Line Reactor / Load Reactor on VFD Output Side for 380V Non-Inverter Duty Motor?

I'm planning on adding a VFD to my lathe, which has a 380V 50Hz motor and built-in transformer. My power is 208V 60Hz 3ØY, the lathe's internal transformer will boost that to 380V (or higher), that would input to a 460V VFD.

My questions: Is there an advantage to running a line reactor / load reactor on the output side of the VFD? Would it be a good idea to do so when running a VFD with an older non-inverter duty motor? Will this protect the windings?

The motor is a high quality (Swiss made) motor, but is non inverter duty. It is a 2 speed 8 pole/2 pole motor. Please refer to my previous thread for details on the motor.

The lathe was made in 1982. I had the motor serviced by a motor rebuilder in 2008, and was told the motor was fine (I assume he tested the windings). I have no way of knowing what the insulation class is on the motor windings. But its not likely to be higher than 600V is it? I'm a bit concerned that if I'm running a 400V VFD and my windings are 600V insulation, that's only a 200V margin of safety for the voltage transients, etc.

Also: I'm assuming an AC reactor on the input side would be superfluous because there will be a big transformer on the input side of the VFD.

Let's break this down...

cinematechnic said:

I'm planning on adding a VFD to my lathe, which has a 380V 50Hz motor and built-in transformer. My power is 208V 60Hz 3ØY, the lathe's internal transformer will boost that to 380V (or higher), that would input to a 460V VFD.

Click to expand...

#1, if you are planning on running it at 50Hz max, 380V input is fine. If you want to run at the 60Hz equivalent, get a 480V transformer so that your V/Hz will match the motor design. In other words, 380V 50Hz is the same as 460V 60Hz, not coincidentally by the way. I know you didn't ask that question, it was a freebie...

My questions: Is there an advantage to running a line reactor / load reactor on the output side of the VFD? Would it be a good idea to do so when running a VFD with an older non-inverter duty motor? Will this protect the windings?

Click to expand...

Line reactor, no, see below (you basically answered your own question later). Load reactor, maybe. Anything you can do will help the motor last longer. More below.

The motor is a high quality (Swiss made) motor, but is non inverter duty. It is a 2 speed 8 pole/2 pole motor. Please refer to my previous thread for details on the motor.

The lathe was made in 1982. I had the motor serviced by a motor rebuilder in 2008, and was told the motor was fine (I assume he tested the windings). I have no way of knowing what the insulation class is on the motor windings. But its not likely to be higher than 600V is it? I'm a bit concerned that if I'm running a 400V VFD and my windings are 600V insulation, that's only a 200V margin of safety for the voltage transients, etc.

Click to expand...

Motor insulation is typically 2.5x the rated RMS line voltage. So a motor designed for 380V (likely 400V nominal) will have had insulation rated for at least 1200V, mostly because the magnet wire mfrs would make it good for 480V so that they can sell it all over. The problem with VFDs is because of a phenomenon called "Reflected Waves" in which the fast little DC pulses that make up what the motor "thinks" is AC, create a capacitive effect in the cables between the drive and motor, creating waves of voltage that bounce back and forth between the VFD and the motor, building up each time until the voltage gets to be up to 2000V in worst cases (even higher in extremes). 1200V insulation, 2000V pulses, you have a problem. Load reactors HELP, but do not completely cure this, by slowing down the rise time of those pulses. However, the length of the cable makes a big difference too, as does the impedance of the circuit. It's difficult to accurately predict the overall influence of all of the factors that go into it, but the generally considered point at which you MUST consider it is 25ft from drive to motor (linear feet, cable length). Less than that and I don't worry too much. But a load reactor also helps with two other issues: bearing damage and drive damage. Bearing damage is similar to the winding issue in that it is due to capacitive coupling, but in between the motor stator and rotor. The pulses discharge across the bearing races, etching them until they become rough, then they fail early. Slowing the rise time of the pulses is again a way to reduce that effect. On the VFD protection issue, the reactor will ALSO slow down the rise time of current in FAULTS that happen down stream of the VFD, i.e. shorted wires, terminations, failed windings etc., so that the VFD can have time to turn itself off before something gets damaged. Basically, a load reactor is "cheap insurance", they don't cost much.

