GFCI with Pentair VS pumps

[P00LNerd]

I noticed the wood was really nice.

Why mahogany?

1. Very weather resistant
2. I had some left over from one of my woodie boat rebuilds
3. Easy to work, I simply T&G'd a length of 1x6 plank, cut it to three sections, and glued it up to make the square panel.
4. Looks nice. My lighting controller is mounted to a similar panel, in a more-visible location.

 

[P00LNerd]

All VFDs create noise that can cause nuisance tripping.

Yeah, but if the claim is that Pentair has suppressed the noise from their VFD's sufficiently to eliminate this problem (see @Newdude), then we have to wonder if it was done only on new designs, or folded back into older designs like this VS Supermax.

One thing I have yet to verify is that the problem isn't associated with wiring damage. I guess I need to just get out there and pull the wiring. It's possible the fence installers damaged the liquid tight and wiring, in this branch circuit, and that is causing the random GFCI trips. I have noticed most nuisance trips correlate to rain days.

 

[JamesW]

the claim is that Pentair has suppressed the noise from their VFD's sufficiently to eliminate this problem

The breaker is what Pentair claims to have fixed, not the VFD.

Pentair just buys the Siemens Breaker and rebrands it with their label.

It is the exact same breaker.

Insulated Gate Bipolar Transistors (IGBTs) are a key component in what makes up a VFD (Variable Frequency Drive).

The input voltage is converted into DC and then into a type of digital AC using switches that create three phase AC with the correct frequency to make the three phase motor rotate at the correct RPM.

It is this choppy process that causes noise that is transmitted back into the breaker and it causes nuisance trips.

The Siemens breaker has some sort of "Shielding" that helps reduce nuisance trips.

What is a Variable Frequency Drive?

Looking for a VFD for the first time? Learn the basics of what a VFD is and the differences between VFD types. Find out what to look for in a quality VFD.

vfds.com

Switching it up: IGBTs

Learn how an IGBT helps to create PWM wave and also how their fast switching relates to carrier frequency, an important parameter you can set on your VFD.

vfds.com

 

[P00LNerd]

The breaker is what Pentair claims to have fixed, not the VFD.

Pentair just buys the Siemens Breaker and rebrands it with their label.

It is the exact same breaker.

Insulated Gate Bipolar Transistors (IGBTs) are a key component in what makes up a VFD (Variable Frequency Drive).

The input voltage is converted into DC and then into a type of digital AC using switches that create three phase AC with the correct frequency to make the three phase motor rotate at the correct RPM.

It is this choppy process that causes noise that is transmitted back into the breaker and it causes nuisance trips.

The Siemens breaker has some sort of "Shielding" that helps reduce nuisance trips.

View attachment 595422

What is a Variable Frequency Drive?

Looking for a VFD for the first time? Learn the basics of what a VFD is and the differences between VFD types. Find out what to look for in a quality VFD.

vfds.com

View attachment 595423

Switching it up: IGBTs

Learn how an IGBT helps to create PWM wave and also how their fast switching relates to carrier frequency, an important parameter you can set on your VFD.

vfds.com

Excellent. Thanks for the info. In that case, I'll drop the VFD swap idea, and look at maybe trying to filter the line between the breaker and pump VFD.

 

[P00LNerd]

The breaker is what Pentair claims to have fixed, not the VFD.

Just re-reading this thread now, I think newdude was saying Pentair changed something in their pumps, to reduce nuisance trips with other breakers:

The pump is recent right ? 1.5 years old or so ? ...There haven't been any reports on new pumps in a several years here so that was also pointing to the problem being solved.

What's the manufacturer date on the pump ? It might have been old stock when you got it, or we can prove it's still a thing.

In any case, I picked up a cheap 20A / 250V EMI filter. This weekend, I'll pull and inspect the wiring, make sure that's not the cause of the trouble. Then when the EMI filter arrives, I'll try to capture waveforms or spectral data with/without that. Hoping that resolves the issue, as I can mount that in a small outdoor box on the back of the existing panel board.

 

[P00LNerd]

The EMI filter came yesterday, so this morning I went out in our swamp-like heat and played a bit. Using a cheap scope without onboard FFT, so I can really only look at waveforms, unless I dig out an old copy of MATLAB to import .wav files and FFT them.

Here's measuring on the breaker, without the EMI filter installed:



Then with the EMI filter installed:



As you can see, small differences, but nothing stark. I didn't bother zooming in further, to see if there were even higher frequency components, as I suspect VFD is all low frequency modulation.

