Reference Design for multiplexing RS232 channels

[larrym]

Does anyone know of a reference design for multiplexing RS232 channels by over sampling each data channel as opposed to separate UARTS for each data channel at both end of the mux??

I have not worked down at this level in a number of years and got rid of all my reference designs a few moves ago...

Larry

Larry Mackey

 

[MikeHalloran]

I think Black Box used to carry multiplexers that could carry multiple RS232 signals over a single twisted pair, or maybe two pairs.
But if the signal on the actual wire/PHY layer adheres to RS232, there is no standard/supported way to multiplex on it.

I have seen a primitive multidrop implemented on RS232, by a creative Mac user who didn't know better. He got an electronic balance and a diluter with a serial interface to coexist, using the Mac's one serial port at two different baud rates in alternation. There was some interference, where some characters at the low baud rate, intended for one device, would be received as different characters at the high baud rate receiver of the other device, but the minimal command sets of the two devices didn't overlap enough for it to be a problem.

Thinking about that led to an attempt at commercializing a multidrop system that used one PC as a master, and up to 32 control/sensor devices that used a PIC at each to organize themselves by device number, decode standardized packets, and act accordingly. It worked okay, but never really found a market, mostly because it needed an external host adapter sort of thing to drive the multidrop lines because not all PC serial ports provided the proper voltage and current for RS232, and some PC serial ports could be damaged by a return signal at what would otherwise be a standard level.





Mike Halloran
Pembroke Pines, FL, USA

 

[IRstuff]

It depends much on what you mean by "multiplexing."

> Time division multiplex -- simply assign fixed time slots for each sender and some overhead for synchronization; the total baud rate is divided by the number of sources that share the time slots

> If all sources are transmitting simultaneously and are uncoordinated with each other, then you would need Wavelength division multiplex -- convert each signal into a different wavelength optical signal and dump everything into an optical fiber, and retrieve and separate the wavelengths at the receiver end.

TTFN
I can do absolutely anything. I'm an expert!
homework forum: //

http://www.engineering.com/AskForum/aff/32.aspx

faq731-376 forum1529

 

[larrym]

We used a overclocking method that sampled each RS232 data channel at ~10X the data rate. At the time I think we used a small PLD that sampled 8 RS232 channels each using a 1MHZ clock which allowed us to pass 38Kbps serial data without any problems. It was a bit of overkill but the circuitry was super simple.

Larry


Larry Mackey