Hi Neil,

Improvements you could make to the existing scheme, before changing topolog=
y,=20
include:

Adding a terminating/load resistor at the receiving end. Add hysteresis at =
receiver=20
(e.g. in hardware with CMOS 4093/40106 Schmitt trigger, or in PIC with 2 x=
=20
comparator setpoints). Add low pass filter (RC =3D cheap). Increase voltage=
 swing to=20
5 or even 12V (add transistor as you say. cheap), with pull-up resistor bec=
omes=20
open-drain/open-collector config allowing allowing multiple senders into 1 =
x receiver,=20
if that is useful.

Beyond that, yes, you can look at differential driving, then twisted pairs,=
 then=20
balanced, with isolation... the list goes on. With RS485 and CAN bus a comm=
on=20
fault is blown driver chips when the live cable carrying the signal wires &=
 power is=20
cut with wire cutters, live, or otherwise shorting of adjacent wires during=
 live=20
connection/disconnection. Not too hard to design around, can be worth consi=
dering.

Brent

On 17 May 2017 at 15:07, Neil wrote:

> I'm doing one-way digital communications between 2 PICs about 12 feet=20
> apart.  ~50 kHz max.  Cable currently has 12V, ground, and one data=20
> wire  (5V direct from a PIC 18F output).  Nothing is twisted yet -- just=
=20
> straight wires.  At this stage, it's functional and works fine as it is,=
=20
> but I'd like to add EMI/noise immunity (against possible external noise=20
> from other nearby signals).  Shielding is noticeably increasing that=20
> cabling cost, so I'm thinking of differential signaling instead. =20
> Looking for a simple way to do this.
>=20
> First thought was to use RS-485 or similar differential transceivers,=20
> but that adds a few dollars, which is higher than I'd like.  CAN=20
> transceivers are lower cost (~$0.40 each side).  However, CAN is a=20
> relatively low-level signal, so I'm thinking that I can just use 2 I/O=20
> pins on the sender to create my own differential 5V-ish signal, and use=20
> the comparator on the receiving PIC to process the differential signal. =
=20
> I would need to add some resistors at the receiving end as the (PIC32)=20
> comparator inputs can only handle up to 3.3V on that side.  From=20
> anyone's experience, would this even make a dent in the system's ability=
=20
> to reduce noise effects?  Additionally, I am hoping to have the sender=20
> be open-drain, and pull-up at the receiver.  Should this be hard-driven=20
> to 5V instead for better noise immunity?
>=20
> Another thought is that I have 12V available at the sender, so I can use=
=20
> that for the signaling level, but that adds a couple transistors.  Or I=20
> can stay open-drain at the sender and pull up to 12V at the receiver,=20
> then voltage divide.  Would it make a difference for differential signali=
ng?
>=20
> Thanks,
> -Neil.



--=20
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.