AC Elevator Drive Audible Noise Help Needed. 6

[scmguru]

We've got a Geared Elevator Hoist that is being run by an Imperial AC VVVF Motor. The Motor is rated @ 1164RPM and the Gear Ratio of the Hoist is 87-2.

About a month after this piece of equipment was installed, it began making a hollow droning sound that is at least 18db louder than the AC Hoist itself.

I recorded the elevator hoist running and captured some WAV file snapshots of the change in waveform.

There is a very obvious and distinct change in the soundwave when this droning begins. I have posted a visual snapshot of this waveform here..

http://onfinite.com/libraries/850891/6ce.jpg

Over the last few months it has become worse and worse and we cannot seem to figure out what it causing it.

I have figured out that the sound is sinewave shaped and it peaks at about 330-338 hz.

I have a recording of an elevator run and this sound

T+0- Elevator Start - Normal Sound
T+9.5s Droning Begins
T+10.5s Droning Peaks
T+12s Droning diminishes
T+15s Droning comes back
T+16 Droning Diminishes
T+17s Droning comes back and doesnt come back
T+24s End of Run.

I would be happy to provide a screen video of this waveform (about 30 seconds) if anyone could give me any help with this.

This is a really serious problem for our building and its residents and our current vendor refuses to acknowledge any kind of problem.

 

[scmguru]

Mark and everyone else, Thank you SO much for all of your insight. This is great info to pass to our consultant.

As I think back to when the hoist was originally installed and look bag at my logs... there were some "problems" with the controller/drive and the tech had to re-program a number of parameters.. I wonder if that is why the problem did not originaly show up when the hoist was installed.

I made a great audio recording as per one of the suggestions here with the mic directly on the gear area and I applied a bandpass filter to cut out the AC motor noise. What I have now is a perfect recording of what is actually going on, the drone is much clearer and obvious.

Also, nice observation re: accelleration of the elevator. There is no drone for the first 2-3 seconds of the run (the accelleration period). The second the car stops accellerating and is in the high speed mode, the drone starts.

If someone would be interested in listening to this mp3 file I made and comment on it, please shoot me a message at audis4***fastmail.fm (remove the *** and replace w/ @)

I'll keep you all in the loop re: the progress made on this..

 

[kontiki99]

Have you considered analyzing the sounds to see what the fundamental frequency(s) are.

That might tip you off to where it's coming from.

If you have recorded sound and it's over sampled enough, you might be able to run an FFT (Fast Fourier Transform) on the wav to produce a spectral map.

There used to be a DSP toolkit built into Matlab that would let you input a .wav and set up DSP analysis using a graphic interface.

The output looks very much like that of a spectrum analyzer.

The frequency with the most power, could be a direct or a multiple (an harmonic) match for the rotation speed of one of your components.

A microphone is a vibration sensor.

 

[scmguru]

Kontiki,

I did a spectrum analysis of the sound and it looks like the loudest peak is between 330-340Hz with another large peak at around 220Hz or so. This is with the bandpass filter applied.

I've posted the screenshot here.

http://onfinite.com/libraries/852424/73c.jpg

The motor is rated at 1164RPM at 208V/60HZ. The gear ratio on the hoist is 87-2.

Does anyone know how I could use this information to isolate where in the mechanical chain this could be occurring?

 

[scmguru]

Also w/o the bandpass filter applied.. there is a major peak at around 22Hz as well.

http://onfinite.com/libraries/852426/b42.jpg

 

[kontiki99]

scmguru,

I'm just winging it here.

Can you correlate any of those frequencies to the number of teeth on one of the gears x gear rotation speed?

I see your peaks at 330, 220 and 160, it seems like there's 3 similar peaks out there at 760, 600 and 630.

 

[itsmoked]

scmguru; I would love to hear the recording!

Upload it here:

http://www.filefactory.com/

Note the link they issue you.
Stick it in here.

Then we can all hear it by hitting the link waiting the annoying few seconds hitting the none-premium 'download' button.

Keith Cress
Flamin Systems, Inc.- <

http://www.flaminsystems.com>

 

[scmguru]

Keith.. Great idea.. Here it is.

file a- no filter recording

http://www.filefactory.com/25650e

file b- bandpass filter (center freq @ about 300hz cutoff 2 octave up and down)

http://www.filefactory.com/7477c8

Screen Capture of Wav File w/ Sound. Note: you'll need the following Camtasia Screen Capture codec to view this..

http://download.techsmith.com/tscc/tscc.exe

Download and install, and then open avi file using windows or other media player

http://www.filefactory.com/068e66

Thanks for the help folks.

