DIY Chatroom Home Improvement Forum

DIY Chatroom Home Improvement Forum (http://www.diychatroom.com/)
-   Electrical (http://www.diychatroom.com/f18/)
-   -   GFCI problem with pond pump (http://www.diychatroom.com/f18/gfci-problem-pond-pump-24530/)

sandsmarc 07-31-2008 12:09 PM

GFCI problem with pond pump
 
I have a 1HP 3-phase pond pump connected to a dedicated 120 volt, 20 amp circuit. The thing about this pump: It is connected to a computerized controller that allows me to variably adjust the speed of the pump. The controller converts single phase to 3-phase and then powers the pump.

The problem: The controller trips the GFCI breaker in the subpanel as soon as the pump starts. As long as the pump is at zero RPM, the controller light up and the GFCI breaker is fine. The moment I dial-up the speed and the pump starts, instant GFCI trip.

Is there a solution to this? Do computerized motor controllers behave badly with GFCI protection?

Thanks!

-Marc

micromind 07-31-2008 10:51 PM

The computerized motor controller is more commonly called a variable frequency drive, or simply VFD.

My experience with 120 input VFDs is very similar to yours. Even though absolutely no actual ground fault exists, apparently the electronics in the VFD can fool the electronics in a GFI into thinking there's a fault, and it'll trip.

Without getting too technical here, it has to do with the fact that the input current to the VFD distorts the sine wave of the power supply. Looking at it on a dual-trace oscilloscope, the waveforms of the hot and neutral don't always match. The further the VFD is from the GFI, the worse this problem is.

Possible solutions;

1) Replace the GFI breaker with a non-GFI one, and install a GFI receptacle close to the VFD.

2) Install an isolation transformer, or a line reactor at the VFD. Either of these will cut back on reflected power, and the waveforms at the GFI will be much more closely matched.

3) I don't have a codebook handy, but you might not need a GFI at all. (Several of the other guys around here will chime in on this one, I simply don't know without looking it up.)

4) The wires feeding the VFD cannot be in the same conduit, or in case of NM cable of SO cord, cannot be run closely parallel to each other for more than a few feet. If they are, the carrier frequency of the VFD output will (through capacitive coupling) induce enough current into the grounding system to trick the GFI.

I guess this turned out to be pretty technical after all, I just can't think of simpler terms; sorry.

Rob

sandsmarc 07-31-2008 11:36 PM

Great info, I'll start attempting a fix. I would prefer the protection if I can get it, even if it is not required, although I'm pretty sure it is now required.

BigJimmy 07-31-2008 11:50 PM

<Yawn>

sandsmarc 08-01-2008 12:03 AM

<someone needs another sandwich...>

micromind 08-01-2008 08:55 AM

It's also possible that a ground fault exists between the VFD and the motor. Usually, the VFD will catch these and go into a fault mode, and not try to start the motor at all, but the VFDs threshold might be higher than the GFIs.

The test is simple; disconnect the power and motor wires from the VFD, then connect the power wires to any two of the motor leads, then turn on the breaker, (just for a couple of seconds). If it trips, you've narrowed the location of the fault. If it doesn't, then there's no ground fault anywhere in the system. The motor won't turn during this test, and can take 120 across its leads for at least 30 seconds without damage.

Rob

Yoyizit 08-01-2008 11:42 AM

Quote:

Originally Posted by micromind (Post 144762)
apparently the electronics in the VFD can fool the electronics in a GFI into thinking there's a fault, and it'll trip.

distorts the sine wave of the power supply. Looking at it on a dual-trace oscilloscope, the waveforms of the hot and neutral don't always match.

