Ask HN: How much electricity does a WiFi router consume? | Hacker News Reader

Ask HN: How much electricity does a WiFi router consume?

32 pointsvpaulus96 comments

Prices of the electricity in the EU recently became sky high, and people should save as much kWh as possible.

I could not find any related and trustable calculation about how much electricity consumes a Wifi router. Especially I am curious about the difference between the idle mode and when I downloading with full bandwith.

I believe there is no significant difference (and it doesn´t even matter at all, while freezers and heatings are on), but it would be nice to know.

View on HN

My Netgear R6350 wifi access point running OpenWrt 21.02.03 and 5 GHz only uses ~4.5 W in casual use, and goes up to ~6.8 W when I run iperf3 full tilt through it.

Measured using a watt-o-meter like this: http://seacourse.dk/wiki/tiki-index.php?page=Home+Electricit...

They can be borrowed in libraries in Denmark, same as books.

Thank you for the tip about the library! I just put a hold on one at my city's library here in Canada.

> I could not find any related and trustable calculation about how much electricity consumes a Wifi router.

On the botton of your router you probably have a sticker about required power supply. Which should tell you about the router max power consumption.

Very basic wifi routers usually requires 12V 1A, which gives 12W (12 * 1). Typical expected power consumption should be around 25% of this - 6W.

This gives you 4,32kWh / month.

The lowest power WIFI router I've seen was requiring 12V 0,5A, and the most powerful - 12V 3,5A.

> I am curious about the difference between the idle mode and when I downloading with full bandwith.

You need to measure it with watt meter. There are too many factors to just guess it.

Full download via WIFI may not be the most power hungry feature of a router.

Out of curiosity,

a) what heuristic are you using to determine that power consumption will be 25% of the power supply sizing, and

b) (less earnestly) how is 6W 25% of 12W?

  > a) what heuristic are you using to determine that power consumption will be 25% of the power supply sizing, and

I'm not the OP, but it's probably about right. Power supplies are cheap and there are variances in manufacturing. Typically power supplies for all consumer electronics are rated for double peak load. And assuming that standard load is half of peak load is probably on the high end - standard load for the processor(s) is probably more like 1/10 peak load, but there could be other components on the board whose electrical load is not workload dependent.

In particular, a WiFi radio uses a comparatively large amount of power but in short bursts. Gigabit ethernet likewise.

https://www.computerweekly.com/blog/Networks-Generation/Powe...

How much of that can be stored in capacitors to smooth out the peak AC load? Is that done at all?

Basically none, because that isn't how capacitors work. Capacitors voltage drop linearly with respect to the amount of stored energy. They are useful for transients in the very high frequency regime, not multi millisecond long periods.

That depends how much smoothing you want and how much you're prepared to pay for it.

Every AC-DC converter necessarily has some energy storage in it to cover the part of the cycle where the voltage crosses through zero. There's capacitors on the input and output of buck converters. But smoothing over more than one AC cycle? Not really much point in doing that.

Thanks for that. I rarely take much notice of PSU's, other than desktop computer builds, where the received wisdom seems to be aim at using around, or just under, 1/2 of the spec'd capability of the PSU.

Seemingly this is optimal for efficiency of the device, and I've never really thought about how or why this claim is made -- opinions are consistent, pricing is rarely significant, etc.

You can measure it yourself. Wattmeters are cheap, i'm running two of them right now. The laptop im writing from is consuming around 20 watts[1].

Generally i would think most heat-generating appliances (microwave, stove, hot water) are the elephant in the room, energy-wise.

[1] https://pomf2.lain.la/f/e2s025vl.jpeg

Remember, P=IV, Power (watts) is Voltage (v) by Current (a). Things like toasters and kettles tend to use around 13 amps, though voltage differs by country. In the UK, we have 230 V, so that's 2,990 W. In the USA 120 V so 1,560 W. In comparison to something like a laptop that may use 3 A but at 12 VDC it's only 36 W.

One of the best things people can do to save money from energy consumption is optimise how they use the high-power devices -- swap your conventional oven for an air fryer, for example. Trying to optimise the smaller items, as we all here know, rarely returns the investment of time and (human) energy.

On the other hand, remember that energy = power * time. Yes a kettle 13A kettle pulls 2990W in the UK and 1560W in the US... but it would take twice as long to boil the water in the US.

