Switching service panels

I’ve had the Sense installed for a while now at the fuse panel inside our house. At the time I installed it we just had our garage remodeled and I had forgotten that the fuse panel inside the house was actually a subpanel of the main panel in the garage.
There are only two branches off the panel in the garage: a workshop area that gets used for a couple hours a month and the line into the indoor subpanel.
Device detection has been just sort of “okay” and I’m looking for ways to improve. Since there’s nothing I could purposely do that would change the ML process, I started wondering if location might have something to do with it. Would it make sense to relocate Sense to the main panel?

I would not bother moving sense to record the usage of a workshop only used for a few hours a month. The detection of devices within the home itself would not benefit from the move.

If you choose to move the monitor anyway follow the suggestions of others to ensure that each CT remains on the same conductor (or same side of the phase) and preserve their orientation too. Also do the same for the red and black 240 V power leads.

Contrary to what the post below thinks about CT and pier lead orientation, I think if you move them that far you’ll lose all your detections.
I changed a single breaker in my panel, that’s all, and had problems.
My feeling is it’s due to change in things like resistance values which will be different for adding length of wire.

Are you sure you didn’t change the “phase” when you were moving breakers. The resistance of the wire is tiny, I would highly suspicious if that changes things. Wire between panels is awg 10 or awg 8 … that has a resistance of 10mohms or 6.3mohm per 10ft, respectively.

Actually both bots are on the same phase as residential is single split phase. But no, I didn’t switch legs for sure. Not only that but I measured with dial calipers how far in the breaker the wire was inserted exactly the same depth. Sense doe t just get it’s meas from the CT’s, it also is measuring from its power leads so any tiny thing makes a difference. I’ve read where people have just turned their ct’s Sideways and not otherwise moved them, this has caused them trouble. I think we underestimate the sensitivity of the monitor.
I just noticed you had the calculations for OHMS in your post. When I was installing my heated floor, I had to measure ohms several times. If the measurements were off 10% it meant the wire was no good. My wire was 180ft long that’s why I think sense can see even tiny differences.

I know residential is single phase … this the quotes ;).
You measured how many mm the wire went into the breaker? I don’t believe sense is capable of measuring this.

Let’s assume 14awg wire… and even 1" difference. That’s 0.00025ohms … 0.25mohms. If you are pulling max current through that wire (15a), you are taking about a voltage difference of 3.75mV. If sense is measuring 0 to 120v that’s 1 part of 32000. Sense has a 14bit ADC (7mV resolution)

I measured how far in mm it went in, yes!
I know, probably not necessary at all but I wanted to duplicate as closely as possible how it had been just to remove as ma t variables of differences there could be. Turns out I did lose a detection so I imagine it was because one was a cutler breaker and the other was a square D and their insides were enough different to cause a slight change.
Do I know that for sure? No, I do not. I’m speculating based on everything else being the same.
Voltages through the wire are low. It’s a 20amp breaker on 12/2 Romeo and the heater uses 540 watts or slightly less (uses slightly more when floor is cool).
I’d love to know the real reason if I’m totally off base. If I have to do anything else in the panel I’d hate to have another episode like the last, or worse.
I see what you have for sense resolution so now I’m baffled why this happened.
Share your thoughts please