 

[scmguru]

Sorry about the bad links folks.. file factory copied the wrong URL's..

Here are the new good links..

The site will make you wait a few seconds and then you click "Don't want a premium download? Click here to access our free download links"

You'll again have to wait a few seconds after clicking on the link and it will then give you the download option(s).

file a- no filter recording

http://www.filefactory.com/get/v3/f2.php?f=25650e

file b- bandpass filter (center freq @ about 300hz cutoff 2 octave up and down)

http://www.filefactory.com/get/v3/f2.php?f=fa7e71

Screen Capture of Wav File w/ Sound. Note: you'll need the following Camtasia Screen Capture codec to view this..

http://download.techsmith.com/tscc/tscc.exe

Download and install, and then open avi file using windows or other media player

http://www.filefactory.com/get/v3/f2.php?f=068e66

Thanks for the help folks.

 

[itsmoked]

ATTENTION!

The ele.wav file is miss-named!!! Change it's type to [red].mp3[/red] if you get an "Invalid File Type" warning.

Nice moaning.. "I want my office moved."

Sounds like a big standing wave..


Thanks scmguru.

Keith Cress
Flamin Systems, Inc.- <

http://www.flaminsystems.com>

 

[scmguru]

Keith.. Thanks for the review. so what do you think? Should I still chase down the drive pulsation? The second file is the best one, with the AC motor noise filtered out, it's obvious the droning happens the whole run except for accell/decell.

 

[itsmoked]

I think I still agree with the gang. Short of the gears causing this the biggest provider of pulsations that can drive an audio vibration would be the motor. I think the Drive's settings could very likely cause this. Screwing with the drives settings seems to be the easiest thing to do too. I would write down alllllll the existing settings and then change any the sharp drive people in here suggest/ed.

Hopefully someone will hassle the sound file download, and then be able to put their finger right on the problem.

It might be helpful if you could note all the drive's present settings in here. I bet someone will see a problem.

Keith Cress
Flamin Systems, Inc.- <

http://www.flaminsystems.com>

 

[scmguru]

Keith, good idea.
Any drive wizards out there care to have me list parameters? afr/asr settings etc? let me know what you need and Ill have my consultant write them down.

 

[jraef]

It's been a while since I worked directly on Yaskawa drives, but they haven't changed that much when I look at the manual for this drive. At a minimum, get this info:

Control Method (OL Vector, Flux Vector etc.) Fn A1-02
We just want to be sure of what the last person set it up for. As I said earlier, if there is an encoder then it should be set to "03" for Flux Vector control, in which case there are a lot of other parameters to look at, too numerous to go into, but suffice to say you will need someone with specific Yaskawa expertise to help you further. I would call Yaskawa directly for help or a recomendation if I were you.

If A1-02 is set to "02" for Open Loop Vector control, then go to the following parameters (cut and pasted from the manual):

C8-08: AFR Gain; Factory setting: 1.00
Range: 0.00 to 10.00
C8-09: AFR Time Factory setting: 50
Range: 0 to 1000 msec
Useable only in Open Loop Vector Control Method ( A 1 - 0 2 = 2), these parameters affect the speed response or prevent the motor from hunting. To improve the speed response,
increase C8-08 and/or decrease C8-09. To stop the motor from hunting, decrease C8-08 and/or increase C8-09.

For the ASR, check out all of the C5 group, C5-01 through C5-08. There is too much to that to decribe everything here, but recording it is the first step. Since this doesn't appear to be happening with load step changes on the fly, it may not be associated with the ASR, which is more akin to droop control than the AFR is.


Eng-Tips: Help for your job, not for your homework Read faq731-376

 

[DickDV]

As far as I am concerned, look for a parameter named "carrier frequency", "pwm frequency", or sometimes "low noise or high noise".

Simply changing this value from, for example, 1khz to 8khz or from 8khz to 1khz would be an easy reversable thing to do and will eliminate any further wondering about motor magnetic noise.

Once this simple test is made, the more complex stuff discussed at length above can be examined if necessary.

Why not do the simple stuff first?

 

[scmguru]

From notes I found on site..

