Resolution of IoTaWatt and calculations

Hi. I am wondering if I will be able to see KwH if I hook to my mains, the way my power meter reads it. I have a 200 watt per hour avg base load in my home that I want to narrow down to exactly when that power is being used. My power company gives me readings in 15 minute increments, so it takes forever for me determine what is drawing the load, especially if it is a fridge that runs for 5 minutes. I know the draw isn’t constant. The 200 watts that is being recorded by the power company is based on a low power factor, so its not as simple as taking amp readings, because my amp readings are way off from my reported KWh readings. I also don’t necessarily want to hook up individual CT’s to every circuit, just to the mains for starters.

So will this give me what I need to find the time frames where the watts are being used? I know amps x volts x pf to come up with watts. Is this the tool I need to figure it out?



If the question is will the IoTaWatt report the same kWh as your meter, what I can say is that in my US 120/240 split phase installations, the IoTaWatt seems to be between 1% and 2%. That said, meters are not always accurate. Newer electronic meters seem to be more in tune to IoTaWatt than older mechanical meters. I had a mechanical meter that was reliably 2% below various IoTaWatt. I got a new electronic meter a few months ago and now they are within 0.5% of each other. An additional challenge may be that you seem to be talking about particularly low usage, which could cause a little more disagreement.

I’m not clear on your usage. Are you saying your average usage is 200watts/hr? In other words, are you using ~150kWh/month? In any event, here’s an IoTaWatt plot of my minimal usage (500w/hr):

That’s the sum of two CTs (one on each main), so you can get more detail about 120V loads by looking at them:

So now I can see a regular load on main 1. I’ve got a CT on my basement and that was my dehumidifier in the basement (July):

so I can subtract that out and a small HW tank that I have a CT on and start to see what’s left:

I can identify most of what’s left because I have CTs there, but just the magnitude and timing of a lot of the major loads are usually identifiable. The key is that you can add and subtract channels and plot the results siting in a chair, as I have just done.

1 Like

Ok. Yes. I’m talking 200w/h and I have a digital meter. Like yours said in earlier posts, it’s resolution is 1kwh, which doesn’t help me detect electricity. My base line represents 20 percent of my power bill though, so my question is what is the resolution as watts decrease. Are we talking about 10 watt units? As the watts decrease, I would think margin of error increases. Can you detect a power factor change when you plug a iPhone charger in? I can’t find a measuring device to help me really dissect low power devices. I’m incrediblely interested in your device if it is reasonably accurate at low wattage measurement.

Looking at the just the mains? I doubt it. IoTaWatt may be a pretty good magnifying glass, but I think you are looking for a microscope. jI can’t reasonably represent that it can do what you are asking. Sorry.

Well, not with a kw of power being drawn, but let’s say something more reasonable, like 100 watts of power being draw from the whole circuit. Looking back at your graphs I see that there are small changes in wattage usage throughout the hour so I think it can detect it. It’s a matter of using the properly scaled CTs for the job. One more question about kw usage.

If the power plug is plugged into only one side of the panel, how do you calculate the voltage cycle for the other side? Do you just assume the opposite wave position based on the measured cycle? makes sense to me and should be very accurate. The reason I ask is what changes are made when you go into 3 phase? Using this approach, you would Need only one more power plug to measure the wild leg? But I think knowing which leg (ABC) you are tied to originally when you plug in the wall plug, the position of the wild leg can be calculated too.

There’s a discussion of that here.

There are two approaches to three phase with IoTaWatt.

Derived reference uses one VT and derives a phase and voltage from phase shifting.

Direct reference - uses three CTs (for three phase).

Direct reference can be used for split phase US systems as well, but understand that US split phase systems at 180deg phase difference aren’t subject to variation in relative phase angle as in three phase systems. So It’s just a matter of voltage. There is a discussion in the Learn section that explains how the neutral line can drift around with varying loads and change the relative angle between phases in three phase systems. The phasor graphic is pretty easy to understand.

Figuring out base load with a single measuring point is really hard but possible if you are willing to put in the time investment. Having many channels of measurement makes it much easier. The resolution of IotaWatt is great for figuring this kind of thing out. If you have CTs on all of the circuits, you can put a set of them on the available channels and monitor for as long as you like and then put a different set on so you can get all the usages. As an example, I have an outlet in the laundry room attached. The panel was mislabeled and I haven’t got around to changing it to something more interesting. But, there was this 4-6W load that showed up for hours at time every few days. Since that circuit only goes to the outlet that is used for the iron, that was weird. I finally asked my wife about it and it turns out she plugs in the portable vacuum there to charge it. She, of course, rolled her eyes at my “spying” on her :grinning:

If your goal is to reduce consumption, absolute accuracy is less important than repeatability and resolution.