For the energy consumed the voltage is close to irrelevant.

In the hydraulic metaphor, which is easy to reason about for nearly everyone around the world, voltage is the steepness of the water pipe. If one wants to fill a tank faster, it's either a steeper pipe (Volts) or a larger one (Amperes) or both.

A 26 A cable and appliance?

If you buy a mains power meter, it should have a "W" mode rather than make you do maths.

I do wish panel-level monitoring was cheaper, though; does anyone have any suggestions for this?

There are ways to read out your meter readings into e.g. a Raspberry Pi (using the P1 port if you have a smart meter, or with a stick-on sensor)

Do you have a link to more info on this? Thanks.

Here's some documentation from Home Assistant: https://www.home-assistant.io/integrations/#search/dsmr

There's this forum thread on Tweakers.net (Dutch): https://gathering.tweakers.net/forum/list_messages/1578510

There are also plug-and-play devices like https://www.homewizard.com/

Other than that, I think the meter types and data standards differ somewhat between countries, so it's probably best to research how it works in your own country/area.

Oh, here's some other, pretty useful-looking documentation from Home Assistant touching on multiple different scenarios: https://www.home-assistant.io/docs/energy/electricity-grid/

Who runs their home toaster for hours on end each day? Even throwing the thing out and never cooking at home with it is unlikely to save much electrical consumption.

If you live in a net heating environment then the cost of running something like a toaster is basically zero. You're getting to cook with and getting the heat in the inside of your home as well. Not as great as a heat pump, but there is a reason why heat pump toasters don't exist.

>If you live in a net heating environment then the cost of running something like a toaster is basically zero. You're getting to cook with and getting the heat in the inside of your home as well

I'd imagine most people run an extractor fan in their kitchen to get rid of smoke, aerosolised oils, etc. in which case almost all of that heat's lost.

Depends on the home. My parents house wasn't designed with one. My house is designed with one. It took me about a month to understand that it acted like a giant leak in the house. When the wind blows across the exhaust it pulls air in or out. Also, it allows rodents and insects to come in & out as they please. So I just plugged the thing up with some old rags.

>I just plugged the thing up with some old rags

Another option is an anti-backdraft cowl and some insect mesh... if you have an air quality monitor take a look some time at what happens to the air quality in your kitchen when cooking (especially frying) - it generally takes quite a dive, particularly if you have a gas cooker.

indoor CO2 goes from like 500 ppm to 950 ppm. It's awful. I avoid using the natural gas oven and range when I can. I intend to install an outside air exchanger, but haven't got around to it.

Trying to reduce power draw from small appliances like a Wifi router doesn't make any sense. Turn on your oven to bake a pie for half an hour (30min at 2kW), that's 10 days of your router's energy consumption.

If you exclude heating, the main culprits for home electricity usages are : kitchen appliances (stove / oven / microwave / dishwhasher), washing machine, dryer.

Well it depends. If you try to go off-grid/rv and already sorted out heating & food prep then every Watt matters.

A quick and dirty method is to read the voltage and current on the DC adapter. That will give you an idea of the upper limit. For example, an old D-Link router I have has a 5V, 1A DC adapter. So the maximum power the device will use is 5W (P =V * I). Really, there is also usually a 30% to 50% safety margin, so it's probably more like 3W in this case.

Probably depends on what your wifi router is made of and how you are applying the electricity to consume it. I'd imagine a lot.

A few bursts from an ESD gun?

That will certainly disable it, but consume it?

Around 10W I'd say. So assuming your router is on 24 hours a day, you're looking at 10*24=240Wh, which is 0.24kWh.

With the electricity price in the EU maybe reaching EUR 0.2/kWh, you're talking EUR 0.04 a day. Even if the electricity price doubled to EUR 0.4/kWh you would be looking at EUR 0.08 a day.

Maybe reaching 0.02 €/kWh? I have to pay 0.0504 €/kWh right now in Germany. Edit: sorry, my numbers are off by one digit. I have to pay 0.504 €/kWh of course unfortunately

Yes, figured as much. The good news is that your WiFi router is still cheap. The bad news is that 0.50 €/kWh is going to make it an interesting winter.

