Login |  Register |  FAQ
   
Post new topic Reply to topic  [ 14 posts ]  Go to page Previous  1, 2

Everything you've wanted to know about TEC's

 Post subject: Re: Everything you've wanted to know about TEC's
PostPosted: Thu Jan 31, 2013 1:30 am 
Offline

Joined: Tue Jan 29, 2013 8:31 pm
Posts: 2
Thanks for your quick reply.

In answer to your questions and comments:

The TEC - Custom Thermoelectric, http://www.customthermoelectric.com/tec ... ec_sht.pdf

I ran some more experiments since my last post to better understand my system.

My first lesson, mentioned by you earlier in this thread, do not run the TEC at full power. This TEC is rated for 24V 15A. It works at that combination, but gives essentiall the same cooling power with much less waste heat at 15.5V 11A.

My second lesson, I made the assumption that my heat load in my test setup was negligible. Dumb assumption on my part. The exposed surfaces in the tubing and water container is significant.

I do not understand your third point -
3 The temperature of the hot side greatly affects the achievable DELTA (dTmax) and MOVABLE heat (QMAX) IF the hot side is lower than the rated temp then it wont beable to achieve the rated delta or Qmax and therefore 150watts will never be possible in your case.

How do you determine the rated temperature for the hot side? My working assumption is that I want to minimize the hot side temperature. Do I need to keep the hot side at a specific delta Temp from the cold side? In my specific example, I am running 10C on the cold side. I can control the hot side temperature by adjusting the radiator performance.

This is step 1 in a multiple TEC system. One set of TEC's will be cooling chips to -20C and a second set of TEC's will be cooling the hot side of those TEC's to 10C. The system will need to be optimized and your TEC calculator could prove quite useful in this process.

Do you have a version of your TEC calculator that works with Windows 7 Pro? I downloaded and tried to run the program, but it did not work (did not respond to keystrokes). During the installation several of the dll's to be loaded were older than the versions on my computer so I did not allow the older versions to be installed.


Top
 Profile  
 
 Post subject: Re: Everything you've wanted to know about TEC's
PostPosted: Sat Feb 09, 2013 9:41 am 
Offline

Joined: Sat Feb 09, 2013 6:42 am
Posts: 3
Image

I can answer some of your question :-

Above is a performance chart from a manufacturer...now manufacturers charts are not deemed accurate mainly because it is almost impossible to recreate the parameters of the chart in real life.
You can ignore the actual values on the axis the important bit is the relative shape of the and the information it shows.

If you look near the top right hand corner it says 27˚C =Th this being (T) emperature of (h)otside.
To produce this chart the manufacturer used a hotside temp of 27˚C. You will never get same output from your TEC even if it was the correct chart without your hotside being held a 27˚C. You will find this impossible.
So if your hotside is hotter or cooler than 27˚C the results will ALWAYS be different. If you cool your hotside below 27˚C you wont achieve what the manufacturer chart says.

Next..... to the right you will see Delta T in ˚C above a short inventory. This is the difference in temp between the hot and cold sides. The different coloured lines represent Q in relation to Delta so you can see straight away that you dont want a large delta between your hot and coldsides (and low deltas are notoriously difficult to achieve.) and you can see you also need the TEC to be drawing a reasonable current to achieve reasonable Q.

I am sorry I havent read back too far and have no idea what your TECs are hooked up to but in a PC if it includes processor cooling you cant hope to start at charts and questioning why your output doesn't match because processors change in heat output vrtually all the time which changes the delta's then the current draw then the Q etc etc

So....what's the point of a manufacturers chart ? You only use a manufacturers chart to asses the suitabilty of one TEC over another that's why they are generally all at the same Th. they know the man on the street is not going to have a lab setup. Then you must test the TEC you have chosen. The heat losses of your setup through hot/coldsides and other bits will determine what performance you actually end up with and it is down to you to test/tweak to get your best performance with your chosen TEC but your cannot reasonably start comparing this back to the manufacturers chart and questioning why you cant get the same performance.....you wont because your Th is not static and you are unlikely to of reduced heat losses to the same degree as a lab.

