Taking a spool of UTP as an example, I have 2 arduinos, one at each end. I want them to exchange data in the form of strings (like k:v pairs, not hi/long or byte). The strings are best left arbitrary in length, with stop byte, but not picky.
I2C seemed the logical 3wire choice, but it apparently is very short range-even though 100kb. Luckily, I don't need speed. 20 bytes or so per message is quite ok for my use case, and if that takes a second or 2, no big deal. I will be using digital pins for I/O on both sides.
Is the issue baud/data rate, 5v won't make it that far, RF interference, or something else?
If you are spending $100 on cable why not spend $2 on differential line drivers with termination resistors so you don't have to worry about signal integrity?
Is the issue baud/data rate, 5v won't make it that far, RF interference, or something else?
With I2C usually it's bus capacitance but all the other listed factors are increasingly important if the length of the bus is increasing. I agree with mikb55.
Sorry if it seemed like I was buying cable. I want to plug 2 arduinos into existing CATx and talk to each other. 8 DI/O on each end.
It is just to test the cable itself (installed or on spool), so low data rate/slow comms are fine. I'm also old enough to admit not remembering differential signaling, though doubtlesslydoubtlessly taught it in school. Great call, and perfect for TP. Thanks!
I'll now look up DC resistance of CATx, and ponder what else I've forgotten...
Sorry if it seemed like I was buying cable. I want to plug 2 arduinos into existing CATx and talk to each other. 8 DI/O on each end.
If you have 8 GPIOs on each end, why not simply send bytes in parallel? It works better if no pull-ups are used as for I2C (which is the worst choice in your case). What distances are we talking about?
MorganS:
Why would 5V not make it to the other end of the cable? Try it at 300 baud.
Twisted pair must be driven with differential drivers for anything faster than a few hundred baud.
Thanks for writing, Morgan. Your first questions answer came from a boring spreadsheet with the (dc) electrical characteristics of the AWGs in CAT 5e & CAT 6, namely 22, 24, & 26ga.not as much an issue at length extremes of the spec for "horizantal", 100m, as it is on each end of a 2000' spool of any gauge.
This is why the remote end is active, well, one reason. So knowing nothing about RS 485, 422, 4__ I'm hoping SoftSerial() at a low baud will work, I just haven't tested.
The fun factor is that during operation, the lines used change. I just need to make sure that is the case.
I also don't know if the differential driver ics out there require high speed clocks, or can only do RS 485, or other restrictions. I may have to make a simple version that can run independent, at a form baud rate, from dirt cheap and, nand, nor, etc. ICs.
