I created a schematic for the MAX3485 with the ESP32, but I have a question about the selection of biasing resistor(R39,R40) values. Some people use 4.7k, 10k, and 20k ohm resistors, but I'm unsure what the exact value should be and how much current is required. Does anyone have insights or suggestions on this? Thank you!
Hi, @leansolution
Why do you need bias resistors.
From here;
1079.pdf (3.1 MB)
You just use Rt = 120R terminating resistors.
Tom....
![]()
My device will be used in an industrial setting, so there may be some noise generated from the wires.
Hi, @leansolution
That is what the 485 comms system is designed for, hence the low impedance feed.
If you use twisted pair you should not have any problems.
What is the industrial location that may be sensitive to 485 emissions, if there are any.
Tom....
![]()
Without bias resistors, the bus floats in idle, giving zero differential and making it noise-susceptible—even if a transceiver has built-in failsafe. External resistors significantly improve long-cable noise immunity.
The 485 is a balanced feed with differential signal, what way any interference injected on to the twisted pair will be in phase and cancelled out by the differential action of the input circuitry.
That is what gives it noise immunity even when no signal is being transmitted.
Tom....
![]()

Tom....
![]()
Think hard about adding bias resistors because they will reduce the total number of devices you can have on the bus.
Have you also considered the grounding issues in an industrial application, will you need an isolated device?
If you want something that is bulletproof consider the LTC2877
For an industrial application where noise is a problem, I would stay with a 5V bus.
Does the MAX3485 require differential pairing for A and B routing when designing a PCB from the chip to connectors? If yes, then what should the impedance be? Is it 120 ohms?
Does the MAX3485 require differential pairing for A and B routing when designing a PCB from the chip to connectors? If yes, then what should the impedance be? Is it 120 ohms?
No.
So have you decided on biasing resistors?
That is quite true!
For an industrial setting, consult a professional for installation of RS485 network wiring. There are potentially far more serious problems than the topic of this thread.
The point of these is to have small differential, when line is IDLE. So anything will do.
Use ~10k values, to save power :)).
You can choose values in a such way, so IDLE voltage diff on your line would be ~0.2V
Then it will be ok: your line is not floating, when IDLE.
The lower the resistor value, the lower is the cable timing characteristics, the more immune it to the noise, the more current it consumes.
BTW i would like to see a workplace which causes noise to an RS485 line. Welding workshop, where you weld and test huge antennas
?
Not true.
Anything that does not guarantee a diferential of at least 200mV will not cause a known receiver output, especially when noise is present.

