Hey, 2 months ago I added my gas and power meter to HA, and for a while now I looked into identifying electric devices that consume either too much power for what they do, or even when I don’t expect them to at all.
How about we share our findings? Might make it easier for eachother to quicker come to better results.
My “favorite” example: I own a pair of 20-year-old 2.1 speakers who have a physical power switch, which I always had stored under “completely off when switched off”. Measuring my whole PC equipment it turned out that the speakers permanently drew 2.5 Watts, even when the switch was off
Another disappointing one, which I unfortunately cannot change: our gas heater (for heating and warm water) draws a permanent 7 watts of electricity, no matter if the heater does anything or not.
By now I was able to reduce our base consumption to 36 watts, most of it drawn by the above-mentioned gas heater, and the router and access point.
The fridge does indeed periodically kick in, yes, and that’s then way more than the 36. But the fridge is pretty new, so replacing it would cost more than it would save, at least for now.
My HA runs on a Pi3, so it does use a bit, but not too bad. Don’t have the actual number out of my head. The additional ESPs are under the threshold my plugin-in powermeter can detect. Except when the LED ring for the small terrarium kicks in, then it jumps to 1 watt
The heating is part of the rented apartment, so I’ll have to live with that.
And my energy prices are changing per hour and not rarely it can change from something around 0.07kr/kWh to up to 10.00kr/kKh in the same day (specially now days in Europe), so finding the right time for powering things can be way more effective than trying to just turn off devices randomly.
Lights are not the problem… But there are exceptions.
Before I started, I always had a feeling that I should find the perfect automation to turn off my lights as soon as they are not used and that would make a huge difference on my bills. That proved totally false when I started measuring everything. Lights are usually around 10W, while a kettle can easily go to 2000W while working, so trying to use the kettle, or open the oven a bit less will make way more difference than hours or days of lights on.
There are exceptions… I have one Nanoleaf Canvas kit with 17 panels and one Nanoleaf Lines kit with 15 bars. Those are wifi based and while in standby (lights of) they are consuming between 5W to 12W (!). This is more than most of my other lights when on at full brightness. Still far from a kettle or the oven, but the lights are supposed to be connected to power 24/7, so I cannot ignore those.
The towel warmers in the bathroom takes way less energy than I expected. Still taking around 80W while warming, but before measuring I was expecting 10x this, especially after measuring other devices that are heating something.
The home office have a cost. I have a docking station for my laptop and two big screens. Even having those on standby most of the time, my Workstation still the second biggest consumer of energy at home. It requires some work from my side to find ways to reduce this, perhaps powering off (not standby) some devices while not in use.
A smart appliance can pay for itself. As you can see in the chart above, the dishwasher is the biggest consumer at home. I usually run it on the eco mode (which is less than half of the consumption of the express mode), but sometimes, when we have guests, it happens to use the express cycle, sometimes even more than once a day. So, my dishwasher takes energy.
But this is a smart dishwasher and a quite on, allowing my automations to only run it when the electricity is at lower prices, so the dishwasher is far from the most costly device (which is probably my workstation).
I have concentrated all the network devices, Home Assistant (RPi4), printer, etc, but also all the chargers, including wireless portable vacuum, my camera batteries, drone batteries, etc. in a walk in closet of my home office (which was a bedroom originally) and I can measure the total power there. It’s far from negligible, so I invested some time to put all the non critical items (mainly chargers) in a single power strip and this is powered only when the electricity price is at the lowest, which saved me some money and didn’t affected the availability of systems.
Well, I probably have more, but those are the ones in the top of my mind.
How is that working? Do you know the price changes in advance? Or is that a completely volatile process on which you constantly need to keep an (automated) eye?
About the homeoffice/PC topic: the reason I found out about the speakers was, that I wanted to check how much my pc and screen draw when they are shut down, but not physically disconnected via a switchable plug e.g. It turned out that both my screen and the pc drop under the measurable value when I’m not using them. They definitely still draw some power, but with a mechanically switchable plug costing 3-5 euros, it would take 2 years to get those costs back.
I’m using Tibber, which in the end is based on NordPool. Every day at around 13:30 I will get the prices for the following day, so usually when it’s around 13:00PM I know the price for the next 12 hours, when it is around 14:00 I know the prices for the next 35 hours.
