Drive/Motor Troubleshooting questions?

Status
Not open for further replies.

Cow

Senior Member
Location
Eastern Oregon
Occupation
Electrician
Had a tough one(for me) today. I have an ABB ACS800 150HP drive running a wellmotor that keeps tripping on a "short circuit" fault, anything from just a few mins. to a few hours apart. Drive manual says the cause can be a cable/motor problem or a faulty output converter in the drive. Simple things first, so I start with busting the T leads loose in the drive, megging the cables off the drive right on through to the motor and figure I'll find the problem lickity split. I should also mention this drive has a reactor between it and the motor, but nothing on the line side of the drive.

It megged fine, 700 megaohms at 1000v(phases-ground) and still climbing when I decide to quit testing. Look in the reactor, open the p-head, everything looks clean and tight, conduit to the motor is in good shape as well. Decide to do a resistance test, with the alligator clips clipped to each other I get .2 ohms. I get the same thing when I clip them to the T leads??? So the resistance is 0 ohms for the motor? Is that right?

I decide to call the local motor rewind shop to cover my bases, see if there is anything else I can check without sending it in for a rewind and he tells me the motor either works or it doesn't. Simple as that.

I then call ABB to see if there is anything that I can check in the drive without sending it in to have the "output converter" looked at. Nope.

I decide to put everything back together(with the reactor bypassed) and then fire the drive/motor back up just to see if it'll run. It does just fine for 4 mins. and then drops out on a "short circuit" fault.

How can I test in the future to know if it's the drive or motor in this instance? Usually it'll meg bad or have something obvious that jumps out at you. But not this one...

The customer wants this well running so they're taking a chance and having a drive expedited out here, I believe. It just kills me I can't narrow this down any better, I hate the guess and replace method, and I definitely don't like the customer spending money this way.

Thanks for the help!
 

mike_kilroy

Senior Member
Location
United States
couple other ideas....

history leading up to this helps: did this just start happening or did it start after a thunderstorm went thru? power outage while running? or has it just gotten worse and worse with time? just had a case where customer had a storm, power outage to plant, then drives on a machine would randomly OC fault.... found in history they had this fault bt not often for the 12 years they had machine. now it is like your frequency. long story short found their grounding at drive was not optimal and they had noise issues; normal noise 'floor' was say 90% of required for OC fault in IPM, so only very random false 'faults.' storm prob took out a couple MOVs in 1 or more drive input sections so noise rose to 99 or 101% so they got fake faults every few minutes. grounding xfmr Xo took noise level back down below threshold and they are running better than ever.

if the OC fault is there 100% of the time, we disconnect the motor leads and enable the drive to see if the OC fault remains, proving it is in the drive. IF I were in your shoes I would do this here and let it 'run' the time it takes to fault to check....

You might ohm from dc bus + to each motor output (both directions on meter leads), and same from dc bus - to each output, to compare all readings; if an output igbt is semi bad it will read < say around 1 mohm; compare all readings and if 1 in particular is low, it may be a failing igbt causing random issue....

Can you have a cable issue? did a mouse eat thru the cable? does cable flex? did it rub against a moving pulley somewhere? get driven over? can you eye ball it the whole length? we just had a case where megger showed fine but wiggling the cable found a spot rubbed thru inside the machine housing against a pulley; they lucked out that each time they meggered it read no resistance, but the drive still randomly faulted on OC; connection shorted only sometimes....
 
Last edited:

kwired

Electron manager
Location
NE Nebraska
Make sure you don't have igbt reflected wave damaged motor or circuit components.

The reactor is probably there at least partially for this reason, but I suppose could be compromised itself eventually.

I kind of understand what goes on here, but not well enough to put in my own words.

