Turning off vs Sleep mode?

[rob4001]

Which consumes less power in the course off a year. Turning on & off vs sleep mode?

[jim]

I don't know how pronounced the difference is: it might be so extreme that the answer is a definite.

However, the frequency with which you turn on and off would be fairly instrumental. If you only turn it on once a year, sleep mode is unlikely to be all that efficient. Multiple times per day, it probably becomes more useful.

[kalniel]

Which consumes less power in the course off a year. Turning on & off vs sleep mode?

If you're comparing on & off vs on & sleep then the former consumes less power, obviously. But if your computer takes so long to boot that you're having to leave it on more than you would otherwise then it's less clear cut.

[Willzzz]

SSDs will save the planet!

Can I have a bigger car now?

[watercooled]

Modern sleep states use barely more power than full off, in some cases I've measured the difference as less than a Watt. But as above, it really depends on your usage of the system.

Personally I shut down, as Windows on an SSD takes a few seconds to boot, and Windows can start to behave strangely if its not shut down once in a while, although that's probably improved since the XP days where I used to use hibernation a lot.

[Funkstar]

There is more to think about here as well. You have on, sleep, hibernate, PC off and off at the wall.

That would generally be a list in order of energy usage. I wouldn't be surprised if there is little difference between sleep, hibernate and PC off on many modern systems. But you may well see quite a difference if you actually flip the wall switch. An ATX PSU/system can suck quite a lot of power when the PC is off. Good example, my Spider Elite is always on, even when my system is off, which means USB is still being powered all the time.

[watercooled]

Yeah and it can vary a fair bit between systems. Older systems tend to keep some parts of the motherboard and USB devices powered with the 5vsb rail, even when the system is 'off'. However, I've noticed newer systems tend to go into a deeper 'off' state after a few seconds, lowering power consumption further and powering down USB devices, dependant on BIOS settings (features like on/off USB charging may bypass the lower power modes).

For example, my Llano system draws roughly 3W AC when in sleep mode and for a short while after shutting down. After a few seconds of being shut down though, my power meter flicks between 0 and 1W (obviously it's not going to actually be zero, but the meter doesn't have enough resolution to show decimals).

[jim]

Modern sleep states use barely more power than full off, in some cases I've measured the difference as less than a Watt. But as above, it really depends on your usage of the system.

Personally I shut down, as Windows on an SSD takes a few seconds to boot, and Windows can start to behave strangely if its not shut down once in a while, although that's probably improved since the XP days where I used to use hibernation a lot.

Mine stays on almost permanently, and I've never noticed any problems. Mind you, I don't use sleep either.

There is more to think about here as well. You have on, sleep, hibernate, PC off and off at the wall.

That would generally be a list in order of energy usage. I wouldn't be surprised if there is little difference between sleep, hibernate and PC off on many modern systems. But you may well see quite a difference if you actually flip the wall switch. An ATX PSU/system can suck quite a lot of power when the PC is off. Good example, my Spider Elite is always on, even when my system is off, which means USB is still being powered all the time.

Presumably though, you are more likely to drain the mobo battery. Not that it's a bit deal in the great scheme of things, changing one battery every few years vs. spending money on constant electrical current.

[Saracen]

This might be a good time to advocate the acquisition of power meters. There was a thread a year or two back, started by Zak IIRC, and it was interesting.

Rob, the only way to be reasonably sure which mode uses what is to meter it on your PC.

I have a plug-through meter that can tell me what any particular device (that's plugged in via it, obviously) is using at any given time, but I also gave a meter permanently on the house mains (a sensor round the output side of the mains meter) and a radio sensor giving me a constant readout of household usage. Watching what items use what power has chanfed my usage habits a fair bit. Some devices use a watt or two in standby, and others use almost as much in standby as they do fully powered up.

Those meters will give you a "right now" reading, or a cumulative reading. So, zero the totals on the plug-thtough meter, contact your PC via it, leave it running in sleep mode overnight for a given time period, and note the cumulative power used.

The next night, repeat the exercise but with the PC "off".

Then compare readings.

Oh, and the plug-through meter cost me about £8 or £9, and the main household RF meter was more like £40. They've paid for themselves over the course of a couple of years by telling me what to keep turned off, and what to not worry about .... and what devices to target for replacement first. One old freezer bit the dust, as I didn't really need it and it was sucking down about 200w, 24 hours a day.

[kalniel]

Presumably though, you are more likely to drain the mobo battery.

Isn't drain on the battery the same regardless? (ie the battery powered bits are always battery powered)

[jim]

Isn't drain on the battery the same regardless? (ie the battery powered bits are always battery powered)

I've always thought that they drew power from the socket where possible, but I've no idea why. Could be wrong.

[watercooled]

I'd take into account the precision of meters even when you're measuring over time; the internal resolution might be higher than the displayed one but you might be measuring to the nearest Watt, provided it's still accurate down in the low single digits.

I think I've mentioned it before, but those clamp-on whole-house meters can be incredibly inaccurate, especially for lower PF loads, generally reading far too high.

[Saracen]

I'd take into account the precision of meters even when you're measuring over time; the internal resolution might be higher than the displayed one but you might be measuring to the nearest Watt, provided it's still accurate down in the low single digits.

I think I've mentioned it before, but those clamp-on whole-house meters can be incredibly inaccurate, especially for lower PF loads, generally reading far too high.

Oh, absolutely.

I wouldn't take them as overly accurate, and an electrical engineeer friend tested mine with some kind on bench PSU and a dummy load and told me "plus or minus about 2%".

I also take the point about dubious linearity, especially at very low levels.

But .... the comparison between readings on two devices can be illuminating.

If one says 3w but is actually 2w or 5w, it's not really the point if the other device reads 30w.
Or, if it's one device in two different states. In other words, I don't trust the absolute reading to be hugely accurate, especially at low levels, but the relative readings is infornative.

For reference, though, if I have a 'standard' load if about 250w, and I turn on two 60w incandescent light bulbs, the reading immediately switches to about 360w, I.e. a 110w increase from a nominal 120w increase in load. One issue is that the meter currently assumes a stable 240v. Hmmmm.


Edit - oh, and a dimmer switch in-circuit, set to zero dimming, with that light bulb test, but presumably, still involving a small loss.

[Willzzz]

Oh, and the plug-through meter cost me about £8 or £9, and the main household RF meter was more like £40. They've paid for themselves over the course of a couple of years by telling me what to keep turned off, and what to not worry about .... and what devices to target for replacement first. One old freezer bit the dust, as I didn't really need it and it was sucking down about 200w, 24 hours a day.

Do you have links for these devices? I have had a look myself, but can't seem to find what I am looking for :/

[kalniel]

I've always thought that they drew power from the socket where possible, but I've no idea why. Could be wrong.

Or could be right It's only powering the CMOS chip which is basically a clock + some memory with BIOS settings so I assumed it wasn't worth the expense of control logic/power transforming down to 3V when a cheap battery will last the market lifetime of the computer, but then Occam's razor isn't always accurate

[kalniel]

Do you have links for these devices? I have had a look myself, but can't seem to find what I am looking for :/

http://www.amazon.co.uk/Energenie-EN...ef=pd_cp_diy_0

No recommendation for that particular product - was just result of a quick search. 'wall plug energy monitor' should do.