AC Drives & Soft Starts biting the dust

This isn't PLC related, but where else can you find the great advice and assistance this place provides?

Can anyone shed some light as to why a customers' soft starts and drives they (and I ) have installed on their cooling tower fans keep failing? Tower was originally installed in 1997 with two 20 hp 460vac motors and two 20 hp Baldor soft starts. The soft starts had adjustable trip circuit breakers. Shortly after startup, I noticed that the middle phase of one motor felt warm, as did the conduit it
was contained in. Electrical measurements revealed no problems with voltage supply, and all three phases pulled very close to the same current. (less than 2% deviation).

Since that time, a soft start has been replaced three times at about two year intervals. I replaced one soft start with a drive in 2004, and the other in 2005. Now, just today, the first drive I installed blew a capacitor. The motors are located about 250' from the drives. The drives are located in an enclosure in a mechanical room that gets pretty hot in the summer. I installed a fan in the enclosure, along with a compressed air vortex cooler to help keep the temperature down. The drives are fed out of a motor control center that feeds a lot of other equipment that does not seem to have problems like this. A ground wire was NOT pulled with the phase conductors out to the motors, but the drives are grounded to the motor control center ground bar as is the drive enclosure.

Anyone have an idea as to what the heck is going on? I suspected at first that the soft starts were junk, but 3 of those and 1 Yaskawa GPD 506 drive later, I'm not sure what the problem could be.

I take it this is a single motor and drive, not one of a group that is bad while the others are fine. #1, look at what has not changed over the course of the timeline. You didn't mention ever changing the motor, or leads. My guess is that the motor is fine, and the problem lies in the length of the run to the motor. 250feet is a fair distance. Are the motor leads a special motor cable, or just 3 (X) guage wires. Finally, what exactly is going wrong. Are they failing in the same manner, and what is the nature of the failure? All these need to be answered.

Russ

Russ- you are correct. motor and leads have never been changed. The leads are just three #8 AWG in conduit. The first two failures were the electronic soft start. I don't know what failed, the soft starts were simply replaced by plant maintenance. The third failure was the adjustable trip circuit breaker refusing to reset itself after a power glitch, we think. After this is when I installed the Yaskawa 20 HP drive. Ran ok for two years, then the capacitor failure. Looks like a trend. I know that long motor lead lengths can cause voltage spikes at motor insulation, can it also cause drive failure?

Randy- the soft start controllers were never used in tandem with a drive. As they failed (about every two years) they were replaced with like until the third soft start failure in 2004. I replaced it with a drive only, (no more soft starts) and it also lasted about two years. Measured FLA current on the motors is 24-25 amps. I have set the accel and decel times at 55 and 45 seconds, respectively.
We do have some serious power glitches here, but I can't correlate
those to drive failures. I agree, we need DickDV.

My first suggestion is to stop blaming the electronics and start blaming the power supply.

What do you mean by "Electrical measurements revealed no problems with voltage supply"? What was found? What measurements were made? How often were they made? How long of a time period were they made for? Who made them? When were they last made?

Soft-starts are not affected by lead length like drives are so it is probably not the motors or their wiring. Capacitors in drives fail because of over voltages, heat, or cheap construction.

My first target of investigation would be looking for capacitor switching transients (usually caused but not acknowledged by the utility) on the line side of this equipment.

Jim - The measurements I was referring to were voltage measurements made after the aforementioned warm phase "B" conductor that the soft start produced. At the time I assumed that there had to be some sort of a voltage drop or supply problem. Phase to phase and phase to ground readings were taken before power up, after power up, and as drive was running. No voltage anomalies were noted, but the only equipment that was used was a Fluke 89 IV DMM. No power quality analyzer or scope was used for in-depth observing, but I did log voltage all night
once with the 89 IV, and did not notice any voltage sag or spikes. What type of equipment would work for detecting these capacitor switching transients?

A snapshot of voltage is not sufficient to diagnose power quality problems. If I were involved I would have a power monitor on-line 24hrs a day (at a point in the system on the line side of both starters) for at least 1 week recording voltages and currents every 15 minutes. If possible I would also look at an "incident recorder" that is triggered by a sag or swell of 15%.

I noticed that the middle phase of one motor felt warm, as did the conduit it was contained in.

Click to expand...

Ding Ding Ding..

1) Measure the current & voltage under load at both ends of the 250ft run.

2) Put a scope on the line and watch all three phases under load.

Been gone, folks. Sorry.

At 460V 20hp, 250 foot motor leads are a serious problem but, to the motor, not to the drive. This system must have a very good motor to have weathered 2 years of being bombarded by huge overvoltage spikes. I would think that a dv/dt filter (made by the reactor people--MTE, TCI, and Hammond) would be the minimum necessary for long motor life. But, that is not the problem at hand.

