Heat Gain from Computers

[ChrisConley]

We've been using the same heat gain from computers for quite a while now in our cooling load calculations, despite the fact that computers are running more efficiently and LCD monitors are much cooler than CRT.

Does anyone have a website or resource that lists 'actual' heat gain from computer equipment into a modern office?

 

[imok2]

Chris, if this sounds interesting to you then go here:

http://www.rmi.org/images/PDFs/Energy/E93-20_EnergyEffComputer.pdf

An energy-efficient computer brings many important benefits to its owners and users, and to the developers and
owners of the building where it is installed. These benefits include in our example:
· direct electrical savings around 80% during office hours
· additional savings, typically about 35% as big, in air-conditioning energy
· further savings in uninterruptible-power-supply losses and wiring losses, raising the total electricity savings
to the neighborhood of $100 a year (or about $500 present-valued) at average U.S. commercial rates where
a UPS is used (but reliability will be much lower if it isn't)
· avoided capital costs for air conditioning equipment, ranging from about $130 for a new building to about
$40 for a routine cooling-system renovation likely to occur in the ‘90s
· halved uninterruptible-power-supply capacity, cutting its capital cost by about $200 where it’s used
· typically about one-fifth smaller plug-load wiring capacity, potentially saving another $200 or so (if the
electrical engineer pays attention to the reduced load!)
· possibly lower transformer costs because of reduced computer-power-supply harmonics
· much larger avoided costs (up to thousands of dollars per workstation in difficult cases) for upgrading old
offices’ wiring and cooling capacity to accommodate inefficient office equipment and its harmonic
generation

 

[imok2]

Chris,The article is pretty old It may not be relevent for you needs!

 

[IRstuff]

I think your supposition that computers are more efficient is belied by the fact that both CPUs and GPUs have bigger and more extensive heat sinks for removing the heat. Why? Because while a single transistor in a 45 or 35 nm technology is more efficient, they use more of them in the newer generation of processors. Dual core and quad core processors have double and quadruple the number of transistors in the math processor path compare to the original Pentiums.

TTFN

FAQ731-376

 

[MintJulep]

Dr. Mohammad Hosni at Kansas State U has done work on this in the past - including an ASHRAE sponsored reserach project which was used as the basis for the numbers included in the most recent ASHRAE Handbook.

The most recent work that I have a copy of is seven years old at this point, and doesn't address LCD monitors at all. I don't know if they have been continuing.

 

[ChrisConley]

Thanks all, I'm going to continue to dig. I'll post if I find anything new.

 

[IRstuff]

Additional stuff:

A late-model HP laptop DV9843 uses a 90W AC adaptor, which is >20W more than what older laptops use.

An HP Pavilion Elite m92002 desktop uses a 350W power supply, older PCs used 150W or 250W power supplies.

An HP Pavilion w1707 LCD display is spec'd at 37W, while the two crts in my office are both tagged at up to 250 W.

So, while displays might have gone down, somewhat, in power, larger and brighter displays will burn a sizeable amount of power. The desktops have not gotten lower in power, but more, as indicated by the power supply requirements.

TTFN

FAQ731-376

 

[MintJulep]

Except that to a large degree the power supply requirements are dictated by the possible need to simultaniously power a whole bunch of crap such as:

Multiple CD/DVD drives, possibly in burn mode, 15 USB devices, Wi-Fi, BlueTooth, two or more monitors, all while running at 100% processor capacity.

Most typical users will never approach the need for the power supply's full output, and even those that do won't do so for extended periods of time.

 

[ChrisConley]

I found the attached spreadsheets listing the power use of monitor's and computers on the EnergyStar website.

They list the average power consumption of an EnergyStar LCD monitor at 28W and the average consumption of an EnergyStar computer at 104W.

A conservative number would be to assume 150W of heat gain from a single computer at 'peak conditions'.

 

[IRstuff]

I don't know about "most," but almost anyone in our office that can cajole IT into it has two monitors. And while the EnergyStar numbers are probably valid for the equipment listed, I know that we've never bothered to consider EnergyStar as a buying criteria. We might incidentally buy one, but there is not a mandate to do so.

Moreover, almost all the EEs amd MEs are running PCs with souped up graphics and central processors for CAD, so they're nowhere near the low side of power consumption.

While EnergyStar is certainly a datapoint, it's only marginally relevant to a generalized heat load design, since even if the initial agreement is to only use EnergyStar equipment, there's no guarantee that future occupants will be held to the same constraints.

TTFN

FAQ731-376

 

[ChrisConley]

Agreed, however the 150W is based on the non-EnergyStar information listed. Using all EnergyStar equipment the number would be 132W.

The major difference in EnergyStar (as far as I can tell) is that they come with an automatic shutdown, or hibernation mode, to reduce usage to 2-5W during periods of non-use.

The second spreadsheet shows numbers from monitors.