Overclock.net banner

1 - 20 of 28 Posts

·
Registered
Joined
·
23 Posts
Discussion Starter #1
I haven't gotten too much feedback from the thread I started in the Experiment section. Wondering if anyone could help comment or give advice.

I am building a well fed (50 deg F) setup to include an 80 gallon reservoir. This reservoir would in turn feed our primary hot water heater to increase energy reuse and savings.

http://www.overclock.net/cooling-exp...ling-loop.html
 

·
Watercooling Enthusiast
Joined
·
4,799 Posts
you're going to have to watch for condensation inside your computer since 50f is hopefully far below the ambient tempt within ur home. you will have to insulate the computer. I would use a heat exchanger to transfer the heat from my computer to the reservoir which would keep fluids separate and ensure a clean loop. can find them here
 

·
Registered
Joined
·
1,148 Posts
I wouldn't use the 80 gallon res, I would use a liquid heat transfer unit such as this Koolance unit as an example :


Koolance HXP-193 Plate Heat Exchanger

That way you can keep the very mineral loaded well water in your house separate from your PC cooling loop. It would significantly gunk up your WC loop and corrode your blocks. I assume you wouldn't use a rad as you want all the heat to go into the water for your house.

I think the biggest problem beyond the condensation issue that was already mentioned is that unless you plan on running your CPU at 100% load all the time it won't make enough heat to sufficiently heat the hot water for your house. If your PC is idle at 35c - 40c that would only be 95 - 100F which is basically warm water. Add the loss of heat during storage and through the pipes and I don't think a PC is a good enough heat source.

For instance see this quote

Quote:

Originally Posted by US Department of Energy
According to the US Department of Energy, a temperature of 120 degrees at the tap is adequate for most household chores with a minimal danger of scalding and maximal energy efficiency. However, that is the temperature at the tap, not in the tank. Tank temperature should be no less than 130 degrees to prevent bacterial growth, such as Legionnaires disease.
Most hot water heaters are 140 to 160 depending on how well insulated the unit is.

I thinking is this wouldn't work very well but creative thinking.
 

·
Premium Member
Joined
·
5,945 Posts
Ratdog, that is the point. The water is first heated by the computers while in the first reservoir tank, before being dumped into the hot water tank.

So a molecule of water might have this happen to it:

Get Sucked from well @50F --> get heated to 70F by computers --> go into hot water tank --> get heated to 140F --> come out of the tap.

He's saving energy because the water heater only had to heat the water from 70F to 140f, instead of from 50F to 140F.

Also, @ OP, I see from your other thread that the contractor didn't want you to mess with the geothermal setup. I don't know what you proposed or what type of geothermal heating you have, but I definitely think you could get that to work.

My system works by sucking water from the well @ 10C, running it through the heat pump (which drops it to 4C) before dumping the water back outside into the ground (A long hose with holes). If your system is anything like mine you could hold that 4C water in a reservoir and use that water to cool the computers. This of course means that the heat from the computers is dumped outside during the winter, which is not desirable.

Another solution would be to use a heat exchanger on the water inflow from the well. Personally, I think this is the best option. Basically, install a heat exchanger in between the well and the heat pump, so that the heat from the computers will heat the water from the well from 10C to 11C (just guessing). This increase in the water temperature will cause the heat pump to run more efficiently. I don't think that your contractor will have a problem with this, as a heat exchanger does not allow the two liquids to mix. The only difference from the geothermal unit's point of view would be a slightly higher loop resistance, and I doubt that would affect the performance of the heat pump.
 

·
Registered
Joined
·
23 Posts
Discussion Starter #6
Oops, perhaps I should have posted my original thoughts from the experiment thread.

Things to note:

The 80 gallon reservoir would be another hot water heater in front of the primary and the primary is gas fired. Essentially, the secondary (reservoir) would be providing somewhat heated water to the primary which would have to do less work to bring it up to temp.

I definitely am interested in the heat exchangers. I thought I would have to put something into the reservoir tank but it looks like those exchangers are the ticket. Thanks RatDog and Stallion!

I guess I'll have to do some reading up on insulating the blocks, etc.
 

·
Unwilling Beta tester
Joined
·
7,188 Posts
Sounds like a good idea, and the temperatures could be pretty good too. I was so enthused (and so bored at work) that I did a quick calculation to see what sort of coolant water temperatures you were looking at.

I used the data from your post / sig (80 US gallon tank, 50°F (10°C) water in, 750W PSU) to run a calculation using both Newton's law of cooling and Q=m*Cp*T to determine the temperature response against time.

Knowing nothing about the thermal properties of your tank I set the convective cooling constants up so that a tank of water would rise from 10°C to the 25°C ambient in 6 hours. I have no idea if this is at all accurate, but it seemed like as good a guess as any.

