IO Bus error UAG-XK300 when VFD is running.

[Ivan Ristic]

Hello.

In my latest project I use USP-104-B10 panel with USC-P-B10 CPU mounted on the cabinet door. IO modules are installed on the back plane in the same cabinet connected with UAG-XK300 extension kit. PLC is controlling speed of 10 VFD devices with analog voltage outputs UIA-0006. I have problem with I/O bus error message showing when I run more than 6 VFD in the same time. Single drives are working correctly but when I switch on more then 6 or 7 drives I get error message and system continue working for couple of second and then communication with IO's is lost. After that, I must switch off power supply for all modules and HMI also to continue. 

I suspect problem is some noise from VFD while running. I try to move cable from expansion kit as far as possible from the VFD and power cables but without success. Also, all motor cables are shielded and shield connected to ground wire on both sides (motor and VFD). All IO modules have ground point connected to common ground (PE). Analog outputs are connected directly to VFD analog inputs and 24 volts common is connected to common input on VFD. I know that analog output module is not isolated but there is no other way. 

I need help because I get to the point where I don't know what to do to fix this problem and I must finish this project in couple of days. 

Thank you.

[kratmel]

Hi,

In this topic i post the pdf link about VFD & PLC

Wiring and Grounding for Pulse Width Modulated (PWM) AC Drives

 

One of possible solution - connect PLC 0V (minus of +24VDC from power supply)  to common ground.

 

[AlexUT]

Error on display show Stk Overflo - there is a "Stack Overflow" programmatic error.

1. Upgrade/downgrade ALL firmware (use "Advanced" option).

2. Load project again and test.

Does this solve problem?

3. If this does not solve problem - send project file to support@unitronics.com

Add problem description and following information:

- PLC model and serial number.

- Firmware versions at PLC.

- Include Country, Company Name. your contact information.

 

- 

[Ivan Ristic]

Hello  Mr. Alex.

Thank you for your help. I am right now at the site and will try to do upgrade.

Since this is first time with Unistream devices I need just short explanation. I packed project file together with firmware on the flash drive. Drive is in the panel. Should I upgrade all (panel, IO, CPU and UserApp) or just IO Units? Here is a screenshot of upgrade menu:

 

Thank you.

 

[Ivan Ristic]

Hello again.

I just upgraded firmware for all (CPU, IO, FPGA and panel) but problem is still present.

When I select 6 or more VFD devices for running, the moment I start motors, all IO module lamps start randomly flashing between green and red and there is message on the screen which I attached in first post.

 

[kratmel]

Do you try to connect PLC  power suply 0V wire to ground for test?

 

[AlexUT]

@Ivan Rustic

It looks that with a number of VFDs > 5 noise level is big.

Than IO BUS emitted noise make communication not stable.

During number of seconds errors' number is rapidly growing and communication is not reliable.

You have to draw and show for community wiring of:

- Ground and Power Supply 0V, include all 0V connection points (include VFD connections).

- How VFD are distributed/installed (distance, ground connections).

- The same for +24V lines.

 

- What cable used for communication.

- If schielded - how schield at each side is connected.

- Are VFDs located at the same enclosure as PLC.

- Did wiring applied in accordance to VFDs; viring guide - part of guide related to installation in proximity of PLC.

- Vhat are VFD models?

Experienced Forum members may ask you more questions.

 

 

[AlexUT]

@kratmel,

My first answer is for your case.

 

[Gabriel Franco]

I definitly wouldn´t install the PLC and its IO cards in the same cabinet of such amount of VFD.

Some advices (assuming there is only one cabinet):

1. Improve grounding inside the cabinet: use shorther, thicker ground cables for all devices, ensure analog cables are shielded and properly grounded.

2. Install a metal plate to separate the IO section from the VFD section.

3. Follow the wiring recommendations of these links: 

https://support.unitronics.com/index.php?/selfhelp/view-article/unitronics-vfd-quick-start-guide

https://www.unitronicsplc.com/Download/SoftwareUtilities/General_Wiring_Guidelines.pdf

4. Change carrier frequency of some of the drives.

 

 

[Ivan Ristic]

Hello.

Sorry for late answer. 

@kratmel  I did test connecting 0V to ground but same problem. I try to separate FG connection of all IO modules from ground and then it was better. Still losing communication but less frequent.

I have also cable for Modbus communication between all 10 VFD. I forget to mention this. When I disconnect cable from RS485 connector on the CPU module then it works. Still from time to time lose connection to IO modules but it's much better. It looks like problem with ground loops together with noise from the drives. All 10 VFD are in the same cabinet. This is first project with this number of drives that I used. I have picture while the wiring was still not finished.

Control wires to VFD are mixed with power lines. Motor cables are shielded, and shield is connected on both sides with ground. Cables go behind back plate through holes, so they are at least not close to other wires. I will try to separate them as much as possible, but problem is that I must finish all till the end of the week. All VFD ground points are connected to common ground on the bottom of the cabinet. Maybe I could try metal plate proposal to separate VFD from PLC IO modules.

One thing that I could do is to move all IO modules and install them on the back of Unitronics panel and extend wires from old position to panel on the door, but I am not sure is this going to help or make more problems because of longer wires?

[kratmel]

I am trying to change your setup...

I had a similar design with 9 frequency converters (unfortunately, there is no photo).

V1210 is used with expansion modules on the backplaine. There were no problems. But all grounding is done in one bus at the bottom.

I placed the power wires to the left, and the signal wires to the right of the frequency converters.

Maybe this link is useful

https://solutioncenter.yaskawa.com/selfservice/viewContent.do?externalId=11736&sliceId=1

 

[Ivan Ristic]

Thank you for help.

