advertisement
from the Automation List department...
Noise Problem with VFD/motor
Engineering and workplace issues. topic
Posted by Joel Young on 13 March, 2005 - 2:21 pm
I have a noise problem associated with a VFD/motor (75HP 460v wye start/delta run) system. This system (motor/drive) runs the blower to a concert hall. The low level lines (microphone lines) of which there are 200+ pick up to varying degrees the trash (noise) produced by the VFD.
We have tried a variety of solutions, including line reactors, drive isolation transformer, moving the VFD to an alternate location (closer to the motor).
We still have the noise...

We have established that the noise is not coming from the power input side of the audio system. We "believe"/"think" that the noise is broadcasted into the low level lines (mics/intercom)...Thses are shielde-ballanced lines, but they still see the trash....

We theorised that the VFD/motor current (50A) was radiating noise into mic lines that were roughly 10-15 feet away, but which ran paralell for 100+ feet. We then decided to test our theory by powering the VFD/motor via a 460v 3PH generator.
...SUCCESS!....but not really....

The audio system was quiet with the generator setup. We reasoned that the VFD/motor power should be fed via an alternate location...The parking garage next door, (3x125 feet of wire run), and away from all mic lines ... We hooked it up... The noise was back...

The "GROUND" issue...
From the beginning, the grounding concept was discussed... The concert hall building is one giant ground loop in my opinion. At every junction box the contractors tied the ground, "the little green wire", to the side of the box. It is certainly concievable that in the maze of electrical systems throughout the building, a "neutral" might be tied to ground (this would be additional to the neutral/ground connection at the service entrance). That could well be the magic wire (if it exists).
The VFD/motor is grounded, and is a likely source of at least part of the problem.
When the VFD/motor power was sourced from the parking garage the ground was left in it's orriginal location,-- tied to the building steel... The ground circuit does have a "copper path" back to the service entrance, but it is by no means isolated.

The audio system ground is (or at least appears to be), an isolated ground.

This e-mail was prompted by a 2002 discussion about floating 3PH systems...
Is it safe (for humans/hardware) to float the motor?... Should I tie the VFD to the motor, as it is at present, and disconnect those units from earth?... Keep in mind that we communicate to the drive via RS485 (232?)...

As I recall, I measured 3-5 amps from VFD/motor to ground....

The VFD is in an odd location, and certainly not optimal for a low noise installation. My current plan is to move the VFD very close to the motor (<20ft.), tye the motor to the VFD, and run a low impeadance ground (#2 welding cable) back to the parking garage earth stake.

Thanks for your time!
Joel Young


Posted by Wayne Shimanis on 14 March, 2005 - 12:51 pm
Mr. Young:
You may want to take a look at Vortex Technologies IL-GTT, ground transient
terminator. The IL-GTT can clear up noise in grounds. A few units placed at
strategic places in your grounding system should do the trick. You can get more
information at http://www.vortextek.com

Wayne Shimanis


Posted by Joel C Ledford on 15 May, 2005 - 12:11 pm
Mr. Young:
Hello, I have long since retired from ATT where I had significant Power Systems responsibilites at several large facilities.

First, Ground Current Flow: Your building should have a grounding system much more complex than a ground rod. For any facilities building, to attempt to do anything more than simple a/c wiring and phone lines with a simple ground rod is well lets say "foolish."

Next, on a very simple level, ground wire configuration: Large building complex's, and I do mean LARGE, may have ground current flow. This will occur if your grounding system provides less resistance than the physical ground/Earth. If, the building designer/engineer did his job at design/construction, the building will have a very complex ground system, especially if during design, he took the intended use into consideration. Current can flow across the ground wiring privided there is unequal electrical voltages in the earth. However, there should not be any current flow across your ground wiring from any electrical equipment. PERIOD! If you have current flowing, then you have an electrical equipment or wiring problem. Period!

Next, lets consider non-linear connections: remember the boyscout crystal radio, well it works simply as a non-linear device which will receive and demodulate the analog components of an amplitude modulated signal. Now, simply, you have non-linear connections and of course, componets, in you current arrangements of wire, electrical units and power generation or conditioning equipment. Period!

I don't know if you are an engineer or if any of the other folks giving you advice are engineers. I am not, but, i do have far in excess of the instruction and experience to offer you this input.

Next, throwing money, etc., all that copper wiring, and proposed new equipment sound like big $ to me. Simple is the first choice.

