Load reactors

Status
Not open for further replies.
With most modern IGBT design drives, the relatively fast switching frequency causes "transmission effect" or "reflected wave" issues, increasing apparent voltage at the motor terminals as the distance increases between the drive and the motor.

If the motor and drive are separated by less that 50 feet (typically), you should be able to use even a poorly constructed motor with no issues. Also, most manufatcurers would rather use some type of filter incorporating a reactor and RC circuit rather than a simple reactor.

For more information, try going to www.transcoil.com and look at the KLC series.

Also, in the drive manual, most manufacturers will give recommendations for cable length and motor protection.

Personally, I wouldn't use the line reactor between the drive and motor at 30 feet.

Good Luck!!
 
Last edited:
ptonsparky said:
We are moving some equipment and came across a VFD with 3 reactors, one on the line side and two (3%) seriesed on the load side. The motor was close to 300' from the drive. It will now be about 30'. What determines if I need 3%, 5% or in this case 6%?
Click to expand...
Sorry, didn't mean to leave off with ONLY the snarky comment, but I got called away...

Higher (5%) reactance solves different problems than 3%. This web page sums it up fairly well.

http://mtecorp.com/load.html

Most likely your customer had an issue that he thought would be solved by the 3% and it wasn't (I'd guess the distance was making him worry), so rather than pay the restocking charge plus buy the 5%, he just added another 3%. Either that or he just got the 3% units cheap, already had them etc. etc. In other words, there is no specific added value to putting them in series, only the total reactance counts.
 
Den7422 said:
With most modern IGBT design drives, the relatively fast switching frequency causes "transmission effect" or "reflected wave" issues, increasing apparent voltage at the motor terminals as the distance increases between the drive and the motor.
Click to expand...
Pton and Den

The high switching frequency is one problem.
As critical, or maybe more so, is the very high rate of rise of voltage (dv/dt).
Cable length and characteristics also play a part in that the distributed inductance, and capacitance cause ringing with the consequent voltage overshoots seen at the motor terminals.

There is no one size fits all answer.

If the motor is a short distance from the VFD, dv/dt will be very high - in some cases I have measured this at more than 2500V/us which can, and has, resulted in very early motor insulation failure.
And the motors need to be designed for VFD operation.
We had a project where several 800 kW motors failed over a period of a few weeks from installation. The main contractor, our customer, bought the VFDs from us. And the motors from the motor manufacturer directly - to save money, of course.
The motors simply didn't meet dv/dt requirements. Our VSDs did. We did bid to supply the motors. I'm that's one I'm glad we lost.

If the cable is long say greater than 100m (about 300 ft), voltage overshoot is a problem that can resulting partial discharge. If it does, inverter output frequency plays a part.
Here's something I put in a technical report:
Voltage Overshoot
At the points it was measured the voltage overshoot is within the limits of Technical Specification IEC TS60034-17 third edition. The limit given for the switching times measured is about 1200V. Using figures from the Gambica Technical Report No1, (second edition) the peak voltages recorded give a very low probability for partial discharge and even if it did cause PD, the number of pulses to breakdown would exceed 10^13. Bearing in mind that the switching frequency is below 1 kHz the motor will never see that number of pulses.
Click to expand...

The short answer is that, if the application works with the two 3% reactors in series, keep them.
 
Jraef said:
This web page sums it up fairly well.

http://mtecorp.com/load.html
Click to expand...
With respect, I don't agree.
Look at motor noise - and it's not the only point one could take issue with.
A five percent impedance harmonic compensated reactor will virtually eliminate the higher order harmonics (11th & up) and will substantially reduce the lower order harmonics (5th & 7th).
Click to expand...
Motor noise would reflect output frequencies.
Power frequency would generally be a few tens of Hz. Maybe 50Hz or 60Hz for most applications. Carrier frequency is usually in the kHz region. Harmonics like 5th and 7th would not be present.
Maybe the author of the article has confused input and output harmonic spectra?
 
