What to do to protect RS485 ICs?

[losturcos]

Hello,
I have two boards, one is main board that does the main jobs, the other for serial information gathering. The connection is like the schematic at the attachment. At some places that has bad electrical conditions, the SN75176 IC breaks down. What can I do to protect the ICs.
Thanks.

 

[antedeluvian]

Re: RS485 IC protection



The TVS/zener combination can be included in the IC by using TI SN75LBC184 or externally as a SOT23 using the SM712 from Bourns and Semtech.

For the PTC I generally use B59980-C120-A70 from Epcos for T/H components and MINISMDC014 polyswitch from Tyco for SMD

Drawing extrated from a TI seminar notebook from 1991

 

[losturcos]

Re: RS485 IC protection

Thanks for your answer,
I will try this solution, the places that there are problems has bad electrical conditions as I said before,
like high voltage or low voltage than normal. And this makes the sn75176 broken down.
If any other hints, I will be pleased. Thanks.

 

[antedeluvian]

I also just came across this brochure

https://www.bourns.com/data/global/pdfs/bourns_tbu_short_form.pdf

you will find a RS485 interface on page 4. Notice the SM712 for transient supression as in my earlier post.

This device can withstand continuous mains voltage. The MOVs I think are for 120V faults, but if you change the values I am sure you can adjust this for 230V operation. Of course the high transient spikes are intially handled by the SM712

 

[losturcos]

I want to ask one more question, at the first reply, there is a circuit with zener diodes.
Does this zener circuit do the same job with the transient voltage suppressors?
Thanks.

 

[antedeluvian]

The response time of the zeners is much slower than the TVS, so if you suspect that there are very fast transients the TVS is the way to go. This is also probably the preferential way to go if you are looking for CE certification.

For slower signals they will both clamp, and the incoming current would be limited by the PTC resistors. Obviously the initial current is higher and then diminishes as the PTC resistance increases (and then opens in the polyswitch case). At the start of the process the zener/TVS has to disspate some power which then diminishes. Since the zeners are normally bigger physically (I am assuming you are referring to the integrated TVSs) and so they can dissipate more power. The zeners also tend to be cheaper and easier to find.

If in doubt use TVS, but I have used zeners for years without any problems (that have been reported).

 

[losturcos]

Thanks a lot!

 

[antedeluvian]

Some more app notes that have just come to my attention on the subject (adding details of additional transient supression)
https://www.bourns.com/data/global/...uit_protection_appnote.pdf#search="0038194856 07ca59485 167485 2485 3485 4485 485 4850 548502 564851 rs485%22

https://www.bourns.com/data/global/...uit_protection_appnote.pdf#search="0038194856 07ca59485 167485 2485 3485 4485 485 4850 548502 564851 rs485%22

https://www.bourns.com/data/global/pdfs/Bourns_RS-485_Serial_PortNote.pdf#search="0038194856 07ca59485 167485 2485 3485 4485 485 4850 548502 564851 rs485%22

 

[losturcos]

Hello,
I have tried the TVS solution, and still having the damaged sn75176s. Do you know any other solutions?
As a recall, the sn75176 ic is damaged at some bad conditions like when bad and some distance cable (i.e. 20 meters) is used.
Thanks.

 

[losturcos]

No idea? I dont understand it, can opto isolation after sn75176s be a solution?

 

[FvM]

I have tried the TVS solution, and still having the damaged sn75176s. Do you know any other solutions?

You didn't clarify, which elements of the solution suggested by antedeluvian you have implemented. TVS-only may be insuffcient.

It would be also interesting to understand about the exact kind of transients damaging your cicrcuit, e.g. if they are common mode or differential one. Optical isolation of the RS485 interface, as required by standards like PROFIBUS will mainly help to deal with common mode transients and potential differences between different systems in industrial automation. But it needs 5V DC/DC converters to supply the isolated RS485 transceiver and the bus side of opto-isolaters.

 

[losturcos]

I have only used TVS. Generally, my system is containing one master and one slave. The communication cables used are flexible cable, and no-extra protection like twisted pair or ground shield. There is a possibility exists that users may use 220vac signal passing near the serial communication cables, this is an out-of-control case. I am now trying a new solution without using external TVS, but 75LBC184 ic which contains inner TVS. I will wait and see what will happen. I am ready to listen more advices to make the system perfect.
Thanks a lot.

Note: Opto solution doesnt seem to be easy solution.

 

[FvM]

Opto solution doesnt seem to be easy solution.

Yes, but it's the best way sometimes.

I'm not sure, if you supplied a ground connection in adition to A and B data lines? I think, it's usual to connect a ground ar least for non-isolated RS485 interfaces, draining common mode voltages somehow.

I also think, that low ohmic series resistors or PTC fuses can give some additional protection for the data lines.

 

[arbj]

add de-coupling capacitors to the supply pin of the IC (see suggested scheme in datasheet), ensure A and B lines are not crossed. PCB layout recommendations should be followed (especially with regard to grounding).

Make sure ambient temperature is within norms.

The IC is otherwise quite rugged and there should not be much problem.

 

[losturcos]

Now I will try whole Bourn solution. I feel it will be an excellent solution. I will inform here after I try it.
Thanks all.

 

[gauchosuizo]

Hi,

I've implemented years ago the IL485. This is a isolated RS485 Driver. You need an isolated DC/DC converter to power the IL485 and at A and B I've connected each a TVS SMBJ8.0. It still works very well.

https://www.nve.com/Downloads/485.pdf

I hope it can help you.

Pablo

 

[lokesh88]

use ADM2490E IC. it is isolated and esd protected internally.

 

[losturcos]

Hello again,
I still couldnt find a good solution for this issue.
I think I am preventing the harm to the 75LBC184(after I added 10nF capacitor), but there are some other problems that can happen sometimes.
First of all; now my first circuit is like this;
between A and B, there is a 10nF/400V capacitor and a resistor of 100R (the capacitor and resistor are series to each other).
And there are biasing resistors of 680R.
By the way, in this system, VVVF inverter exists in the system. The communication cable is about 30 meters long.
Actually, communication cable is random 2 cables in a 24pin flexible cable (0.8 mm). No twisted pair exists and no ground shield
exists also (out-of-control, as I have said before).
Now; what is happening is, we start the system, and it seems to work perfect, no delays, no interrupts. After 1 or 2 or 3 days,
it starts working improperly: momentary communications can only happen, and when one of the 220VAC circuits in the flexible cable
is on, it can completely stop working. And when that circuit is off, it returns to improper working state.
Then, when we restart the whole system by switching the power off and on, the system again starts to work good.
And again after some days, it becomes improper state again.
What should I do to solve this?

 

[losturcos]

As a recall, my system is generally working from 15 meters to 50 meters.
And again generally two devices; one master and one slave.

Thanks

 

[losturcos]

Nobody has an idea?