Now, find the source of current flow on the ground wiring. Next, consider an office grounding or ring ground system for any DC applications. Next, try to determine how much DC flows on the grounding system and then try to determine how much AC is flowing on the Grounding system.

Try to contact the designer/builder of the building or maybe if available get your hands on the drawings/electrical for the building and see just what you do have.

Finally, I came across this for some strange reason, so guess it is ment to be. My wisdom to all of you as i go on my marry "old" way, before you start trying to be like the local mechanic who fixes your automobile problems by parts replacement, be a good technician test and isolation the problem, then spend your money to fix it. Power conditioners, etc., are always most effected by noise generated by bad wiring, wiring design, and bad thinking. Have a good day, and by the way, none of this may solve your problem. You may be exceeding the design of your envirnoment and simply need a new design.... I wonder if the problem is really the equipment. without prejudice and safety is your responsibility, Remember Lock out/tag out. And, oh, maybe, you need a consultant or someone to come in and use their thinking, I can't remember how many problems have been solved just by having someone else look at it....bye, joel


Posted by Steve Myres on 14 March, 2005 - 12:54 pm
Could you use a line reactor on the line side of the VFD, (or maybe on the feed lines from the drive to the motor, if you stay with the long run configuration)? It is also possible to get shielded power cable for the run from the drive to the motor.

--
Steve Myres, PE
Automation Solutions
(480) 813-1145


Posted by Curt Wuollet on 14 March, 2005 - 12:55 pm
It does sound like switching harmonics are radiating from the drive wiring. If so, reducing the wire length fron the drive to the motor may help. You might also check into a different drive or a different means of speed control for the blower. Some drives with very fast switching devices generate harmonics well into the RF region. You might also check the shielding and actual balance of the input cables. If the noise goes away when you disconnect the long cables, you are probably going to have a hard time getting rid of this without rerouting them far away from the source. Something else you might consider is not using long cable runs, replacing them with a wireless setup. The cost on these has come down quite a bit.

Regards
cww


Posted by Prakash phatak on 14 March, 2005 - 1:45 pm
It looks like the noise is EMF noise on output line and you must use VFD cable from a recommended mfg. I have used Olflex cable on the output side and it helps a lot to reduce the noise.


Posted by William Hinton Sr. Electrical Engineer @ Delphi on 14 March, 2005 - 1:45 pm
Joel,
You may not find the source of your problem where you are looking. There are two possibilities you may consider checking out: Is your power transformer a wye-wye? Does the problem only exist when the motor is running DELTA?

A wye-wye power transformer has no isolation or harmonic filtering and a wye transformer has no harmonic filtering. If the problem occurs only when the motor is running delta, the delta motor may be attempting to run at its base speed (60 Hz) and 3 times the base speed due to the third harmonic (180 Hz) etc. I suspect the noise is from harmonics. The delta motor is also filtering the odd harmonics divisible by three causing heat and vibration and excess bearing wear.

Delta power transformers are natural harmonic filters, have no neutral or problems associated with this type of noise. Ungrounded power systems can have unstable phase voltages which are filtered from phase harmonics and have stable phase voltages by adding a Phaseback voltage stabilizer from http://www.applied-energy.us

There is a great paper on this at the above website, simply follow the link.

I hope this helps,
William Hinton


Posted by Shah Nawaz on 15 March, 2005 - 10:15 pm
If by reduction of cable length would solve your problem which would, then you defenitely need a dv/dt filter, dont confuse it with load side reacror. For any more help, shahiiee@hotmail.com


Posted by Hakan Ozevin on 15 March, 2005 - 10:21 pm
> We have tried a variety of solutions, including line reactors, drive isolation transformer, moving the VFD to an alternate location (closer to the motor). <

Did you try a B class EMC (RFI) filter?

> The audio system was quiet with the generator setup. We reasoned that the VFD/motor power should be fed via an alternate location...The parking garage next door, (3x125 feet of wire run), and away from all mic lines ... We hooked it up... The noise was back... <

Taking the VFD to a distance will not solve the problem, because high frequency interference will be transmitted with the power cables to every place having power connections.

> The VFD/motor is grounded, and is a likely source of at least part of the problem. <

Nope, your problem is RFI coming from the VFD.

Do the following:
1. Buy a "B" class RFI filter and connect it in between the mains and the VFD. They should be very close to each other and cables should be shielded.
2. Connect the earth terminal of the RFI filter with a low impedance to the shortest ground point. If possible establish a direct earthing.
3. Use shielded cables between motor and VFD.
4. Decrease the pulse frequency of the VFD to the minimum.
5 (optional, last chance): Buy a sinusoidal filter and connect it in between VFD and motor.

