• # Real power consumption of an ESP8266 on a net adapter

Hi,

Did anyone of you ever really measured power consumption of an ESP8266 on a 5V net adapter? Even with or without deep sleep options in place?

I have seen many posts on the web mentioning something like that the average power consumption of an ESP8266 is around 60 mA.
60 mA * 5V ≈ 0,3 Watt and therefore you don’t even have to think of your energy bill. And that figure is of course negligible on yearly basis.

However, I have always learnt that power consumption of an adapter/transformator based device is NOT determined by the power consumption of the device, but by the power consumption of the net adapter.
For my esps I have several net adapters ranging from 500mA output to 2.3A output.
However, the minimal power consumption is always 100 mA according to the sticker on the plug.
It says something like “Input 100 - 240 V, 100mA - 300 mA; Output 5V 2.3A”

This means that even in minimal power consumption that net adapter is using 230 V * 100 mA = 23W
On year basis that is ≈ 250 kWh and for a kWh price of approx. 20 cents (Netherlands), that is 50 Euros!!
I see posts where users proudly mention to have 10 to 26 (!) Esps in their house all running 24 hours a day.

I don’t have such an energy meter which you put in the wall outlet and where you can put the power plug of your device in, but I’m planning to get me one. To measure Esps (1) on net adapter with deep sleep, and 2 constantly on.

So did anyone really measure and can he/she share his/her results?
(not battery results, but net adapter results!)

• ESP8266-01 with 2 relais and a cheap chinese 500ma 5v Adapter:
Both relais off = 12mA = 2,76W
Both relais on = 22mA = 5,06W

Gruß M.
Edit:
I forgot to say that the measured mA are on the 230V side!

• It says something like “Input 100 - 240 V, 100mA - 2.3 A; Output 5V 2.3A”

This means that even in minimal power consumption that net adapter is using 230 V * 100 mA = 23W

I’m not an professinal electrician, but a can’t belive that even a creepy china usb adapter will burn 23W doing nothing.
Have you ever touched a 30W bulb? It would not burn your finger, but it’s warm. And i don’t thing that your net adapter is getting hot like a 30W bulb!
I think the label is wrong. If you calculate the other side 240V @ 2.3A = 550W means that you could start fry your breakfest eggs cause of the heat.
I would say that it means that 100mA @ 5V is consumed for transfer 230 -> 5V. This is 0,5W
(0,5W * 24hours * 365 days) / 1000 = 4,38kWh * 0,20€ = 4,38€

With full load (2,3A @ 5V = 11,5W)
(11,5W * 24hours) * 365days / 1000 = 100kWh * 0,20€ = 20,15€

I’m not an electrician either. I’m simply reading what’s on my net adapter.
That’s why I ask whether people really measured, not calculated (like you also do). I did see from several articles on the web that those cheap chinese net adapters have efficiencies of 50-70%. That makes me even more curious.
Two example: http://www.righto.com/2012/10/a-dozen-usb-chargers-in-lab-apple-is.html
and
http://www.lygte-info.dk/info/usbPowerSupplyTest UK.html
But there are more articles to be found on the web.

I bought an “Energy consumption monitor” during lunch and I will check my ESP8266’s, and my raspberry pi 3, bananapi and RPi B+.
And I will report back.

• @Harry-van-der-Wolf

What you can do:

• plan your home automation system with care. Don’t measure stuff where you don’t need it, really. Each data point comes at a price and energy consumption makes a significant part of the overall bill.
• buy quality power supplies with Energy Star Label rather than cheap wall warts
• use power supplies with a suitable supply range, i.e. don’t use a 2A supply for a 100mA draw. Most supplies reach the optimum efficiency factor at a certain power draw. As a rule of thumb I’d say the maxium draw of the device should be 60-70% of the maximum supply.
• switch off/unplug stuff you don’t really need

http://www.elektroniknet.de/elektronik/power/gesetzliche-regelungen-fuer-leerlaufleistung-121618.html

"It always takes longer than you expect, even when you take into account Hofstadter's Law.", Hofstadter's Law

• You are reading correctly, but interpreting it wrong.
The 100mA is on the 240V side… have to go now, i’ll explain it further later on.

• I did 2 approx. 8 hour measurements now.

1 nodemcu 1.0 on a 600 mA adapter.
The nodemcu holds 2 ds18b20 temperature sensors (vreezer & cooler) and a PIR motion sensor.
During 7 hours the constant power consumption was 0.8 Watts.
On a yearly basis that is ≈ 7 kWh for 0,20 Euro per kWh (Dutch rates) ≈ 1,40 Euros.

1 nodemcu 0.9 on a 500 mA adapter for 9 hours.
This nodemcu holds one Keyes relais to control my central heating.
relay off ≈ 0,7 W
relay on ≈ 1,2 W
Extrapolated and calculated with the “night off” period and a normal day on/of schedule at 1-4 °C outside temperature, I come to an average of 0.9 W.
On a yearly basis that would be ≈ 1,75 Euro (Dutch rates).

All not very much and nothing to worry about.

@Rene-Arts : I know that the 100 mA is on the 230 V side. That’s why I calculated the 230 V * 100 mA = 23W. Fortunately it is using far less.
I have about 15 loaders in my house: for my esp8266 nodemcus, my banapi, my pimatic RPi3, my 2 RPi B+ (media streamers; not always on), our phones, my 2 photo cameras, and some more.
All have slightly different specs typed on the adapter. Some require less mA on the 230 V side while delivering more output on the 5V side than others. All my adapters are good quality adapters as I also know from 2 cheap “Action” adapters that they produce a very high annoying beep.

