Will PCs hit the max on normal home circuit?

[lilrayray69]

I've wondered this for a while as I've seen PSUs increasing to higher and higher watts. Your typical home circuit is a 20 Amp which has a max of roughly 1900 Watts. I've seen PSUs now up to 1250W...so I'm just wondering what they'll do if/when PC PSUs go over that?

I guess they'll try to focus on power efficiency to make sure that doesn't happen...and I'm sure people who run multiple high end rigs, like bitcoin miners, probably get their own dedicated circuits...

But do you think PCs will hit that wattage cap any time soon?

 

[Zap]

I think most "typical" home circuits are only 15A. What voltage are you using for your calculations? Normal mains voltage is 120v, give or take as it doesn't seem to stay a fixed number.

Volts * Amps = Watts

120v * 20A = 2400W

120v * 15A = 1800W

Assuming you can load it to exactly 100% and the PSU is exactly 80% efficient, if I'm doing my math correctly...

(1500W /80%) * 100% = 1875W

 

[fffblackmage]

Nope, not gonna happen. If anything, computers are using less power as they're becoming more efficient. You'd have to have an insanely beefy high-end computer with insane overclocking to get anywhere close to maxing out the home power circuit, and even then, the beast barely hit 1500W.

 

[SecurityTheatre]

Nope, not gonna happen. If anything, computers are using less power as they're becoming more efficient.

I would like to dispute this. My old 386 had a heatsink the size of a bottlecap. My 286 did not have a heatsink at all and used 120W power supply.

A Pentium 4 had a heatsink the size of a stack of post-it notes.

Modern chips have one the size of your fist. Until 6 or 8 years ago, many GPUs didn't use active cooling. A passive heatsink was normal. 15 years ago, it was common to have no heatsink at all on the GPU.

This pretty well indicates an increase in power use. Heat dissipation is a pretty good indicator of power use.

 

[Golgatha]

I would like to dispute this. My old 386 had a heatsink the size of a bottlecap. My 286 did not have a heatsink at all and used 120W power supply.

A Pentium 4 had a heatsink the size of a stack of post-it notes.

Modern chips have one the size of your fist. Until 6 or 8 years ago, many GPUs didn't use active cooling. A passive heatsink was normal. 15 years ago, it was common to have no heatsink at all on the GPU.

This pretty well indicates an increase in power use. Heat dissipation is a pretty good indicator of power use.

I'm sure that performance per watt has increased faster than power consumption.

 

[Evadman]

I think most "typical" home circuits are only 15A. What voltage are you using for your calculations? Normal mains voltage is 120v, give or take as it doesn't seem to stay a fixed number.

Volts * Amps = Watts

120v * 20A = 2400W

120v * 15A = 1800W

Assuming you can load it to exactly 100% and the PSU is exactly 80% efficient, if I'm doing my math correctly...

(1500W /80%) * 100% = 1875W

A normal outlet in the US is 15 amps, there are usually a few in a home that are 20 amp (microwave, dryer are 2 popular ones) The maximum sustaned load that may be placed on a circuit is 80% of the rating. So a 15 amp circuit may only pull 12 amps sustained. Pulling more than that for more than a short time should trip the breaker. So max sustained for an entire circuit would be 120*12 or 1440 watts. At 80% efficiency, that leaves 1150 watts for the computer.

Most people have all their computer equipment plugged into one circuit so that 1440 watts has to cover a computer and monitor minimum. Circuit breakers will however allow an inrush current for a very short period of time which is an event with most electric motors. Inrush usually lasts about 10 milliseconds. Most breakers will actually allow over their rating for a while until it heats up and trips. (pulling 16 amps instead of 15 is possible.)

