RS-485 Bias: a harebrained solution?

[bob_miller]

Sorry about the null post; somehow it got away from me...

I have stumbled upon what I hope is a method of biasing the
4-wire rs-485 bus out of a cFP2020 without the use of an external
power supply and without defeating the isolation of the interface
through the use of said power supply:

There appear to be a raft of 3rd-party rs-485 interface/isolation products
which are *powered* by the handshake lines. Sooooo, why not *bias* with
these lines, i.e connect RTS+ to TX+ throught 620 ohms and RTS- to TX-
through 620 ohms? It doesn't matter what value RTS (or should I be using
CTS?) is asserted to as long as it is stable. I don't otherwise need these lines;
the FP1001 interface at the other end doesn't implement them anyway.

It is possible a somewhat lower resistance should be used since the
signal driver itself should be 120 ohms, but I can't think of an electrical
reason why I can't do this...

Thoughts?  Advice? Someone out there in NI land's got to be a serial guru....unless
I should be asking in a different forum.

Bob

[Marni]

Hello Bob,

I don't know how well the CTS/RTS lines are regulated, so you might have instability with that. Also, you would have to make sure that you were using the output driven by the FieldPoint. Please let me know if you have any questions. Have a great day!

Sincerely,

Marni S.

[bob_miller]

OK, I will probably try it, and do some probing to check on the behavior (and to
assure that I pick the right handshake pair). I will report back on the result.

Bob

[bob_miller]

Follow-up:

Probing the RS-485 pins on the cFP2020 revealed that the RX wires are biased
by the '2020 (they showed a steady 220mV of differential when nothing else was
connected but 100 ohms across TX and RX pairs at each end). No other lines on
the interface were anything other than 0V relative to the GND pin when the port
was idle.

I also found no way in LV of manipulating the handshake line manually, meaning that I
appear to be only able to configure the port to use the lines, i.e. enable handsaking.
This implies (if I wire the RTS pair to the CTS pair to keep the port working when
handshake-enabled)  the bias would toggle -- not a desirable behavior!

However, I also found that the bus appears to work fine with only the built-in bias
on the RX side. Therefore, I plan to not bother biasing unless I run into problems in the
future. I want to make the system robust enough to handle the occasional trashed
transmission anyway and I value the protection of the interface isolation more highly.

Bob