I want to make serial communication with a device which works on 0-5 Volt TTL level. But my serial chip uses 3.3 Volt logic. (this is an UART communication)

I know that I need a step down at least for the TX pin of device (which will be RX on my serial chip). I wanted to go with normal resistor voltage divider but I'm assuming it will cause problems in high speed transmissions. Can you recommend a good, easy to use chip for this task?

Also, do I need to have a voltage step up from 3.3V to 5V for the device's RX (my serial chip's TX)? I heard that anything above 2.4 volts should be considered high. What do you think?

And I have question about whether the converter should be inverting or not? How can I determine this?

  • \$\begingroup\$ While you dont say whether this is for I2C communication, I would suggest you to take a look at this question and it answers. Especially the linked document may be helpful. \$\endgroup\$ – PetPaulsen Sep 4 '12 at 9:54
  • \$\begingroup\$ @PetPaulsen - I2C doesn't use TX and RX, sounds more like UART. OP also doesn't talk about a clock. \$\endgroup\$ – stevenvh Sep 4 '12 at 9:59
  • \$\begingroup\$ Yeah this is about UART...sorry for not being clear about it! \$\endgroup\$ – Sean87 Sep 4 '12 at 10:08
  • \$\begingroup\$ What serial chip are you using? If it's something like the FT230X, those chips are specifically listed as being 5V tolerant and compatible even when using 3.3V I/O levels, so you wouldn't need any conversion at all. You might check if your chips are similar. \$\endgroup\$ – Jim Paris Sep 4 '12 at 17:15
  • \$\begingroup\$ It's some TI microcontroller which in datasheet says 3.3V \$\endgroup\$ – Sean87 Sep 4 '12 at 17:47

The 2.4 V is the minimum for a high level TTL output. A TTL input needs at least 2.0 V, to give you a 400 mV noise margin. But that's for TTL devices, or TTL compatible, like HCT-CMOS. An HCMOS device will typically need 0.7 Vdd as minimum for a high level. For a 5 V supply that's 3.5 V, so you probably won't make it. You can use an HCT single gate IC as level shifter.

The resistor divider probably won't cause much problems, except for very high speeds. Since this is less than 100 kbps UART, chip resistors shouldn't be a problem.

  • \$\begingroup\$ Thanks. Yes its UART communication. I will try the prototype with resistor voltage divider, but can you name a decent chip for this job? \$\endgroup\$ – Sean87 Sep 4 '12 at 10:07
  • \$\begingroup\$ @sean - What's the bit rate? \$\endgroup\$ – stevenvh Sep 4 '12 at 10:07
  • \$\begingroup\$ The baud rate is either 57k or 19k \$\endgroup\$ – Sean87 Sep 4 '12 at 10:09
  • 1
    \$\begingroup\$ The problem with voltage dividers is they have a relatively high output impedance. Combined with stray capacitance this rounds off the edges of the pulses. Also if the line idles high (as most logic-level serial does) then they constantly waste power. \$\endgroup\$ – Peter Green Sep 7 '17 at 11:28
  • 1
    \$\begingroup\$ If you want a small chip to take a line from 5V to 3.3V you might consider the 74LVC1G32. \$\endgroup\$ – Peter Green Sep 7 '17 at 11:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.