PowerFlex40- Devicenet - F71

[Karim Halawa]

All power flex inverters are suffering from F71 fault randomly across the Device net network which cause Operation to stop.

As a resolution from our side,

- We’ve downloaded network parameters and power flex configurations.
- We’ve checked the terminal resistors and sure about their values and functionalities.

After deep exploration and questionnaires from our SMEs they recommended to replace all cabling system for the device net , erect new earthing system and installing new active harmonics conditioner to compensate generated distortion across the network.

Also , I replaced the power supply and removed the resistance installed on PLC card side and let the resistance installed on one of the powerflex in place .

Could you please confirm on the root cause of the problem and the action required in order to solve the problem ?

 

[Ken Roach]

- We’ve checked the terminal resistors and sure about their values and functionalities.

What is the resistance value of your termination resistors ?

removed the resistance installed on PLC card side

How many termination resistors are there on the network total now ?

There should always be two termination resistors, each of 120 ohms value.

 

[Karim Halawa]

What is the resistance value of your termination resistors ?
120 Ohm .

How many termination resistors are there on the network total now ?
one fitted on PLC devicenet card and one fitted with powerflex and another one fitted with remote device net IO .

Any way i removed the resistance from PLC card side but the problem still exist

 

[Ken Roach]

DeviceNet termination rules are very simple.

One 120 ohm termination resistor goes on one physical end of the network.

One 120 ohm termination resistor goes on the other physical end of the network.


It does not matter what kind of device or connector they are near. They just have to be at the far physical ends of the network.


If the PLC is more than about 6 meters from the resistor that's connected to the PowerFlex drive, then you still don't have a properly terminated network.

This is because 6 meters is the maximum length of an unterminated dropline.

It sounds like you've already had somebody with tools and experience examine your network. Have you followed any of their recommendations ?

Replacing all the cabling sounds like a dramatic change, but I don't know if it was proper type or properly screened or properly installed.

 

[Ken Roach]

DeviceNet really does work well when it's properly installed.

But when you cut corners or ignore those rules... it's a fiend to solve the problems.

 

[Karim Halawa]

Yes , I suffer from this problem since 4 months .
Although the system had been running trouble free for 3 years .

What is the new changes lead to this ?

 

[briancr]

How many termination resistors are there on the network total now ?
one fitted on PLC devicenet card and one fitted with powerflex and another one fitted with remote device net IO .

Any way i removed the resistance from PLC card side but the problem still exist

The way I read this is that you have three resistors in in your network.

Draw a sketch of your network indicating where the resistors are placed. this will help a lot

 

[Geospark]

Besides resistors...

I like that Ken,

I've often referred to it as "DevisiveNet".

Could you please confirm on the root cause of the problem and the action required in order to solve the problem ?...

...What is the new changes lead to this ?

You tell us!

You really are asking a lot of us here without providing anywhere near enough information. We are good, but we're not that good!

We cannot "see" what you can "see". We can only "see" the "picture" that you "paint" for us. In other words, you are our eyes and ears here. When posting for assistance, and especially when dealing with a problem which may be as a result of many different causes, it is important to include as much detail as possible in your posts so we can start to assist you more quickly. This way we can get to answering your questions instead of having to ask you more questions...

How many nodes (devices) in total are there on the network? For "drives" how many?

What is the layout of the network - 3 years old, most likely daisy-chain wiring?

What are the lengths, roughly, of the longer segments of the network?

Which two nodes represent the physical ends of the network - look for single terminations?

How is the network powered?

Were any nodes added or replaced in the last few months?

What is the current baud rate?

Are the AC drives using Pulse Width Modulation?

The resistors...

Yes, if there were three physical resistors on the wiring then that would not have been correct. But just because you have removed one does not automatically mean the remaining two are correctly positioned. You may have removed a valid end of line resistor and left one other that is still not correct. You must figure out the correct physical ends of the network and terminate there and there only.

When you say the resistors are correct, as in they measure 120 ohms, that is simply verification of the resistors themselves. To quickly check that they are terminating the network correctly you should perform an inline test. Measure the ohmic reading between CAN_H (White) and CAN_L (Blue) when the network traffic is idle and you should get around 60 ohms. This is because you should have two 120 ohm resistors in parallel. This should be measurable anywhere between properly terminated resistors. The more nodes there are the more this value may be seen to drop off a little but it should not be below 50 ohms when idle. If you're not getting a correct reading then either the resistors are incorrectly positioned or you have a wiring fault.

Make sure the 24VDC power supply for the DeviceNet network is isolated to just the network. It should not be powering anything else. If someone had inadvertently powered something else off it recently then this could potentially cause noise issues.

Voltages are also important but I won't go into measuring all that until you've verified some or all of the above.

Remember, be more detailed in your replies. We cannot do it for you and we certainly won't be guessing based on scant information.

Regards,
George