I will reroute IO modules "FG" wires tomorrow and test. I also use left side channel for most VFD power wires but there is couple of drives supplied from the right side channel. Signal and control wires are in the right channel but there is also in the left. I mixed them because of space which probably make problem with noise coupling. I will try to separate them as much as possible and see if this fix the problem.

Regarding ferrite beads (rings), maybe this can help but I can not get them soon. 

I could also try to install metal plate between drives and upper part with IO modules as proposed by @Gabriel Franco just I am not sure how to install it? At the end, I could change switching frequency of the drives. Now is 8 KHz if I remember correctly. I can put it to 6 or 4 KHz if this could help?

Tommorow I will test this things and see what happens.

Thank you again for help.

[DanT]

Hi;

As a note :

Take all Grounds to 1 Point( localized area)  as Kratmel   illustrates.

Do not Daisy Chain them. The noise just gets amplified as each ground is added  ( more resistance) to the daisy chain, and can find a new less resistive route route( maybe thru the device)   along its way.

A single point ground  will greatly reduce the chances of these electrons going back thru devices. They want to get "home " the easy way, less resistance the better.

Milli Ohms do add up -- to trouble.

DanT

[Joe Tauser]

You've got low voltage and high voltage wires in the same wireway (3rd one from the top).  You need to separate out all those wires - not just the field grounds.  

Wiring connect to VFDs is super high in harmonics and the current present makes the wiring into a very effective AM transmitter.  Any wiring within a few inches will pick up a high frequency signal and transmit it into the backplane of the I/O modules via capacitive coupling.  This will in turn combine with the data signals on the backplane, resulting in errors on the controlling CPU.

Joe T.

[Ivan Ristic]

Hello.

I finally fixed problem. Yesterday and today, I rerouted power and control wires so now all high voltage/current wires go on the left side and control wires goes on the right side. Also, I connect all common ground wires to one point on the bottom of cabinet as proposed by @kratmel and @DanT. I tested all 10 VFD running at the same time for almost one hour and communication with IO modules is perfect, without any single error (blink) of status LED.

Thank you all for helping me solve the problem. You are great.

Best regards.

Ivan

[Ausman]

It's good that you've successfully learnt about separation.

With all that running (now) ok in the cabinet, I'd put a logging temperature system in place to check internal temps for the first few days/weeks/months of actual use.  Either a separate logger, or program the plc to do it.

Personally I look at the pics and see way too much heat being generated, unless there's a heat exchange system in there somewhere that's not shown in the pics.

cheers, Aus

[Ivan Ristic]

Hello.

There is no heat exchanger. I install fan on the cabinet controlled by thermostat. Also, there is one more thermostat connected to PLC input to trigger alarm if there is over temperature in the cabinet. I could record temperature as you proposed and "see" inside conditions in the cabinet. Thank you for suggestion.

Best regards.

Ivan

[Ausman]

I now see the fan bottom right (how did I miss that? 🙂 )  and I can just make out plastic top left that I assume is a filtered inlet/outlet top left.  I hope that you have set it up so that your fan is making positive pressure in the cabinet, and intake air is well filtered.  And the client knows to keep filters clean!  You can probably tell from all of this that I absolutely hate any dust ingress, from both a maintenance and reliability viewpoint.

cheers, Aus

[Ivan Ristic]

Hello.

Yes, fan is at the bottom right and exhaust filter is top left. Fan is pushing air into the cabinet. Regarding cleaning of filters, I am not sure how often they will clean it (probably once every 5 years  as usual). I agree with you about the dust. I like clean cabinets without dust and spaghetti wires.

Thank you all for help. Best regards.

Ivan 

[kvlada]

Greeting respected colleague. I have vast experience with C-tier and low-cost equipment. Every time I work with A-tier brands like Unitronics or Yaskawa it feels like cheating.

My personal record is 14 VFDs in a single panel, all of them Chinese Xinje and ENC. There's one nasty 132kW drive nearby, also cheap Chinese, pulling 300 Amps when the motor is starting. Power cables to motors aren't shielded, and power and signal cables are wrapped like spaghetti in conduits... As much noise you can get. Command & control of VFDs is digital via 57600bps Modbus RTU, and connected to Unitronics PLC via Ethernet through through several (again cheap) Serial-to-TCP converters. Despite everything, UniStream works flawlessly. No comm problems at all!

So here are a few tips from a wastelander:

1. Signal cables: must be shielded. Connect cable shield to one side only (side with better ground) if there are problems. Connecting ground to both sides can sometimes create a ground loop.

1a. If ground is sketchy, like no protective earth at all, because some idiot only connected neutral from transformer and called it a day, then the solution is to put an isolation transformer for 24VDC PSU, network and signal circuits.

2. Use Ethernet whenever possible because it's magnetically isolated. No ground loop problems. IO-Link if your boss is rich. Profinet if your boss is Elon Musk rich.

3. Avoid analog at all cost. Voltage and current loop I/Os are expensive, both in money and stress. PT100 are necessary evil, but cheap 3-wire 10€ transmitters to 4-20mA are a thing.

3a. Cheap transmitters from 4-20mA to Ethernet are your friends.

4. Connect all grounds to a single point. Every VFD has a ground terminal or a metal frame, that's enough.

5. For longer runs, use a signal cable with twisted pairs, so that you can later add a Modbus remote I/O island instead of a simple junction box. Because plant engineer will always forget that there's 36 extra I/O points in the crawlspace.

6. Fast Ethernet works okay over low quality shielded telephone cables. Be quiet and always carry a DIN-rail Ethernet switch in your bag.

7. TVS for low voltage and OVP protectors for SELV are your friends. Because you never know if grounding and lightning protection is done properly. It could be your panel is mounted in a wooden shed and powered by a 3km single phase run.