RS-485 Termination with TVS and Current-Limiting Resistors

Thread Starter

Gipper

Joined May 20, 2021
6
I have a question regarding proper RS-485 termination for a system that would be at the "end" of the network. The circuit requires additional TVS protection with a series current limiting resistors for when the TVS is conducting. What is the proper termination scheme? Having the 120 Ohm resistor in front of or behind the TVS protection?

The two configurations are depicted below (Termination #1 vs Termination #2). Disregard the actual part numbers.

rs485_termination.png
 

Thread Starter

Gipper

Joined May 20, 2021
6
Hi, MrChips.

The current limiting resistors are required, unfortunately. In a high voltage event, the TVS diodes will conduct, and having no series resistance will allow hundreds of amps to flow through the diode destroying it. This is for aviation lightning protection and is a common configuration for TVS implementation.

In my own searching, I've seen both these configurations, but I do not know which one is appropriate.
 

dcbingaman

Joined Jun 30, 2021
439
When I worked for the Avionics industry we generally used #2. I really don't think it will matter much though for the RS485 bus. That is pretty robust no matter which way you do it. The bus expects a 120 ohm impedance at the end, so it should be #2. #1 ends us having 120+33x2 or 186 ohms which is not the correct termination for RS485. Though again like I said, the bus is very resilient so doubt it matters much.
 

nsaspook

Joined Aug 27, 2009
8,759
I use this part for high EMI applications: https://www.analog.com/media/en/technical-documentation/data-sheets/ADM3095E.pdf
Certified Level 4 EMC protection on RS-485 A and B bus pins IEC 61000-4-5 surge protection (±4 kV) IEC 61000-4-4 EFT protection (±2 kV) IEC 61000-4-2 ESD protection ±8 kV contact discharge ±15 kV air gap discharge Certified IEC 61000-4-6 conducted RF immunity (10 V/m rms)
https://www.analog.com/en/products/adm3095e.html#product-overview

They do have the next level:
https://www.analog.com/media/en/technical-documentation/data-sheets/ADM3095E-EP.pdf
Fully certified DO-160G EMC protection on RS-485 bus pins: Section 22 lightning protection Waveform 3, Waveform 4/ Waveform 1, and Waveform 5A pin injection, Level 4 protection DO-160G Section 25 ESD protection: ±15 kV air discharge
 

Ian0

Joined Aug 7, 2020
3,457
Definitely #2 For the reasons given by @dcbingaman
If you replace the 33Ω resistors by 265V rated PTC thermistors, with a similar resistance then it won’t blow up the resistors when there is an over voltage situation.

And thanks for drawing my attention to the LT2862 - I was looking for a device like that.
 

Juhahoo

Joined Jun 3, 2019
214
Blowing something up during operation isn't really protecting anything, If your TVS shorts or your current limiting resistors burn, you have nothing left, the system is down anyways no matter is your IC safe or not. The only way to protect critical systems is that in any situation system will resume and is "self healing".

This is something to consider, source: https://www.semanticscholar.org/pap...rbac/c722d13cc2900201cc1bc1ea68722a4798c6bb3d

1632402363897.png
 

Thread Starter

Gipper

Joined May 20, 2021
6
Thanks for the input.

The intention is not to blow anything up, it is to avoid damage for specific "lightning" waveforms. TVS and series resistor wattage calculations are critical.

Looking into the Analog ADM3095, I found another part in the family that has a demo board with DO-160G lightning protection, which is what I'm going for. It does not require TVS diodes, but it uses a scheme similar to configuration #1.

EVAL-ADM2795EEPBZ (Rev. 0).png
 

Thread Starter

Gipper

Joined May 20, 2021
6
When I worked for the Avionics industry we generally used #2. I really don't think it will matter much though for the RS485 bus. That is pretty robust no matter which way you do it. The bus expects a 120 ohm impedance at the end, so it should be #2. #1 ends us having 120+33x2 or 186 ohms which is not the correct termination for RS485. Though again like I said, the bus is very resilient so doubt it matters much.
So, does the termination only matter to the transmitter? If that's the case, would a 54 Ohm resistor and two 33 Ohms (~120 total) work for proper termination in #2?
 

MrChips

Joined Oct 2, 2009
24,391
RS-485 communication is intended to be bidirectional. You cannot assume one node to be a transmitter unless you know this to be the case.

Termination need to be the same at both ends except for the placement of bias resistors.
 

dcbingaman

Joined Jun 30, 2021
439
So, does the termination only matter to the transmitter? If that's the case, would a 54 Ohm resistor and two 33 Ohms (~120 total) work for proper termination in #2?
It is a give and take from what I can see. With the 33 ohms and 120ohms using #2 you end up with 64% of the signal following the transmitter. That should not hurt anything. But if you go with 33 ohms and 54 ohms, you only get 45% due to the voltage divider. I would stick with 33ohms and 120ohms and #2. But a few people have chirped in with really good solutions on this thread that I think you should consider. Nsaspook and others have much better ideas IMHO.
 

vijaygmad

Joined Feb 25, 2019
1
Great question and replies. There exists a third configuration from the layout perspective. The TVS is right next to the (RS485) terminal. Followed by the termination resistor after the terminal. Finally the two series resistors that protect the transceiver from high currents. In this order TVS prevents line surge. Termination on the ends are not affected. Series resistors do their job. Hope this helps.
 

du00000001

Joined Nov 10, 2020
81
Remove the series resistors from schematic #1 !
If the SMAJ33 blow up from the test pulse, use SMBJ, SMCJ or SMDJ (I do not know the specs of the specific test pulse - in industrial automation SMAJ usually suffices).
TVS diodes are made to absorb several KWs (value increasing with TVS size - from A to D) for very short times, so they are not prone to blowing up when properly sized. But the series resistors might blow up
 

Ian0

Joined Aug 7, 2020
3,457
. But the series resistors might blow up
That’s why I use PTC thermistors as protection, not resistors.
No size of TVS will survive a misconnection to a high DC voltage, or mains.
If you use a 265V rated thermistor, The TVS+ thermistor circuit survives a misconnection to a 230V mains supply
 

Johnfoxwell

Joined May 23, 2021
4
Linear Technology data sheet for the LTC2862/3/4/5 states that pins A,B,Y,Z feature ESD protection to +/- 15kV (human body model). At the end of the data sheet they show their typical circuit for IEC Level 4 protection against Surge.
 
Top