I also assumed that there was no flow through the tank during a 12 hour time window. This is meant to be a worst case, overnight folding / "I'm so involved with this game that I haven't left my seat for 12 hours" kind of situation.

I ran four cases; 750W, 500W and 250W heat load with convective tank cooling, and an ultimate "worst case" 750W heat load with no cooling (perfectly insulated tank). Yes, I realise that you will not be putting anywhere near 750W into the water, or even 500W for that matter. See attached graph for results.

The graph shows an initial sharp rise on the first 3 cases. This is due to the water warming up through convection to the ambient combined with the heat dump into the water from the PC. As you can see, even with 750W of heat being dumped into the tank it will equalise at 28°C. Not bad at all.

The worst case scenario shows that with 750W going into a perfectly insulated tank, you will still only reach 35°C after 12 hours. Getting a little warm but still not too bad, and you will never see this in reality.

This brings me to the subject of insulating your 80 gallon tank. If you have a hot ambient and a lot of flow through the tank, I would insulate it as well as possible. This will help to slow down the warming of the water due to ambient temperatures, and the high flow will get rid of the warmed water well enough. If, however, you have a low flow or a cold ambient, I would not insulate the tank at all. The less insulation the better the convective cooling, so you would get closer to ambient after about 3-4 hours of heat dump with no flow through the tank.

I hope this helps a bit. Good luck with the project, I will be very interested to see how it turns out.
LL
 

·
Registered
Joined
·
23 Posts
Discussion Starter #8
@charliehorse55

I will double check with the Geo people but I probably need more information on how their system is planned. Once I have some hard facts on their setup I can see if they are willing to work a system like you proposed.

@GingerJohn

Thanks for the calculations! Per the question of insulated tank I would prefer to have it release as little heat as possible to influence the ambient. The reason being that my saltwater fish tank is located in the same mechanical room and I do not wish to raise the temp of that water. I was told that the mechanical room would most likely be a few degrees above the rest of the house so no cold ambient there. The computers will not be on unless used and a 12 hour gaming session would be a rare likelihood. (<sigh>, gone are the days...)

I imagine that the average water use will also positively affect the replacement of any heated water. Typical day as follows:

Up at 6:30 A.M
2 showers @ 10 minutes each (100 gallons)
work
get home
100 gallons - Bath at night (4 nights a week - big tub)

Game 3-4 hours (best scenario) on water cooled from morning shower, hand washing, food prep, bath etc.

Go to bed and the cycle repeats. So whether showering or bathing once a day. Most, if not all, of the water in the storage tank should be replaced with fresh and cool well water.

Used this site for water calcs

http://www.thewatergeeks.com/Water-C...ion-sp-16.html
 

·
Registered
Joined
·
23 Posts
Discussion Starter #9
So thinking strongly about using the exchanger to keep the loops separate but I'm wondering if how to cycle the heat best into the storage tank. Would there be exchangers I could submerse in it? If not, should I insulate the exchanger and put a small pump on the storage tank side in order to circulate the water and better capture the heat from the computer loop?
 

·
Watercooling Enthusiast
Joined
·
4,799 Posts
Quote:

Originally Posted by RizzayDJ View Post
So thinking strongly about using the exchanger to keep the loops separate but I'm wondering if how to cycle the heat best into the storage tank. Would there be exchangers I could submerse in it? If not, should I insulate the exchanger and put a small pump on the storage tank side in order to circulate the water and better capture the heat from the computer loop?
you could set up a small loop inside the reservior. use a submersible aquarium pump that will just intake from the reservoir and pump it to the heat exchanger. on the other side of the heat exchanger u could have a short outlet tube to get the warmer water away from the exchanger. you would want the pump or whatever u decide to use as intake as far away from the exchanger as possible and preferably on the bottom of the reservoir. think of it as setting up a fish tank
 

·
Registered
Joined
·
1,148 Posts
Quote:

Originally Posted by RizzayDJ View Post
Oops, perhaps I should have posted my original thoughts from the experiment thread.

Things to note:

The 80 gallon reservoir would be another hot water heater in front of the primary and the primary is gas fired. Essentially, the secondary (reservoir) would be providing somewhat heated water to the primary which would have to do less work to bring it up to temp.

I definitely am interested in the heat exchangers. I thought I would have to put something into the reservoir tank but it looks like those exchangers are the ticket. Thanks RatDog and Stallion!

I guess I'll have to do some reading up on insulating the blocks, etc.
Ahhh.. ok sorry I wasn't following what you were trying to do.

I still think there might be some issues. You would need to circulate the water in your first hotwater tank through the fluid exchanger multiple times for it to pick up enough thermal energy to actually heat 80 gallons of water significantly.