C5-01- ASR Proportional Gain 1 --- 20
C5-02- ASR Integral Time 1 --- .200 sec (Factory setting is .500)
C5-03- ASR Proportional Gain 2 --- 20
C5-04- ASR Integral Time 2 --- .500 sec
C5-06- ASR Primary Delay Time --- .004sec
C5-07- ASR Switching Frequency --- 0
C5-08- ASR Integral Limit --- 400

I am going to work on scheduling a Yaskawa tech come out and make sure this is tuned properly and will followup with any progress we make.

 

[scmguru]

Also saw this in the drive technical guide..

A. d 3 - 0 1 : Prohibited Frequency 1 Factory setting (each): 0 . 0
d 3 - 0 2 : Prohibited Frequency 2 Range (each): 0.0 to 400.0 Hz
d 3 - 0 3 : Prohibited Frequency 3

These parameters allow programming of up to three prohibited frequency points for eliminating problems with resonant vibration of the motor/machine. This feature does not actually eliminate the selected frequency values, but will accelerate and decelerate the motor through the prohibited bandwidth.

B. d 3 - 0 4 : Prohibited Frequency Deadband Factory setting: 1 . 0
Range: 0.0 to 20.0 Hz
This parameter determines the width of the deadband around each selected prohibited frequency point. The factory setting is " 1.0 ", which establishes a deadband of ±1.0 Hz.
EXAMPLE:
Vibration encountered between 30.0 and 36.0 Hz.
SOLUTION: Set d3-01 = 33.0. This is the center of the problem frequency band.

Set d3-04 = 3.0. This will cause the drive to reject all
frequency command values between 30.0 and 36.0 Hz.
A frequency command in the deadband will be converted to the bottom value of the deadband, e.g. a command of 33 Hz would result in a run frequency of 30 Hz.

---
I'd be willing to bet that an experienced drive tech could use dial out the problems we're having..

 

[Marke]

Hello scmguru

There are two possible issues here that involve the drive. One is the possibility of a system resonance which is frequency dependant. This can be a real problem in some installations and is a system issue where the system is a combination of the drive, motor and driven load.
A small change in frequency will highlight whether this is the problem If it is, then it is a matter of avoiding the resonant frequencies.

The other possibility relates to the internal feedback systems within the drive. If the feedback parameters are incorrectly set, the drive can go unstable and hunt. Many drives can be easly set to V/Hz which eliminates the closed loop systems. If the drive can be easily set to V/Hz mode and this eliminates the problem, then there is probably an issue with the drive setup.

There can also be an interaction between the carrier frequency and the motor characteristics and this can be easily checked by altering the carrier frequency, but in this case, due to the noise occuring only during run, I would not expect this to be the major issue. - do not discount it though.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[ozmosis]

Just a word of warning on changing the carrier frequency. I'm not overly familiar with the Yaskawa G5 product but some drives will automatically derate output current if you increase the carrier frequency. I agree with DickDV, try the simple things first but just be aware that if the current rating of the vsd derated that this could create other problems.
I'm also not too sure setting a skip/jump frequency on an elevator application is a good thing either. Even though you are using a variable frequency drive it is only using a relatively small number of speeds and is therefore not ideal to 'vary' the speed as a skip frequency would invoke. Sure, use the VFD to find the problem but knowing Yaskawa's experience with elevator applications, I would be surprised if it is a setting up issue with the drive.

 

[jraef]

Not to beat a dead horse, but the fact that you have an encoder for feedback, yet the ASR is programmed, means to me that the initial design was meant to use the encoder and someone came along after the fact and switched it to Open Loop Vector. The Yaskawa drive has no interaction capability between an encoder and the OLV option. Encoders are expensive. The original design engineer would not have put one in without a reason.

The encoder can however be used in either FVC or V/Hz mode. V/Hz may make sense in that regard because of what Marke said. The point is, we can second guess this forever but in my mind, you need to pressure the system supplier to answer these questions. It would not be the first time a technician came up with a solution in the field that had unforeseen consequences later, then his boss backs that decision to avoid a callback by telling you "It's normal, live with it."

Eng-Tips: Help for your job, not for your homework Read faq731-376

 

[scmguru]

I had our building manager ask our elevator techs (who were working on another issue) if the drive/controller was running in open loop vector. The modernization supervisor said "he didn't know", their "troubleshooter" told him, it is running in "AC Mode"...

Does "AC Mode" correspond to any drive modes of this drive?