Data by breaker manufacturer

Has anyone polled users for what type of breaker box they have (GE, Siemens, Westinghouse, …), and also if they have any hums / buzz noises. I am having a disappointing discovery rate (but maybe I’m just impatient (6 devices in 2-3 weeks). But if my box is noisy itself because of the way the breakers attach, it would be good to know.

I have a siemsens box and the breakers do not screw down, just snap in. I could see how this could be a poorer design for “quiet” operation.

That is an EXCELLENT discovery rate…took me a year to get there. Patience.

1 Like

Yes that appears good rate, I have been on for 3 weeks and only have 1 item discovered. I disagree with the patience, they did not advertise results in a year so the time is not as advertised at all. All residential panels will have snap in breakers. the screw or bolt in ones are all on commercial heavy duty load centers.Siemens are one of the better ones in my opinion. From being an electrical contractor for 20 years

I rate the panels as (it is basically from high price to lower but cost is not how i ranked them)

#1 - SquareD QO series and Cutler Hammer,
#2 Siemens/ITE and SquareD home series
#3 GE and others at the bottom of the list.

This is based on the breakers themselves mainly, I have seen the GE and others have a load of 30 A on a 20 A breaker and not trip which will cause the buss to burn.

It is always good to have an electrician to check the panel approx every year or two (more often as they get older) to check the buss for signs of bad connections. Most if not all home load centers have aluminum buss bars and the breakeshave steel or an alloy for the connections. The dissimilar metals can cause issues over time.

2 Likes

@brian4,
Gonna post this Sense Q&A on noise and what makes a device detectable.

1 Like

Great video , thank you.

So, the CEO says something right at the end that if I understood I would disagree with. I think he says that with voice data someone can type in what the data represents, and that you cannot do this with electrical data…

Some truth to this with operation signal of my TV , but the on signal of my TV, XBOX, or other device I can purposely turn them on and off. I could imagine that the XBOX like the Nissan Leaf has hundred of different power down signals , but still. In the simplest view, I can say bathroom lights and toggle them 50 times for training…

Still happy with my sense, just wanting more discovered…

You ask a good question - I think there are two bits of context embedded in Mike’s answer that you might be missing:

  • machine learning works best with a huge tagged datasets, where tagging refers to labeling the spacial or temporal feature to be learned (face in bounding box or sentence in an audio clip). Tagged datasets have enabled the breakthroughs in machine learning we’re seeing today. But tags have to be applied close to the actual features to be learned - in Sense’s case, that means consistent tagging within a half second or so of the transition.
    ImageNet: the data that spawned the current AI boom
  • there are tons of tagged speech datasets where the tags have been added after the fact (same as image tags) by humans. But how would you, as a human tag an electrical transition ? Highlight a little portion of the waveform and add the tag info ? It’s not good enough to just repeat the on/off cycle 50 times in row.

A post was split to a new topic: Device detection questions

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.