I'm not sure that the dv/dt issues on short cable runs would be much of an issue so much as common mode problems. Most of the drive manufacturers (ABB, SqD, AB, etc) usually have NO installation recommendations in their manuals at close drive/ motor proximities, yet want shielded drive cable and output filters as the distance increases.

I agree that motor construction is a concern (MG1-31 vs MG1-30), and that choosing a decent motor is always a preference.
 
Good to see some healthy information and discussion on this.

Most point in others answers are very relevant. Output reactors requirements is really depend on the cable type (dielectric material) and length. For the run you have described now being shorter, good quality drive cable with balanced earthing conductors and overall screen with XLPE insulation (and correct cable glands EMC complaint) will remove the need for the reactor in my opinion. But if already installed the reactor will not cause any grief to the installation moreover reduce motor heating.
 
Den7422 said:
I'm not sure that the dv/dt issues on short cable runs would be much of an issue so much as common mode problems.
Click to expand...
Here are a couple of measurements from two drives systems on the same site and identical in every respect - except cable length.

P253Hz.jpg


P453Hz.jpg


Note the different time scales.
For P2 the dv/dt was 2000V/us.
For P4 it was just over 5000V/us.

This is typical of what I have measured elsewhere. A short cable run gives high dv/dt and low or no overshoot. On long runs, the dv/dt is lower and overshoot is greater.
 
Well got a question on VFDs why does the length effect the wave shape in simple terms ?


Is a line reactor just a choke coil or is there other types ?


I can see a line reactor on the line side to keep a back feed or harmonics down .


Why is a line reactor needed on the motor load side for what reason ?


Can you explain the term ac ringing of long length conductors from vfd to motor load what is meaning by that statement ?


just interested in the output and the inputs of how we wire and use VFDs in a correct manor and the finer points best to yas
 
Besoeker said:
Here are a couple of measurements from two drives systems on the same site and identical in every respect - except cable length.

P253Hz.jpg


P453Hz.jpg


Note the different time scales.
For P2 the dv/dt was 2000V/us.
For P4 it was just over 5000V/us.

This is typical of what I have measured elsewhere. A short cable run gives high dv/dt and low or no overshoot. On long runs, the dv/dt is lower and overshoot is greater.
Click to expand...


VERY interesting report Besoeker, thank you. What were these measurements recorded with??? What were the cable lengths?
 
ohmhead said:
Well got a question on VFDs why does the length effect the wave shape in simple terms ?
Click to expand...
The cable has distributed inductance, capacitance and resistance.
Cable length determines the values and thus behaviour.
A fast transient such as the switching edge out of an inverter can cause an LCR circuit to ring.[/QUOTE]
 
76nemo said:
VERY interesting report Besoeker, thank you. What were these measurements recorded with??? What were the cable lengths?
Click to expand...
They were measured with a Tektronix DSO and downloaded via a serial link to a PC. The file format is compatible with Excel which makes presentation in a report relatively easy.
From memory, cable lengths were around 250m for P2 and 15m for P4.
The inverters were driving 60kW submersible pumps. The site had four and the waveforms for all four were different.
 
Besoeker said:
They were measured with a Tektronix DSO and downloaded via a serial link to a PC. The file format is compatible with Excel which makes presentation in a report relatively easy.
From memory, cable lengths were around 250m for P2 and 15m for P4.
The inverters were driving 60kW submersible pumps. The site had four and the waveforms for all four were different.
Click to expand...



Good input. Thanks Besoeker:wink: What model was that TEK meter?
 
76nemo said:
Good input. Thanks Besoeker:wink: What model was that TEK meter?
Click to expand...
It wasn't a meter. It was a digital storage oscilloscope.
We have a bunch of them. The one I carry around for routine measurements is a TDS 210 with a serial port.
 
Status
Not open for further replies.
Top