Here is a link that probably describes better than I can
http://motionsystemdesign.com/motors-drives/reflected-waves-damage-ac-motors-1197/

You did not mention the voltage but 480 volts has more problems than 208/240 because the spikes are much higher voltage when it is 480 volt base voltage You will not detect this damage with 1000 volt test voltage.
 

mike_kilroy

Senior Member
Location
United States
nice article. my experience with voltage spikes that exceed CIV is the corona ionizes the air trapped in pockets in the motor insulation, which then erodes the insulation, depending on amount of air and how far exceeding the CIV, in minutes to months. But it is a progressive thing destroying the insulation until it fails and shorts. I have never heard of it causing an 'almost bad' situation so I think even the 1000v megger test is 100% valid still....
 

kwired

Electron manager
Location
NE Nebraska
nice article. my experience with voltage spikes that exceed CIV is the corona ionizes the air trapped in pockets in the motor insulation, which then erodes the insulation, depending on amount of air and how far exceeding the CIV, in minutes to months. But it is a progressive thing destroying the insulation until it fails and shorts. I have never heard of it causing an 'almost bad' situation so I think even the 1000v megger test is 100% valid still....

My experience is 1000 volt meg test results in passing motor insulation - to ground anyway. Now maybe in all those cases insulation failure was from winding to winding or even more likely turn to turn, not going to find that with a megger. Drive is sensitive enough to detect this and turns off, even though it is not something obvious to see. Connect motor directly to line and it becomes real obvious there is winding damage.
 

Besoeker

Senior Member
Location
UK
nice article. my experience with voltage spikes that exceed CIV is the corona ionizes the air trapped in pockets in the motor insulation, which then erodes the insulation, depending on amount of air and how far exceeding the CIV, in minutes to months. But it is a progressive thing destroying the insulation until it fails and shorts. I have never heard of it causing an 'almost bad' situation so I think even the 1000v megger test is 100% valid still....
My experience also.
I'm inclined to think that it's a fault in the drive electronic controls.
Failure of power circuit components is usually catastrophic, evident physically, an non-reversible. If it breaks, it stays broken.
 

mike_kilroy

Senior Member
Location
United States
My experience is 1000 volt meg test results in passing motor insulation - to ground anyway. Now maybe in all those cases insulation failure was from winding to winding or even more likely turn to turn, not going to find that with a megger. Drive is sensitive enough to detect this and turns off, even though it is not something obvious to see. Connect motor directly to line and it becomes real obvious there is winding damage.

agreed too. Megger of 1000v or even higher of no value to find turn - turn shorts: sometimes you can hear this on a drive as a squeel as current is pumped to the shorted phase, often you can see it by measuring inductance L-L and comparing to the other 2 readings.

I think a history leading up to the faults would help too. I also would not rule out a cable with kink/short/damage until it was looked at the whole length. Since the motor connections were removed from the drive, and the inductor removed, it is not a loose connection there; it still might also be a loose connection on a motor lead at the motor as that was not said to be checked yet.
 
Last edited:

mike_kilroy

Senior Member
Location
United States
I meant to add also that history of this failure could help - as example, is this a 15 year old drive? It may use di/dt detection for OC fault on input to dc bus; if so, this could be an indication of failing capacitors on Dc bus also.
 

Jraef

Moderator, OTD
Staff member
Location
San Francisco Bay Area, CA, USA
Occupation
Electrical Engineer
I meant to add also that history of this failure could help - as example, is this a 15 year old drive? It may use di/dt detection for OC fault on input to dc bus; if so, this could be an indication of failing capacitors on Dc bus also.
I don't think an ACS-800 can be 15 years old yet... I think it was released in 2004? Maybe 2003 at the earliest, at least here in the US.

But I too agree; the best initial tool in troubleshooting is a good time line.
 

Cow

Senior Member
Location
Eastern Oregon
Occupation
Electrician
At this point I haven't been back to do anymore investigating, and they haven't called me to tell me the new drive has shown up either.

We were thinking the drive was around an 04 model, so Jraef is pretty close. I'm only showing the last 10 -15 faults in the drive. Those faults were all logged in the last couple hundred hours or so, with the oldest 4-5 being overcurrent faults and then last 10 or so being short circuit faults ranging from 4 mins apart to several hours apart. This thing runs around the clock and the area is known to have poor power quality, spikes, lightning, you name it. These wells I'm sure, catch all of it, running 24/7.

