Two days ago, Sense announced it had found a new device in my house. Yay! Right? Always happy to see that.
But, immediately, I noticed something puzzling: Sense knew it was a Panasonic! How did it know that? Only one way I can think of: it’s a smart TV, and Sense somehow found it on my network.
OK, that’s fine, but then comes the problem. Sense has assigned it a load value of 84W. According to the meter graph, it actually uses about 38W, even with the cable box on/off at the same time. Apparently, Sense is using a lookup table to get the watts instead of using an actual measurement.
Doesn’t that mean the 84W assigned to the TV is introducing a false parameter into all energy calculations?
Hey there @c72rg. Thanks for sharing. This looks like an issue with Network Device Identification (NDI), Can we take a look at your monitor to see what’s happening here?
Nothing on your end at this time. Someone from our engineering team is remotely looking at your monitor to see what the issue is. I can let you know where we net out.
From our side of things, we have very good device model for your TV (device seems to be turning on and off appropriately.) However, It seems that when the TV turns on, a bunch of other entertainment devices turn on as well in quick succession. Since everything turns on together, what you’re seeing is the cumulative consumption for the entertainment devices plus your television. With all of these devices turning on in quick succession, it’s very hard for us to separate the TV from other entertainment devices. It looks like your television consumption is actually around ~83-84W.
My recommendation would be a Kasa HS-300, which i use to monitor my entertainment center and office. With the smart plugs we integrate with, you can name the device plugged into each individual outlet and that information (and data) will be visible within Sense. When setting up you television with the HS-300, you would just combine the original television device with the smart plug device. This would solve the issue moving forward.
When the TV and its 2 associated devices are on, the Sense meter (moving graph) screen reports a total increase of ~42W, not 84W as seen on the balloon screen.
I have checked this by stopping all other loads in the house and turning the Panasonic tv and its 2 associated devices on and off. The balloon says 84W, the meter says 42W.
There are only three devices that energize when the entire entertainment stand is activated, and they register together a sum about half of 84W, at least according to what my Sense system is reporting.
This sounds more cool than problematic. Is there a Sense blog about NDI? I work in television and we use NDI for broadcast-level devices, so I’m curious how it works for detecting consumer electronics.
TVs and many other DC-type devices are very difficult to detect. I didn’t even bother with mine and put the whole entertainment center behind a HS-110. I have three other LCDs, two laptops, and numerous phones and tablets that haven’t been detected. I’d Sense is detecting DC electronics, That is cool.
Nice feature? Perhaps, but on my system, NDI is having trouble assigning watts properly to devices identified this way. If NDI can’t do that, it is not a nice feature.
@c72rg, in my mind, this shows how poorly knowing on/off translates into power prediction. Experiments with my TV show how much power usage depends on content, not merely whether it is on or off.
Yes, the brightness of the scene has an influence on momentary watts used by a TV, especially if it is one of the very large screen TVs. This is very distinctively shown on the meter screen, as the wattage line will be very jagged. It should make identifying TV sets easy.
Two things. With recent large screen OLEDs, I believe that much more energy variability comes from decoding and up scaling in the processor, than screen brightness. And jagged wildly variable power usage makes it difficult to both detect on/off plus also to predict power usage of the TV.