I don’t know where to start… Everything seems meaningless when the biggest power consumer draws many kWh and unless you want to melt alive, you need it on (AC). This was July:
I turned the AC down >2-3C in August and totaled 2,720 kWh. We are in October and I am projecting over 2,000 kWh. Last 12 months was 23 MWh (23,000 kWh).
I am about to install current metering for all the 240V appliances (2 x HVAC, Dryer, Oven) to get more resolution in what is driving up consumption but it is pretty obvious it is the AC. I am planning on building ESPHome based devices to monitor AC performance as one year I was well in the 3,000 kWh mid month because the AC was faulty (both coils had to be replaced due to leaks!) and monitoring inlet/outlet temp would have caught the issue sooner.
Side note… the dashboard was not designed for large numbers… maybe we can get rid of the single decimal that doesn’t really matter much once the numbers are past a certain point, no?
@EdwardTFN - I have 2 AC systems, one for each floor. Both have new coils, and one also has a new compressor so I am guessing (hoping) all is fine. They appear to be working properly. The best way to know in the long term is to monitor inlet air temp and humidity and outlet temp and humidity (right before and after the coil in the attic). The delta should be within a certain number of degrees. I am looking for the most appropriate sensor and other parts to accomplish this but may not be done before the cold weather negates the need for AC. I am also about to install an energy monitor that can monitor usage in the 240V 200A panel outside (not the indoor one with all the circuits) so I will see what each 240V appliance (ACs, Oven, Dryer) uses. I am not certain but I believe each AC unit can draw around 5kWh.
Of course, another aspect of saving energy, when it comes to using AC, is to ensure you have a “tight envelope”, meaning that the house is tightly sealed so that hot air from outside is not able to enter the home. Insulation is also important. I have found some weak spots with a thermal camera but can’t do much about it without major work on the house. The easy stuff are seals around exterior doors, windows, etc.
Example of air leaks around front door taken during cold weather:
What surprised me (somewhat) is the power usage of all my smart home stuff. My Pi for HA, my Celeron based NAS, my NVR with a spinning HDD, all the routers, two security cameras (the others are only on when we’re out), the wifi AP, the main powers connected switches and heater control modules, standby TV and network connected satellite receiver, etc etc, add up to about 150W.
While objectively not much, my home office PC alone pulls that when idle after all, but it’s still 1300 kWh a year. At the current electricity prices here (about 0.12€ / kWh) my smart stuff costs me around 160€ per year. Not that bad, but it shows how even all the small things with supposedly ‘insignificant’ power consumption will add up in the end.
Does the art mode really differs from like watching TV? I guess the image changes time to time, but does it reduce the number of LEDs in use or the screen brightness?
Actually a few things surprised me. For example
about 30 W pulled by office standby devices.
Gas boiler which may take 350W
But what actually shocked me is almost 0.7W drained by dumb dishwasher when doing nothing. Dumb washing machine takes “only” 0.5-0.6.
On the other hand consumption of Samsung tv in standby is unmeasurable low.
Also I thought I can remove smart switches from equation m. But not, those missing 3-4W was caused by 8 shellies
Are you sure that is not caused by the startup current of a pump?
Bosch dishwashers have a no-waterleak feature. If the machine leaks it shuts off the supply water at the end of the hose. The inlet hose is a special one with cables inside and the shutoff valve at the end. It needs continuous power to work.
I was surprised how much energy the tumble dryer actually uses.
With little kids we need it but in December and January it was nearly the same amount as our tiny C0 electric car.
We have a pv system and while manually optimising consumption I learnt that the peak power order of big consumer is:
Washing machine 1.5 kW
Car 1.5 kW
Dishwasher 2 kW
Tumble dryer 2.7 kW
I still need to try the sankey chart for real-time power distribution to get further insights
Peak power is important for wires and circuit breaker, not for power consumption. For example, my washing machine uses a lot power (not sure, maybe ~2kW) when heating, but it can finish a typical cycle* just with ~0.3kWh. However, when I increase temperature, it can rise even to something like 1.3kWh.
(Power consumption of my wash machine is reported by itself.)
*) My wash machine recommends using 30°C for ordinary washing instead of 40°C, due to some bubble technology, which reportedly can achieve the same result with lower temperature using some oxygen.