For example, here is the dripping curve for Square D : http://download.schneider-electric....ves&p_File_Id=80048294&p_File_Name=730-01.pdf

Basically, it says that the breaker should trip between 20 and 80 seconds if pulling 150% of the rating (22.5 amps on a 15 amp circuit). Note that at 500 seconds, the trip curve converges with the rating. So if you pull more than the rated 15 amps for 500 seconds, the breaker should trip. There is a bit extra under the curve allowed depending on the breaker, so the breaker may not trip if you pull a bit over 15. looks like about 18 amps on this is the high limit before a breaker is no longer compliant.

Interesting side note. 745 watts is equal to 1 horsepower. That means it is flat out impossible for a vacuum cleaner, shop vac, or anything plugged into a 15 amp outlet to create more than 2 horsepower. Those 120v 5 HP shop vacs are all marketing gimmicks and are based upon the inrush current, which can be 2.5 times higher than normal operation. Sure, by a electrical definition, the motor is pulling 50 amps for a few milliseconds but for everything after that, it is pulling a maximum of 12 amps and creating less than 2 HP, which is an electrical code requirement in the US.

 

[SecurityTheatre]

I'm sure that performance per watt has increased faster than power consumption.

Sure it has. Performance has gone up 100,000x as power consumption went up 4x.

But it's still notable that power consumption of the average home desktop has increased notably during the last few decades. And that was the nature of the question.


And i have tripped breakers running 3 desktops, a laser printer, 6 monitors and a dozen misc peripherals a on the same 15A circuit. But a single machine... unlikely.

 

[lilrayray69]

Yeah I know most outlets are actually 15A, but you can easily get a 20A outlet.

Most people don't have a circuit dedicated to a PC. So say max you can get around 1850W out of the circuit, add in the monitor, lighting, TV, etc...You could hit it fairly quick with a high end system.

I was mostly wondering if your common desktop PC for sale might hit a point that it can't run on a normal users home circuit, given everything else on that circuit. Though sure for now it's not like your average consumers going out and get a Titan SLI desktop system..

I've never tripped the breaker but I have a 650W PSU.

 

[Sheep221]

A normal outlet in the US is 15 amps, there are usually a few in a home that are 20 amp (microwave, dryer are 2 popular ones) The maximum sustaned load that may be placed on a circuit is 80% of the rating. So a 15 amp circuit may only pull 12 amps sustained. Pulling more than that for more than a short time should trip the breaker. So max sustained for an entire circuit would be 120*12 or 1440 watts. At 80% efficiency, that leaves 1150 watts for the computer.

Most people have all their computer equipment plugged into one circuit so that 1440 watts has to cover a computer and monitor minimum. Circuit breakers will however allow an inrush current for a very short period of time which is an event with most electric motors. Inrush usually lasts about 10 milliseconds. Most breakers will actually allow over their rating for a while until it heats up and trips. (pulling 16 amps instead of 15 is possible.)

If breaker is 15 A, than it will trigger at 15 A(+/- small difference for initial peak load on some devices). Breakers which would allow no more than 12A would be useless when the circuit requires the 15 A and probably more for a short periods of time.

 

[frowertr]

While most outlets might be 15 amps, the conductor ran to the outlets was probably 12/3 which you can run 20amps on. Of course, if the OCD (circuit breaker) is only 15 then that would be the max for the circuit. BTW, the only difference between a 15 amp 120 and a 20 amp 120 is the hot blade is turned horizontal. The internals of a 15 receptacle is exactly the same otherwise as a 20 amp receptacle.

One thing we should say is that the wattage on the PSU reflects its output power in DC and not what it pulls from the wall. For example, I have two Litecoin mining rigs. Each rig has four MSI 7950's. I use one Seasonic 1250 PSU in each rig in order to power run them. At max pull from the wall each rig pulls 4.5 amps at 240v which is 1080 watts of energy. If I had that on a 120V 20 amp circuit I'd be eating up half its capacity right there (9 amps * 120 volts = 1080 watts).

But that is a pretty heavy duty load. Normal day to day computers won't come close to that.