You mention controlling your hotside temp with a radiator...hmmmmm...I fear you aren't at least not in the manner your hoping, I believe you dont have sufficient control to get the static performance your hoping for especially if your processor's in the loop.
Unfortunately if you are cooling TEC/s with a radiator then using those TEC/s to cool the hotsides of further TEC/s which then cool your system.......I regrettably predict that that you will never achieve the performance you appear to be hoping for, I have never believed stacking to be of any use. It compounds all your heat losses, your deltas etc not to mention you will have to seriously upgrade your rads, I think you have overlooked the heat output of the top TEC's hotside when you plan to cool the coldside of the bottom set to -20˚C. You are far better to use high performance TECs (greater than 127 couple) of a reasonable high voltage which you then step down conservativelly by WIRING them into series and parallel goups to achieve the correct voltage/amps for each TEC. I believe this is what you have already been told by someone else. Ths is what they meant. Two TECs or groups wired in series leaves current draw the same and halves voltage per TEC/group, Two TECs or groups wired in parallel leaves voltage the same per TEC/group and doubles current draw.

If you list the company & cat.no for both sets of TECS you plan to use (please dont say they all the same.) or a copy of the manu. stats I am sure Ultra will advise.

Unfortunately people buy TECS then try to mould them to their plan....You need to work everything out first including voltages etc then find the TECs that have a fighting chance of achieving what you want and then test them to see if you were right. Then install them to work.


Top
 Profile  
 
 Post subject: Re: Everything you've wanted to know about TEC's
PostPosted: Sun Feb 10, 2013 12:20 am 
Offline

Joined: Sat Feb 09, 2013 6:42 am
Posts: 3
I have had a little look at what you have done in your setup.
According to the manufacturers graphs at 15.5v you should be drawing much less than 11 amps how did you come up with this figure, if you measued it , fair enough...They have less resistance than they should have. Have you correctly torqued the TEC's? or perhaps they were just low grade....
Did you know that when buying TEC's you can , if you pay more, have premium TEC's nearer to the advertised quality ?
This is particularly handy when using multiple TEC's. When manufactured all TEC's vary a bit they are tested and graded and ultimately some fail the grade. If your using multiple TEC's ideally you want them all the same you don't want them with differing resistances (and therefore differing running characteristics.) you can ask for them to be matched for resistance and also ask for them to be within a tolerance of resistance.
If your mounting several TEC's between single hot/coldsides you can ask for them to milled so they are all the exactly the same height. This is important because when you tension the TEC's (apply torque on the screws.) compressing the TEC's increases the resistance so if the TEC's arent all the same height they wont increase by the same amount. This is the reason why you should not overtension TEC's....stick to the manufacturers limit.

I notice you plan to have the coldside at -20°C (realistically unfeasible.) with the hotsides cooled to 10°C by the top TECs (not sure how feasible this is.) the delta will be 10-(-20) = 30. So, looking at the Qc to Amp chart on the left of the manufacturers charts, with you thoeretical current at 11A and a delta of 30 the working Q will only be 100w not your intended 150w. To further compound the problem your hotside is cooled to below 27°C so you wont even get a working Q of 100w it will be less. Of course with the actual rated current at 15.5v (according to the graph on the right.) which is nearer 9.5 amps you will get a Qc of nearer 80w which again has to be adjusted down for the cool hotside. In short these TEC's probably wont achieve what you want.


Top
 Profile  
 
 Post subject: Re: Everything you've wanted to know about TEC's
PostPosted: Fri Feb 03, 2017 12:47 pm 
Offline
Site Admin
User avatar

Joined: Sat Nov 06, 2010 11:28 am
Posts: 150
Ultrasonic2 wrote:
Everything you've wanted to know about TEC's

By Ultrasonic2 & Friends

Attachment:
50x2x4N4.jpg


Visit the You Tube Channel
The You Tube channel contains the latest information and covers most of the information covered here
https://www.youtube.com/user/Ultrasonictwo


CONTENTS

* How is a TEC Rated ?
* I just bought a 540watt TEC from E-bay so it can cool a load of 400watts?
* Why do i need a cold plate and a hot plate. Why can i just place the TEC directly on the CPU ?
* What can i use for a Cold / Hot plate and can i use Aluminum?
* How much pressure should i apply to the TEC?
* Why can't I air cool my TEC and get good performance
* Can I run my TEC off my PSU?
* Which power supplies are preferred when it comes to TEC-ing?
* What is a TEC
* What happens if the TEC’s loose power and my PC is still running
* How can i control the power going to a TEC
* I want to TEC cool my PC BUT i DONT want to go below ambient to avoid condensation.
* Do i have to insulate my Board and Block
* Can a TEC cool my load as well as Phase Change can ?
* I’ve heard that TEC’s are very inefficient is this true. And what about phase change ?
* I have a cooler with a TEC that draws 75watts of power so it’s Qmax or Pmax is 75watts right ?
* TEC's in Parallel or Series ( stacking, Cascade)
* So how much Qmax do i need.
* So how many radiators do i need.
* What TEC should i use and how much power will it use?
* I have noticed that the TEC’s use less power the longer I leave them running ?
* Where can i get a TEC calculator app to do all the math for me ?
* So i downloaded your calculator and entered the relevant data then went out and tried it but i didn't get the temps your program said i should why ?
* What's best a Direct Die cooler or a Water to Water chiller.
* What TEC should i use and how much power will it use?
* Where can i buy TEC Direct Die blocks and Water to Water chillers?





