1
\$\begingroup\$

I have a problem where my 485 transceiver just get damaged at field, when it's at laboratory everything runs fine, now reading about it I saw that 485 requires a 120Ohm impedance on both ends and bias resistor, what I used was 120Ohm on each drop and 4k7 pull-up on A, 4k7 pull-down on B on every drop. We use 2 transceiver MAX485 and Texas' SN75176, both are getting damaged, when I change transceiver, the communication doesn't happen.
I need some advices about resistors, good quality transceivers, etc.
Edit:
We use a 4 pin connection, 5V, GND, A,B, while 5V isn't connected. There's a common gnd between them. Also, for protection we use two 5,6V Zener at A and B, but normally on web searchs I see they using TVS.
Schematic
The BUS is a PCB that only carries A,B,GND and 5V, there's a connector at the end and it's not connected, only connected when I want to probe what is being transmitted.
My power supply is a 5V/3A flyback SMPS. The field condition is a aluminium "crate" which my PCBs are in, sometimes we use a PC too. It's stay at sun and rain 24/7.

\$\endgroup\$
  • 4
    \$\begingroup\$ Please show the relevant schematic. \$\endgroup\$ – Eugene Sh. Jul 6 '18 at 14:07
  • \$\begingroup\$ What are your field conditions? What are your power supplies? \$\endgroup\$ – Chris Stratton Jul 6 '18 at 14:36
0
\$\begingroup\$

Use TVS diode and make sure the capacitance is low, you didn't put any part numbers on those zeners so there is no way to check the leakage and capacitance.

The next problem is you don't need the pullups and pulldowns on each device, only on one end. The rest only need a 120ohm resistor. The cable also needs to be matched to 120ohm.

You might have a voltage drop between devices and with the pull ups and/or pulldowns you are getting too much current for a given transmitter.

enter image description here

Intefacebus.com has a good schematic for reference on how to terminate the buses.

\$\endgroup\$
  • \$\begingroup\$ Sorry for that, the Zeners are 1N5339, now about the resistors, why 620 ohm both? The extra current brought by the pull-up/down on each line could damage my devices? How should I put the TVS diodes? About cables, Twisted pair normally have 100 Ohm impedance, should I insert a 20ohm at the end, or just a 120 ohm at both points of cable? \$\endgroup\$ – Marlon Franco Santos Jul 6 '18 at 16:06
  • \$\begingroup\$ You don't need 620, you size the pull ups and pull downs for how many devices you have. The pull ups and pull downs could be creating to much a load for the driver to handle. Read the link for how to size the bus. \$\endgroup\$ – laptop2d Jul 6 '18 at 16:09
  • \$\begingroup\$ Thanks for the edit, I was searching the website thoroughly, I'll read the link now. \$\endgroup\$ – Marlon Franco Santos Jul 6 '18 at 16:20
  • \$\begingroup\$ meta.stackexchange.com/questions/126180/… \$\endgroup\$ – laptop2d Jul 6 '18 at 16:21
  • 1
    \$\begingroup\$ At this point, don't worry about matching the cable impedance, 100 ohms is "close enough." Getting "damaged" is not descriptive enough. Id the whole chip burnt? Are the IO pins melted? Does the "damage" magically disappears at the lab, or is it because the chips are replaced? \$\endgroup\$ – Guill Jul 12 '18 at 21:21
0
\$\begingroup\$

I do not understand what you mean by "damaged." Assuming that you mean they don't work, I would say that the problem is that in the field, the line is not properly terminated (missing the 120 ohm resistor). If the communication does not work, this in itself would not be the cause of the damage (overheating?) of the chips. You need to verify that the current load on the output pins is not greater than they can safely source (this would be a design problem)!

\$\endgroup\$
  • \$\begingroup\$ About this and your other answer, for what they say to me, the chips just "start working", or "MAX doesn't work with SN75176". They just replace all the chips and then, everything start working again, I'm getting chance to put my hands on it now, the first problem I had everyone claimed the chips were damaged, couldn't communicate with each other it was just other thing fault. So I appreciate everyone answers and as soon as I get more problems in my hand to study I'll update this post. \$\endgroup\$ – Marlon Franco Santos Jul 17 '18 at 14:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.