You know, Mr. Mind, saying that GFI's can be fooled by nonsinusoidal currents is pretty controversial in some circles.
I've given up trying to sell that idea. . .
:huh:

Stubbie 08-01-2008 01:23 PM

Quote:

Originally Posted by Yoyizit (Post 144878)
You know, Mr. Mind, saying that GFI's can be fooled by nonsinusoidal currents is pretty controversial in some circles.
I've given up trying to sell that idea. . .
:huh:

Not at all... not anymore than rf can trip a gfci. The problem you have selling the idea on your other posts is that there wasn't any vfd's involved or rf interference.

If any of those posts had read "can a vfd trip a gfci?" You just might have someone listen to you. It is very difficult to respond in a positive manner to someone who gets all their information off the wiki or internet sources then misapplies that information on a thread.... Or is not relevant.

Yoyizit 08-01-2008 03:09 PM

Quote:

Originally Posted by Stubbie (Post 144901)
Not at all... not anymore than rf can trip a gfci. The problem you have selling the idea on your other posts is that there wasn't any vfd's involved or rf interference.

If any of those posts had read "can a vfd trip a gfci?" You just might have someone listen to you. It is very difficult to respond in a positive manner to someone who gets all their information off the wiki or internet sources then misapplies that information on a thread.... Or is not relevant.

I concede some of what you're saying, and I will try to keep my position narrowly focused and on point. And I definitely have to be more literal. No glossing over details.

In my own defense, I have done very little writing for general audiences. That will have to change if I am to be tolerated on this forum.
:huh:

I've thought about this some more.
Here's some stuff from a previous post of mine. Sorry for the font, I can't fix it in the edit mode.
It seems I already knew this about people, I just didn't apply it to this forum. But I should have. . .

". . .you might need to fully explain and fully justify your methods, materials used, troubleshooting strategy, everything.

Knowing yourself that you are an expert doesn't count for much if the customer doesn't know you, which applies to almost everyone on this forum.

Ash 08-01-2008 04:32 PM

Some filtering circuit which has caps between ac and ground may trip RCDs. Simple modulating circuits are not likely to do it

Check this possibility and the possibility of real fault in the computer or the pump

chris75 08-01-2008 07:16 PM

I'm betting on an ACTUAL faulty pump, I know its a long shot with all the fancy equipment being used to control a pump thats in a body of water connected to a cord with a termination to the pump that would never ever leak, but o'well....

wbabbitt 06-22-2009 08:28 PM

GFCI Trips when VFD controlled motor is energized
 
2 Attachment(s)
I converted a single phase 12 0z in, motor used for calibration of tachometers. The original setup was a two speed, two winding, [four total] permanent split capacitor, synchronous [yes I mean synchronous in the true sense of synchronous, rotor and all].

HISTORY The device never worked without problems from day one, which was years ago, and they hobbled along with it until they ran into me. By not working the problem was two fold, not enough torque, on some, speed fluctuation on some, with new caps or not, and they really needed three testing speeds to couple their device to, 300RPM, 600RPM, 900RPM.

It originally had only 600 and 900 rpm by using this two winding single phase motor on top of all the rest of the complexity. I then had to meet these tolerances: plus or minus 1 rpm, all three speeds min 12 oz inches of torque, with a four position rotary switch, off, 300, 600, 900.

I had a very good company modify one of their stock non feedback drives, no room for tach or encoder, and from my first attempt, of about plus 2 or 3 rpm, they tweaked the special order drive to get me 299.5, 599.9, and 900 on the nut. Measured with a NIST calibrated laser tachometer at 16 oz inches of torque, all was wonderful.

Then they plug them into circuits with a GFCI, and the damn thing trips every time. Nothing was said this was a consideration up front, but now it is, so to build more, I have to come up with a solution.

The motor meggers with a 1000 volt megger at above 5000 megohms, the drive will not trip with the motor leads disconnected, but screw the motor on the chassis with a correctly wired three conductor so cord, 36 inches long, and grounded to the frame, where the motor sits on top, and it trips the GFCI right now.

