VFD Noise on Incoming Line

JRoss

Lifetime Supporting Member
Join Date
Sep 2006
Location
Dillsburg, PA
Posts
184
Hi everybody. I have a customer with four 350HP Mitsubishi VFDs, each driving four 25HP motors. Each drive has a line reactor on the output side (between the VFD and the four motors) and a DC Reactor.
The system is a pavement testing rig and the 16 motors drive the rig up and down a track. One of their engineers contacted me because they upgraded some of their data acquisition equipment and are now seeing noise on some of the sensor lines whenever the rig is moving.
The old DAQ system had integral low pass filters where the new one does not, and they suspect that may be why they are now seeing the noise. However, they want to know if there is a way to reduce or eliminate the noise at its source before adding filters back in, as these could degrade the quality of the data they are collecting.
So my specific questions:
1. Would installing a line reactor on the input side of the drive help?
2. If I install an input line reactor, can I leave the DC reactor in place?
3. Are there any other suggestions for reducing the noise produced by the drives?

Thanks,
Jeremy
 
One big culprit is the method of wiring the motors to the drives. A LOT of people pay almost zero attention to that, despite the fact that the output conductors become, essentially, FM radio transmitting antennae. This is especially prevalent on moving machinery where they will use flexible cord and not bother to use those with braid shielding because it's more expensive. I would start by thoroughly investigating that aspect before bothering with anything on the line side.
 
As jraef said most of the issue is likely from having output cables that are not shielded.

VFD cable is used on drive systems for several major reasons one of which is containing noise when using shielded VFD cable without it all the noise created from the firing of the IGBT's in the drive and the rectification circuit is going to travel all over that machine and possibly to other nearby equipment if there is a suitable path.

Also from your post if a reactor is on the output side of the drive it's termed a load reactor just to reduce and confusion.

You normally want to see a line reactor on the input side to add some impedance to the system.

My money is on your output cable. Start there before you do anything else.
 
Thanks for the clarification on the terminology.

The rig is huge, so the control panels are mounted directly on it, and it has a power rail like on a crane. VFD output is wired to the load reactor, then to a distribution block, then to individual overload protection, then to the four motors. Wires from the motors are going through conduit much like you'd see in any fixed installation. Assuming this is all grounded properly, doesn't that provide the shielding?

The DAQ doesn't run through the rig at all. Sensors are imbedded in the pavement and run directly back to a control room beside the main testing area. They're seeing the noise on pressure sensors, I believe.
 
Thanks for the clarification on the terminology.

The rig is huge, so the control panels are mounted directly on it, and it has a power rail like on a crane. VFD output is wired to the load reactor, then to a distribution block, then to individual overload protection, then to the four motors. Wires from the motors are going through conduit much like you'd see in any fixed installation. Assuming this is all grounded properly, doesn't that provide the shielding?

The DAQ doesn't run through the rig at all. Sensors are imbedded in the pavement and run directly back to a control room beside the main testing area. They're seeing the noise on pressure sensors, I believe.
From my experience, pressure transducers are, for some reason, especially sensitive to RFI coming off of VFDs. I once had to prove it to someone by removing a 5HP VFD from a control panel back plane and, because the leads were long, move it around to show how the location of the drive was affecting how much interference was getting to the pressure transducers. The cables inside of properly grounded STEEL conduit shouldn't be emitting, that's true, but even the short pieces going from the drive to the reactor and the reactor to the splitter will radiate RFI. Is the enclosure steel? Does the enclosure door happen to be open when you are seeing the interference? Are you POSITIVE that everything has GOOD ground connections?