Also: I'm assuming an AC reactor on the input side would be superfluous because there will be a big transformer on the input side of the VFD.

Click to expand...

Yes, exactly the same, no need for a line reactor + a transformer, it's either / or.

Jraef said:

#1, if you are planning on running it at 50Hz max, 380V input is fine. If you want to run at the 60Hz equivalent, get a 480V transformer so that your V/Hz will match the motor design. In other words, 380V 50Hz is the same as 460V 60Hz, not coincidentally by the way. I know you didn't ask that question, it was a freebie...

Click to expand...

Jraef, thanks for the very informative answer. I did remember about the V/Hz ratio, you mentioned in response to my last thread regarding this motor. Fortunately the transformer built into my lathe is a very high quality Swiss made unit (the entire machine is Swiss made down to the switches!) It has taps going up to 500V. So I'm fairly confident I'll be able to boost the 208V up to around 460V.

So I take it from your answer that if I'm going to run the VFD up to 60 Hz, I should use the transformer taps for 460V? I don't envision needing to run much faster than 60Hz. I can easily get high speeds by using the belt/countershaft combos. Also the motor has two speeds and on the 2 pole setting (speed 2) I get all the speed I need. I normally run 90% of the time in the 8 pole setting (speed 1)

The plan is to run at 50 Hz most of the time, and drop down to lower Hz when I need to turn larger diameters, or for trepanning. I'm hoping I can get the motor to run smoothly at 25 Hz which will give me about 100 rpm at the spindle in the lowest gear combination.

The plan is to live most of the time with the F/R 1/2 switch set to F-1 (8 pole setting forward) and the knob removed, VFD set to "motor 1". When I need high speeds I would shut down, insert the knob and set the switch to F-2, (2 pole setting forward), power up VFD and set to "motor 2" then do the work.

It may be a little inconvenient, but I would be using "motor 2" less than 10% of the time.

The cable run between the VFD and the motor will be around 6 feet. But it seems to me from your answer that the load reactor might make my old lathe motor have a longer life on VFD power?

consider an alternative since you want long life.....

since you do not appear to NEED the additional 20% HIGHER HP rating of this motor by running it at 460v/60hz, DON'T.

Tap your xfmr to simply supply 380v and set your vfd to 380v/1500rpm/50hz as base speed with 60hz as max speed setting. You still get nameplate HP rating going up higher to 60hz. but you don't add that extra 20% more voltage stress to it.

Similar effect to putting a small (swag 1.5% output load reactor) between the vfd and motor..... a 3% one would probably drop your output voltage spikes 35% or so....

mike_kilroy said:

consider an alternative since you want long life.....

since you do not appear to NEED the additional 20% HIGHER HP rating of this motor by running it at 460v/60hz, DON'T.

Click to expand...

Interesting... The lathe has a flat belt which slips and limits torque. I assume the belt is 32 years old like the machine. I flipped it back in '09 so the less worn side contacts the pulleys (smooth curved iron pulleys). Replacing the belt would require either taking the headstock apart or having a new belt fused.

So I really can't use full nameplate horsepower. I haven't yet run into a situation where torque was an issue though.

Are you saying that with a 380V 50Hz base setting that the output side load reactor would not be needed?

BTW motor's nameplate rating is 670 rpm 0.8 HP (8-pole) / 2850 rpm 2.0 HP (2 pole) @ 50 Hz. I used an online calculator and determined that the motor produces more torque at the 8 pole setting.

I definitely want this motor to last... its a unique motor: Swiss made, metric frame, both base mount and flange mount, and has a sleeve with needle bearings on the output shaft. The Schaublin brake and clutch fits onto the front of it. It might be a real pain to find a replacement motor that wouldn't require modifying a lot of stuff to make it work. Not to mention the cost. If it was as simple as fitting a new inverter duty motor I would have done it years ago.