I am moving to the Netherlands next week and my electric contract says €1.09/kWh…

Replied to the wrong comment earlier! Sorry

Daily variable prices are around €1 per kWh, longer term (usually 3 or 6 months) variable between .7 and .8. Considering the current market expect it to rise more.

Do note: you get quite a lot back in tax returns which are calculated at payment at the end of the month, semester or year depending on the contract type. In most countries these are already deducted in the kWh price so there is a discrepancy there comparing for instance Germany with the Netherlands. Expect around €0.5 to €0.7 depending on the contract in the end.

I'm not going to hold my breath but thanks for the small bit of hope. Right now, Vatenfall is quoting my payment at £400/mo for a new build 89m2 apartment. Considering my 4100/mo salary and rent, it doesn't leave a ton left over to pay for other basic necessities like food, mandatory health insurance, transportation to and from work (20 miles away) and other living expenses.

I leave next Monday, so I'll find out more really soon!

Quite impossible, except if you run a steel mill or something. End-user prices for electricity have been >0.20Eur/kWh for around 10 years now, currently approaching >0.40Eur/kWh (sometimes even more).

See e.g. some comparison website like verivox.de (beware the dark patterns, use "Postleitzahl" 20095 for Hamburg center).

Edit: and those are 2022 prices, usually most suppliers have a 12month fixed price which changes on 1st of January. So all the current price hikes in power and gas will only be priced in in 2023, current offers usually don't include those yet.

> End-user prices for electricity have been >0.20Eur/kWh for around 10 years now

Nah most of europe had it under 20ct per kwh until before covid

https://strom-report.de/medien/electricity-prices-europe-202...

Yes, but the parent post specifically referred to Germany. And due to taxes, renewable subsidies in the range of 3 to 10ct/kWh (which every household pays (but not all industries)) and sky-high generation prices due to prematurely switching off cheaper power plants like nuclear and coal in favour of far more expensive gas (even back then), Germany has had far higher household electricity prices than most of Europe.

You are right, I edited my comment.

Some British businesses have been quoted more like £0.50/kWh. https://twitter.com/GWalker9/status/1566171433780346882/phot...

~20p a unit PLUS ~30p a unit "energy price adjustment".

Just to add another data point for Belgium: my estimated price in November will be 0.74 euro/kWh. This is almost a tripling from my current rate, and the company can't offer a fixed price anymore.

And this is still below the estimated market price of 0.89 euro/kWh, because I'm buying my power from a co-operative that attempts to cover most consumption of its customers with its own wind/solar production.

Is there also a delivery charge for the electric in Europe? e.g. in the US I pay $0.10 per kWh but double it again for the "delivery charge".

A wifi router is not likely going to make a huge difference in your electricity budget. If you want to make a dent the most cost efficient ways are typically replacing lighting with LEDs, and improving the insulation of your house. Heating and cooling typically consumes 30% of your electricity budget.

After that I would take a look at your refrigerator and water heater.

My Unifi U6 APs draws 4.2W from the PoE switch as of writing this (light traffic, but 5-10 IoT devices connected). I'd guess add 10% PSU inefficiency and 10% margin for error, and you're looking at 5W.

A full router obviously draws more, but it depends on your internet connection: Our old DOCSIS3 router (AVM Fritzbox, full router mode+wifi) ran quite hot and is known to have thermal issues. The new place has some shitty DSL (32MBit/s effective speed), and the router I use (older Telekom Speedport in PPoE modem mode) barely gets warm. Didn't measure, but I suppose something like 15-20W vs 5W.

I can only echo the other recommendations: Get a kill-a-watt/powermeter/energiekostenmessgerät. BUT be careful, they are often not very accurate in the low ranges, and don't confuse apparent power and effective power (I know I usually do). You also want one that can show the total energy consumed over a given time frame, e.g. for your fridge or freezer or dryer.

You can also use smart sockets, since some of these have a powermeter as well (same caveats regarding low consumption apply; also these draw another 0.75W). Just make sure to configure fridge/freezer sockets to be always-on to avoid nasty surprises. That's what I'm doing. Since I'm not a fan of cloud-based solutions mine run Tasmota and feed the data into Home Assistant.

Oh, and: Especially if you have an older dryer (non-heat-pump), measure that or outright get a new one. Our heat-pump dryer payed for itself after 3-4 years (comparing advertised consumption pre-buying and the old dryers measured consumption; even if it takes 5-8 years imho it's still worth it). And it can be controlled via WiFi, so theoretically we can sync it with our (planned) photovoltaic installation.