Q. How is a TEC Rated ?

Qmax Umax Vmax Imax Amax dTmax Pmax

Th
This is the temperature of the Hot side of the TEC . The Th is very important when evaluating all of the below ( dTmax, imax …. ) because as you increase the TEC’s hot side all of them will increase. Knowing what the Rated Th was is very important to make sure your comparing TECs on an level playing field. Th is normally rated at 300kelvin or 26.85c
A TEC rated at a Th of 26.85 with the ratings of dTmax=69 Imax=11.3 Umax= 24.6 Qmax=172 will be dTmax=77.7 Imax=11.55 Umax= 25.7 Qmax=194.5 when rated at a Th of 50c

dTmax
This is the maximum Delta ( temperature difference ) between the hot side of the TEC and the cold side. dTmax is achieved while NO heat is transferred between the two sides. As soon as you start applying a load to the cold side the delta is reduced.

Imax/Amax
This is the Current a TEC will draw while achieving dTmax. When a load is applied to the cold side of the TEC the current draw will exceed this value

Umax/Vmax
This is the voltage that achieved dTmax

Pmax
This is the amount of power a TEC will draw while achieving dTmax in watts or Umax*Imax. Ebay sellers like to list their TEC’s based on Pmax not Umax. This is very important to understand as listing TEC’s based on Pmax makes TEC’s appear far more powerful than they are if your assuming the TEC is Rated on Qmax, which is the industry standard. You can however ruffly convert Pmax to Umax by multiplying Pmax by 0.6161

Qmax
This is the amount of heat that can be transferred from one side of the TEC to the other while applying Imax (not Umax). Qmax is NOT the amount of heat that can be transfer from one side of the TEC to the other to a delta of 0. The amount of heat that can be transferred to a delta of 0 while applying Umax is actually greater than "Qmax" by approximately 5.5%. At this point Imax has also increased by about 30%
For example if we took this TEC dTmax=69 Imax=11.3 Umax= 24.6 Qmax=172
and we were to rate it to a delta of zero it would be dTmax=69 I=14.7 Umax=24.6 Qc=181.5
Image

---

Q. I just bought a 540watt TEC from E-bay so it can cool a load of 400watts?

A
NO
E-bay sellers tend to advertise their TEC's on Pmax and or at a Th temp greater than 27c. In the case of the 540watt TEC advertised on E-bay it is rated on Pmax AND at a Th greater than 27. Doing so jacks up the advertised value from 270.5Qmax to 540watts. Now since the 400watt load we are applying to this TEC is MUCH greater than the 270 it is able to move, this TEC is unable to cool a load of 400watts
Attachment:
!B6z9hFw!Wk~$(KGrHqUOKjkEyP+QVwDHBMydKZIhHw~~-1_12.JPG


---

Q. Why do i need a cold plate and a hot plate. Why can i just place the TEC directly on the CPU ?

A.
You HAVE to cover all the TEC with a hot and cold plate for 3 main reasons

A TEC has no inherent heat spreading ability. This means that if the TEC isn't covered by a heat spreader the percent sticking out wont be able to cool anything. This will obviously lead to reduce performance or failure of the TEC if the hot side isn't covered.

A TEC needs a far amount of pressure pushing it together otherwise magnetic forces will result in poor life expectancy.

Most Water blocks and Air coolers don't have perfectly flat bases, in most cases this is intentional. However this is a very bad thing when applied to a perfectly flat TEC. A water block maybe able to cover a 50x50mm TEC but all of it's surface may not actually be touching the TEC let allown applying pressure to it !

If using a TEC greater than about 50x50mm (you should be using a 62mm one) Then it's unlikely that you be able to use a perfectly flat cold plate because there are normally some capacitors or something in the way so it's best to machine a step into the base so the cold plate can clear these.

