Paradigm shift: Eliminating waste heat

In a comment on my recent blog entry, Mads points to water-cooled systems from Sun.  In another comment, John mentions a project he’s involved in that may or may not get the goahead to harness heat from a data centre.

Following Mads’s link, I see Sun claims its watercooled systems to 40% more efficient than some alternative – presumably one that would be equivalent in functional terms.  Whilst 40% may be better than nothing, it’s still a helluva lot of waste.  You could say you waste 50% more than you save.  What is missing from Sun’s pages is any suggestion of harnessing that energy and putting it to good use.

What Sun should produce, and what could make the decision much easier for John’s project, is systems in which capturing and re-using waste heat is integral.  A plumber should be able to able to plug them in to a normal heating system, just as they would another heat source like a boiler or immersion heater.  That is a paradigm shift in computer design, and it is indeed manufacturers such as Sun who are best-placed to take the lead in it.

On a smaller scale, manufacturers of desktop and home computers could perhaps do something similar.  A computer with builtin water cooling, that could be plugged in to the cold water supply feeding a domestic water tank.  So the computer’s waste heat goes into the household or office heating/hot water, and less energy is required from conventional sources.  The cooling system is plumbed in, but the computer’s innards need to be accessible so components can be changed whenever necessary.

In the case of desktop-replacements, that’ll work best where the water tank is within bluetooth range of the desk for connecting peripherals.  But if the idea takes off, we’ll soon see it incorporating longer-range options, such as a terminal that just plugs straight in to a wireless router (whose heat should really also be captured), or wired mini-boosters.  The principle is simple: concentrate energy use in the plumbed-in components, and minimise it elsewhere.

Posted on February 13, 2008, in energy, environment. Bookmark the permalink. 4 Comments.

  1. Using the spare heat is a nice idea but not really practical. It might be spare heat but its not actually hot enough.

    Air coming out of a rack of servers might hit something close to 40C. You pass this, lukewarm, air over a heat exchanger which produces even more lukewarm water, which you send to another heat exchanger to create something you wouldn’t even be able to call tepid. That second heat exchanger has to take enough energy out of the original water that its cold enough to go back around to the computer and extract more energy.

    The basic problem is that speed of energy transfer is proportional to the difference in temperature. Now if you could get your computers to operate safely whilst chucking out air at >100C or water at >80C you might be on to something.

  2. Not hot enough isn’t a problem – as witness the instances where computer waste heat is used to heat a building.

    If for example a CPU runs at 50-60 degrees, the waste heat in a domestic water system might for example warm water from 5 to 30 degrees in a domestic water tank over a few hours computer use. Maybe even more if the water remains unused all day; less if it’s in high use (like running several baths in succession). Thus the conventional immersion heater (or whatever) has less to do and requires less energy.

  3. A couple of considerations: Desktop PC’s located at the user’s feet is a pretty decent thermal mass, overcomming one of the major problems with most building heating — it uses forced air, and warm air rises while a CPU keeps the heat closer to the person…

    HOWEVER that’s problematic in the summer 😉

    Better to think of using a HVAC system for HVAC, and reduce the heat from a desktop unit to zero.

    How to do that? Reduce the power to almost zero.

    How to do that?

    ThinClients on the desktop burning 4-5 watts RDP (remote desktop protocol) to a virtualized terminal server in the Data Center (virtualized so you can run several different servers on one physical hardware box). At that point you can discuss tieing that ONE 450 Watt 24/7-on server into the heating system instead of trying to tie hundreds of individual desktops into the system.

    ThinClients at home then could be used (instead of laptops and full-bore desktops) to access services at work.

    For most web surfing and e-mail, a terminal server (in my case, Windows 2003) is a much faster machine than either my standard desktop at work (which I’ve given up for a thin client for the past 6 months). I don’t drag a laptop back and forth between work and home because I can RDP to the terminal server from the computer at home (not yet a thin client) and be right back into the same applications I was using an hour before at work…

  4. Oh yes, for home use it would make bigger bang for the buck if your clothes dryer, dish washer, shower and refrigerator’s waste heat were reclaimed/recycled ala a heat exchanger as compared to your desktop pc…

    Not that I’m against the concept — I see it as very valuable at the Data Center/Server Room level. I just see other areas at the desktop/home level where you could make more dramatic leaps with your shift…

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: