Naked Science Forum
Non Life Sciences => Technology => Topic started by: JOHN Gamel on 05/08/2010 08:30:06
-
JOHN Gamel asked the Naked Scientists:
Doctor Chris, here's a question for you:
Â
When I put my computer to "sleep," does this consume more energy than if I turned it off?
Â
John Gamel
What do you think?
-
I don't have the exact numbers...
in terms of power consumption, off means no consumption, sleep means very little consumption. That's for sure.
When your computer is in sleep, your CPU, harddisk, graphics card, I/O, etc, are all basically off and does not consume much, if at all, power.
The state of the computer is essentially stored in the RAM, which runs at approx. 15W/Gb. There is also some overhead associated with the motherboard, power supply and some spinning fans.
Reference here: http://www.pcpowerandcooling.com/maxpc/cases.htm
You can see how much you are saving by putting the computer into sleep.
Note: figures depends not only on the hardware specs, but also usage. If you are not crunching numbers, your CPU does not use the peak power figure.
-
I don't know that how to consume energy in sleep mode and how to cheak Online Psychic (http://Online Psychic)
OBVIOUSLY NOT A VERY GOOD PSYCHIC BECAUSE YOU DIDN'T FORESEE THE BAN THAT HAS BEEN IMPOSED ON YOUR ACCOUNT FOR POSTING SPAM.
-
JOHN Gamel asked the Naked Scientists:
Doctor Chris, here's a question for you:
When I put my computer to "sleep," does this consume more energy than if I turned it off?
John Gamel
What do you think?
.
Cooling needs to be on or monitored and maintained, harddrive motors are off, video functions are off no lighting, but the motherboard's chips set bias voltage must always be present in sleep or standby the minimum current draw here to. The wattage drawn is now down to a minimum.
I believe that the CPU needs some functions to maintain an index of the users last known interface of operation, so when awaken, it can place everything that was, is now the same, to the users.
So at least the Fetch, Store, functions need to be always active and it should not stop swapping or sharing capacitive memory charge, otherwise everything locks up. "CRASH"
There may be some other function in the internals of the CPU that work in conjunction also to make this happen.
Think of it as an Automobile sitting at an idle, there are still moving parts, it still uses petro, but not much to maintain the idle, until it is awakened.
I found some figures here with a know resource.
"EnergyStar
To receive the EPA's EnergyStar rating desktop computers must use two watts or less when in standby mode and four watts or less when in sleep mode. Laptops must use one watt or less when in standby and 1.7 watts or less, when in sleep mode. On-mode EnergyStar ratings are based on a formula that calculates watts with a monitor's megapixels."
Read more: How Many Watts Does a Computer Use? | eHow.com
http://www.ehow.com/facts_5084931_many-watts-computer-use.html
-
JOHN Gamel asked the Naked Scientists:
Doctor Chris, here's a question for you:
When I put my computer to "sleep," does this consume more energy than if I turned it off?
John Gamel
What do you think?
I think in sleep mode it is very less consumption of energy.
-
This is a good question.
The components that are operational will be the static components, such as RAM.
I've been running a heavily overclocked watercooled PC for a number of years, that's almost on 24/7 I've got a power meter here, I'll plug it into the socket of the case and report back, I think sleep power consumption will be dependant of the motherboard and possibly, if used an external program
-
This is a good question.
The components that are operational will be the static components, such as RAM.
I've been running a heavily overclocked watercooled PC for a number of years, that's almost on 24/7 I've got a power meter here, I'll plug it into the socket of the case and report back, I think sleep power consumption will be dependant of the motherboard and possibly, if used an external program
RAM isn't that Random Access Memory, I always thought it was dynamic.
And ROM is Read only Memory which was explained to me was static.
-
RAM is indeed reliant on a continuous (if small) trickle of current to retain its contents.
I would estimate that the largest power requirement of any single PC component is the power supply itself.
Modern switch-mode PSU's are a good deal more efficient than earlier types, but inefficiencies do not scale down linearly for current drawn.
That is, heat dissipation per Watt delivered will be worse for 30 Watts on a 300W PSU than compared with 240W delivered.
-
RAM is indeed reliant on a continuous (if small) trickle of current to retain its contents.
I would estimate that the largest power requirement of any single PC component is the power supply itself.
Modern switch-mode PSU's are a good deal more efficient than earlier types, but inefficiencies do not scale down linearly for current drawn.
That is, heat dissipation per Watt delivered will be worse for 30 Watts on a 300W PSU than compared with 240W delivered.
Stepping aside from the original question...
Addressing your statement peppercorn...
Looking at Modern Switch mode PSU concept, compared to (early PSU) that you are referring to linear PSUs.
I can not remember when any linear supplies were ever used in a consumer desk top PC.