I was interested in that some time ago, as I host a datacenter^H^H^H^H^H^H the upper shelf of a cupboard is the place where I self-host.

I bought a wattmeter and plugged it between the UPS and the power socket. In addition to the UPS there is a 7 years old tower computer ("the main server" - without a GPU but without any optimisation either, and a few disks including platter ones), a PoE 8 ports switch, a RPi and a ER-4 Ubiquity router.

I was really surprised that the draw is a consistent 60 W. This is about 100 € per year in France.

I recommend getting a plug in energy meter. You can get an accurate reading for any device with it.

I'm currently using mine to figure out why were still using 500W at 2 in the morning.

Maybe this wont help you much, but I am currently looking into the shelly 4 pm, which goes directly into the fuse box and can measure the power consumption on 4 fuses (and trigger them)

I have been wondering about that. Are these meters built on the assumption of a purely resistive load, or will they give accurate readings even with a switch mode power supply? My very limited understanding says these draw current in a very different pattern than a resistive load. The meter would actually have to integrate voltage×current to be accurate, I think.

Most of them report power factor, which requires measurement of reactive loads.

Thanks, that is helpful to know. However, a switch-mode power supply is quite different from a reactive load. Cheap ones tend to draw power only near the voltage peaks, whereas better ones improve on that. See [1] and [2]. So the question remains how well these cheap power meters cope with such a nonlinear load.

[1] https://en.wikipedia.org/wiki/Switched-mode_power_supply#Pow...

[2] https://en.wikipedia.org/wiki/Power_factor#Power_factor_corr...

I picked up a power / energy meter a couple of years ago, when we put on solar panels, and spent a good couple of weeks putting it inline to various appliances to identify peak draw or 24h draw (depending on the type of appliance).

They're about AUD$30, and highly informative devices - excellent for answering this kind of question.

About 10 Watts or 7-8 kWh per month, give or take.

If you have electric heating, you need heating, you have a thermostat, and the router is located reasonably close to where you need the warmth, it will cost you nothing. If all these conditions are met, the same electricity is producing routing and heating and nothing is wasted.

Only if you have resistive electric heating. If you have a heat pump, you lose out on efficiency (remember that for every J you put into a heat pump, you get >1 J of heating out!)

I've often thought that electric heaters were rather wasteful, when one could get some processing for "free" (in energy usage terms), and still heat the room.

Mine bitcoins in winter.

Heat pumps are more efficient than space heaters though

I use Mi WiFi Range Extender AC1200 (on AP mode), which consumes 0.3 A (the label only list the ampere instead of watt). I live on 220V AC country, so:

- 220V * 0.3A = 66W

- google "how many hour in a month" = 730 hours

- 66 W / 1000 W * 730 hours = 48.18 kWh / month

I am a bit skeptical about the power that this router use, should be just using around 7W instead of 66W, I am thinking of buying a power meter to get the real power usage. and will change my router if this router really consumes too much electricity

That sounds like way too much. It's more likely it makes use of DC power at 5V or 12V: the fact it has an integrated adapter from AC to DC only means it's harder to figure this out.

A first AC/DC adapter I grabbed has an input of 0.1A at 230V and output of 0.5A at 10V. If your device has got a similar ratio (there is nothing to say it does, though :)), you are looking at 3A at 5V or 15W max.

Agree, listing only 0.1 A without saying at what voltage is is not informative. Recently bought a power meter and it says it actually only consumes 2.4W. I posted my correction on the replies above

Here's my correction. My power meter has just arrived, it says my wifi router only consume 3.0 W at start and then drops to 2.4 W, so:

- google "how many hour in a month" = 730 hours

- 2.4 W / 1000 W * 730 hours = 1.75 kWh / month

10-15W for a wifi router with a cable modem, less if it's only wifi, but that's the general ballpark (this is measured with a calibrated smart-plug)

It vary quite a bit. I have seen them go from 6w (ubnt aplifi hd) to 24w (A asus router, forgot the specific model) in idle.

A cursory google for "wifi router kWh" comes up with 7.3kwh in a month, from energysage.com. Does that suit?