Once you have eliminated ALL of those more likely possibilities, you could next try installing an EMI/RFI filter on the line side. The rectifier on the front end of the VFD will cause the incoming line cables to radiate RFI as well. This is especially problematic on SCR front end rectifiers, a design on larger VFDs that eliminates the need for a pre-charge resistor circuit. Some mfrs build RFI filtering into their drives, but not all. Yours may be one that didn't.

http://www.transcoil.com/Products/KRF-EMC-EMI-RFI-Filter.htm
 
Last edited:
Why not connect a scope before you start changing thing around.
then you may be able t see just where the noise is coming from.

in your opening you stated that you had 4 350 HP VFD's running 4 25 HP motors each.
that seems odd to me way overpowered or something. I just wanted to make sure I understood it right.

one other thing a long time ago I had I think about 20 HP running on a VFD. every time the pump would run the flow interments would jump all over the scale no way to get a good valid reading when the pump was running. I changed the switching frequency on the out put transistors on the VFD. tried a few different settings found the one that worked, problem solved never got a call back.
 
The DAQ doesn't run through the rig at all. Sensors are imbedded in the pavement and run directly back to a control room beside the main testing area. They're seeing the noise on pressure sensors, I believe.

While eliminating noise at the source is preferable, you may want to also examine the instrumentation wiring. For example, if those sensors are not using twisted wire back to the DAQ, they may be susceptible to inductively-coupled noise from the motors. Best practice is twisted and shielded, with a single termination of the shield drain wire. That would also include any "excitation" or separate power supply to the sensors, if applicable.
 
An input line reactor should help. You probably need to verify sizing and effect on the DC reactor with the supplier, but they should play well together. Line reactors are a standard option with most drives. You will get a very slight voltage drop to the drive, and some power loss and heat, but that is likely insignificant compared to your other problems.

Others have discussed shielded motor leads - a god idea, but I have lots of successful installations without that. I'd save that for plan B.

You should make sure that all the analog signal cables are shielded twisted pair, and that the shield is grounded only at the control panel end.

You may be able to solve the signal noise by putting a capacitor across the input leads to the I/O card - usually at the card terminals. In effect you are creating a nose filter. I've used a 220 micro-Farad cap with success, but you can play with it - caps are cheap.
 
Customer is several hours from me. I helped upgrade the drives about four years ago, and haven't been back since, so most of what I have is going from memory and/or drawings. The customer sent me an email with the problem, and I'm just trying to gather some expertise before giving them some suggestions! Thanks to everybody for chipping in.

I don't know if the conduit is grounded properly, but we can certainly check.

Yes, I stated the architecture correctly. There are four (4) 350HP VFDs, each driving four (4) motors for a total of sixteen (16) motors. I had thought the motors were bigger, but the drawings state 25HP. They are custom constant torque motors. All the motors run at the same time, all the time, no exceptions.

In addition, there is an encoder on one of the motors in each set of four, which is wired to an encoder card on the VFD and used for speed control. The encoders are also parallel wired back to the PLC for position monitoring.

I know next to nothing about the DAQ system, since that is a totally separate system, though I can certainly ask them questions.

Gary, what do you suggest we scope?
 
A few thoughts on this application you take them for what they are worth the are offered to try and help.
I am not trying to put anybody down, but I have been at this for well over 30 rears these are my observation on this.

It was mentioned that 350 hp vfd with 4 25 hp motors was to allow for the start of the motors while the vfd is running.
That doesn’t add up. The rule of thumb for an across the line start of an ac motor is the inrush can exceed 10 times the full load of the motor. With this configuration all 4 motors start at the same time you would need at least a 1,100 hp vfd and even that may not be enough. You only have a 350 hp vfd. The output load reactors my help but you could never count on it they are not designed for that.

As for the use of line or load reactors, in this case I don’t think either will do much good. Reactors are designed to work in a given defined current / frequency range. In this case they would be designed to work with a 350 hp vfd but the actual load will be much less my guess would be that load would be about 60 hp. Based on the 75% rule. With a load that size the reactors would have very little effect I would not look to closely at them. The fact that they originally designed the system with that combination of 350 hp vfd and 4 25 hp motors I would think they anticipated problems in the system from the start. From the statement the vfd’s were upgraded tells me it may not have ever been working correctly you don’t upgrade unless you are having problems.

