One Line Reactor Multiple VFDs in Parallel

Base the size of the common line reactor on the cumulative INPUT current ratings of the drives. Input current on drives is different than the output current, which is the rating basis of the drive. Then if the drive has a DC bus choke, the input amp rating is often LOWER than the output amp rating. But small drives like you are looking at might be the other way around, because they typically DON'T have bus reactors, they use all caps for smoothing, and the caps make the RMS input current look higher at low operating speeds. So for example I'm looking at a drive that is rated for 11A continuous output (3HP 230V), so too small to have a bus choke. The max. input current is 13.8A. But a 15HP of the same drive, which DOES have a DC bus choke, is rated 62A output, but only 56A max. input. If your small drives use single phase input, the input current goes up a LOT more than the output current rating. So watch those numbers, it can make a difference in your reactor sizing.

Also because you can have longer cumulative conductor lengths, I like to bump my common input reactor size up by around 10% or so. I don't get excited about it if it's only 7-9%, I just like having a little "fudge factor".

Thanks very much for the reply

So for a common line reactor configuration as long as i size it correctly i could use say 4-6 drives on one common bus that has a line reactor on the feeder to the bus. I understand what your saying about input amperage being different then the output current.

So getting back to my other thread, sorry i should of put them both together.

When using a common line reactor setup and you want to decide if you even need the line reactor or not

Using the 10 X rule of them on the trasnformer

Should you still look at each and every drive individually or the sum of the KVA of them all

For example, say i have 5 drives i want to run in parallel off one transformer.

They are all 10KVA for the sake of it. Should i say my transformer max size without line reactors should be 10X one single drive or 10 X the total drive

aka.

10KVA*10 or (5X10KVA) * 10

Im trying to understand the implication of the harmonic contribution on the line from one drive vs many drives. Are the harmonics mainly caused from say the largest or smallest drive or is there a harmonic contribution from each and every drive and it all sums up... if that makes any sense.

Also since we are on this topic, how are linear loads affected by line reactors?

So say on a single bus i have 5 small VFD's and maybe one Direct online motor and maybe one soft start.

Obviously the line reactor in series is going to give me another voltage drop and other then hurting my power factor a little bit i cant see or think of any other adverse conditions to running a linear load on the bus that has a common line reactor also?

Any thoughts on that?

Thanks again for the replies and ideas.

I always ran on the conservative side - without line reactors the smallest drive is the limiting factor - so the largest transformer is approx 10 x the smallest drive.

With additional line reactors you increase the impedance to individual (groups) of drives so the small drives appear to be attached to a smaller transformer (higher impedance). Basically you are limiting the current to the drive.

J Raef may have more ideas.

Scott165,

You might want to check with the reactor's manufacturers, to see if they can point out any problems.

I am worried about heat build-up inside your enclosure. VFD's have a small but very real heat outut. This is many times larger than contactors, and is there for the entire time the drive(s) run. Check the VFD manufacturers' catalog data for the actual amount per drive, and then start to figure out how you can get rid of this "new" extra waste heat. Otherwise you will have the drive(s) fault out on excessive bus temperature.

You speak about using an existing cabinet and fitting these drives into the available space. Please pay careful attention to the required space above, below, and sometimes on the sides of each drive. If you have other components in the panel that are inside these spaces trouble will occur.

Drives typically require the control wiring to be run in separate wireway(s) and not close to the output (and input) power wiring. If this is impossible, you may need shielded cables for not only the controls but likely also for the output wires to the motors (the good cables are specifically designed for this application).

Pay careful attention to the drive manufacturer's specification for overcurrent protection on the drives input wiring.

That's some of the stuff for you to think about. --JMM

Hi JMM

Thanks for all the thoughts.

I should of been more clear. The Cabinets were designed with ari conditioning and were deisgned around the drives.

This is a new installation that has not been commissioned yet. The only addition i am considering making is when to add the line reactors or not.

For the small drives can i add a common line reactor and when multiple drives are in parallel with a common line reactor what is the limiting factor to decide if you need one or now.

For example say 4 or 5 drives on one distribution block being fed by one transformer.

The rule of 10, is it 10 combined or just the smallest one that is the deciding factor like i was mentioning above. I cant see to fine the answer on this except for the one person above that was kind enough to mention he believes it is the smallest one that is the factor.

I was also reading that the major cause of line harmonics is the low output inpedance of transformers?

Does anyone know what a typical power factor for a VFD is?

What is the typical input impedance of a VFD?

