Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Register Log in
  • Congratulations waross on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

VFD to PLC Comm Loss Intermittent Issue

Status
Not open for further replies.
Stk8

Stk8

Electrical
Aug 25, 2017
8
0
0
US
I'm dealing with a natural gas compressor that's controlled by a Schneider ATV71 VFD and a M238 PLC that's been intermittently shutting down due to comm loss over the CANOpen bus. It's been going on since a lightning strike some time ago, according to our customer. After the strike, the drive and PLC were both replaced, but the compressor would occasionally go down due to comm loss. I'm fairly new to VFDs and PLCs, and new to my company, and I've inherited this problem. Some things that have been tried are replacing the shielded CANOpen cable, replacing DC power supplies, cleaning and verifying grounds, slowing the baud rate, and checking/replacing Line Terminating resistor on CANOpen terminals. I also checked for a Time Out parameter, but it seems this drive only has an adjustable Time Out for Modbus, not CANOpen. We're still getting a comm loss fault 2-5 times a week on that machine, and I'm hoping for some suggestions.

I'll be visiting the compressor again this week and have plans to replace the PLC, replace most of the wiring inside the enclosure(which houses both the VFD and PLC), and connect a direct ground to the CANOpen shielding to try to further eliminate any noise from the VFD. Are there other measures I can take to try to eliminate the issue?
 
Replies continue below

Recommended for you

Sort by date Sort by votes
Well. You HAVE done about everything in the order that I'd do them. I'm a little confused though in that you say:
After the strike, the drive and PLC were both replaced
Click to expand...
and now you're saying you're going to:
have plans to replace the PLC
Click to expand...
.

Is that twice??

At this point you need to put a scope on the comm line to see if there are anomalies and if so what they are. Bad rise times? Noise? Large common-mode voltage (caused by a blown lightning arrestor or something similar).


Keith Cress
kcress -
 
Upvote 0 Downvote
That, indeed, would be the second time the PLC has been replaced. My company has no tools available to test the comm line, so I'm currently stuck with trial and error, instead of troubleshooting off of concrete information. The possibility of the original replacement PLC having an issue was the next most logical step. It's far less than ideal. On that note, are there any recommended scopes that are light on the wallet and good for use out in the field?

 
Upvote 0 Downvote
If it were Modbus, or some other protocol that uses RS485 as a transmission medium, I'd be grabbing something like this, that normally picks up a lot of the communications issues.

CANOpen has its own heartbeat arrangement to detect loss of communications, exactly how its implemented may vary though. If you've replaced all the hardware and still have the problem, then the issue is more likely to be the configuration, which one would expect would have been copied over when the original equipment has been replaced. Its possible to buy CAN sniffers in the same manner as the RS485 equipment, but its not as easy to hook up. Here is a USB-CAN interface device, I've seen (and used) some diagnostic software to monitor the CAN stuff before, but can't remember what it is unfortunately. The penalty for using most of the protocol monitoring equipment (like the USB-CAN) above, is that it won't pick up any hardware level issues, which is where itsmoked's recommendation of a scope comes in.

CAN as a communications medium is generally quite reliable, I'd expect to have far more problems with the equipment than 2-3 times a week if it was to do with the cables or terminations (which you've probably concluded also, having replaced it all). Its also possible (although I'd expect its less likely) that the faults may coincide with other things in the plant (some large equipment switching or similar), you'd likely need timestamps of the faults to further investigate that possibility.

Having said all that, I'd expect that for fully operational equipment prior to the lightning strike, that they would have been experiencing the issues prior to that, or other issues would also be experienced after the lightning strike, but not noticeable compared to a loss of comms fault.

I like the Tektronix scopes personally, although they're not what I would expect to be 'light on the wallet', particularly not the battery powered ones.

EDMS Australia
 
Upvote 0 Downvote
There may be some other issue that is causing the report of comm loss.
The lightning strike may have damaged something other than the VFD and the PLC.
I would put a scope on the power lines and on the output of the control power transformer. I would be looking for a small persistent arc somewhere in your system.
Is your power system grounded, impedance grounded or floating?
Anecdote. I forget the details but I do remember going around in circles for a couple of days because the diagnostics were reporting a fault that was not in any way related to the actual fault.
The point is that some other issue may be falsely reported as a comm loss.

