RS485/RS232 Multiprotocol Transceiver IC

[fuseshut]

Hello,
I am using this LTC1387 software selectable multiprotocol chip with an RS485 to Ethernet converter. I am worried about the receiver inputs only being +/-25VDC. I have seen many RS485 transceiver chips with +/-60VDC on the front end of the receivers. IS there anything I can do to protect my receivers if say a voltage of 60V is produced? Resistor in series on the receiver line. A 80V rated schottky diode in parallel would not work because it will still blow the socks off the receiver.
Thanks

 

[itsmoked]

Why are you worried about the greater than +/- 25V? RS232 is generally short range stuff which usually infers no large "mystery voltages" The RS232 spec actually states 50ft max. (I think)

If you are concerned about a specific threat by a known or expected device then we can address that. If you are just shotgunning don't waste your money here, it will work fine.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[fuseshut]

Hi itsmoked. Not shotgunning!!! Actually I am worried about a sensor that I hook up on the RS485 side. A lot of the RS485 ICs have fault protection of +/-60V. I would like to have the same protection on this multiprotocal chip. And you're right: 50ft max for EIA/TIA 232 and 4000FT max for 485

 

[itsmoked]

Well well.. Someone who is actually going to use one of these.. Amazing. lolo

I think of using them but always end up having connector problems, etc., then having it just not make sense at some point.

Well heck, why do you want 60v?? Both these techniques are supposed to occur below +/- 15V!!

If you want to heap on more protection(I would not) then get out your 21V Transorbs and run them to the ground point, both low and high sides.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[fuseshut]

LOL. I hear ya. Your right about the DC voltage levels. It is more of overdesigning or the worst extremes. I am not going to worry about this. I am going to stick with the existing chip!

 

[itsmoked]

Atta boy! PCs have carried what, 300 million RS232 ports to all sorts of strange things in weird environments. I have never heard of any particular problems besides the usual pinout screw ups.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[Comcokid]

It's a good idea to add protection to those longer run serial lines. I used to live down in Florida, and it was amazing to see how many products could not handle lighting. I don't mean those direct strikes but those strikes that are a block or two away that radiate like an EMP pulse. I replaced a lot of RS-232 drivers - especially on longer RS-232 cables. Transient supressors largely eliminated the problem.

 

[Comcokid]

Fuseshut - I just saw another post of yours that indicated you are doing ROV applications. While the EMP issue may not apply much to your applicaton, you need to consider that your RS-232/485 runs in parallel with your AC power in your ROV cable. Big transients in the AC current (shorts, motors, etc) may couple directly into your RS-232/485 lines.

Then again, since it may take a long time to deploy and retrieve a ROV, the operators may end up operating in a storm. Even though salt water is conductive, a long cable connected at each end to exposed metal may be a great lighting conductor!

 

[kf6bbl]

Once upon a time I designed an RS485 system with 128 nodes that were on 4 different AC circuits, so isolation was taken into consideration.

One approach was to transformer couple all RS485 lines, similar to the approach Ethernet takes. This provide galvonic isolation to thousands of volts. Only downside is that you need to have constant transitions in your data, so the data needs to be NRZ encoded or similar. For this reason I opted not to go this way.

The other approach is simply use a fuse. Each line Tx+, Tx-, Rx+, Rx- gets a fuse. A small 0.1A fast blow fuse in series with each line has no effect on the data, adds no significant resistance, and will add enough protection to prevent the circuit from a major fault or huge differential voltages. There are also self healing fuses, but not sure how fast they are.

-Bill
CE Designer Forum

http://www.cedesigner.com