Personnl Computers Heat Rejection

[schola]

Dear friends:

In this case, we have to perform a cooling load estimation, where 75 PC are required and will be used by workers in continuous labor during eight hours daily job.

Normally we used to assign 320 watts heat dissipation/PC.

Do you think taht this is a realistic value?


Thanks,

schola

 

[IRstuff]

Might even be a tad low. My monitor is rated at 1A-2A draw alone

TTFN

 

[darkvibe]

I use 300W for computers in a dorm or office and not a high end workstation with large monitor and large power supply. I use 500W for them. ASHRAE numbers are really low so i just guessed.

 

[HVAC68]

There was an article written in ASHRAE Journal June, 2000 which was part of ASHRAE Technical Committee research. The figures indicated are much lower than the above. This article, which was available as a free download then, is no more available and can be bought through the ASHRAE store.

The values indicated in that paper are as under :

For the computer - 55Watts / 65 Watts / 75 Watts based on average/conservative and highly conservative values respectively.

For the monitor - 55 Watts/70 Watts / 80 Watts depending on the size of the monitor starting from 13"-15" to 19"-20".

The article also furnishes lot of information on other office equipment such as copiers, printers, etc.

The name plate wattages varied from 165Watts to 759Watts.

HVAC68

 

[jraef]

I think those numbers are exceptionally low. That might explain why the article is no longer available. I have a 450W power supply on my PC, and I have a power monitor on the input (cuz I'm into that sort of thing). I am currently using almost 400W on average, not including the monitor because it is on a separate circuit. When you think about it, a PC is not doing any other work outside of it's environment so all of that power is dissipated as heat into the work space. Being that you never know what the peak power use could be, the safe bet would be to use the rating of the power supply as the total heat rejection. I don't know of any PCs sold now with anything smaller than 300W power supply any more, most are bigger as they get more capable. 500W with the monitor seems like a good number to me. Then don't forget to look at printers, scanners, copiers, fax machines, network routers etc. etc. If everyone has their own, all those little power supplies add up.

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more."
Nikola Tesla

 

[friartuck]

The heat gain from a PC depends upon the equipment it contains and the likely useage.

I have done some testing using a current transformer which measures the load taken. Modern PC's have a higher heat gain because of the current technology on the main processors. These run really hot...but every so often, the technology changes and the motherboard voltages change and so on, and the load goes down again.

The label on the PC power transformer is not a good indicator of the load. A 400W power transformer has the ability to have say the DVD, CD, hard drive, LAN card etc all running at the same time...but the likelihood is that they will not be.

My readings taken with a Pentium 4 processor gave a figure of 150W..but this varied.

I would estimate that a figure of 350W would be more than adequate, but it would be nice if ASHRAE produced some more figures.

Friar Tuck of Sherwood

 

[MintJulep]

I have a copy of the article HVAC68 cites. The values presented are actual heat loads measured.

One of the main points of the article was that heat dissipation is much lower than the nameplate rated power consumption.

Both the article and the research papers that it summarizes are available from ASHRAE.

I am currently using almost 400W on average

"Currently" implies an instantanious reading, so there is a built-in contradiction in your statement. How many watt-hours does your computer suck down in an hour?

 

[IRstuff]

Let's get real, that article is at least 5 yrs old. That is almost 3 processor generations ago.

The power consumed by the processor was less than 1/4 of that of the current generation of processors. Likewise, there's more memory, faster peripherals, wireless, bigger harddrives. Whatever the values were 5 yrs ago, they have little relationship to the current crop of computers.

TTFN

 

[HVAC68]

Quote

"The article is no longer available"

Unquote

The article is still available, but at a cost of US$8.00. It was available for free download for a brief period - or maybe I downloaded it when I was still an ASHRAE member - can't remember.

I have always considered 200Watts for my calculations, and touchwood - haven't gone wrong so far. Well, that isn't proof that I am right - maybe there are other research papers.

I agree that the article I referred is almost 5 years old and probably the research was conducted a good 6 months prior to that - The article refers to Pentium processor - but, newer processors have already come and maybe the heat emissions are higher. Is there any other newer research done on this subject ?

HVAC68

 

[IRstuff]

Well, case in point, 5 yrs ago, a P2 laptop with a smaller display, slower processor, wimpier battery had about 5-6 hrs of op time.

My current P4 laptop drains a bigger battery in less than 2 hrs.

TTFN

 

[katmar]

The trend towards LCD screens must be helping???

 

[MintJulep]

The power consumed by the processor was less than 1/4 of that of the current generation of processors

Likely true, but the point to note was that the heat disapation was much lower than the nameplate power consumption.

The label power consumption and power supply are rated for peak conditions. If you are running FEA or some other processor intensive application than you may be operating near peak for extended periods of time, but the vast majority of office computers do not tax the processor's capabilities at all. Open your task manager and take a lood at the CPU usage on the performance tab.

 

[IRstuff]

200W power supplies are no longer the norm, that tells you that the actual consumption is going up.

While the actual consumption may be lower than the current nameplate values, they are still probably substantially higher than the equivalent conditions 5 yrs ago.

TTFN