Watts not zero when breaker is off

[peppersass]

Some of my channels show 1 or 2 watts when there's no draw from the load. This can happen even when the breaker is turned off. It happens periodically on 120V circuits but not terribly often, and it usually doesn't last more than one packet cycle (8 seconds) before returning to zero watts.

But it happens much more frequently on 240V circuits with neutrals where I'm using CTs on each legs wired in series (same polarity on the legs, opposite polarity at the channel connector). And it can persist for long periods. In fact, I'd say 1-2W is displayed more often than 0W.

I can't determine the amount of current flowing because the reported current in the Live Data display appears to be incorrect for small loads. For example, a 1 watt load might show 0.60A in the current column. I suspect that's really .006A, which would likely round up to 1 watt at my nominal 121V. I've noticed that when the current for one of these "off" loads falls to .02A the watts go to zero, so my guess is that .04A is the threshold because if it's really .004A then power would be real close to .5W, which is probably the round-up or round-down threshold. The fact that the current for channels 17-32 is a repeat of the current for channels 1-32 further complicates the troubleshooting (I reported this bug a couple of days ago.)

I'm using micro-40 CTs on all circuits.

My question is this: Why is any current flowing at all?

Answering my own question, I assume there's noise on the CT leads. I haven't extended the CT leads. All leads are the stock length. They're much longer than needed, so I coiled up the excess wire and bundled all the coils together (I'll be posting photos of this soon.) It's possible that in some cases the coils are picking up stray RF (or they might be choking it to a lower level.)

Would tweaking the CT type number down a notch get rid of the noise without throwing off the accuracy?

If not, is there anything else I can do to address this problem? I read a post in the ECM Support area about a similar problem, where Ben recommended swapping the CT leads. But these had been extended with CAT5, and it sounds like the ECM connections are different anyway.

Any suggestions would be most welcome.

[Teken]

You have 3-4 choices here from simple to more time consuming.

1. Place a jumper wire on the channel on the GEM that is giving you the problem. Details are listed in my projects thread.

2. Fine tune the values for that channel within the GEM which you have indicated before.

3. Using the GEM Network Tool you can make current / amp adjustments following the guide which is also listed in my projects thread.

4. Place a ferrite choke on the circuit.


Encrypted By: Phoenix Security Solutions

[ben]

The resolution is 1W steps so if you have any noise, you're guaranteed to see 1-2W (nothing less). Try what Teken suggested above as it does sound like noise being picked up.

[peppersass]

Changing the offset moved the current on some of the zero loads to zero amps, but that didn't necessarily stop them from sometimes displaying as 1W. And sometimes the amps would change to 02A or .06A. In that case, the power might randomly display as 0W or 1W. It did appear that changing the offset reduced the frequency of 1W being displayed, but I'm not sure.

Jumpers were somewhat more effective, but didn't completely eliminate the problem. I installed them on 8 channels where the loads can go to zero (many of my circuits have small vampire loads and aren't likely to go to zero.) After installing jumpers, I set the offset back to zero. That didn't seem to change anything, so that's where I left it.

I'm sure it's a noise problem, and it's it's not on all the circuits. Some zero-watt loads consistently always up as 0W, while others occasionally show up as 1W. All four of my 240V loads occasionally display 1W when there's zero load (three use two CTs and one uses one CT). It looks like they do it a little more frequently than the 120V loads.

Teken, have you actually tried ferrite chokes? If so, what was the part number and source?

Seems to me that several miniature ferrite beads on each of the two leads might do the trick. That said, I don't know the frequency of the noise, so selecting the right choke will probably be a shot in the dark. I may have to drag a scope down to the basement to see if I can figure out the frequency or frequencies of the noise.

I might try shortening the leads, too. There's an awful lot of excess wire between the panel and the GEM.

[Teken]

Hello,

A couple things I didn't see in your reply being done. I would humbly suggest you first measure the line voltage where the GEM is plugged into. Measure the line voltage with a high quality DMM with RMS capability and ensure the it's the same and adjust as needed.

Once done, adjust each channels values using the set up page in the GEM. Using known loads like a incandescent light bulb for small, medium, and a heater to reflect high resistive loads adjust the CT value so it's close to what the load is.

I would definitely uncoil the wires from the CT's as it acts like coiled antenna in some instances and induction can happen if done near the service panel.

I will need to follow up with you about the ferrite choke as it was something I had in my junk drawer and simply threw it on to see.


Encrypted By: Phoenix Security Solutions

[peppersass]

As it turns out, I adjusted the voltage yesterday using a Fluke 189 DMM that displays true RMS voltage. Although the voltage fluctuates, on average it was within a two-tenths of a volt or so of the live data value shown by the GEM. I bumped the PT adjustment up by one to get it a little closer.