The motor connections and all the conduit and flex to the motor looks good. This is wellhouse with a motor in it, so there is no other driven machinery to cause vibration, etc.

I didn't think to try running the drive with no T leads on it though, that's easy enough to try.

Thanks for all the tips so far, I appreciate the help!
 
Last edited:

Cow

Senior Member
Location
Eastern Oregon
Occupation
Electrician
Installed a Yaskawa IQPump controller on Wednesday with only one issue so far, the transducer quit yesterday. But so far, so good, no "short circuit fault" on the drive thankfully.

This is the first time I've ever programmed a Yaskawa, with the pump quick setup guide this programmed a little easier for me than the Allen Bradley's I've done in the past. Customer likes the nice PSI reading on the screen too. Super easy for the customer to adjust as well. We'll have to see how it holds up, I know a pump company around here sure likes 'em.
 

mike_kilroy

Senior Member
Location
United States
Installed a Yaskawa IQPump controller on Wednesday with only one issue so far, the transducer quit yesterday. But so far, so good, no "short circuit fault" on the drive thankfully.

hmmmmm..... if by 'transducer' you mean the position feedback device (normally encoder), and it 'failed' yesterday, that probably was your issue all along! An intermittent encoder!

Depending on the drive, some do not detect a faulty encoder; those drives would instead exhibit "output short" or "inverter output stage fault.."

If the old ABB DID have encoder loss detection, it is a toss up whether it would throw out a "loss of feedback" or "output short."

Then there are ways to get bad data yet not cause encoder loss alarm, so even then, it would just output "output short."

So there is a decent chance, with this information, that the old ABB drive is still good! Don't throw it away; keep it for a spare (after either testing it yourself or having someone else test and verify it is still good).
 

Jraef

Moderator, OTD
Staff member
Location
San Francisco Bay Area, CA, USA
Occupation
Electrical Engineer
hmmmmm..... if by 'transducer' you mean the position feedback device (normally encoder), and it 'failed' yesterday, that probably was your issue all along! An intermittent encoder!

Cow said:
Installed a Yaskawa IQPump controller...

I'd hazard a guess he meant the pressure or flow transducer providing feedback to his PID loop.
 

kwired

Electron manager
Location
NE Nebraska
hmmmmm..... if by 'transducer' you mean the position feedback device (normally encoder), and it 'failed' yesterday, that probably was your issue all along! An intermittent encoder!

Depending on the drive, some do not detect a faulty encoder; those drives would instead exhibit "output short" or "inverter output stage fault.."

If the old ABB DID have encoder loss detection, it is a toss up whether it would throw out a "loss of feedback" or "output short."

Then there are ways to get bad data yet not cause encoder loss alarm, so even then, it would just output "output short."

So there is a decent chance, with this information, that the old ABB drive is still good! Don't throw it away; keep it for a spare (after either testing it yourself or having someone else test and verify it is still good).

Why would that create a "short circuit fault" which would be a fault code related to the output and not to the input devices? What is the point of having a long list of fault codes if they each did not have a fairly unique meaning?
 

mike_kilroy

Senior Member
Location
United States
it would create an output short fault if the encoder all of a sudden had a discontinuity and said the rotor was in a significantly different location. That would cause the high performance internal current loop to demand instant max current output current, maybe even the opposite polarity, causing the output rate of rise di/dt short circuit sensors to see it as an instantaneous output short since the current rose in perhaps 1 update cycel time - 250usec is my guess for that vintage ABB drive - which looks exactly like an output short. I've seen this exact scenario many many times......
 

kwired

Electron manager
Location
NE Nebraska
it would create an output short fault if the encoder all of a sudden had a discontinuity and said the rotor was in a significantly different location. That would cause the high performance internal current loop to demand instant max current output current, maybe even the opposite polarity, causing the output rate of rise di/dt short circuit sensors to see it as an instantaneous output short since the current rose in perhaps 1 update cycel time - 250usec is my guess for that vintage ABB drive - which looks exactly like an output short. I've seen this exact scenario many many times......

No disagreement here:cool:
 
Status
Not open for further replies.
Top