To give you an example (with some assumptions), I have a swimming pool that is ~16k gallons and has a 400,000 BTU heater. With an ambient temp of 75F It takes it an hour to raise the temp of that swimming pool 2 degree's F (~.5c)

So if we think about that in terms of your 80 gallons. The heater puts out 400000 BTU/h or 117228.4 watts for 2F. Divide by 200 = 586.14 watts to raise 80 gallons 2F in 1 hour. I know this isn't perfect because there are all sorts of variables in surface area of a pool vs closed loops, insulation factors, heat transfer differences (gas heat vs heated liquid) efficiency of transfer (some gas heat goes up the flue vs liquid to liquid probably more efficient) .

Anyway, so rough math. A 980x stock does 130w TDP, and a GTX 480 is 250w TDP. Go with quad SLI and you have 1130w at full load. Forgetting OC for a minute, your PC would need to run at 100% load to raise your 80 gallons 4F in one hour. This is horribly inefficient and if you let the PC go idle it would do almost nothing. Again, this is assuming a closed loop and nobody turning on the hot water. It also assumes 100% insulation and no heat lost to the atmosphere through normal convection.

My point simply is, yes it could work, probably wouldn't do much to make your water heater more efficient and if anything would just make your WC loop work great due to the low ambient of 50F on the liquid transfer block.

Good for OCing, not so good for heating your water for hot water.

woof
 

·
Premium Member
Joined
·
5,945 Posts
Quote:

Originally Posted by RatDog View Post
My point simply is, yes it could work, probably wouldn't do much to make your water heater more efficient and if anything would just make your WC loop work great due to the low ambient of 50F on the liquid transfer block.

Good for OCing, not so good for heating your water for hot water.

woof

Why not just calculate it with the heat capacity of water:

Water = 1 L = 1 KG = 4187 J/C
80 Gallons in L = 303 L
303*4187 = 1,267,949 J to raise the 80 gallon container 1 C.
A Joule is 1 Watt for 1 Second.

Assume average computer power usage of 300w:

= 4,226 seconds to heat the water 1C
= 1 hour and 10 minutes.

Also, RatDog, I don't know what you mean when you say "probably wouldn't do much to make your water heater more efficient", if he is using electric heat the water heater will have it's power consumption reduced by the exact amount of power the computer is using. This effectively means that he isn't paying for the electricity to run his computer. (Free folding at home anyone?).
 

·
Registered
Joined
·
1,148 Posts
Quote:

Originally Posted by charliehorse55 View Post
.

Also, RatDog, I don't know what you mean when you say "probably wouldn't do much to make your water heater more efficient", if he is using electric heat the water heater will have it's power consumption reduced by the exact amount of power the computer is using. This effectively means that he isn't paying for the electricity to run his computer. (Free folding at home anyone?).
but it isn't 100% efficient loop so you are losing energy along the path, and you have to add more energy to run the 2nd pump. So no, you are not getting free computing cycles. Also, if you use a gas heater as we do in the states, gas is much more thermally efficient than electric and cheaper. So you are actually increasing your costs by trying to heat the water with electrical power (strictly looking at the water side, I know you would be burning that to run the pc).
 

·
Premium Member
Joined
·
5,945 Posts
Quote:

Originally Posted by RatDog View Post
but it isn't 100% efficient loop so you are losing energy along the path, and you have to add more energy to run the 2nd pump.
How does a 15w pump compare to a 300w average computer? I think that he could get 85% efficiency out of this loop, which would save him roughly $250 a year in electricity bills (more if he has more than 300w average power going into the loop).
 

·
Registered
Joined
·
1,148 Posts
Quote:

Originally Posted by charliehorse55 View Post
How does a 15w pump compare to a 300w average computer? I think that he could get 85% efficiency out of this loop, which would save him roughly $250 a year in electricity bills (more if he has more than 300w average power going into the loop).
Let me try it this way.

A high powered electric water heater at 3800 watts and a high power PC quad SLI at 1000w.

With my 1000w PC it would heat the water 1c every 21 min. but the hot water heater would heat the water 1c every 5 min. Hot water heater would heat that 80 gallons in 3.5 hours to 55c. The computer would heat the 2nd 80 gallons to 55c in 15 hours.

Water heater at 3.5 hours using 412 kw/h in a month. The computer uses 465 kw/h in a month to do the same work heating the 2nd 80 gallons to 55c. At 30 cents a kw/h you are burning $15 / month more than needed or 180 / year MORE money.

This is assuming you turn off the computer as soon as the water hits 55c. It also assumes the heat transfer of the PC is as efficient as the hot water heater which it isn't. Other factors such as thermal transfer weakens as the delta in temps decreases and unless he insulated the loop I would be amazed at 85% efficiency would also come into play.

