On Fri, 31 Jan 2025, Paul Koning wrote: > > FWIW I was able to get reliable serial communication under Linux of up to > > 3.5Mbps with off-the-shelf proper PCIe UART hardware clocked at 62.5MHz > > despite that line drivers used with said hardware (soldered onboard) were > > spec'd for up to 1MHz only[1]. This was with plain PIO interrupt-driven > > operation, but then the UARTs used had decent FIFO sizes of 128 characters > > and the FIFO trigger level for the interrupt was reasonably set. > > > > Finally at 4.0Mbps data corruption reproducibly triggered, but it was > > garbled rather than lost characters, so I conclude the reason was either > > line drivers finally giving up or the transmission frequency hitting the > > bandwidth limit of the serial communication cable used. > > Was that with an actual RS232 port, i.e., a device using RS232 signal > levels, or a "TTL" logic level serial port? I'm guessing the latter.
I'm not sure what you mean by 'a "TTL" logic level serial port', please elaborate. Do you mean signalling used between the UART and line drivers by any chance, such as with a serial connection made between UARTs without actual line drivers in between? I've only seen such serial connections between onboard devices, such as a SoC's onchip UART wired to an FTDI-like device soldered next to it on the PCB for a USB serial console, which seems an industry's recent workaround with development hardware for the usual lack of serial ports with modern general-purpose computers. I don't expect this to work very well over a cable unless very short. > In my high speed experiments, I found that the limit for RS232 data > rates comes from the relatively slow rise/fall times implemented in > RS232 drivers. If you have a port that uses logic levels the > transitions are likely to be much faster so loss of signal integrity > occurs only at much higher speeds. With the RS232 drivers I have used > (MAX3222), 250 kbps is roughly the upper limit. The serial port hardware I refer to uses a UART wired to a Zywyn ZT3243F line driver, which according to the manufacturer's datasheet signals at ±5V minimum transmitter voltage levels and accepts up to ±25V receiver voltage levels and: "Meets or Exceeds the EIA/TIA-232F and CCITT V.28/V.24 Specifications for VCC at +3.3V ±10% and +5V ±10% Operations." While the transmitter voltage levels are not the highest recognised by the standard I do believe this line driver does comply with RS-232. As I say the datasheet explicitly says: "Guaranteed data rate 1000kbps," and according to my findings quoted above it is indeed the case (and well beyond). [Yes, I got it wrong by writing 1MHz rather than 1Mbps, a mental slip I suppose.] NB I've also used the TI TRS3122E line driver, suitable for operation with 1.8V signalling per my requirement, and it is also documented to handle "data rates up to 1000kbps, while maintaining RS-232-compatible output levels." I haven't got a chance to go beyond 230400bps with this device though, but these two samples do suggest that supported operation at 1Mbps isn't that uncommon for currently available RS-232 line drivers. I've looked up the MAX3222 datasheet and it does say 250kbps max though; I guess it's older technology then? Does this answer your question? Maciej