As for the single encoder for 4 motors. That’s unusual but in the right set up it could work.
It was stated that the original motors were designed as torque motors. From that I take it that the original design was to run the motors in torque mode as torque follower with speed limited.
But first understand no 2 motors will ever run exactly the same speed even when connected to the same power source. Small difference in manufacturing will make them run at slightly different speeds. It doesn’t matter if they came off the assembly line one after the other. In this case you have 4 motors all trying to run at different speeds. Remember the vfd is in speed mode trying to regulate speed of all 4 motors but it only sees feedback from one motor so it doesn’t really know the speed of the whole system. But if the other motors are running faster or slower and each of the others will be at a slightly different speed. Even if it’s just .5 rpm difference the motors will end up fighting each other with the vfd trying to regulate speed
.
If I read this system correctly each motor drives a separated traction wheel to move the whole machine down the bed. I assume they have 8 wheels on each side a total of 16 with 4 vfd’s 2 on each side to allow the machine to travel. This will compound the differenceal speed problem because no matter how big or wide the span is, movement on one side will affect the other. Try to run just one side by its self and see how much the other side moves you will see what I mean.
I think the original designer had planned on 8 separate drives on each side. On each side one drive would be setup as a speed regulator drive (Single Encoder Feedback ) the other 4 drives would be setup as torque follower configuration.
With this configuration they share the load without fighting the speed regulator motor / vfd.
With 4 wheels running on the rail or bed they each will have to run at different speeds. I have seen this with new machined wheels on a rail with about 50 ‘ of travel they are racking binding up the system.
This configuration has be working for over 50 years. One of its original uses was with DC drives on a common shaft in multi unit printing presses.
The motor / drives don’t have to be the same size. Early in my career I had to work on a system that had a 200 hp main motor and 75 hp aux motor all connected on a common 2” shaft both motors ran at the same speed and at the same load. This was the standard up until a few years ago when they came out with Digital Line Shaft
Synchronization now that is the standard for multi unit presses and other applications as well.

With the vfd constantly changing output frequencies to correct for mismatch motor speed it no wonder that you have all kinds of crazy noise bouncing around in the system the wiring would act like a transmitter causing all kinds of problems. The conduit would be a good ground but not a good shield. Conduits are generally connected to the structure at multiple points that’s good for grounding but a good shield should only be grounded at one point if you have 2 points grounded then you have a current path in the shield forming an antenna to transmit the noise to everything close.
The best way to see what’s really going on is to put a scope on while it’s running. Look at different points
Also insure that the DAC wires are properly shielded as well ( Bonded at one point only)
 
The system was installed with this architecture in 1999. We upgraded the drives in 2013 to add regenerative braking which reduces wear and tear on the mechanical brakes and gives more controlled stopping. They also wanted to avoid obsolescence problems in the future. We also added the load reactors and individual motor protectors that were not part of the original design.

The original DAQ system didn't pick up the noise, likely due to built in filters the new equipment doesn't have.

Gary, where would you have me scope the system? What signals should I monitor?
 
Jeremy
I have a sope and you are welcome to use it. I have not used in a long while just didn't have the need. it's in storage right now I need to get it out and find the leads and make sure it works for you.
check back with me on Monday.
not having individual motor protectors on the original leads me to question their thinking
Any drive manufacture will tell you they are needed even back them, but there is a lot about the original system that I would question Shielding it the first place to look you may consider replacing the motor leads with shielded vfd motor cables
 

Similar Topics

J
Is it possible that VFDs can cause noise to a degree of intermitent fuse blowing on Servo drives fed from the same immediate power source? We...
Replies
4
Views
7,552
Hello, I'm trying to delve a little into rs-485 communications for a couple projects of mine. Until now I've been using a delta vfd and a delta...
Replies
2
Views
49
I have an issue with Power Flex 525 during running processing, the VFD stopped suddenly while the PLC and VFD connection ok, VFD does not have any...
Replies
1
Views
89
Has anyone setup communications with Red Lion 3.0 MODBUS to Baker Hughes Centrilift GCS VFD? Thanks
Replies
0
Views
70
Hello, I've been trying to learn this a while now and still have not found out how this works. I have an Omron CJ2M PLC and an ABB ACS 355 VFD...
Replies
1
Views
191
Back
Top Bottom