I have to ask because I am a bit confused:
Power is power, temperature is temperature, and it is a direct relationship between each other though unit conversion.
What I interpret from what you are saying, is that a Switcher (SMPSU) is working harder and dissipating more heat at 30 watts being drawn on a SMPSU than 240 watts drawn on the same SMPSU?
-
What I interpret from what you are saying, is that a Switcher (SMPSU) is working harder and dissipating more heat at 30 watts being drawn on a SMPSU than 240 watts drawn on the same SMPSU?
No. The key phase here is "Heat dissipation per Watt delivered"...
Heat dissipation per Watt delivered will be worse for 30 Watts on a 300W PSU than compared with 240W delivered.
By heat dissipated I mean main electricity converted (wastefully) to heat inside the PSU. By Watts-delivered I mean delivered to the PC's electronics at the standard voltages (3.3v,5v,12v,etc).
As a ratio the 30W scenario will be higher (worse) than the 240W scenario.
Obviously, (as you imply) in pure BTU's at 240W the PSU will be pumping out several times more, but that was not my point.
Looking at Modern Switch mode PSU concept, compared to (early PSU) that you are referring to linear PSUs. I can not remember when any linear supplies were ever used in a consumer desk top PC.
Admittedly, I can not remember that time either. This pdf (http://www.ece.msstate.edu/courses/design/2008/pwrsupply/Deliverables/Microsoft%20Word%20-%20problemstatement_final.pdf) gives the date for the first SM-PSU as circa 1972, somewhat earlier than I had assumed.
It's possible that linear PSUs were in service for a number of years after the birth of the x86 series though, due to issues with noise inherent with SM.
-
from: peppercorn on Today at 03:07:44 pm
Heat dissipation per Watt delivered will be worse for 30 Watts on a 300W PSU than compared with 240W delivered.
By heat dissipated I mean main electricity converted (wastefully) to heat inside the PSU. By Watts-delivered I mean delivered to the PC's electronics at the standard voltages (3.3v,5v,12v,etc).
As a ratio the 30W scenario will be higher (worse) than the 240W scenario.
Obviously, (as you imply) in pure BTU's at 240W the PSU will be pumping out several times more, but that was not my point.
Evaluating the ratio between heat disipated per watt drawn or delivered, just running the linear or Switcher without a load, would disipate heat. So as you load either PSU down, this ratio at the start will weigh like you say and change exponentially as the load increases?
This wasn't the point of the original question, it is your point? That is where I was confused. Sorry.
Admittedly, I can not remember that time either. This pdf gives the date for the first SM-PSU as circa 1972, somewhat earlier than I had assumed.
It's possible that linear PSUs were in service for a number of years after the birth of the x86 series though, due to issues with noise inherent with SM.
When you mention x86 series are you addressing the CPU series or the developed consumer home IBM PC where intel and microsoft had the 8088 cpu and the developed consumer home PC?
A linear power supply at 100 watts has a heavy stepdown transformer, I know that would not be in a home PC desktop.
-
Evaluating the ratio between heat disipated per watt drawn or delivered, just running the linear or Switcher without a load, would disipate heat. So as you load either PSU down, this ratio at the start will weigh like you say and change exponentially as the load increases?
This wasn't the point of the original question, it is your point? That is where I was confused. Sorry.
http://en.wikipedia.org/wiki/Power_supply_unit_(computer)#Energy_efficiency
"The energy efficiency of power supplies drops significantly at low loads. Efficiency generally peaks at about 50-75% load."
- Hence "higher waste heat per kW delivered".
Look at the numbers on the 80plus site that should give an idea of the efficiency curve (zero load to full load).
I think this is pertinent to the original post - to make the distinction between peak efficiency and low load efficiency.
When you mention x86 series are you addressing the CPU series or the developed consumer home IBM PC where intel and microsoft had the 8088 cpu and the developed consumer home PC?
A linear power supply at 100 watts has a heavy stepdown transformer, I know that would not be in a home PC desktop.
I had made the assumption that when talking about computers in sleep mode, we are discussing desktop PCs, but of course this could be misleading.
As you mentioned the 8088 was the CPU (forerunner to the x86's) used in the first IBM standard 'pc' circa 1981: I've just looked it up and, fair-enough it says,
IBM 5150
"POWER SUPPLY Built-in 63.5W switching power supply unit".
http://en.wikipedia.org/wiki/Power_supply#Computer_power_supply
The first generation of computers power supplies were linear devices, but as cost became a driving factor, and weight became important, switched mode supplies are almost universal.
I would like to know whether the 5150 was the earliest mass produced 'desktop' computer to use SM though (?).
-
Peppercorn, Thanks for taking the time clearingup my understanding of the other posts.
The first generation of computers power supplies were linear devices, but as cost became a driving factor, and weight became important, switched mode supplies are almost universal.