A few observations:

That fine ripple seen in the images taken at 1 ms/div would be about 11 cycles / 1 ms = 11 kHz. The vertical scale is 50V/div, so the minor grid dots are 5V/ea, and the amplitude of that 11 kHz component is thus about 5V (= 10 V pk-pk) = 7 Vrms). Installing the filter knocked that component down a bit, from 10 V pk-pk to maybe 5 - 7 V pk-pk. Nothing huge, but maybe enough to make a difference.

Some of the other bigger glitches, such as that 15 V kick right at the tail end of the visible waveform in image 3, were also reduced. Maybe this makes a bigger difference.

The major "knees" after each waveform crest were mostly unaffected.

I had rigged this up in open air, nothing I could leave standing. I think I'll put this filter in an all-weather receptacle box, and nipple it onto the breaker panel, so I can re-install the GFCI breaker and see how it does. I also still need to pull the wiring out of it's underground conduit run, to make sure this isn't a simple case of damaged wiring causing the GFCI trips.

 

[JamesW]

One thing you can do is to use a clamp ammeter and clamp around both power wires at the same time, which should result in a reading of 0 amps because the readings should cancel out.

 

[P00LNerd]

One thing you can do is to use a clamp ammeter and clamp around both power wires at the same time, which should result in a reading of 0 amps because the readings should cancel out.

I'll have to check the bandwidth of my clamp-on ammeter, if it's even published. That could be a good method of measuring common-mode noise, if the thing has at least a few hundred kHz of bandwidth. I suspect most RMS ammeters are filtered way down, though, so most of the noise would not come through.

I used to feed the voltage output from clamp-on ammeters (eg. Fluke i30) into an oscilloscope, but it was always for measuring current change while slowly sweeping input conditions, so not looking for any change much faster than a few hundred milliseconds. Here, we're looking at 11 kHz = 0.09 milliseconds.

I really don't know much about VFD's, but suspect most of their operation is in the realm of 1 kHz, plus or minus 1 decade of bandwidth.

 

[P00LNerd]

Pulled wiring this morning, and it’s all clean and good. Installed EMI line filter in an IP65 box tied to the back of my load center mounting panel, and wired it back up to the CH220 GFCI breaker. Will let it run at least two weeks before calling the issue fixed.

The reason I had suspected wiring was that I noticed the GFCI mostly seemed to pop on bad weather days, almost like water getting into the wiring might be the cause of it. I don’t know if water on a poorly-coated motor winding could cause GFCI faults, though a VFD, or if even my conclusion that the trips were related to weather was more coincidence than correlation.

Anyway, we’ll check back in 2 weeks, if nothing exciting happens first. In the meantime, I can probably turn some attention to plumbing up the new SWCG.

 

[P00LNerd]

Well, we experienced our first false trip this morning, with the EMI filter installed on the line.

I was thinking about this yesterday, and realized that they offer these filters with many different L/C values, and that I could probably pick one much more suitable than the first one I grabbed off Amazon. But, since I already need to lay out a new and larger panel board to hold a SWCG controller and some timers, I am just going to use it as an opportunity to switch everything over to Siemens.

It irritates me that Siemens doesn't make a nice tall/narrow 6-slot lugs load center like Cutler Hammer (CH), Eaton (BR), and Square-D (QO) all do, but I'll just make the new panel large enough to hold the Siemens 4x2 = 8-slot load center.

When I get a chance, I'll get the schematic and values off the EMI filter I tested, and maybe even run some quick calculations to suggest new/better values, for anyone else who runs into this problem and wants to pick up the effort later.

 

[JamesW]

2025 TFP Donation Campaign

7/23/2024, 11:15am: We have achieved our $10,000 goal! Thank you again to all who have made this campaign so successful! The benefits of donating during the campaign (see below) will still be active until August 31st, 2024. Welcome to Trouble Free Pool’s 2025 Campaign! We are seeking to raise...

www.troublefreepool.com

 

[P00LNerd]

Dang... got CircuPool controller hung on the panel, and the line filter, and the utility receptacles, and the new receptacle for the SWCG... and the original load center. That is one densely-populated little square of mahogany.

I have to install one more thing, and then I'll get a photo, just for comic relief. This is all temporary, I'm going to tear it all out and redo with a Siemens panel after closing the pool this fall.

 

[P00LNerd]

... and as promised, one densely-populated little square of mahogany.



Yeah, it's goofy. This wasn't planned, but grew organically when the need to add SWCG and a line filter (VFD on GFCI). It will be redone in the off season, with a new load center, and probably a roof over the whole thing.