VFD Line and Load Reactors / Filters

[adamscb]

Forum,

Are there general rules of thumb in the industry for when to use line/load reactors? I've always known that they provide protection against harmonics for the incoming power supply, and protection against high dv/dt for motors, etc. With that being said, is there a point in which the advantages of using filters is not worth the extra cost/space in MCC starter buckets? (most of the VFD's in my plant are built-in to the bucket, along with some older styles that are wall-mounted). Below are some bullet points in which I'm already aware of, with some specific questions I have.

Line

• I've heard that if a certain percentage of a transformer's load is controlled by VFD's, line reactors need to be installed. What is the general industry standard for this?
• What are the benefits/disadvantages of using 5% line reactors, instead of 3%? Is the only downside cost?
• I've read online that with motors above 20hp, you should use 5%. Is this just because of the load increase?

Load Reactors

• From my understanding, the levels of protection for motors on VFD's (going from weakest to strongest) is load reactors, dv/dt filters, sine-wave filters, and active filters. From my understanding these are based on cable lengths of the wiring to the motor - the longer the cable length, the more you travel to the right on that list.
• Same question with line filters - are there any advantages/disadvantages to using 5% instead of 3%? Is the only downside cost?

Thanks for the help!

 

[petersonra]

I would suggest this. There are a bazillion vfds out there. I don't know what percentage of them have any of these items on them but I would guess it is probably single digits. There are cases where they're needed but unfortunately it's not always easy to tell when you really need them or not.

I would suggest that for line reactors how close you are to a big Transformer or automatic power factor capacitors that the utility uses might be more of a indicator of when you should use LINE reactors than anything else.

I don't think you'll probably ever need load reactors with modern vfds.

Dv/dt filters are probably not something you'll ever need either.

The thing with stuff that's on the load side of the vfd is it is highly dependent on distance between the vfd and the motor. The longer the distance the more likely you will need some kind of protection from reflected waveforms. Again this is much less of a problem with modern vfds.

I don't believe that there is any good rule of thumb you can use to determine when you might need these things other than to follow the manufacturer's recommendations on such things. Personally I think there is a good chance that if you use them extensively you will spend more money on them then you will on the small vfds that are the most common these days and you will never get any kind of return for the money you spent. Small vfd are inexpensive enough that if they fail they fail and you replace them.

 

[ptonsparky]

We deal with a just few VFDs per year. Line reactors because of KVA of POCO transformer size, switching of PFC by utility and load reactors for distance or larger, older, non VFD motors.

The rest are voodoo and we haven’t had reason to use them as yet.

 

[Jraef]

Here are the generally accepted rules:

Line Reactors are what I call “cheap insurance for expensive parts”. Without them, your drive Front-End components are exposed to the raw and often dangerous line issues coming from a utility. I could give you an hour long talk on what those are, but the Readers Digest version is that because of how rectifiers function, transients in the line can create dangerous conditions for the rectifier components. Adding reactors adds what’s called an “inductive time constant” to those transients that slows them down, which allows the drive to handle them better.

Then as to 3% or 5%, that depends a little on the drive design and a lot on your ratio of short circuit current to fault current. . If a drive comes with a DC Bus Choke, that performs some (but not all) of the same functions, so on those drives, 3% is fine. 5% is, in my opinion, rarely needed, unless you have a very large source transformer (that ratio issue). So the rule of thumb is; if your drive has a DC bus choke, you must add line reactors (3%) if the source kVA is more than 20X the drive kVA (20:1 ratio). With no bus chokes, the rule is 10X (10:1 ratio). But also, you can get the same benefits from having one reactor ahead of multiple drives, so if you have a cluster of drives on a wall or in a panel all coming from the same feeder, you can save on the installation cost by using one reactor sized for the collective drive input Amps. There is a down side to 5% reactors in that the added impedance results in a significant (but not large) voltage drop, which can then limit the output voltage of the drive, which can end up not allowing you to run at full speed/torque.

As to load reactors/filters, I agree, load reactors are almost never needed and some mfrs will give you tested distances from drive to motor that you can use to decide when, some of those can be hundreds of feet depending on size and the motor used. If no tested values are given, the only safe bet is 25ft max. So when they are needed, I usually recommend using at least a dv/dt filter instead. That already has the load reactor as part of it and adds some filtering as well. Sine wave filters are necessary only when the distance gets extreme, ie 2,000ft or more, but can also be used to quiet motor whine in applications where people get annoyed by it, where older non-inverter duty motors are used and can’t be replaced, and now are recommended by a lot of the submersible pump suppliers, because the flat cables they use can lead to other problems since you lose the cancellation properties of twisted cables.

 

[Besoeker]

Talking 3-phase applications........