I would like to know whether the 5150 was the earliest mass produced 'desktop' computer to use SM though (?).
Got me there???
I remember as far back as the intel 8080 and IBM 8086 CPUs I seen the IBM 5100 PC but what supply ??? In a desktop Apple had their Motorola CPU Supply "\???.
Tandy Corporation (Radio Shack), introduced its model in 1977 TRS-80 model one it used a Z80 CPU but the PSU was external plug in which I think was Linear supply By Model III
they integrated the PSU in 1981
DigiBarn Systems: IBM Personal Computer 5150 and VisiCalc for the ...The IBM PC (known inside IBM as the model 5150) was launched on August 12, 1981
AS far back I can find it but I wll still look...
If I try to find a replacement supply for the earlier model desktops, so far I only find hot swaps.
This only tells me that the supply has been replaced with upgraded supply designs. E.G. A linear design may have a hot swap to a SM design.
SMPSU Introduction to the industrial world wasn't it in 1976?
To 1981 that gives it 5 years for implimentation to get rid of flaws.
-
SMPS has been a feature of avionics since the 1960's, as they are a big saver of mass, and are able to be placed into small spaces, and not need complicated cooling arrangements. This reduced space and power requirements a lot, a very important thing in the aircraft arena. This carried though to the PD , as a linear power supply for a PD would more than treble the size of a PC, and increase the unit mass considerably. The early minicomputers had linear power supplies, which made up more than half the mass of the unit, and were responsible for half the power consumption.
-
Seems that everyting that eventually gets to the consumer. Trickles down from Avionics and Space administation technology. We get a watered down versions.
SMPS has a flyback transformer correct?
The the B&W TV has a high voltage section to the CRT and that concept was also A Switcher Mode PSU. What decade was that? First half to mid 20th century?
Application of the low dc output SMPSU on the early Desktop computers, presented and affordable to the everyday Joe.
I can only find the 1981 models 8080 and the 8086 CPU driven computers. All the others marketed prior had the external low wattage AC to DC linear adaptors.
At least what I can remember from my early days and I have now searched for.
If I remember the rating of a 100 watt dual output linear power supply (+/- 5v) and (+/- 12v) was as heavy as a Red Clay Brick and a half.
-
I didn't know the exact number but some amount of memory is used by the computer when it is in sleep mode.
-
SMPSU Introduction to the industrial world wasn't it in 1976?
To 1981 that gives it 5 years for implementation to get rid of flaws.
I'm interested where you found 1976 as the 'magic' date (?) for the SMPS.
The pdf ref. I mentioned earlier gives it as '72, but this itself is a ref. of a (now) broken link, so I can't guarantee its authenticity.
The first computer I ever used was a already oldish VIC-20 that I had a play around with the ol' BASIC on, but I never opened it up. I dissected an old Apple-II some years later, but remember little about its psu.
SMPS has been a feature of avionics since the 1960's, as they are a big saver of mass, and are able to be placed into small spaces, and not need complicated cooling arrangements. This reduced space and power requirements a lot, a very important thing in the aircraft arena. This carried though to the PD , as a linear power supply for a PD would more than treble the size of a PC, and increase the unit mass considerably. The early minicomputers had linear power supplies, which made up more than half the mass of the unit, and were responsible for half the power consumption.
That's interesting to hear Sean. The article I found claims that the first commercial high-speed mains voltage transistors became available in 1967. I suppose avionic apps may have had SM devices earlier as the supply voltage in aircraft is nominally lower (24v?).
I would also hazard a guess that was one of the prerequisites of having a 'miniaturised' power supply that could fit inside a desktop computer rather than being a heavy (not-to-mention inefficient) 'brick' on the floor.
-
Yea the need to give up some things to gain in other areas, where the gain is needed most I guess relative to the loss.
Especially losing inefficiency :-)
Sorry Peppercorn when I what posting last night I remembered this but was not sure where I seen it... I could not find the link for some dumb reason...
I did not save the link prior search and apply its reference. I asked and guest-estimated the year. Nothing magic about it.
We were speaking of the era of PC's.
I remember seeing touching in 1972, where, hand held scientific function calculators, Casio, Texas Instrument and HP, and the IBM 360 card punch system and that Ram core was like a 1 foot cube interwoven wafer with small toroid beads at each three way intersection of weave section.
In 77 intro to Microprocessors class had the intel 8080 CPU and I had a hard time with that class. Since I did have a small exposure to discreet analog circuits, concept of digital was really new to me.
I CORRECT myself... earlier posting should of been comparing the intel 8088 to the IBM 8086 SORRY again...
My source in question, I found it, and it brushes a little insight to Governmental Uses:
"A pulse-width modulation control chip invented in 1975 spurred the development of switching power supplies, leading to a power control IC industry that today is measured in billions of dollars."
...more (http://powerelectronics.com/power_management/pwm_controllers/power_pwm_single_chip/)
http://powerelectronics.com/power_management/pwm_controllers/power_pwm_single_chip/ (http://powerelectronics.com/power_management/pwm_controllers/power_pwm_single_chip/)
That's interesting to hear Sean. The article I found claims that the first commercial high-speed mains voltage transistors became available in 1967. I suppose avionic apps may have had SM devices earlier as the supply voltage in aircraft is nominally lower (24v?).
The featured conception of SMPSU, I believe was not dependent on Transistor technology. TV High Voltage section had the concept of SMPSU. The military had to have implemented this concept in its prior applications pre transistor years.
Transistor technology enhanced its growth reducing its expense until ICs made discreet electronics smaller and more affordable
-
I remember the earlier SMPSU in the early PCs.
One bad characteristic was if you shut down the computer supply and turn it on a second later.
It would become a voltage doubler and spike.
Sometimes releasing the life force that drives the otherboard ICs.
(Releasing the smoke inside these chips, academic joking with the science professor he told us that smoke is what makes them work once it is released they are broken haha.)
After a some years maybe in thelae 80's or early 90's, not sure when, engineering had employed a fail safe in the newer model supplies.
After the supply is turned off it will not respond to a turn on until it discharges the primary voltages at shutdown I think it was a 10 or 15 second wait period.
-
The pdf ref. I mentioned earlier gives it as '72, but this itself is a ref. of a (now) broken link, so I can't guarantee its authenticity.
Can you display the URL to this broken link. I may be able to track the cashe
-
The pdf ref. I mentioned earlier gives it as '72, but this itself is a ref. of a (now) broken link, so I can't guarantee its authenticity.
Can you display the URL to this broken link. I may be able to track the cashe
http://www.steve-w.dircon.co.uk/fleadh/mphil/history.htm
Original claim that I read was in:
http://www.ece.msstate.edu/courses/design/2008/pwrsupply/Deliverables/Microsoft%20Word%20-%20problemstatement_final.pdf
-
The pdf ref. I mentioned earlier gives it as '72, but this itself is a ref. of a (now) broken link, so I can't guarantee its authenticity.
Can you display the URL to this broken link. I may be able to track the cashe
http://www.steve-w.dircon.co.uk/fleadh/mphil/history.htm
Original claim that I read was in:
http://www.ece.msstate.edu/courses/design/2008/pwrsupply/Deliverables/Microsoft%20Word%20-%20problemstatement_final.pdf
Wow that is something glad you had found it.
We the everyday public gets to see the cutting edge of technology 10, 15, or 25 years after its used up presents.
I found this in an uncommon corner it is not so much a focus on history but points out some interesting things, stating and relating this topic used in 1962
http://www.repairfaq.org/sam/smpsfaq.htm#smpstsps
I am still trying to find a relationship between the older Vacuum tube Television's high voltage section flyback area and switcher power supply characteristic function.
-
http://www.repairfaq.org/sam/smpsfaq.htm#smpstsps
Indeed, a handy and enlightening little article - Thanks!
I am still trying to find a relationship between the older Vacuum tube Television's high voltage section flyback area and switcher power supply characteristic function.
Off the top of my head - both would require a HF chopped waveform to drive them and require a fast slew rate to as an intrinsic characteristic of their operation.
-
JOHN Gamel asked the Naked Scientists:
Doctor Chris, here's a question for you:
When I put my computer to "sleep," does this consume more energy than if I turned it off?
John Gamel
If you've not given up completely on an answer to your original question, John - here's some 'ball-park' figure care of the web:
http://www.dell.com/downloads/global/corporate/environ/Opti_GX620_DT.pdf
DELL Optiplex GX620 desktop claims:
1.2 Watts for Sleep mode (pretty amazing if true esp. for a desktop!)
http://www.wisegeek.com/how-much-electricity-does-a-computer-use.htm
"In standby mode, a computer uses approximately 6 watts of electricity and the monitor’s electrical consumption drops to almost nothing."
-
My commodore pet (strictly a CBM) has a linear PSU based on a big heavy transformer. Since it has no "sleep" mode it uses pretty much the same power no matter what you are doing.
-
My commodore pet (strictly a CBM) has a linear PSU based on a big heavy transformer. Since it has no "sleep" mode it uses pretty much the same power no matter what you are doing.
You might be able to reduce its power consumption a bit by slowing down the clock and removing power from the monitor. The early versions used static RAM, so it might actually be possible. Mind you, as it takes little time to boot up, you might as well just turn it off!
If operating systems were not so horribly bloated, there might be less need for standby/hibernate/sleep schemes. (I'll probably get some heat for that comment.)
-
Underclocking a 1MHz computer; I bet not many people do that.
(and you are right about the bloatware)