels from the 3.3V 'outputs' even if it is a traditional TTL circuit. * 3.3V output 5V input, there is no problem for reasons as above * 3.3V input 5V output Here it can be different. The problem is the diode at the input of the circuit. At over Vcc+0.6V (3.9V) it starts to conduct. It will typically withstand a current of 20..30mA. Worse still, once this is exceeded, the parasitic thyristor will trip. The only way to disable it will be to disconnect the power supply. I have described the case of the most typical input. However, the design can be different, e.g. a Zener diode at the input of the circuit. If the output is less than the permissible 20mA, for example, then even a million volts is not terrible. In practice, due to the MOSFET transistors on the outputs, PROVIDED that directly connecting the 5V output to the 3.3V input will not damage it. Naturally, I DO NOT RECOMMEND such solutions, but a MOSFET is what it is and when too much current flows through it, it will heat up, when it heats up, it will increase its resistance stabilising the current at some level. I emphasise that IT DOES NOT WORK THAT WAY, but such a solution may work (under certain conditions) reliably for several decades in other conditions, the circuit will fail after, for example, a week of operation. The current can be limited with a series resistor. You do not need a divider, a converter (e.g. 74LVC) just a resistor with all its advantages and disadvantages. Converters are required when signal changes are fast (e.g. SPI) in UART where speeds rarely exceed 1Mb/s a series resistor is sufficient. I guess you can count what kind and what will be the maximum allowable rate of voltage change?

labo · Answer

. This was the essence of my question. * I2C, 1-Wire and others with OC(OD) you pull up to +3.3V and there is no problem. Unless you have a module on 5V and there is already a pull-up resistor there, which brings it to a higher point. . No. You know, when you're retired, you don't always want to know everything or investigate everything anymore. Well, unless it turns out to be needed. Which isn't to say you can't explain, this is an open forum after all, and it can be read by people who could use the knowledge as much as possible. Greetings

Anonymous · Answer

. No. Calculate what current will flow,at a given resistance, when the voltage drop across it is 5V - 3.3V + 0.6V. Will you exceed 20mA? If not, you can pull it up, as I wrote, even to a million volts and nothing will be damaged. What's different, will it work? 5V circuits accept 3.3V levels without a problem, what would be different if it was a 1.8V connection to 5V. The 1.8V high level would be treated as a low level for 5V. * You pull up I2C, 1-Wire and others with OC(OD) to +3.3V and there is no problem. . Unless you have a module on 5V and there is already a pull-up resistor there, which brings it to a higher point. . No. You know, when you're retired, you don't always want to know everything or get to everything anymore . Calculating the current flowing through a resistor at a certain resistance is knowledge at the level of primary school, grade 7 or 6 (I think it's 7, I can't remember). Once you have calculated the resistance, knowing the input capacitance and any parasitic capacitances, you can calculate the time constant and therefore the maximum rate of voltage change, i.e. the baud rate. My retired colleagues who are interested in electronics have no problem with this. Young people, right out of electronics school, are a different matter; for the vast majority of them, such trivial calculations are an impenetrable wall.

khoam · Answer

. The need for a level converter is indisputable in this case, the tolerable input voltage range of the processor has nothing to do with it.

Anonymous · Answer

Are you saying that the ESP will be damaged when connected to the DHT22, which is powered from 5V? In what way? The recommended 4.7k resistor will force more than 20mA to flow when connected to 5V? How is this possible? Or will the voltage levels be incompatible?

khoam · Answer

. It all depends on the particular copy of the ESP. The ESP8266 technical documentation is unambiguous in this regard, and clearly states an allowable voltage of 3V3. Yes, I know the internet is teeming with various experiments to confirm the possibility of control from 5V. People are always looking for a challenge and the cost of a converter is none.

kaczakat · Answer

Officially, the ESP8266 is allowed 3.3V+10% on the pins. The CEO of Expressive has admitted that it tolerates 5V, which is in line with user observations. Does this affect the circuit well? Probably not, which is why it is better to use logic level converters. However, in most cases nothing will happen if you make a mistake and apply 5V to the pins (even if you supply the ESP with 5V). I wired it that way myself once by accident. But you can damage the processor by supplying even 3.3V, so it is better to use a converter. I have provided links to CEO statements here: ( and this is the manufacturer, contrary to the answer in the comment). The converter not only lowers the voltage for the ESP, but will also raise the voltage for the 5V device, so it's harder to have a communication error - something that is a 1 can be considered a 0 and vice versa.

Anonymous · Answer

. It certainly doesn't say 3.3V anywhere in the documentation, just VCC+iles there and GND-iles there. So read the documentation ALL over again! Note the max current that causes input failure. See how the input is built and you will understand why you are wrong and misleading.

khoam · Answer

. Yes, you should read that documentation before you even continue to post anything on this forum on this issue. .

Anonymous · Answer

I think I have explained why. Surely 5V from a 1A power supply will damage the circuit, from the gate output will not, although, as I've written before, I don't recommend that. In the case of limiting the current through a resistor, the issue is simple. I think I've described this in sufficient detail too. Added after 3 [minutes]: . Educate yourself mate, as I can see that you are completely unfamiliar with the design of the inputs and outputs of digital circuits, even worse, you don't distinguish between voltage and current. To make the issue of current and voltage easier to understand, why is it that 230V from the power grid very often kills a person, while tens or hundreds of thousands of volts from a stun gun does not? Added after 3 [minutes]: . And you have confirmed that it is not 3.3V but between 2.5 and 3.6. am I wrong? You have the note in front of your eyes but can't see what it says. What input voltage Vil? is acceptable? 0V or as I wrote 0 (GND) minus how many (here 0.3V)? The note is very poor, there are no maximum allowed parameters but the design of the inputs in digital circuits is similar.

khoam · Answer

. The trick is to be able to read with understanding, not to write mindlessly.

Why connect DHT22 directly to ESP8266 NodeMCU V3 3.3V pins without level shifter?

Post #1

Post #2

Post #3

Post #4

Post #5

Post #6

Post #7

Post #8

Post #9

Post #10

Post #11

Post #12

Post #13

Post #14

Post #15

Post #16

Post #17

Post #18

Post #19

Didn't find an answer? Ask Artificial Intelligence

Topic summary