I have twisted the incoming power by using same size conductors, and same problem, took everything out of the chassis and laid it on a table, same thing, I am not a drive expert, so what could I possible try? Isolation transformer on the incoming 110 to board? The motor is supplied with manufactured 230 volts three phase, good balance on voltage and current all three legs. Drive is 115 single phase in. Company that made them no longer does special orders.

The box that holds the motor on top is metal, an inch and 3/4 deep, it about 8 inches wide and a foot long, so plenty of room with a chassis mount bread board drive, but I need a low profile bread board type chassis mount drive if I have to change out the drives, but there should be a solution to this.

Any ideas? I see this is an old post, so if anyone is dealing with these issues, and can toss me an idea or something I would surely appreciate it.

I have photos and will look to see if a photo can be uploaded if not my email is service @emrrepair.com. Attn Will

Thank you in advance for reading through this mess, and putting thought to it.

spark plug 06-22-2009 09:01 PM

VFD controlled motor tripping GFI brkr.
 
Quote:

Originally Posted by micromind (Post 144762)
The computerized motor controller is more commonly called a variable frequency drive, or simply VFD.

My experience with 120 input VFDs is very similar to yours. Even though absolutely no actual ground fault exists, apparently the electronics in the VFD can fool the electronics in a GFI into thinking there's a fault, and it'll trip.

Without getting too technical here, it has to do with the fact that the input current to the VFD distorts the sine wave of the power supply. Looking at it on a dual-trace oscilloscope, the waveforms of the hot and neutral don't always match. The further the VFD is from the GFI, the worse this problem is.

Possible solutions;

1) Replace the GFI breaker with a non-GFI one, and install a GFI receptacle close to the VFD.

2) Install an isolation transformer, or a line reactor at the VFD. Either of these will cut back on reflected power, and the waveforms at the GFI will be much more closely matched.

3) I don't have a codebook handy, but you might not need a GFI at all. (Several of the other guys around here will chime in on this one, I simply don't know without looking it up.)

4) The wires feeding the VFD cannot be in the same conduit, or in case of NM cable of SO cord, cannot be run closely parallel to each other for more than a few feet. If they are, the carrier frequency of the VFD output will (through capacitive coupling) induce enough current into the grounding system to trick the GFI.

[I guess this turned out to be pretty technical after all, I just can't think of simpler terms; sorry.]

Rob

No. This is the best explanation for the reason the GFI breaker trips when there is no [apparent] Ground Fault. The working principle of the GFCI, whehter it's a breaker or receptacle is; That it "reads" the current in both the HOT and NEUTRAL. When it detects an imbalance (leakage current) it trips. Thus protecting a person against forming a path to ground!:no::yes::drink:Don't Drink and Drive!!!

wbabbitt 06-23-2009 01:22 AM

Gfci tripping from VFD
 
Thanks I appreciate the advice, the problem is these things are for the National Weather service, they have their own cords, and the Government never will listen to NO or REASON, or anything else, if I sent them every post which pretty much says the same thing, they would say, find a way,

This VFD has no particular chassis ground, it is like a breadboard with a terminal block in, and out, no PE like the Vtech nema one enclosure units I usually use, so there is NO chassis ground on the drive, and no where I can see to put one, I know they don't like bridging the neutral with earth, I am going to find some kind of snubber or something to reduce the back emf, back whatever, and see if something like that helps, trouble is I don't know hat to put across the incoming 115, but hopefully some drive expert will tell me, but hey thanks for replying, lots of brains on this site, and really responsive people, much appreciated, Will

Yoyizit 06-23-2009 11:07 AM

Quote:

Originally Posted by wbabbitt (Post 291458)
The motor meggers with a 1000 volt megger at above 5000 megohms,

So it's reactive leakage current >4 mA that's tripping the GFCI. Capacitance to ground somewhere of 0.09uF or more?


All times are GMT -5. The time now is 07:40 AM.


Copyright 2003-2014 Escalate Media LP. All Rights Reserved