0

I've built a UART transceiver with a FC684205R and FC684205T pair. Both use TTL logic levels and I thought it would be a straight forward case of [optic lit = logic high].

However this wasn't the case, the signal off the receiver seemed to be inverted when simply applying 5v to the transmitter Vin. I tried to invert the signal coming out of the receiver module however the transistor was only held open momentarily, which I figured was because the receiver module couldn't provide enough current to drive the NPN (S8050.) So I inverted the input to the transmitter module.

When probed by my oscilloscope the receiver module output a very nice UART signal. But when I disconnected the oscilloscope the amount of noise and garbage I got was crazy. The "resting" voltage was about 2.00V and I got peaks of 5V and lows of 0V, outputting a load of nonsense to my PuTTY terminal.

I read somewhere that these modules use NRZ encoding. However when crossing the Vout with a 47uF & 15pF and GND the signal cleans up a lot (however not as clean as when attached to the oscilloscope.) It's workable but I was hoping someone could explain to me what might be going on here.

  • Why did adding the capacitors or oscilloscope in parallel eliminate this NRZ looking encoding?
  • How can I improve my circuitry as I would like to get speed better and cleaner than I can at 9600 baud.
  • I'm using this as a way of transmitting UART data across a workshop (20m) to a CNC machine that is used as a llasma cutter (the plasma cutter is blasting out a lot of RF noise that was knocking out the vanilla USB solution.)

enter image description here

JRE
  • 67,678
  • 8
  • 104
  • 179
Chris Robertson
  • 49
  • 3
  • wow, that slightly um... unusual schematic must have taken a lot of effort to draw! Thanks for doing that! Next time, save yourself a bit of effort and use the schematic editor built into this site, or the schematic capture software of your choice :) It's actually faster, I promise. – Marcus Müller Nov 17 '19 at 20:11
  • There's one built in to the site!? what an age we live in – Chris Robertson Nov 17 '19 at 22:17

2 Answers2

4

The receiver is meant to receive signals with at least 0.1 Mbit/s data rate, it expects data with encoding like manchester to have about 50% duty cycle for the light being on or off. The bits sent from UART at 9600 bps is just too slow and even baud rate is set to 115200 bps it would still work poorly as there is no encoding. The data has too long on and off periods to be received properly, depending on the data sent. So if you must make it work, you could do the encoding in software so that hardware can send continuous encoded data stream at 115200 bps

Justme
  • 127,425
  • 3
  • 97
  • 261
0

TOSLINK and all the optical standards like this indeed use Bi-Phase-mark not NRZ

https://en.wikipedia.org/wiki/S/PDIF#Hardware_specifications

This will satisfy any baud rate used for 24 bit audio with 192 bit status words

When you implement this, then your failure will not seem so ironic. ;)

" Optical Jacks are virtually unaffected by noise when transmitting and receiving signals between digital audio equipment, enabling high-quality audio recording and high speed signal receiving"

. . . I remember designing my 1st Bi-Phase telemetry design for my SCADA network in the late 70's before same-topic App notes and the Internet existed. e.g. https://www.maximintegrated.com/en/design/technical-documents/app-notes/3/3435.html

Tony Stewart EE75
  • 1
  • 3
  • 54
  • 182