Good luck.


Posted by Bob Desrochers on 15 March, 2005 - 11:56 pm
Joel,

Incorrect grounding of the motor to the VFD will allow the system to act as a noise transmitter. I suggest a slightly different grounding method to eliminate the noise problem. An appropriately sized grounding conductor installed in the same raceway as the phase conductors with one end terminated in the motor doghouse and the other end terminated at the grounding terminal of the VFD. Splices or compromises in conductor integrity are NOT allowed anywhere along the length of this conductor. Bonding of enclosures along the raceway to the motor should use a separate grounding conductor which is used in a typical fashion. The VFD then needs another grounding conductor back to the power source. After this grounding activity is completed the only other item that may need changing is carrier frequency. Large VFD driven motors are usually easier to quiet down than the small fractional ones. Good luck!

Bob


Posted by Bob Peterson on 16 March, 2005 - 12:00 am
One trick that MIGHT help is to feed the motor with twisted cable such as SO cord rather than seperate wires. This can reduce radiated noise in some cases.

Be advised that this approach while it might be effective might also be in violation of local electrical codes and thus might require special approval from the AHJ.


1 out of 1 members thought this post was helpful...
Posted by Phil Corso on 16 March, 2005 - 12:04 am
Responding to Joel Young's plight:

Thus far you have supplied a great deal of qualitative observations. How about some quantitative info. Some of the questions that would provide additional insight are:

1) Is there a noticeable difference in "noise" between the start (higher current) and run (ower current) operation?

2) Floating the motor's carcass is not only unlawful, it is downright dangerous. Is it grounded now? To what?

3) Have you tried to operate the motor with a conventional controller, i.e., DOL?

4) What were the results of voltage and current measurements? At various locations?

5) Although you fed the motor from an alternate source, have you tried replacing the motor? Another VFD?

6) What is power source to motor configuration? 3-ph, 3-w? 3-ph, 4-w? In magnetic or non-magnetic conduit? With or without grounding conductor in close proximity? PF correction capacitors? At source bus? At motor VFD? At motor?

7) The ground-path current you measured is unacceptable. How was it measured? In the motor's grounding (bonding) conductor?

Regards,
Phil Corso, PE {Boca Raton, FL, USA}
[tal-2@webtv.net] (Cepsicon@aol.com)


Posted by Gerald Beaudoin on 16 March, 2005 - 4:35 pm
I would certainly NOT let the motor float as this could present some very serious safety issues. I like your idea to tie the motor right back to the ground rod with an insulated low impedance conductor. Wonder if you have tried changing the carrier frequency of the drive?


Posted by LINCman on 31 March, 2005 - 3:40 pm
Joel,

Are you sure that the entire cable run from the drive to the motor is in metralic conduit? If an any point the cables are in exposed cable tray, PVC conduit or non-metallic flex conduit, the output conductors will be acting like a powerful antenna.

If it is in metal conduit the entire way, and that conduit is grounded, then I agree with the poster who suggested a Class B EMI/RFI filter.


Posted by Adam Anderson on 22 March, 2009 - 7:11 pm
Filters won't do too much to reduce the noise.

By design the VFD's are chopping up approx 700VDC supply into a complex waveform to control the motors. This is done in conjuction with motion feedback from the AC motor. Using a Filter on the output of the drive simply means that the feedback loop is modified. The VFD will compensate for the filter and you will see only a small reduction in Noise.

The power levels mean that there is quite a bit of EMI radiated from the cable to the motor. You may be actually picking up nearfield noise (ie the magnetic component) of the EMI. so simple screened cable will not be sufficient. You need magnetic screening, ie steel conduit as well as screened cable.

Secondly earth wire is insufficient due to the "skin effect" of the noise waveforms. You need flat copper or flat braided earth straps, between the Motor/VFD/Earth.

At the end of the day only reducing the distance between the VFD and the Motor will have any great effect.


Posted by David Herrell on 26 March, 2009 - 3:57 pm
For many VFDs, harmonic switching noise can be coupled out of the VFD power input into the building power wiring. Once harmonics are coupled into the building power, they can appear on sensitive equipment throughout the building. This problem may be made worse by a VFD with a poorly designed input filter and a small number of switches between the input AC and the DC link.