OK, I’m very new to OEM actual reporting engine. I have read that you have your own graphing app that gives finer detail than the oem provides. OEM requests updates not done in less than 10 second intervals. OEM reporting online reporting may not give me the precision that I need vs iotawatt’s built in graphing? You were saying that in the graph, a channel can be subtracted from the total, was that in OEM or the onboard graphing? Comparing the emonTX V3 to the IOTaWatt, we have more CT inputs, onboard logging, onboard graphing, and wireless connection out of the box to OEM, does that summarize it?



It’s the same basic graphing app, modified to get data directly from IoTaWatt.

IoTaWatt keeps the most recent 12 months at 5 second resolution. Thereafter it is 60 second resolution.

That’s done by defining outputs on IoTaWatt using the “calculator”. Once an output is defined, it can be plotted in the graph. Outputs act on any data in the log, so when you define one, it is immediately available for any historical data as well as new data.

I’m not an expert at the EmonTX series. They work in a different way, through Emoncms, either local or remote. The way I look at it, those devices have evolved over time to serve some diverse energy monitoring requirements, like temperature and humidity, on/off and direct pulse meter reading. They are part of a system. IoTaWatt is more of an electricity monitor stand alone type unit. So while those things you mention are true, it doesn’t cover the reciprocal comparison. Depends what you are looking for.

I’m not sure I’d completely agree with that. I came across several big anomalies with retail plug-through power meters while I was trying to track down my baseload. In this picture:

I’ve got two plug-through power meters measuring my microwave oven in standby (just being a clock). They both claim to know about PF and one clocks it at 83W with a PF of 100% while the other reckons it’s 2.5W witih a PF of 0.2. For the TV in standby, they claim 17W Vs 0.9W respectively.

As the OP points out, it’s all about PF. Some power supplies that have both a high power mode and a standby mode can appear very reactive while they’re in standby mode. If your monitor doesn’t measure that accurately, you can end up chasing down ghost loads that aren’t really contributing to your baseload at all, but sure look like they are.

Here’s how my energy monitor sees the microwave while in standby:

The current signal (Red) has some harmonics in it, but it’s sine-ish and very close to 90° out of phase with V (Green). In that case, even a small introduced phase error can make a significant difference to the result and you can end up chasing down loads that really aren’t a problem at all.

Although that doesn’t come close to explaining how one of those meters claims 83W and a PF of 1. It’s completely lost its mind. 83VA is about the right answer for Apparent Power, but it’s pretty much all reactive.

My comment was assuming a reasonable level of accuracy on a whole house monitor and by turning off circuits and/or unplugging individual items.

If power factor of a device is bad, a decent meter might be off by a lot at very low current, but probably not so much at higher levels. Even if it were, I would always double check before making the decision to replace something. As an example, I have checked all my stereo radio/amplifiers and they are certainly different with some idling at as much as 50W when on but at a low volume. That does add up if they are on all day. I removed one that was particularly bad and replaced it with an unused one with much lower idle power for a location that is on most of the day.

I know the lure of high accuracy and I have a pack of 10 very high accuracy energy meter ICs in my parts bin waiting for me to put them on a board and hook them up. But, they been there for years and I still don’t know when I’ll get to it. So when the IotaWatt went on sale I bought one, since actually measuring something with reasonable accuracy is probably better than waiting for the perfect solution.

It sounds like you got much further in your quest and I am sure you are getting great insights from the data you are collecting, which is excellent.

1 Like

I reviewed the code and I like the way the IoTaWatt is layed out. Looking at the schematic, It looks good. For that reason, I have placed it on my giftster Christmas List to see if the wife will pick one up for me.

My average KWH is .7 over a 24 hour period most days. We have a 40 circuit panel, so measuring as a whole is important because I won’t be able to hook up to every single circuit on a leg. Thank you Bob for answering all my questions. dBC for sharing your experience measuring power, I experience the same inconsistencies with kill-a-watt meters from different era’s. What will make me more upset than anything else is if I find out between my wired in smoke alarms in every room, my builder grade ground fault breakers and my door bell transformer, it’s using most of my baseline power requirement. All of those things can’t be measured with any plugin device and must be measured at the panel, but requires specialized equipment to do so because of the power factor.