The PWM output to the VFD output to the motor has very sharp rising edges. That sometimes causes problems for the winding insulation. I've seen well in excess of 3000V/uS at the motor terminal box. Or, in the case of submersibles, the surface terminal box. I have seen a few failures of motors of up to 760 kW (>1000HP).
As one wag put it at a pumping station where there were a few failures, 760kW motors are not to be considered as consumables.
Anyway, that's one significant application for output reactors.

The most common input configuration is a three phase six-pulse plain diode bridge where each diode will typically conduct for 60deg electrical. As the current moves from one diode to the next the one going off is reverse biased and reverse current flows for a little time and the supply is temporarily shorted. This can cause supply notching. Input reactors are a common fix for that if it is problematic.

 

[drktmplr12]

It might be worthwhile to test harmonic content at the point of common coupling. and a couple other select points to see if the harmonic content is something that you need to worry about. Jraef is right in that line reactors are cheap insurance.

We put line reactors on anything above 3 HP. Anything 100 HP and higher gets an 18 pulse phase shifting transformer.

 

[Besoeker]

some mfrs will give you tested distances from drive to motor that you can use to decide when, some of those can be hundreds of feet depending on size and the motor used. If no tested values are given, the only safe bet is 25ft max.

It isn't just about distance. Connect the motor close to the drive terminals and you see eye-watering dv/dt.
Then there is the motor itself. Before the advent of VFDs, it was common to have random wound motor windings. These were prone to insulation failures. I've seen a few of those and you probably have too. A VFD retrofit is liable to fry those unless you fit chokes or some other means of limiting the dv/dt.

 

[Besoeker]

It might be worthwhile to test harmonic content at the point of common coupling. and a couple other select points to see if the harmonic content is something that you need to worry about. Jraef is right in that line reactors are cheap insurance.

We put line reactors on anything above 3 HP. Anything 100 HP and higher gets an 18 pulse phase shifting transformer.

Good man!
An active front end is another solution.

 

[Jraef]

It isn't just about distance. Connect the motor close to the drive terminals and you see eye-watering dv/dt.
Then there is the motor itself. Before the advent of VFDs, it was common to have random wound motor windings. These were prone to insulation failures. I've seen a few of those and you probably have too. A VFD retrofit is liable to fry those unless you fit chokes or some other means of limiting the dv/dt.

Yes, I should have emphasized that more than I did. If the motor was not MADE to be run via inverter, a filter is almost mandatory. But we have a lot of 230V 3 phase systems here and the insulation is the same if it were 480V, so they can handle that a LITTLE better if the distance is kept short. So the only time I recommend filters on those (typically machine tools) is if the motor is old and/or so specialized that replacing it is going to be difficult and costly, if even possible. When a machine has a motor deeply embedded in it and that motor is some custom frame/mounting design, a sine wave filter is a must. When the man-hour cost it takes to get AT the motor exceeds the the value of the motor, the filter can pay for itself in avoidance of down time.

 

[mike_kilroy]

Another good point is to understand Just what that 1.5, 3, and 5% means. It is erroneously called impedance. It is rather the percentage voltage drop across it at its nameplate rated current.

So pop a 5% on a low line with no taps and you may be in trouble.

Sent from my SM-G900V using Tapatalk

 

[GoldDigger]

Another good point is to understand Just what that 1.5, 3, and 5% means. It is erroneously called impedance. It is rather the percentage voltage drop across it at its nameplate rated current.

So pop a 5% on a low line with no taps and you may be in trouble.

Sent from my SM-G900V using Tapatalk

Do keep in mind that the voltage drop will be across a primarily inductive impedance, so the reduction in voltage seen by a near 1 PF load will be significantly less than the difference in voltage between the terminals of the line reactor.

 

[Jraef]

Another good point is to understand Just what that 1.5, 3, and 5% means. It is erroneously called impedance. It is rather the percentage voltage drop across it at its nameplate rated current.

So pop a 5% on a low line with no taps and you may be in trouble.

Sent from my SM-G900V using Tapatalk

Right. That voltage drop is usually insignificant at 3% but 5% puts you at risk. So as an example let’s say you have a 480V service and at the VFD terminals, you are at 95% voltage after the drop due to distance, so 456V. A motor will be rated for 460V, +-10% so it can handle down to 414V and still provide rated torque at rated current. So adding a 3% reactor ahead of the drive results in a maximum output voltage of 442V at the motor and the motor is fine with that. In this example even a 5% reactor would only drop it to 433V.

But if your voltage drop from service to drive is already at an otherwise acceptable 10% (432V), then you add another 5%, you are now at 410V and out of spec for the motor, whereas a 3% reactor would have had you at 419V, still in the acceptable range.