An appropriate filter on the VFD power inputs, designed to eliminate input switching harmonics resulting from the VFD AC to DC input switching, may help eliminate the problem. It may be less expensive and easier to place the audio equipment on some form of filtering, like a true double conversion UPS, which would significantly reduce the VFD (and any other source) harmonics from the audio equipment.


2 out of 2 members thought this post was helpful...
Posted by Conor on 25 August, 2012 - 5:35 pm
I am by no means an expert on this topic and I am reading many great comments from my peers on how to reduce the effects of EMI or the source of the issue, but I feel my comments below might be useful, although perhaps too late, in this discussion.

All of my comments can be verified and referenced from any number of IEEE papers, books by industry consultants like Henry Ott, EMI Mitigation manuals from various VFD manufactures (SEW tells the story very well), an o'scope, Pearson Wide Band current probe and a simple lab bench test stand with a motor, isolation mount and VFD or servo drive (the higher the bus voltage the easier to see effects).

The IGBTs and their subsequent switching times (as well as any switching power supply, DC, servo, PC, etc) will generate some level of EMI. The dv/dt effect, change in voltage versus time change is one of the factors. A 170 volt bus (rectified 120v) with a delta t of 1 usec will have a lower EMI effect on the "system" than a 680v bus with a dt of 200 nsec. The dt component will result in a frequency of fr (rise frequency) = 1/pi/t, therefore 200 nsec = 1.6 Mhz. Next we should consider the coupling effect these dv/dt transitions cause, because they can couple via stray capacitance between other victim conductors, ground planes, motor frames. Mutual inductance and galvanic (conducted) coupling also play a part. Now we have this high frequency (HF), due to "tr", current circulating in our system.

Over the past 10 15 years we have seen a big change in the use of more higher bus (most DC power supplies are no longer 120 volt inputs, but 230 and 480, not to mention servos and VFD drives) devices and these devices now use much faster (more efficient) switching devices. Therefore the same "marginal" control system wiring, shielding and bonding practices which wroked previously, will no longer work effectively.

Circuit theory tells us current will return to the source and in this case the source is the power structure of the VFD. We also know current will flow in any parallel path provided, but proportional to the resistance or in this case impedance (AC current, so don't expect to measure with your DMM). So now we can assume it is to our best interest to "manage" the route of the current, limit it's effect on "victims" and also reduce it's frequency & magnitude.

Three common mistakes (my opinion) made in modern control systems (other than CE countries) is the lack of shielded cable use, the incorrect termination of the shields of these cables and the continual use of "antiquated" painted panels in control enclosures.

Also, there should be a difference noted in a control system ground (UL / safety, little green wires) and HF bonding component of a control system. The ground is to protect us from harm and the HF bond is to create a low impedance "equipotential plane" for HF currents to flow without large voltage potentials being created.

Okay the stray HF is circulating to return to the source, so we should provide a low impedance path(s) for it to travel on. Let's try to keep it (or the largest part) out of our "ground". One good way to accomplish this is to use good braided, shielded cable AND properly terminate the shields on both ends. The braid provides a low impedance path, perhaps the lowest impedance of any path available, from the motor frame to the chassis of the drive (or to a back panel mounted shield clamp). The shield - a bonded conductive barrier - also provides a Faraday shield to reduce the coupling effects along the length of the motor cable.

Another point to note is a bonded conductive barrier between two conductors running in parallel can provide the equivalent of 6 8 inches of separation and the coupling effect is reduced by the square of the distance of isolation.

Now we have the largest part of the current routed back the source. In order for the HF current to route back into the VFD power structure, we need either HF decoupling capacitors (sometimes provided in the VFD), an Input Line Filter (LF) or a grounded transformer secondary. The decoupling capacitors and the LF integral capacitors act as a short to HF currents and an open circuit to power frequencies (DC, 50-60 hz), routing the HF current back into the VFD, it's source.

If neither of these two routes are provided, then the closet, grounded transformer is the route to get the current back onto the power phase conductors and routed back to the appropriate source (VFD). It is easy to imagine if we choose to save money by not installing LFs, isolation transformers or VFD's with HF decoupling capacitors and the transformer is hundreds of feet away, then the route / path from the transformer to the source can be "polluted" with all sorts of HF "noise" and exposed to a variety of victim devices or conductors.

To reduce the impedance for shielded cable terminations, the only termination acceptable should be a 360 degree or circular bond (conductive gland clamps, shield clamp systems, etc). The primary issues are impedance and "skin effect", which basically forces current flow to the outside surface of any conductor or group of shorted conductors. Skin Effect basically states for a given conductor type (magnetic permeability and resistivity) of a given length, the higher the frequency the current carrying "layer" of the conductor becomes thinner.