Bill
--------------------
"Why not the best?"
Jimmy Carter
 
Upvote 0 Downvote
Stk8; I'd suggest you pick up one of the hand-held oscilloscopes. One of the biggest problems with using scopes in the field is understanding how using the scope and its ground will affect or be affected by the circuit being studied. Using a handheld avoids this almost entirely as it isn't grounded like a typical desk scope. There are scads of them now under $500. For testing things like this comm issue you only need qualitative tools. "Is there noise?" "Is there distortion?" Etc etc. So more expensive high fidelity scopes are not called for.

If you want to hunt some down and provide us links we can give you yea-nae opinions on them.

Keith Cress
kcress -
 
Upvote 0 Downvote
waross said:
Is your power system grounded, impedance grounded or floating?
Click to expand...

All of the components are grounded to the cabinet, which is Grounded to Earth

waross said:
I would be looking for a small persistent arc somewhere in your system.
Click to expand...

What would you see on the O-Scope in this case?
 
Upvote 0 Downvote
Upvote 0 Downvote
Remove ground connection of PLC and supply it from 24V from batteries, with enough capacity to cover a week.
If no comm loss anymore, very probable is a problem in supply and / or grounding.
Check if CAN_GND was to ground connected.
 
Upvote 0 Downvote
Stk8; Both those scopes would be more than enough even for high speed CANBUS. As for sample depth... If you want to actually see/understand the CANBUS messaging you would instead want a network analyzer or "snooper". Some scopes provide this but often at painful prices. I was recommending a scope to check signal health NOT message examination. That said both those scopes plug into a wall and can present the issues/dangers I mentioned.

I'd look for scopes that specifically state that you can hook the probe to the hundreds of volts one might see in controls. I'd probably look for them stating CAT II, CAT III, or CAT IV input and battery powered. Lots of field work can be very hard to get plug-power to.

A 10MHz scope would be adequate.

Keith Cress
kcress -
 
Upvote 0 Downvote
Waross said:
Is your POWER SYSTEM solidly grounded, impedance grounded or floating?
Click to expand...

It is solidly grounded.

iop95 said:
Remove ground connection of PLC and supply it from 24V from batteries, with enough capacity to cover a week.
If no comm loss anymore, very probable is a problem in supply and / or grounding.
Check if CAN_GND was to ground connected.
Click to expand...

Unfortunately my PLC is powered by 120VAC, but as of yesterday it's being fed from a new transformer, and the CAN_GND is directly connected to a ground.
 
Upvote 0 Downvote
I think CAN_GND connected to ground it's not a good idea.
VFD have big caps in input filter and in transient conditions may "move" localy ground potential in respect to CAN_HI and CAN_LOW, for short time, enough to alter CAN voltages and report comm loss.
If both, VFD have communication GND isolated from their chassis, may try to keep both GND together but isolated from ground (chassis).
 
Upvote 0 Downvote
iop95 said:
Your PLC accept 24Vdc also.
Click to expand...

The transistor output model is powered by 24VDC, but we use the relay output model, which is powered by 120VAC. The suggestion is much appreciated though. [smile]

itsmoked said:
Connecting the PLC and VFD shields together can cause high frequency noise to run down the shield. I'd try it disconnected from the VFD side.
Click to expand...

Really? Even with just a single CANOpen line directly between the VFD and PLC and only a few feet between them? Also, my VFD connects with an RJ45 port. Is it possible to disconnect the shield in this case short of stripping the end and reterminating with a new RJ45?
 
Upvote 0 Downvote
Status
Not open for further replies.

Similar threads

PegLeg03
  • Locked
  • Question
VFD to PLC via CANopen issues
Replies
3
Views
27
Heaviside1925
Heaviside1925
Sujj33
  • Locked
  • Question
VFD, Motor, Vibration
Replies
3
Views
44
electricpete
electricpete
MitchusMaximus
  • Locked
  • Question
General VFD Enclosure Cooling Question
Replies
19
Views
36
Heaviside1925
Heaviside1925
Umer9
  • Question
Loss of Excitation on a Grid-Tied Generator 1
Replies
3
Views
40
waross
waross
systemsaver
  • Question
how to set PID control in VFD that operate a centrifuge pump transporting juice being concentrated 1
2
Replies
23
Views
80
systemsaver
systemsaver
Back
Top