 

[fffblackmage]

I would like to dispute this. My old 386 had a heatsink the size of a bottlecap. My 286 did not have a heatsink at all and used 120W power supply.

A Pentium 4 had a heatsink the size of a stack of post-it notes.

Modern chips have one the size of your fist. Until 6 or 8 years ago, many GPUs didn't use active cooling. A passive heatsink was normal. 15 years ago, it was common to have no heatsink at all on the GPU.

This pretty well indicates an increase in power use. Heat dissipation is a pretty good indicator of power use.

Guess I'm too young. I'm recalling the more current progression from Nehalem to IB, and soon, Haswell.

 

[Zorander]

No.

Just can't see PCs exceeding the power usage of air-conditioner and electric outdoor tools.

 

[pandemonium]

I would like to dispute this. My old 386 had a heatsink the size of a bottlecap. My 286 did not have a heatsink at all and used 120W power supply.

A Pentium 4 had a heatsink the size of a stack of post-it notes.

Modern chips have one the size of your fist. Until 6 or 8 years ago, many GPUs didn't use active cooling. A passive heatsink was normal. 15 years ago, it was common to have no heatsink at all on the GPU.

This pretty well indicates an increase in power use. Heat dissipation is a pretty good indicator of power use.

True, however, PCs can do substantially more today than they could 15 years ago. In terms of power requirements, productivity and efficacy have dramatically increased.

One benchmark test would be to run programs from 15 years ago with a modern PC (assuming the legacy software works properly) and see how much power is being drawn. PCs back in the day didn't have much option for throttling or stages. So really, the big performance upgrade here could be when the computer is idle.

As for the OP's original question, no, I don't believe they will.

We're at a major turning point right now, and I'll explain why. Smartphones and tablets are driving sales for highly efficient, passively cooled hardware. What I would guess is that highly-powerful PCs will become extremely niche (it already has become more and more niche every year). In ~10 years, the typical home "PC" will simply be a tablet built into desks or folded out from drawers, walls, tables, chairs, etcetera. From there, in maybe ~25 years, interfaced via holographic displays if there's a push for this type of user interaction. The market for server-based services is becoming increasingly more prolific. Essentially I see small-powered, end-user tablets/displays being driven by server farms via internet connection. Your gadgets, appliances, and even your HVAC will eventually be interconnected via the web to a login-based controller system that we'll be able to switch on/off or adjust by cybernetic implants. Of course, that'll be in the next 75+ years; pending legislation and social adoption rate (we're always afraid of change).

 

[Ayah]

No.

Just can't see PCs exceeding the power usage of air-conditioner and electric outdoor tools.

Think computers before ICs. They were the size of apartments and small houses.

 

[Sheep221]

While most outlets might be 15 amps, the conductor ran to the outlets was probably 12/3 which you can run 20amps on. Of course, if the OCD (circuit breaker) is only 15 then that would be the max for the circuit. BTW, the only difference between a 15 amp 120 and a 20 amp 120 is the hot blade is turned horizontal. The internals of a 15 receptacle is exactly the same otherwise as a 20 amp receptacle.

One thing we should say is that the wattage on the PSU reflects its output power in DC and not what it pulls from the wall. For example, I have two Litecoin mining rigs. Each rig has four MSI 7950's. I use one Seasonic 1250 PSU in each rig in order to power run them. At max pull from the wall each rig pulls 4.5 amps at 240v which is 1080 watts of energy. If I had that on a 120V 20 amp circuit I'd be eating up half its capacity right there (9 amps * 120 volts = 1080 watts).

But that is a pretty heavy duty load. Normal day to day computers won't come close to that.

Yep, the EU electrical system especially allows to draw insane 3700W from one outlet, you just connect some power strips and you can connect dozens of performance computers just like that.

 

[fuzzymath10]

My biggest problem is my printer. My breaker often trips when it comes out of standby. For some reason, it does not trip from a cold state, so unfortunately I have to manually switch it on rather than let it sleep.