1) Check the VFD manual. It may require a line reactor, in which case it could be inside. Or not. The line reactor keeps current pulses to diodes and bus capacitors within design limits, if required (basically improves input power factor). A transformer MIGHT do some, even all, of the job, but cannot be counted on to do it unless you know the "percent" reactor required and can compare that to the "percent" impedance of the transformer you have.

2) Load-side reactor will take stress off the motor windings, the amount depending on the inductor value. It will ALSO reduce voltage to the motor, with more reduction at high currents (loaded condition). That isn't so nice.

My suggestion is that a load-side inductor should be relatively small in impedance.... very low "percent" impedance. All it needs to do is to knock back spikes, which are well over 60 Hz in frequency, and need relatively little inductance to cut them back. You don't want any effects at 60 Hz, because they are pretty much all going to be bad.

Forrest Addy recommended a few turns on a slice of pipe for each wire, and he had a point. Not much inductance, just enough (if it doesn't saturate) to knock off spikes, and just lossy enough not to reflect them.

Motors, especially older motors, are pretty tough. They took the time to install insulation right in those days. The worst stresses will be coil-to-coil, and they will generally have good insulation there, if it isn't cooked into carbon by now.

Wire varnish insulation is pretty worthless, even some of the highly touted stuff. UL doesn't even count it, they consider varnished wire to be bare. All it needs to withstand is a few times the nominal voltage for one turn of the coil.... maybe 15 to 25 V max. Spikes may test that, but a good vacuum varnish job should raise the voltage quite a bit. The spikes are generally only going to affect the first 5 to 10 turns of the coil. After that there is too much inductance. So a 300V spike might end up at around 50V between wires. Could be more, of course.

But older motors, with neatly wound, tied and vacuum-varnished coils, will be pretty good for the wire-to-wire insulation needed to resist spikes. Old-time switchgear produced plenty of them too, after all.

It may be expensive to build a load side filter from off the shelf parts to deliver a proper sinewave to the motor, from the VFD, but it is relatively simple.
See this post I made a few days ago:. http://www.practicalmachinist.com/v...-blown-i-think-ez-trak-dx-292309/#post2391764

---
Unless you put a voltage reducing transformer between the vfd and the motor, the voltage stress on the windings is the same, regardless of voltage or frequency, the motor will see square waves at 380v x1.41 = 535 volts)
Current flowing through the windings will smooth out to be a sine wave but the voltage stress will be 535 volts at the carrier frequency of the VFD, which is probably 4-8Khz by default.
It is this high frequency square wave that pushes current though the capacitance between the windings and the rotor, sending current through the bearings.

johansen said:

It is this high frequency square wave that pushes current though the capacitance between the windings and the rotor, sending current through the bearings.

Click to expand...

Which is one of the things that a proper inverter rated motor takes into account. Bearings may be insulated from case, rotor may have a drain contact, etc. Windings can be shielded to prevent that, my boss has a patent from his time at Emerson on a system to do that.

Inverter rated motors also are generally supplied with a rating on minimum rpm for sufficient cooling.

johansen said:

Unless you put a voltage reducing transformer between the vfd and the motor, the voltage stress on the windings is the same, regardless of voltage or frequency

Click to expand...

Careful with generic statements; Not so in this case

mike_kilroy said:

Careful with generic statements; Not so in this case

Click to expand...

i explained the math for 380vac input to the vfd.
here's a particularly bad example.
think 200 volts per division instead of 2 though.
http://johansense.com/bulk/vfdout1.JPG

mike_kilroy said:

Careful with generic statements; Not so in this case

Click to expand...

Your last post is not relevant here.

I said what I said because the OP is deciding on INPUT VOLTAGE TO HIS VFD. He is deciding whether to tap his INPUT xfmr to 380v or 480v.

As you know, if he taps to 380 then his vfd DC bus will be 530v, his spikes to motor will be 1000v.

If he taps to 480v, then his vfd DC bus will be 675v, his spikes to motor will be 1360v.

This makes your general statement wrong: of course if OP input to vfd voltage was the SAME, then setting the base speed to lower will NOT lower the spikes - he still will have the 1360v spikes.

