0

This is (an isolated, differential) capture, 1µs/div, of some Profibus communication over about 10m of cable. The first image is typical for about 90% of packets and seem fine. Most are +/-5v with slight deviation from device-to-device as they reply with data. Transitions are ok with a little ringing evident: Good Profibus Data

But this is what happens when one particular Profibus device replies: Bad Profibus Data

This output is complete garbage and generates Profibus errors. Only from this device... other devices (same bus) work fine. Which seems to infer that the cable must be ok and that this device can listen just fine, since it does try to reply.

From this capture, is it discernible whether the issue is an open terminator at the device, or a damaged transmitter? To me, it looks the transition timings are still good but it's swamped in reflections or some crazy additional inductance, leading to peaks as high as +/-10v. What is going on here?

Edit: suspect device is a Beckhoff brand BK3150 model "PROFIBUS Compact Bus Coupler." There are several, and the others work fine; just one in particular sends garbage. Speed is 1.5Mbaud. Connector is DB-9, such as this.

rdtsc
  • 15,913
  • 4
  • 30
  • 67
  • which device is it? some datasheet. What is the speed? What type of connector? – Marko Buršič Aug 10 '20 at 18:51
  • 1
    Since the peak amplitude is about doubled it looks like a missing termination. –  Aug 10 '20 at 19:55
  • I removed two BK3150's, the suspect one and a working one, and bench-tested them with a Profibus-Master-Simulator. Both were able to respond cleanly. Now I'm wondering about ground-loop in that end of the cable. – rdtsc Aug 10 '20 at 21:57
  • 1
    It looks as if the terminator is missing on Rx and the Tx and cable are mismatched. I confirmed that by Falstad simulation using 220 ohm cable, 75 ohm single end driver and load at a few MHz or 1.5 MBPS BiPhase on 10m cable – Tony Stewart EE75 Aug 11 '20 at 02:26

1 Answers1

1

It's hard to tell. What I have learnt from Profibus is that usually the problem lies elsewhere, but not near the device with comm. errors. It can drive you crazy, so forget about that specific device, search entire bus. Measure the resistance A to B, A to GND and B to GND. If you have the possibility, get the Siemens Profibus tester BT200, it's the best tool I have tried so far (tried: DSO, Profitrace).

Some non-Siemens connectors have a GND pin, spike like shaped to grab the shield - well they damage A and B wires, too. The ones on the photo are problematic, the wires can break when you mount it. The best so far are Siemens Fast Connect.

If you undo the wires, you have to redo the connection by cutting a piece of cable and strip again. Original Siemens stripping tool is highly recommended, we don't strip any cable (profibus or profinet) without orig. tool - too many times we spent hours, days to find the problem like you.

I do suggest you to redo all the bus, with correct tools.

Marko Buršič
  • 23,562
  • 2
  • 20
  • 33
  • Just a note, the customer had Siemens come out and redo the entire bus. They were not told exactly what the fault was, but at least one section of cable (not the suspect one) was replaced. – rdtsc Aug 03 '21 at 13:04
  • @rdtsc I am glad that you solved the problem, in the exact way I had to redo it. – Marko Buršič Aug 03 '21 at 13:37