Thanks again for the continued group thought

Scott165,

Your reply implies that the drives have already been specified (brand, catalog number) and you are only looking at the power quality the cabinet imposes on the upstream supply. I am not the most knowledgeable person on this forum--have only installed about 100 drives, all low-voltage in sizes from 1/2-200 HP. The old electronics term for a reactor is a "choke", because it limits the inrush current to a device. Assuming all your drives, or groups of drives receive their input power at the same instant (when power to the panel is turned-on), then the reactor will do its job with a group. The input section of the drive may have active or passive components to filter out the harmonics it imposes on the incoming line. But without knowing the brand and size of drives you plan to use, I am in no position to know this.

The AC input power within the drive goes to a rectifier section to charge a capacitor bank and supply the DC bus. At low loads, the DC bus voltage is very close to the AC line's peak voltage (1.414 x the RMS voltage), so the rectifiers conduct power for only a small portion of the input sine wave. At high loads, the bus voltage is somewhat lower, so the current flows for a much larger portion of the input sine wave.

Engineers calculate the resulting input current (not voltage!) waveform as a sum of the portions of current for the fundamental (line) frequency and all the multiples (harmonics) of it up to 33 or higher. At full drive power, the percentage of harmonics is fairly low, but at very low drive power this percentage can be surprisingly high (such as over 500%). A linear load is one in which there are no harmonics, with the current waveform being a sine wave--this is typical for an incandescent light bulb, a heater, a motor, etc. A non-linear load has harmonics--this is typical for any device with electronics such as an LED bulb, a VFD, a computer, etc.

A transformer supplying the non-linear loads has to be able to absorb all these harmonics (or at least a large portion of them), which creates a high circulating current withing the transformer and higher waste heat than would be expected for its nameplate capacity. Thus, the manufacturers rate the transformers with a "K-factor" which is its ability to absorb the non-linear loads without failure--the higher this K-factor, the better it will handle non-linear loads. By adding line reactors, you are not only reducing the inrush current, but you are also absorbing a portion of the harmonics the drives impose on the power line and thus reducing the non-productive circulating currents and extra heat the upstream transformer has to handle.

I think the factor of 10 you have mentioned is the suggested limit of non-linear loads to one tenth of a non K-factor rated transformer's total load (a ratio of 1:10).

You asked about power factor. Power factor is the trigonometric COSINE of the phase angle between the voltage and current waveforms for a linear load. I learned "ELI the ICE man" as a way to describe these--for inductive (L) loads the voltage rises first and then the current; while for capacitive (C) loads the opposite occurs. Motors and transformers are inductive loads, while long power lines or circuits are capacitive loads. However, VFD's are non-linear loads and cannot be described as having a power factor, even though they supply a motor.

Look at a good transformer manufacturer's technical literature regarding sizing transformers for non-linear loads and for VFD's. Although fairly expensive, you may want to consider a drive isolation type of transformer ahead of the panel. This post has not gotten into the topics of drive output waveforms, carrier frequency, impedance mismatch between the motor and its supply conductors, use of motors designed for use with VFD's (insulation has a much higher voltage rating), induced currents on the motor shaft (transmitted to ground through the bearings), and others.

You asked about transformer and drive impedances. I have already replied that the drive causes the line harmonics and the transformer reduces them. The transformer impedance is directly related to its internal heat losses and inversely related to its ability to pass through fault currents. A drive without any internal active filtering is inherently a low impedance device. This will redirect you back to the topic of using a 2% or 3% impedance line reactor. This reactor's impedance is based on the reactor's rated load current, so a reactor has to be sized for all the currents of the drives it supplies, and not by the input current of any single drive in a group.

I'm sure others will be able to find and correct possible errors in my post, but I hope this helps you.

--JMM

Thank you very much sir

I guess i never really considered that non linear loads dont really have an impedance angle. What i was thinking was trying to determine how to model the impedance of a drive. That is why i was asking about power factor. I do tend to interchange the words power factor and impedance angle because they are effectively the same thing as long as you remember which is positive and which is negative when it comes to the angle and leading and lagging. Yes i once upon a time learned Eli the Ice Man.

This is all very interesting stuff here actually. I really wish i knew about such isolating transformers as you speak of and this K factor business.

My 10 times factor is just something i read somewhere. No idea if its accurate or not or if it even relates to K factor.

Someone just mentioned to me that if the transformer is 10 times the drive in KVA then use line reactors.

Thanks again for taking the time to write.