Now see what my RPi3 does simultaniously running pimatic and my video monitoring system motion using a USB cam also doing subsequent video conversion on 720p captured content and rsyncing that to my webserver.

• ESP8266-01 with 2 relais and a cheap chinese 500ma 5v Adapter:
Both relais off = 12mA = 2,76W
Both relais on = 22mA = 5,06W

Gruß M.
Edit:
I forgot to say that the measured mA are on the 230V side!

Strange though that your measurements are much higher than mine. Is it the cheap chinese adapter? the relays? the used “energy consumption monitor”?
However, even a continuos 5 W power consumption would lead to ≈ 9,5 Euros per year (Dutch rates).

Still curious what others will measure (if they do).

• However, even a continuos 5 W power consumption would lead to ≈ 9,5 Euros per year (Dutch rates).

At the rate 0.2 Eur/kWh - you pay about 8.76 Eur per year, if device run nonstop.

With Power supply efficiency of about 75%. The account will be one esp about 0.87 Euros per Year.

Pimatic = Smart Home

• Real measurements are mostly better than calculations. Problem in this case is that most power meters are very inaccurate at low power consumptions (say, 50% error or more around 0.5/1W).
There is a nice comparison between different power meters: https://nl.hardware.info/reviews/1460/6/vergelijkingstest-9-energiemeters-tabel-met-testresulaten
Here you can see that some even cannot measure below 5W, others are very inaccurate.
I have a Brennenstuhl PM231E, which is not very expensive though quite accurate at low power (if not dead on accurate it’s at least an indication). I bought it at Conrad.

About the 100-300mA: that is the maximum drawn current @100-230VAC. So an indication for the efficiency at full load would be:
Input: 230VAC * 0.1A = 23W or 100VAC * 0.3A = 30W (lower input voltage might give a lower conversion efficiency).
Output: 5VDC * 2.3A = 11.5W
11.5W/23W gives around 50% efficiency… Not very good, though not too bad on full load either.

I did some measurements with the HKLM 230VAC-5/3.3VDC supplies, they are quite ok (also see reviews on the internet). Most switching supplies use up to 100mA when not used and efficiency is reasonable. Nothing to really worry about as long as the devices are useful
Currently I’m only buying well built supplies as I am more concerned with safety (fire/shock) than price. Added benefit is that both power factor and efficiency are mostly better as well.

• There are many on the web.

At the rate 0.2 Eur/kWh - you pay about 8.76 Eur per year, if device run nonstop.

With Power supply efficiency of about 75%. The account will be one esp about 0.87 Euros per Year.

I do not exactly understand what you mean here. The efficiciency of the adapter or the power savings onthe esp?
If I measure 0,8 W continously over 7 hours I see no power efficiency on the esp.

• Real measurements are mostly better than calculations. Problem in this case is that most power meters are very inaccurate at low power consumptions (say, 50% error or more around 0.5/1W).
There is a nice comparison between different power meters: https://nl.hardware.info/reviews/1460/6/vergelijkingstest-9-energiemeters-tabel-met-testresulaten
Here you can see that some even cannot measure below 5W, others are very inaccurate.

I don’t have one of them. I have a Koenig now.
The inaccuracy is also what bothers me. That is also why I’m questioning the differences in @Michael-Rudek measurements and mine.
And that’s also why I mention that in case his measurements are correct, a continuous 5W consumption leading to 9,5 Euro/year, is still acceptable.

• 5W would be quite something, but indeed, depending on the saving they realize it is quite acceptable. Don’t forget most relays have both a Normally Open and Normally Closed output: if the relay is used to only now and then switch something off then you should connect it through the NC output. Then you only need to actuate it while it needs to be opened.

• It indeed depends on what you use it for whether the costs are acceptible.
As mentioned: one of my nodemcus has 2 ds18b20s connected to measure my freezer and cooler. If my (vertical door) freezer is not correctly closed and everything melts I have a way more expensive bill than that 10 euros a year, apart from all the mess.
It already happened once and that is why I have installed that esp. (cooler is not so important)

• This cheap power meters are OK if you want to know what your Dishwasher, Washing machine etc. is consuming.
They are using hallsensors which are not very accurate.

If you want to measure small devices like this small 5v Plugs, it’s better to use a Multimeter “inline”.
The measurements i made in the second post are made with a calibrated Gossen-Metrawatt Multimeter.
By the way, I’m electrician
Or in German language : Energieanlagenelektronikerfachrichtungbetriebstechnik

• 5W would be quite something, but indeed, depending on the saving they realize it is quite acceptable.

You could use latching relais (aka bistable or impulse relais) (Stromstoßschalter). This way, you’ll get a more efficient design regardless of the switch state.

I think there are ready to use modules available http://wiki.seeed.cc/Grove-2-Coil_Latching_Relay/ or you can try to roll your own https://forum.arduino.cc/index.php?topic=362423.0

"It always takes longer than you expect, even when you take into account Hofstadter's Law.", Hofstadter's Law

• The 5 Watt figure is purely hypothetical. Most relays wont consume anything close to this. I guess most need 10-100mA at 5V to stay actuated. Nothing to worry about.

Posts 17Views 20611

Looks like your connection to pimatic forum was lost, please wait while we try to reconnect.