Failed drive caps on an application that does not have braking is almost certainly an input power quality problem. Spikes, etc. are almost impossible to trap ( the one that got the cap took two years to happen!) but, if you are going to try to record anything, I would record the DC bus level. This will give you an indication, not only of the height of the pulses but also of their energy content. If the AC supply is truly a problem, you will see the DC bus bouncing around from the spike energy. I would expect an occasional High DC Bus Fault to occur on the drive, as well. The solution to a noisy AC supply is usually a 5% line reactor in the power leads. I don't know if Yaskawa already has reactors in their standard furnish but, I would add some being careful not to overdo it or you will start to have low voltage problems under heavy motor loads.

You may also want to check on whether the power supply is one of those nasty floating delta systems. If it is, rather than use input line reactors, use a drive isolation transformer with a wye secondary and be sure to ground the center of the wye. Floating systems with one lead grounded can cause all kinds of ugly drive problems as well as serious safety hazards to maintenance personnel.

Finally, your system should have a continuous copper ground from the motor frame to the drive frame. Yours doesn't and you don't seem to be seeing nuisance faults but, consider yourself lucky on that score. The lack of this bonding jumper doesn't impact your drive problem or the life of the motor but it usually reduces nuisance tripping. It's probably too late to add it at this point so run without the ground bond. In future installations, insist on it.

Sorry I have been gone a bit too. First you need to set the carrier frequency on the Yaskawa drive to 2k due to your cable length. Secondly, your failures are probably due to the motor free wheeling BACKWARDS and the motor being started.

I have seen a few failures like this. Additionally a 3% line reactor both at the input to the drive and at the drive output will help to filter out some of the inherent line voltage spikes and some spurious noise that can get the drive. Your cap failure was 90% most likely caused by HEAT or EXCESSIVE bus voltage.

As DickDV said, get a ground wire to the motor from the cabinet the drive is located in. This will eliminate 99.99% of common mode voltages.

Interesting topic...

New to group. Have found the topic interesting. I work in plant with 8 colling towers. Free wheeling fans were pretty much eliminated on newer towers through the drive train. I agree with a power monitor on the load for a 24 hour perioid. I may suggest on a warm day with a steady load onthe tower. A surge tester or motor analyzer will detect winding and cabling issues if done from the starter. (disconnect CT's and VFD). Just thoughts

Headroom

Just a thought from the maintenance side.

I notice that the motors are 20hp and the soft starts were also rated for 20hp. I have seen several instances where closely matched ratings between drives and motors result in frequent drive failures. 20% or greater headroom on the max power ratings on the drive eliminates many of the repetitive failures.

Of course the drives are setup for the existing motor (kw, I, ...)

regards,

Thomas

Thanks for the replies, guys. I replaced the drive that blew the

capacitor on Thursday, 4/20/2006. I used a Yaskawa P7 drive. Using their startup procedure, I stepped the drive from 0 to 60 hz in 5hz increments, logging output current, output voltage, and DC BUS voltage. Voltage and current seemed stable, but the DC bus voltage looked squirrely. By that I mean at 10 hz, it read 704-707vdc, but as I ramped the drive in 5 hz increments, the drives display of bus voltage became more erratic. For example, at 15 hz the value was wider but still pretty stable (699-705vdc). At 30 hz the range was 694-705vdc, a slightly larger range, but still the drives' display was still pretty stable. However, as I went to 35 hz, the DC bus voltage on the display became erratic. And when you get to 60hz, the bus voltage range is 645-710, with a lot of bouncing around. Does this indicate a bad power source?
We have a grounded 480v wye system.

Your bus voltage will NEVER be smooth on a VFD unless you are NOT running a motor. The levels of ripple you have seen are with in normal running conditions. There is actually a monitor in the drive for excessive ripple on the bus. It will trip you off. When you are running the motor, remember you are turning transistor off and on into an inductive load. The bus voltage will droop and spike. As you are doing this 2 thousand times per cycle and you are doing 5 to 60 cycles per second, that is a lot of on/off per second. Some digital scopes cannot capture the waveforms generated. You also have pulsating current flowing. If you were to put a current probe on the AC input and look at it with a scope it is 2 pulses much like you see with rectifing AC to DC.

You state you have a 480 system. What is the actual voltage at the drive input terminals? The drive is rated 460 volts +10/-5%. That is 506 VAC max on the input. The DC bus voltage is 1.4 times the input voltage or roughly 707 Volts. If the line is high, the bus will be closer to the over voltage trip level. Constantly on the high side decreases capacitor life. If you cannot get the line voltage down to at least 490 or lower, definately get a line reactor on the input. They are rated in % and that is the amount they reduce the line voltage at typical FLA levels. If you are at 505 and cannot get it lower, I recommend you use at least a 5% reactor. The cost is minimul when you consider the cost of a drive.

Have you grounded the motor frame yet?

Have you measured the line votages to ground? They should be about 60% of the line to line voltage.

The motor is not grounded, the drive will tell you if the motor has a grounded lead. (Section 6 of the P7 users manual)