If it were MY machine, and I knew what he has said about the load and use, I would drop the damn tap to even LOWER than 380 - put the sucker at around 320vac output, set base speed correspondingly lower to 42hz/1260rpm and call it a day with a motor that probably would NEVER have insulation failure due to vfd spikes of only 450v. It would only have (320/380)^2 or 71% of nameplate HP rating, but from his comments, that is plenty.

just to help clarify, here are a couple real pictures of motor voltage spikes with and without inductors between motor and vfd...

i had misread your post, and i should have said regardless of voltage or frequency of the VFD output.
OP had already figured out tapping the transformer for 380v.

btw, a 5% line reactor should be sufficient to connect delta caps across the motor.. but i'm a bit suspicious of those line reactors that sell for 10$ a pound. i have a feeling they aren't what they say they are, and might want to check their temperature rise @ no load if using them for an LC filter.

First, I want to say thanks to all that posted on this thread for the very informative responses.

I thought it might be helpful if you can see what I'm working with. Here are some photos:

First photo is the transformer. The secondary taps are 380, 400, 420, 440, 480 and 500V.. The label on the primaries are concealed by wires, but I believe it was labeled 220V.

Photo 2 shows the complete drive assembly. The complexity of the clutch/brake arrangement can be seen. That assembly drives the countershaft via V-belt, and the countershaft drives the spindle via flat belt (which can just barely be seen in this photo). With the foot pedal released the clutch is engaged, when you press the pedal the brake engages and the motor spins freely.

Photo 3 shows the motor. The plate on the front is a Schaublin part that mounts to the flange of the motor and the brake pad mounts to it. The special sleeve on the spindle with the needle bearings can be clearly seen. The rotating part of the clutch/brake assembly rides on this.

It should be clear from the motor photo that an easy motor swap could only be accomplished by finding a motor with the same dimensions for the base mount, the flange mount and the spindle diameter and length. Also, I'm not sure how I'd go about getting that sleeve off the spindle. I never removed it when I did my "mini overhaul" on the machine.

johansen said:

...btw, a 5% line reactor should be sufficient to connect delta caps across the motor.. but i'm a bit suspicious of those line reactors that sell for 10$ a pound. i have a feeling they aren't what they say they are, and might want to check their temperature rise @ no load if using them for an LC filter.

Click to expand...

Sorry, I'm not clear on "connecting delta caps across the motor". Can you explain.

Keep in mind that since the motor is two-speed (2 pole / 8 pole) I will need to connect the VFD output to the 2 speed switch. I can rig it so the switch cannot be operated while the VFD is powered (by taking the knob off). But without the switch I would only be able to connect to on of either the 2 pole or 8 pole windings of the motor.

So the arrangement would be like this:

(Mains) 208V 60Hz 3Ø -> (Autotransformer) -> 380V 60Hz 3Ø -> (VFD) -> (Load Reactor) -> (2 speed switch) -> 2/8 pole cables -> (Motor) 2/8 pole windings

Also, Mike Kilroy, since I can't go lower than 380V on the VFD input, wouldn't I be able to limit the voltage on the output side of the VFD and accomplish what you suggested that way?

Update 2015...

Recently I achieved a long-time goal of mine and moved my Schaublin 102N into a commercial space where I will have access any time I want. The space also has 3ØY power. I measured the power (no load) at 215V line to line.

So this VFD project is finally going to happen. The VFD is on its way (I ordered a GE AF 300 G11 model 6K-G11-4-3-002-X1B1 from Dealer's Electric for a very reasonable price) and it should arrive Tuesday.

I still feel that connecting the VFD between the transformer (which will boost 215V to 387V) and the two speed switch will work. I plan to remove the knob on the 2-speed switch so that it cannot be inadvertently switched while the VFD is powered up. To switch speeds I would power down, insert knob to switch and then power up and tell the VFD it's connected to "motor 2". I will probably be on "motor 1" (the low speed 8 pole setting) 95% of the time.

