Profibus VFD Slowing Down, Then Speeding Up?

Without knowing exactly how your drives are connected, I can't say for sure why its doing what it doing, but here are some things to look:

1) does your drive have feedback from the motor to tell it how fast it's going?? If so, check the cable/encoder. Maybe you have a loose wire.

2) assuming the whole system is working properly before this, is there a chance the the conveyor belt is catching/slipping?? If the belt is slipping, it will cause this type of symptom.

Please let us know what you find, good luck.

MidniteFighter

Well, the conveyor system communicates on a profibus token network. They are in series, or daisychained if you will. There is software on a controller, prorgrammed with Visual C#/Basic, to tell the conveyors when to run. The sorter starts first, and then the other conveyors will start one-by-one upstream. And each prior is also a small percentage slower...

Each of the drives have an inverter in some form, whether it be a Danfoss VLT 5000 panel display, or one that mounts directly to the drive and is programmed via a handheld version of the VLT 5000 (LCP2 handheld display).

We do utilize encoders, but they are mainly used with update photoeyes for product tracking to ensure a proper divert off the sorter. Now the software does check for variance between the ticks sent from the encoder and the speed reported back from the inverter of the drive. The only thing is that if this were to occur, the software will then shut down that particular belt and any VFD upstream from that point.

I have checked wire connections, but the problem arises far and few between, making it difficult to "pinpoint" a loose wire if that is the case. We have a device to test the network cables, but hit a wall since the issue seems to happen when we're caught off guard. Can't test it when it's working properly...

I suppose I was hoping someone here has had a similar incident and would know what the root cause may be.

Oh, and the belts aren't slipping. They accelerate and deccelerate fairly smoothly, with a total of 6 conveyors that are infected. None of them have stopped completely either.

Thanks for the help!

Is data available to indicate current draw of motors during phantom sags?

BWire:

I'm trying to get the guys on first shift to log the readout on the VLT display. I unfortunately have not been here when the problem occurs, other than one time. At that point I didn't have much time to look before it mystically went away.

Thanks

Why not log the output, with measurements from drive current, and all sensors/enoders into a .tsv file?

This would give a result which could be easily graphed - and may show that the problem occurs much more frequently, but in a non-visible form. You don't have to be there to see it.

Can you please explain how to go about setting this up?

It is normal in Profibus systems to have some form of data logging - for example, the number of pairs of shoes produced - to be held and available for the administration of the business.

Adding new files for maintenance purposes is done in the same way, but taking readings much more frequently. That is, where you see instantaneous values on a screen, these would be appended to the file, and be accessed at any time for scrutiny.

There is a new tool available here which may be useful.

I believe I will have to look into creating a separate program to monitor & log this information.

The software created for our facility mainly keeps track of package information (what it is, where it's going). It does log other information about the operation. This would include things like scanner performance, device alarms(jams, overloads, etc.). Nothing as far as power is concerned.

Would there be a conflict if two separate programs were accessing the same profibus network?

Thanks for all your help!

And thanks go out to everyone else who participated in this thread.

It sounds like the encoder on the sorter could be missing a few pulses intermitenly, you could try swapping it with one from farther upstream. You should also check the supply voltage for this encoder, same advice move the power supply upstream & see if the problem follows. The variance may not be enough to trigger a full shut down. Depending the actual software the varience detection may just be useing a timer & looking for a minimum number of pulses per second, to monitor possible slippage/stoppage the output from the inverter would proportion either the timer or pulses.

Of course all this is conjector with out a bunch more details. I've seen a similar intermitent hunting caused by encoders before.

Garthh,

I have checked the supply power for the encoder and did not find any problems. I've even checked the ticks being reported back to the controller.

spread out the heavier shoes they are bunched up.

Initially I would doubt incoming power problems. The VFD's have a fair amount of capacitance so if the incoming voltage was fluctuating that badly I would think you would see it elsewhere more quickly. However there are power line monitors which could be rented and used to monitor the lines for a few days to verify.

I'm not clear if the VFD's are running in closed loop with encoder feedback or not. But if they are you could try running them in open loop (V/Hz) mode to find out if it's the encoder or the encoder module on the VFD.

Do all the drives downstream from a particular drive slow down or just one? If they all slow down it is probably because the controller is looking at the upstream VFD reported speed and taking a percentage of that. So the first one to slow down would be the one with the problem, the rest are probably just doing what they are told.

Good Luck!

Shawn

I like Gartt's idea. Like the old XMAS lights, if you have identical units up-stream, just swap one component at a time until the problem follows the swapped component. Personally, the Dan Foss VFD would be my first suspect. You could data-log or video the suspect VDF output frequency to verify electrical vs. mechanical slippage.

Shade trees & apple trees.

Short intermittent problems are the worse type,Have you added any new equipment near any of your feed back devices? Look close see if any other machines are starting stopping at same time.Had a metal detector once false trip when a 125hp dc motor would start turned out to be a loose conduit connector on motor supply conduit. Could be encoder power supply,as suggested by other posters.I would take a long hard look at all cables involved with controls.Loose shield connection on encoder cable etc.

No, we haven't made any changes to the setup.

Sounds to me like it is a mechanical problem.

A belt is probably stretched due to wear or age. Due to different stresses at certain times it will try to move sideways and bind on something.

Look for signs of wear or fraying on the edge of the belts, or residue near any guides.

The other possibility is bearings starting to wear out.

Thanks for the comment.

All the belts are tracked/tensioned properly. Two are endless 9/16" gates round belts, three are 20" wide longitudinal ribbed belts, one is an endless 22" wide belt -90° curve, and the shoe sorter.

Upon further reading of your symptoms. If the whole system speed is referenced to the sorter, and the sorter is slowing down, Your problem is happening at sorter. I still think it is a physical problem such as:

• misaligned or bunched up product
• misaligned or dirty sensor
• something as simple as a worker putting his hand in front of a sensor could have the the same result.

Hi there

I would look into the software. You mentioned there are certain conditions that might slow down or stop the conveyor. Try to latch these conditions with a non-used marker or variable (depending on language) to analyze them later.

Supply voltage is not an issue (would be very surprised). We are running Inverters of all different makes in fairly unstable V-conditions

I would transfer all the parameters in the working location drive to the "infected" loaction drive.

This overwrites any parameter changes done by curious fingers in the drive.

This is definitely a drive issue.

Cheers.