A point to note is a 16ga, 8ga and 2ga copper conductor 1 meter long will have the same impedance above 8k hz. Since most control devices will have carrier frequencies (internal clock)above this, not to mentioned the semiconductor rise time component (tr), most of us need to rethink what is acceptable for bonding critical components together to create a "equipotential plane" for control devices to reference. We can also quickly determine little green, short or long, ground wires and painted (non-conductive) back panels are no longer useable.

Output chokes on VFD motor conductors and Low Pass DC filters can also be used to lessen the effect of EMI in control systems and the impact on nearby systems.

-We can manage the route of EMI and attenuate its effects on sensitve devices

Create a equipotential bond plane to bond all devices to. A conductive back panel bonded to a conductive machine frame using a conductive (wireway, conduit, flat braid, ect), can be very effective. Remember you are dealing with HF currents and "skin effect" or skin depth" must be considered, so do not rely on a welding cable or #2 copper conductor. These will appear as high impedance paths to the HF current. Since skin effect forces the current to a thin layer of the conductor, wide, thin and as short as possible (length of a conductor still increase the impedance) will be the best bond source.


Posted by Kevin on 10 July, 2014 - 8:46 am
Dear Sir,

I was doing some trouble shooting and observation on inverter operation today. We have 3 inverters in a control panel to drive 3 motors (of water pumps)

The 3 inverters were not able to drive the motor, each time it was given command to operate, it just displayed ground failure. All the 3 units are the same.

We isolate the item one by one, start from motor, then motor cable, then ground cable of the control panel. We measured the insulation and continuity. All of above gave good reading. So this could not be the problem of grounding. So we suspect noise from the inverter may be the problem.

We decided to put additional ferrite core on the output side of the inverters. We made it to 3 pieces on the output side from the original of 1 piece. We then operated and observed the outcome.

Problem/observation:
inverter displayed a ground failure when it was given command to drive. In this case the motor cable is PVC single core with length of 2m, in a PVC flexible conduit.

we decided to change the cable to a 2m flexible neoprene cable with 4 core without conduit. It still displayed ground failure and would not drive.

We than change the motor cable to a 7m neoprene flexible cable of 4 core without conduit. This time, the inverter worked well with this longer cable. What is the reason behind this, anyone can help?

I was taught that the motor cable should be short to prevent EMI noise problem. However in this case the result shows otherwise. Please advise.

Thanks for your input and sharing.
Kevin


Posted by Curt Wuollet on 11 July, 2014 - 12:13 am
Sounds like an insulation breakdown/harmonic problem to me. But it would really take more information to do more than guess. Like voltage, brand of inverter, insulation class of motor, exactly how wired, etc. Bear in mind that what is a ground for 60 Hz. or DC may not be a ground for harmonics of the carrier frequency. And the high frequency content can cause "leakage" which is simply due to the capacitance.

And _I_ was taught (in the school of hard knocks) that the first thing you do for VFD EMI issues is to use shielded cable intended for VFD use with Polyethylene insulation. I suspect a chunk of that rated for the drive voltage class, properly terminated, would fix your issues. Normally short runs aren't that picky, but some lengths are just not favorable with standing waves. And "any old wire" is risky even if it works in the short term.

In fact, I'd bet lunch on it. Why not? you've tried everything but that.

Regards
cww

Your use of this site is subject to the terms and conditions set forth under Legal Notices and the Privacy Policy. Please read those terms and conditions carefully. Subject to the rights expressly reserved to others under Legal Notices, the content of this site and the compilation thereof is © 1999-2014 Nerds in Control, LLC. All rights reserved.

Users of this site are benefiting from open source technologies, including PHP, MySQL and Apache. Be happy.


Fortune
Real Programmers don't write in PL/I. PL/I is for programmers who
can't decide whether to write in COBOL or FORTRAN.
Advertise here
advertisements
View free setup and multi-vendor EtherCAT demo videos online
Time to incorporate data handling, web HMI and motion in one system!
164-page eBook free download - EtherCAT Applications Guide
Control.com is the largest Automation community on the web. Learn how to advertise here now...
Servo, steppers, analog, digital & web HMI - Fully Integrated!
our advertisers
Help keep our servers running...
Patronize our advertisers!
Visit our Post Archive