If you just turned the computer off and used two electric hot water heater to heat the 160 gallons I would use 824 kw/h vs 877 kw/h. The only way this is MORE efficient is if he uses his computer 15 hours a day at 100% load. Then you could say he is saving the 412 kw/h he would have used to heat that second 80 gallons or about 123.00 / month.

Assuming he sleeps and works (or goes to school) he probably uses his computer ~ 6 hours a day at 30% load. Therefore about 6c / day. But that water started at 12c so he raised it to 18c. But it sits the rest of the day and overnight warming to ambient or ~23c. While his computer accelerated the time to ambient it probably would have gained the 6c just by sitting in a 23c room overnight. Therefore zero net gain from the PC.

Back to my original point, good for cooling a PC, not so much for heating water for use. I didn't even factor using gas instead of electric to heat that water or the fact it cools over time and would have to be reheated, or that every time the water is used the whole system is hit with 12c water again.

If the goal is to super cool your loop, this is great. If the goal is to try and save some money by pre-heating your water from PC heat, this would have little to no benefit. Things are not 100% efficient so that 300w or 1000w heat load on the PC would not be perfectly transfered to the water tank, and that tank wouldn't be perfectly insulated.

To me this is like hybrid cars. You spend 7k more to get a Prius and while you get 50+ MPG the savings in gas means you have to drive it for 12 years to make up for the additional cost of acquisition. The cost of the 2nd tank, liquid transfer unit, 2nd pump, etc would cost so much that even with your 250 figure which I debate, it would take 4 years or more to get cost even.
 

·
Unwilling Beta tester
Joined
·
7,188 Posts
Doesn't matter which way you cut it, you will be recovering some heat energy from your computer. It may not be a lot, but if you assume that you are running a computer anyway, so long as you are putting more heat into the water than it takes to run the pump (>15W) then you are good.

I agree it is not an efficient way to heat the water, but then you are not JUST heating the water. Cost benefit might be a little bit iffy, but hey, what are you doing on this site if you are worried about spending money on a computer


I think it is a great idea, if it works out well I might try it myself.
 

·
Registered
Joined
·
1,148 Posts
Quote:

Originally Posted by GingerJohn View Post
Doesn't matter which way you cut it, you will be recovering some heat energy from your computer. It may not be a lot, but if you assume that you are running a computer anyway, so long as you are putting more heat into the water than it takes to run the pump (>15W) then you are good.

I agree it is not an efficient way to heat the water, but then you are not JUST heating the water. Cost benefit might be a little bit iffy, but hey, what are you doing on this site if you are worried about spending money on a computer


I think it is a great idea, if it works out well I might try it myself.
LOL.. that was exactly my point. Great for a project, but don't expect big savings. Sort of like compression fittings, fun but not much real value.
 

·
Registered
Joined
·
23 Posts
Discussion Starter #18
Thanks everyone. The goal here is definitely not to save a great deal of money but rather to supply the best overclocking environment, yet capture the expended energy in the form of heat. I do have 2 watercooled computers (at the moment
).

Computer A
8600 overclocked
2 x 8800 GT SLI overclocked (working on volt mod)
3 hard drives

Computer B
8400 overclocked
1 x 8800 GTX overclocked
2 hard drives

Keep in mind that they are both running a total around 14 fans and so it gets quite noisy in the office. Also, the ambient temp has risen 5 to 6 degrees in the room with the door shut, easily. This is uncomfortable and also an issue for the HVAC which has to deal with one room operating at a 5 to 6 deg variance from the rest of the upstairs. The office will not be a different zone and so you end up having a very cool upstairs if you try to correct the difference. Again, money lost and inefficiency because the HVAC is actually working harder to compensate.

I would be happy enough making things quiet and not sweating when I game. Saving the heat is a bonus...
 

·
Registered
Joined
·
1,362 Posts
Sounds excieting to make a geothermal loop. Got plans to share?

 

·
Registered
Joined
·
23 Posts
Discussion Starter #20
I'm still open to the idea of tapping in to our geothermal but the company we are working with is hesitant to added extra load. I suppose if I can give them some rough thermal calculations they may be more apt to work with me but so far they have stated loop and equipment has been sized to the home and are nervous to expand.

Another thing to keep in mind is that we ordered a kill switch for Xcel (power company) to utilize in their savers program. They will offer us %50 off our total heating bill (measured with a separate meter) or 20% off our entire electric bill. As proposed by someone else, if we hookup to the geothermal heat pump we would run the risk of Xcel controlling the power to that pump and shutting it off at determined times. The max amount of time they could turn it off would be 12 hours but I'd rather not leave them in control of computer usage.
 
1 - 20 of 28 Posts
Top