1747-SDN (SLC 500) DeviceNet error 78 with random devices

Hi,

customer woke me up today with strange problem.

He got machine controled by SLC 500 with DeviceNet network (some inverters and Remote IOs, 14 or something like that). After five minutes of work the error 78 (Slave device in scanlist table does not exist, solution from manual: Add the device to the network, or delete the scanlist entry
for that device) with random devices occurs. One time it's inverter, and other time it's Remote IO. They reset fault by powering down the whole cabinet.

When machine isn't working everything is ok, situation only occurs when it's working.

They checked all connections, voltage level etc.(they don't have expierence with it but checked everything against Rockwell instalation instructions) Those seems ok. They instaled seperate DC power for the network. There where no changes in network recently ?

Any idea ? I think that there might be something wrong with the sccaner, but I don't have any expierence with CeviceNet on SLC 500. (I work with it only on CLX and CompactLogix).

Best regards,

Matt

My first step would be to physically isolate and examine all the devicenet connections. Make sure they are not near any of the VFD output wires.

Make sure you have at least 20vdc at the red and black terminals at all of the nodes and that the data lines and shields are all properly terminated. I know the d-net bible says 16vdc is enough, but real world experience has proven to me that our PF700 D-Net adapters like no less than 18.5vdc to work reliably.

After ruling out electrical noise or poor connections, I would replace the scanner. I have had one scanner lose its scanlist several times but it would lose the whole list, not just a device or two. Power cycling would temporarily fix it, but it would happen again within 24 hours. Finally I replaced the scanner and it hasn't given us any trouble since.

I'm not at the site actually. All I know is from phone call.

They checked network with suggestions in installation instruction. Voltage is 24 V with little ups and downs ( +/- 0.1 V) but I don't know in how many places they checked it. As I wrote they added additional DC power dedicated for DEviceNet. They also tightned all screws so I assume that everything is ok with psychicall side of network.

It's going to be 350 km ride on Monday, so I'm looking for suggestions now

zitam said:

I'm not at the site actually. All I know is from phone call.

They checked network with suggestions in installation instruction. Voltage is 24 V with little ups and downs ( +/- 0.1 V) but I don't know in how many places they checked it. As I wrote they added additional DC power dedicated for DEviceNet. They also tightned all screws so I assume that everything is ok with psychicall side of network.

It's going to be 350 km ride on Monday, so I'm looking for suggestions now

Click to expand...

Seen this before on a palletizer.
Turned out to be one connector was intermittent eventhough terminals were tight.
You may need to inspect every connector and wires on the DNET cable to check for corrosion.
Also check End Of Line resistor.

Yes, I did not mention the power supply rating, but if they are Devicenet rated they should be okay. Supplies that are not devicenet rated may not handle the surge required when all the devices are powered up at once.

When you get there, I would put a meter on every single node to verify that voltage, check terminations, ensure the end node resistors are there.

Our devicenet networks pretty solid. Aside from the one flaky scanner issue, and one firmware issue with the PF70 series, our problems have always been physical problems, including a case of having cables too close to some noisy motor leads.

Most likely, there is a frayed wire somewhere, but I suppose it is possible to have a failing device barfing intermittently.

I have not had the need to get any deeper than that, but some folks here highly recommend some more advanced diagnostic tools. In your case, I would think that carrying a devicenet meter on your trip would be sound advice.

Here are threads that you may find useful:

http://www.plctalk.net/qanda/showthread.php?t=26456&highlight=devicenet

http://www.plctalk.net/qanda/showthread.php?t=59339&highlight=devicenet+tools

Pay close attention to posts by Ken Roach. He is very experienced with Devicenet.

I've DeviceNet meter so I'll surely take it with me.

The think that confused me most is fact that fault only occurs when line is working. When it's off DeviceNet is working without error.

Hope for some other advices so maybe my trip would give some results

zitam said:

The think that confused me most is fact that fault only occurs when line is working. When it's off DeviceNet is working without error.

Click to expand...

This is the reason I think you are looking at a noise issue, possible from one or more of the VFDs...when drives start running, the noise they produce is getting onto the network cabling...When the drives are stopped, there is little or no noise being produced so you have no problem.

It could also be some other sort of device that is actively switching on and off, generating noise impulses...

I told them to start (one by one for 15 min) every drive they can in manual mode, and hopefully one of them will give error after few minutes. But the results of that will be tommorow.

After avery talk with them I geting more sure that they checked everything with physicall side of network. Maybe there is some problem with connectors as one of you suggested.

Shielding

Sounds like shielding problem. What sorts of media are being used? Flat cable, small round shielded, large round shielded? Flat to round? Where are shields for cables made up? What type of media is power supply attached to?

Shields need to be made up at one end only and it is preferred to shield at the power source or closer rather than farther from it in the case of flat to round and power supply is on flat cable.
DC Com should be grounded at the power supply. Flat cable does not have shield so when you're linking a round to a flat the shield from the round must be attached to the DC Com wire of the flat. Since the DC Com is tied to ground at the power supply the shield is grounded. This one is often missed and requires explanation to electricians.

I've seen your symptoms a few times and each time it was improper shielding. Often times devicenet problems are a combination of things. Power supply, terminations, and end resistors have already been mentioned. I'm adding shielding in detail since it is often overlooked.

As I said I've seen the line for my own eyes, so I don't know what kind of media they are using.

Thanks again for every suggestions.

They're using thick cable for main line and thin for drops. DC power is supplied by 110 V from transformer.

Network is shielded, and it's grounded in one place in cabinet near the scanner.

They replaced scaner and it doesn't help. They checked with oscilloscope for noises in signals and are sure that they eliminate all of them. They also checked connectors at the VFD's.

They made a test with running the drives in manual, and single working drive don't give an error. When they run more of them the error occurs.

Now they are going to change the cable.