Looking forward to seeing what others say

One of the reasons to install a line reactor is to prevent 'line notching' on parallel connected drives, as you recharge the dc bus capacitors. The parallel drives all have to fire close to the same time, matching the incoming line frequency. The drives on your common bus will all interfere with each other, to some extent, and the upstream line reactor will somewhat accentuate that voltage dip, and perhaps make the situation worse than it needs to be.

For drives that small, I never bothered with line reactors for power quality, so I'd be inclined to leave them out, unless the non-linear load is a large proportion (>40%) of the total load on the next upstream transformer.

The other practical use for line reactors is to support the local line voltage during transient events. I had older (like 2005 vintage) Rockwell drives drop off line, while magnetic motor starters stayed latched in, on the same bus. One other VFD with a line reactor also stayed on line, just tripped off the one VFD that I didn't have room to add the reactor to. I did find room to add the (3%) reactor.

For a six pulse diode front end the PF is about .826

For fancy low harmonic AC drives it approaches unity. (But they are expensive drives with active front ends).

interesting.

Thanks very much for the info

I question the basic design quality of any drive/rectifier front end that destroys itself in normal operation, especially if the application instructions don't require an input reactor, isolation transformer, unless perhaps applied on an extraordinarily inflexible (stiff) power system. Low voltage system supplies up to 2,000kVA are routine in industrial environments (50kA @ 480V for instance), so most reasonable quality equipment is designed for easy application in that environment.

However, line reactors to protect external components is common application practice, perhaps over extended out of an abundance of caution. I witnessed the incineration (from the inside) of some 2000kVA transformers with 100% small PWM loads, (A-B 1336 Drives). This around 2001, before the effect of harmonics on transformer laminations was more widely understood. Grouping drives downstream of a reactor defeats some of the purpose of the reactor in the first place.

I worked for a while for a small company that made a “bulletproof” VFD which came with a line reactor whether you needed it or not. In 3 years I experienced exactly zero field failures. But that made our drives more expensive than anyone else, and they went out of business. People want the cheapest thing they can get, so drive mfrs feed the market demand by making small drives as cheaply as possible, then “recommending” they use line reactors “if necessary.” The fact that they are essentially almost always necessary is not their concern, it’s up to the end user to make that decision.

Thanks for the info Jraef

Would you say the smallest drive on the bust is the limiting factor or the sum of drives together?

Limiting factor for what? Line reactor inductance, current rating? If you put two VFDs behind one reactor, then the each of these drives will present harmonics to the input section of the other. The line voltage drop after the reactor is increased, and stands a good chance of reducing the voltage enough to trick the phase passing timing circuits in the other VFD. Better to install no reactor, after examining the true consequences and need for the reactor; probably not needed at all, but you haven't provided enough detail.

All drives have a rating for available short circuit current from the source, easy to calculate a conservative value for most power system configurations. If you do not exceed the short circuit rating of the drive, and drive components fail internally, it is certainly not your fault.

If someone makes a drive that cannot be applied as designed, they also won't last long in the market place.

What i mean is

In reality EVERY drive from the main utility transformer is in parallel with each other.

The problem i am trying to cure is, say you have a single bus bar and 5 VFD's attached to it.

Say they are 5 kva each

Because of limited cabinet space, is it possible to put a line reactor on the feeder that is feeding the bus that is rated for 25 kva instead of 5 individual reactors on each VFD

That is all im getting at.

In regards to my question on limiting factor, what i meant was....

If you used the 10 times the transformer rule discussed above in determining if line reactors are needed or not

AND

You were trying to use one line reactor (common reactor ) for the same (5) 5hp drives

using the 10 times rule

Is the answer (5 *5KVA) * 10 = max trasnformer size or....

Is the answer (1*5KVA) * 10 = max transformer size

Thats what i was meaning by limiting factor.

The line reactors are needed if the predominant load on the supply transformer is non-linear. Otherwise they are of little practical value.

The drive manufacturer's instructions will very clearly indicate what short circuit rating the front end is rated for. If your supply transformer allows more current than the drive is rated for, then you might need to add a current limiting reactor, which is not necessarily a harmonic smoothing reactor.

Adding a line reactor for grouped drives is good for other loads on the transformer, however, it may cause front end cross-talk (due to notching) on the drives downstream of the reactor. This is highly dependent on the drive design, and requires intimate knowledge of the ac line coupling & measurement circuits used to control the drive.

You need to install a Line reactor for each drive independent from all others. They also need to be sized for each drive respectively.