Varistor selection for RS485 2

[bosma]

Can you please help me select the correct varistors to try and use on our RS485 communication circuit. I believe we are running into a induced voltage on the lines causing them to loose communication. I was thinking about installing a varistor in the circuit on each of the two lines to ground to try and dissipate any of this excess voltage. I believe the max voltage on rs485 is 5 volts so the closer we can get the better, without causing unintended faults. I would like to modify one of our varistor boxes we have on hand and remove the 8 varistors that are currently in it, and replace them with 4 pairs of different size varistors. I think a 5V, 6V, 8V and a 10V would be a sufficient test group to use. I have already gone to mouser.com and selected a couple of varistors that I would pick but I am not a expert in this area, so that is were I could use your help. I am not sure how the clamping voltage, capacitance, and amperage are going to effect the varistor and it effectiveness. I have added the shopping cart that i picked out.

http://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=716b46c9fc

 

[Skogsgurra]

If you are having common-mode voltages large enough to activate a varistor, then a varistor will not help. The voltage will still be there. Only reduced somewhat.

Check why (and also IF) there is excess voltage and then eliminate the reason for it.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[MacGyverS2000]

This sounds like a cart chasing the horse scenario... as Gunnar suggests, you need to determine why you're getting these voltages on the line and remove their source.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[DRWeig]

Gunnar is right. Varistors are to kill transient things, not to eliminate induced common-mode voltage.

RS485 is differential-mode, make sure you haven't accidentally connected to a source or ground -- the Tx and Rx should be floating. If voltage is being induced, shield cable and ground shield properly.



Good on ya,

Goober Dave

 

[bosma]

Thank you very much for the help.

 

[IRstuff]

If it's noise, then more shielding and more grounding will help. Are the inputs actually getting damaged? You can check by grounding the inputs and measuring the input-low current, anything larger than about 1 mA.

TTFN

FAQ731-376

 

[VEBill]

Common mode chokes might help if the interference is RF or high frequency noise. But that's less likely than some of other causes and cures described above.

 

[danw2]

There are two great resources for troubleshooting for RS-485.

1) A number of app notes (check out troubleshooting & grounding notes) at

http://www.robustdc.com/application-notes/

2) B&B Electronics' RS-485 technical page:

http://www.bb-elec.com/technical_library.asp

 

[LionelHutz]

RS485 is differential-mode

Not really. RS485 is 2 signal wires, each referenced to a common. The signal is transmitted in the difference between the wires. We already had a thread discussing this before. I think we agreed it was common mode wiring with a differential for data or something like that. RS485 is not a true 2-wire differential signal like say POTS or 10base-T.

The device will be using earth ground as the common if you only have the 2 wires going to the device. And earth ground as a reference is prone to local ground differences causing problems such as you are describing. Well, unless you have left an isolated port floating and in that case you need to run a common wire to it.

The Robust DC notes are an excellent read. Personally, I would recommend grounding the circuit in only one place and using isolated devices everywhere else. Run a shielded cable with 2 twisted pairs. Use one pair for the signals and one of 2 wires from the second pair as the common.

 

[DRWeig]

Thanks for the correction, Lionel!

You're right, the data is differential but there is a common reference. My bad.

Good on ya,

Goober Dave

 

[Skogsgurra]

I do not see why it would be wrong to call RS485 differential mode. The information is in the difference between the two signal wires. There is no useful information in each single wire.

So, can anyone give a good reason for RS485 not being differential?

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[LionelHutz]

Gunner, I'm just making the point that RS-485 is a 3-wire network, not a 2-wire network. You need to take care of the common connection and ensure the signal wires to common voltages stay within spec too. As you already posted, you must do so by designing the network correctly, not using MOV's to force the data wire voltages to stay within spec.