I gave some thought to replacing the wiring with inverter duty wiring and that job would be quite time consuming. There are two sets of wires (6 line wires and ground) going to the motor from the two speed switch. So I think the best bet would be to put the load reactor on the output side of the VFD ahead of the 2-speed switch and keep the original wring.

Can anyone give any advice on selecting a load reactor? Here are the motor specs:

Mfr: RCB Elektro Apparate AG Bienwil Am See, Suisse
Type: 33,07
Volt: Y 380
Watt: 1.5k / 0.6k (2.0 / 0.8 HP @ 50 Hz)
Amps: 3.63 / 2.3
RPM: 2850 / 670 @ 50 Hz, nameplate rating
RPM: 3593 / 899 @ 60 Hz, tested at no load with strobe tach

BTW, does anyone know of a source that sells inverter duty wire cut to length (rather than having to buy a 100' spool)?

Also: Is there any benefit to feeding the inverter a slightly higher voltage (transformer has taps at 400V, 420V, 440V, 460V etc) and limiting the output to 380V?

OK, guys, the GE AF300 G11 VFD arrived yesterday. After verifying the connector type I ordered the keyboard extension cable from an eBay seller. (BTW can you believe the part number for the correct cable is NOWHERE to be found in GE's user manual - I had to look it up in a Fuji document).

I'd still very much like to hear from anyone with advice on sizing the output side load reactor. Is it ok to just go by maximum motor kW (or HP)? My motor specs are above. I checked the transformer that will be boosting the input power from 215V to 380V+ and it is rated at 2.5 kVA.

cinematechnic said:

OK, guys, the GE AF300 G11 VFD arrived yesterday. After verifying the connector type I ordered the keyboard extension cable from an eBay seller. (BTW can you believe the part number for the correct cable is NOWHERE to be found in GE's user manual - I had to look it up in a Fuji document).

I'd still very much like to hear from anyone with advice on sizing the output side load reactor. Is it ok to just go by maximum motor kW (or HP)? My motor specs are above. I checked the transformer that will be boosting the input power from 215V to 380V+ and it is rated at 2.5 kVA.

Click to expand...

The only difference between a line reactor and a load reactor is where it is located, and the current rating. The LINE side current is going to be lower than the load side current, because the VFD corrects the power factor of the motor, but on the load side, the reactor must pass both the active and reactive current through it. So bottom line, you size a LOAD reactor based on the motor FLA. Your motor appears to be a 2 speed, but you are likely going to use the higher speed and the VFD instead, so size the load reactor for 3.63A minimum. Small reactors like that are probably going to jump from 2A to 4A, so get the 4A.

As Jraef said but also know there are a lot more losses on the output side - on the input side you have nice sinewave current AND voltage; on the output side you nice sinewave current but SQUAREWAVE voltage - the inductor will not like that. A 4amp load thru a 4amp inductor on the input side will get nice and warm, a 4amp load thru a 4amp inductor in the output side will probably overheat and melt down.

so either pick one designed for the high PWM freq losses on the output or oversize it. If you must use a cheap regular input inductor, pick one rated 3% (not 5% or higher), for say 8amps instead (this will make it about a 2% inductor which is probably better for you anyway).

example: TCI, LLC - KDR Line Reactor

I would pick a KDRA2P

I have no relation to this company. in fact, I sell custom reactors for input and output on vfds all the time, but for 1pc it is best you just go to a commercial source like this or equiv.

mike_kilroy said:

example: TCI, LLC - KDR Line Reactor

I would pick a KDRA2P

Click to expand...

Mike,

Thanks for the help. That company has so many line reactor options it's confusing. Took me a while to find that part number. I may get the same unit but in the DIN rail version since my lathe has a DIN rail.

To be clear, based on your reccomendation, the spec I need is:
480V
"Open" (what does that mean?)
Output
3 HP
4.8 NEC Amps

TCI forgot to address one important issue on their website: How do you buy their products?

Also, I noticed that the part number you give is a UL "Listed" as opposed to "Recognized", is that really important for someone that doesn't have to meet any regulatory issues?

Thanks!