You should probably not simply take the values on the box. Just like a computer, routers also have idling/high processing periods that affect power needs drastically. It's a good idea to take some smart plugs (like TP-Link's HS110, also supported by Home Assistant) to make these measurements.

I was curious about this as well when I purchased a UPS that shows the wattage used in real time. I connected only my cable modem and 1st Gen Google WiFi router to the UPS and it registered as between 1 and 2 watts. So, it seems like it's a fairly small amount.

Rule of thumb: if it gets hot it burns power. More hot, more power.

Be more judicious with electric kettles, immersion heaters, electric ovens, lower washing machine temps, outside air drying over electric dryer, LED bulbs.

The other factors are too small to worry about.

Well that depends on how big your area is and even if your router is set to eco mode.

I suggest, if you have an appartment, by reducing the signal strength first, which will consume less power.

But in the end of the day, you will need to measure it.

The actual WiFi radio signal only uses about 1/10th to 1/30th of the total energy consumption, if you assume that the device as a whole uses around 10W.

There's not a whole lot to save by turning down the signal strength.

I am moving to the Netherlands next week and my electric contract says €1.09/kWh…

Edit wrong reply but I’ll leave it

at that price you might as well invest into solar balconies

If the power usage is correct on my UPS, my whole setup draws about 70W for: - USG Pro 4 - Cloud Key gen 2 - Two basic POE Switches - 4 access point - 5 cameras - 1 raspberry pi - 1 VSDL Modem

> and people should save as much kWh as possible.

What really matters is cutting peak hour consumption, not total kwh. At least around here during night-time electricity is almost free compared to peak hours

My router uses 8 Watt from the mains, regardless of how much traffic it's handling. It also has an ECO mode that doesn't actually lower the power consumption.

Heating, using electricity to heat anything is a mild environmental crime. For rest of the thing, all appliances are efficient enough and still improving.

Heat pumps are quite efficient . So it’s probably still a lesser crime than burning stuff to heat something.

It wouldn't take much, pretty sure any mainline AC voltage direct to a router would consume it in a fraction of a second. Hope this helps :)

I think vaporising it totally would take multiple kilowatts.

Anything with a wall wart will have negligible current draw. You're way better off focusing on appliances and heating/cooling.

Tri-band LTE router with WiFi 2.4 only takes 4W, under speedtest 6W. Enabling WiFi 5GHz adds extra 1W.

"Prices of the electricity in the EU recently became sky high"

I will nominate for Nobel Prize of Economics politicians that voted for the mapping of electricity on the most expensive source (now gaz).

This is an EU reform trying to make an electricity market.

I mean, elecricity is fungible, and demand is relatively inelastic. Why on earth would a solar farm sell it to me at a much lower price then the gas powered plant?

Often the markets for renewables, especially wind, are set up like futures markets so the price the generator receives is less elastic. e.g. Contracts For Difference: https://www.emrsettlement.co.uk/about-emr/contracts-for-diff...

Normally this operates as a subsidy, but now it's returning money to the taxpayer: https://www.current-news.co.uk/news/cfd-costs-to-be-paid-bac...

Gas should have been phased out a long time ago, for geopolitical and environmental reasons.

It's much better than coal. Most places don't have enough hydro, biogas or geothermal, and nuclear ran out of favour. There is currently nothing else that can deliver a base load.

For sure. I understand perfectly well what led up to this. Nuclear also ran out of rivers to cool them in France. Now we will see a fast transition instead of slow and steady transition away from gas. It will be "interesting".

A dependency on russian gas should never have been established.

Coz the solar farm cant match supply to demand whereas a gas plant can.

Yes, that's what GP said (or what i understood).

Basically the cost of electricity should be the average cost of the network's generators. Or, as it is hard to calculate, the operating cost of the most expensive generator we have to put online. Peakers (pulsed coal and gas) are the most expensive to operate, so the price of electricity is based on their costs.

That we use gas and coal for baseload, however, is the dumbest thing EU (and German) legislation put us through.

And so? I need a KW of power for the next hour, I'll buy from whoever will sell it to me at the cheapest price. Why would anyone sell it to me at less than the highest price (gas) - 0.01€? As long as demand is inelastic I believe that it's unavoidalbe that price will settle up to the price of the most expensive unit sold.

Reply on news.ycombinator.com