Now that I think about it, I suppose the fluctuating voltage could be all or part of the "noise" that's driving the problem, and that adjusting the CT sensitivity might help. That said, a few days ago I briefly tried adjusting the CTs of the zero-load channels to try to zero them. But I haven't tried adjusting to a known load. I'll give that a try.

It's a bigger project to trim all of the CT leads and I've been holding off as it's non-reversible. There's very little chance I'll need to move the GEM, though, and I guess I could extend the leads if that happened. I'll start with the affected channels and see how that goes.

And yeah, I'd appreciate any info you can give me on the choke. Got plenty of stuff like that in my junk box, too.

[Teken]

As it turns out, I adjusted the voltage yesterday using a Fluke 189 DMM that displays true RMS voltage. Although the voltage fluctuates, on average it was within a two-tenths of a volt or so of the live data value shown by the GEM. I bumped the PT adjustment up by one to get it a little closer.

Now that I think about it, I suppose the fluctuating voltage could be all or part of the "noise" that's driving the problem, and that adjusting the CT sensitivity might help. That said, a few days ago I briefly tried adjusting the CTs of the zero-load channels to try to zero them. But I haven't tried adjusting to a known load. I'll give that a try.

It's a bigger project to trim all of the CT leads and I've been holding off as it's non-reversible. There's very little chance I'll need to move the GEM, though, and I guess I could extend the leads if that happened. I'll start with the affected channels and see how that goes.

And yeah, I'd appreciate any info you can give me on the choke. Got plenty of stuff like that in my junk box, too.

I personally would not trim or cut down the CT leads. I would just uncoil them and ensure proper clearance and routing of the CT wires that exit the service panel.

I am unsure how you have the wires routed or suspended but in my install I used some stick on cable holders to assist in restraining the wires and to ensure they were not touching or over lapping any other breaker wires.

I hope the trouble shooting guide in my install thread on page five was helpful for some of the reference material.

[peppersass]

Here are some photos of my installation. I'm planning on a more extensive writeup with more photos, but here's a preview showing how I routed and dressed the CT leads. I like the idea of using standoffs to keep the CT leads off the AC wiring, but it's a lot more crowded in my panel than yours, with lots of wires running up the sides of the box. Not sure standoffs would be completely effective but would like to give it a try. Can you post a photo and the source of your standoffs? I don't recall seeing them in your project post.

Here how the CT leads are dressed inside the panel:

untitled-20_resize.JPG (210.36 KiB) Viewed 2714 times

Half the leads exit through a hole in the bottom right corner of the box and half exit through a hole in the right side just above that.
Here's what the leads looked like before I coiled them:

untitled-16_resize.JPG (205.79 KiB) Viewed 2714 times

Lots of extra wire! The GEM/DB mounting location isn't ideal. You can see several AC runs passing behind the GEM (it's mounted on standoffs). I could move them to the right to eliminated that, but they'd still be close to the GEM. They don't seem to be causing RFI to the GEM, though I haven't yet done precise accuracy measurements to confirm that.

I suppose I could build another wood panel to the right of the GEM, which would allow me to stretch out the leads. Probably a good idea because for years I've planned to invest in a standby generator but never got around to it. The transfer switch will need panel space, too. Only downside is mounting a new panel on concrete, which I've not done before. Will have to learn how to do it right so I don't crack or pierce the foundation.

Here are the leads coiled up:

untitled-22_resize.JPG (253.16 KiB) Viewed 2714 times

Note the outlet installed on a dedicated breaker for the GEM/PT/DB.

I've hit the 3-photo limit, so the next one is in the next post.

[peppersass]

Here are the packed into split cable tubing as neatly as I could do it:

untitled-25_resize.JPG (209.89 KiB) Viewed 2712 times

I routed the cables so that the leads had to travel the minimum distance to the corresponding terminal block in the GEM. I thought this would keep it neater inside the GEM. So, half the leads on each side of the main panel go through the bottom hole in the panel and the top hole in the GEM, and the other half of the leads on each side go through the opposite holes. This resulted in multiple sets of split tubing and awkward junctions at each hole in the panel. If I had it to do over again, I would route all the leads from the left side of the panel through one hole in the panel and one hole in the GEM, and all the leads from the right side of the panel through the opposite holes in the panel and GEM. Then there would be just two split tubes between the panel and GEM.

[Teken]

Just Wow . . .

When I said my pre install was pretty messy you got me beat! The third picture I believe definitely needs to be uncoiled from that massive bundle.

I don't believe any sort of stick on stand offs will fit your panel given the tight environment you have in place. Based on the photo it looks like you did a really good job in bundling them all up from the top down.