A thought on serial ports

Some time during the last year or two there has been a change in my attitude towards serial connections: I used to think of them as a relic of a bygone time, when personal computers came with RS232/DE-9 interfaces and you could find them on all sorts of peripherals.  A relic used by scientists and instrument makers for reasons, probably, of expedience. And if your instrument comes with a serial connector, you sigh, go look for a serial-to-USB converter and hope that it’s not one with an unmaintained and buggy driver that for all time will make you doubt in your data.

It is true that my deep negativity stemmed from the time I was receiving GPS signals in an airplane and worried a lot about USB buffer lag impacting the time stamps. I still think my worries were founded in this particular application. It is also true that for full-blown precision instruments providing an ethernet interface (or maybe something else?) out oft he box would be highly preferable to the unholy rat’s nest of serial hubs, USB converters and ethernet hubs (not to mention proprietary software to run them) that you find in too many scientific installations. 

But since I’ve been playing with electronics and development boards, dealing with UART feels completely normal and appropriate. I own a bunch of converters (USB-to-UART, USB-to-DE-9) with FTDI chips, which work very well. And while it’s true that I’m far from understanding everything about serial communications, and am still leery of the reliability of timings below the ms level, it’s a lot more convenient and manageable than other options. 

To be continued…