 

[imagoon]

If breaker is 15 A, than it will trigger at 15 A(+/- small difference for initial peak load on some devices). Breakers which would allow no more than 12A would be useless when the circuit requires the 15 A and probably more for a short periods of time.

He is correct. 15 amp breakers will trip if you pull over the 80% continuous load. If you need say 14.9 amps for a long period of time, you would need to upgrade to a 20amp breaker and 12 gauge copper.

 

[AdamK47]

I've tripped the circuit breaker with my PC before. Of course a few other things were on as well (TV, set top box, laptop, printer, and some lights). I've also fried poorly connected wires that went to a light in one of the closets on the same circuit. Good thing that didn't start a fire.

 

[nickbits]

While most outlets might be 15 amps, the conductor ran to the outlets was probably 12/3 which you can run 20amps on. Of course, if the OCD (circuit breaker) is only 15 then that would be the max for the circuit. BTW, the only difference between a 15 amp 120 and a 20 amp 120 is the hot blade is turned horizontal. The internals of a 15 receptacle is exactly the same otherwise as a 20 amp receptacle.

I think "probably" is an overstatement. My house, which was built in 2003, sure isn't wired that way. Most everything is 14/2+ground.

 

[imagoon]

I think "probably" is an overstatement. My house, which was built in 2003, sure isn't wired that way. Most everything is 14/2+ground.

Same here,

The only 12/2 in my house is the stuff I pulled to update outlets.

 

[frowertr]

I think "probably" is an overstatement. My house, which was built in 2003, sure isn't wired that way. Most everything is 14/2+ground.

Hence the word "probably". My home was just built 6 months ago and the electrician ran 12/3 everywhere but put in 15 amp breakers for bedrooms and 20 amp breakers for most everything else (bathrooms/living room/microwave/refer/garage/etc...).

 

[Concillian]

My biggest problem is my printer. My breaker often trips when it comes out of standby. For some reason, it does not trip from a cold state, so unfortunately I have to manually switch it on rather than let it sleep.

When laser printers heat up, they use quite a bit of power. Even a modestly sized home laser printer will pull 300-700 watts when it's heating up.

 

[JimmiG]

I would like to dispute this. My old 386 had a heatsink the size of a bottlecap. My 286 did not have a heatsink at all and used 120W power supply.

A Pentium 4 had a heatsink the size of a stack of post-it notes.

Modern chips have one the size of your fist. Until 6 or 8 years ago, many GPUs didn't use active cooling. A passive heatsink was normal. 15 years ago, it was common to have no heatsink at all on the GPU.

This pretty well indicates an increase in power use. Heat dissipation is a pretty good indicator of power use.

High-end GPUs are using more and more power, Titan is at 250W, Radeon 7990 @ 375W

However CPUs are using less and less, at least on the Intel side. It is more difficult to cool a tiny chip with densely packed transistors putting out ~80W of heat than a much larger chip at the same TDP, which is why coolers remain the same size or even get slightly larger.

Those 1000W PSU's are ridiculously overkill. A typical overclocked i7 system with a Titan uses just 400-500W at the wall.

 

[Torn Mind]

PSUs don't "use power" in of themselves expect to run the fan. Power use is completely dependent on how much the components are drawing.

Also, you use a crappier proxy variable(heatsink size, LOL) to approximate power use over TDP. Every Watt eventually does get converted into heat, and the cooling products MUST be able to dissipate however many rated watts on average. The fact is that CPUs rarely go over 140Ws at stock is a sign that CPUs have been "stuck in netural" heat generation-wise since the Pentium 4 era and that is only going down now.

 

[creation portal]

I only 12/2 in my house is the stuff I pulled to outlets.15 amp breakers will trip if you pull over the 80% continuous load. upgrade to a 20amp breaker and 12 gauge copper.it's far cheper than buying conventional type circuit breakes and besides a voltage regultor wont allow you max up your pc unless you configure it befor.