Attachment:
cold plates.JPG


Here are some cold plates for a 62mm TEC. The face not showing (the bottom in the pic) is the large surface where the TEC is clamped down via the bolt holes. The left plate is the Cold plates and you can see the Step down that has been machined in it to clear caps and so on. The right plate is the hot plate it has a step down to be compatible with most AM2 and 775 mounting hardware

---

Q. What can i use for a Cold / Hot plate and can i use Aluminium ?

A.
There was a detailed study undertaken quite a few years ago now that showed that a copper block of about 10mm (3/8) performed the best. But obviously the absolute best block thickness would very depending on your application. When copper is used manufactures generally recommend a minimum Cold and hot plate thickness of 3mm to keep bowing to a minimum.

As for Aluminium you can certainly use it if you want too. People don't use it in PC cooling for 2 reasons

* Aluminium and copper dont mix well leading to galvanic corrosion

* Copper conducts heat almost Twice as well as Aluminium so your results are ALWAYS better when copper is used.

If you do use Aluminium you'll need to use thicker hot and cold plates because aluminium is softer than copper, so it bends more easily.

Contributors (Ira-K, flak-spammer)

---

Q. How much pressure should i apply to the TEC

A.
You should apply about 300 PSI to your TEC to work out how to apply this pressure see the equation below
T = (C x D x F x in2) / (# of screws)

Using the following equation you can solve for torque per screw:
T = torque per screw (in-lbs)
C = torque coefficient (0.20 as received, 0.15 lubricated)
D = nominal screw size (4/40 = 0.112, 6/32 = 0.138, 8/32 = 0.164)
F = Force (lbs / in2)
in2 = Module surface area (length x width)

This per module

C= 0.20 (who oils their bolts)
D= 0.157 (4mm)
F= 300 (150 is the minimum recommended)
in2= 3.875 (50x50mm TEC)
# of screws = 2

18.25125 = (0.20 x 0.157 x 300 x 3.875) / (2)

So to achieve 300psi on the TEC i'd have to tighten each bolt up to 18.25 inch pounds. i'll probably round that up to 20inch pounds. This is actually a very SMALL amount of pressure and will require a special tool to apply. It's my experiences that a TEC with a higher couple count ( higher voltage ) can not withstand as much force as a lower voltage TEC AND the Pressure applied really isn't as important as widely though.

---

Q. Why can't I air cool my TEC and get good performance

A
It's not a simple answer really
The reason that people dont air cool TEC's is largely due to the inability to keep the TEC cooled well. A TEC may require a large amount of electricity to move the heat from one side of the TEC to the other. If the cooler is unable to remove this heat efficiently enough then this can result in the hot side of the TEC just getting hotter, so any possible gain from having a delta between the two sides of the TEC are lost. The cold side might still be 20c lower than that of the hot but the hot side may now be 80c instead of 60. Resulting in the same performance. But with a significant increase in cost

Another reason why people dont Air cool TEC's other than the above is the possible benefits of having a TEC is relatively low if not worse when compared to a normal water cooled system. A normal water cooled system also has a lower running cost and you dont have to worry about condensation.

However you certainly can air cool a modern overclocked quad while reducing its temps if you use a big enough TEC and heat sink, maybe a aire cooled water chiller

---

Q. Can I run my TEC off my PSU?

A.
The reason why people don't recommend you use a normal PC's PSU is that a TEC will draw ALOT of current. This current will exceed the rated value of the molex connector (about 8 amps) leading to excess heating and possible fire. ALSO alot of PSU's have multiple rails and it's not advisable to connect all the rails together. However if you connect a Single TEC to a PSU using multiple molexes then it's doable

---

Q. Which power supplies are preferred when it comes to TEC-ing?

A.
Meanwell's are pretty much the only way to go. They also have a voltage adjustment of about 10% each way which is good for getting the right voltage. The correct voltage for your TEC will depend on many factors.
Attachment:
Meanwell_SE_200__4dcb5e4606601.gif


---

Q. What is a TEC

A.
Thermal Electric Cooling
I wont go into how it works as the Wicki page is pretty good on that
http://en.wikipedia.org/wiki/Thermo-electric_cooler
http://en.wikipedia.org/wiki/Peltier...Peltier_effect

---

Q. What happens if the TEC’s loose power and my PC is still running

A.
A TEC is made from two dissimilar metals pushed together with a current passed between them. This means when the power is off the TEC will still conduct heat, but very poorly. If your machine is not running at full load when the power goes out it may have no impact on the stability of your machine. With two TEC’s I can perform all tasks expect game with the TEC’s off.

---

Q. How can i control the power going to a TEC

A.
The only realistic way of doing this is to PWM it (Turning the current on and off every fast like a switch). This is what the manufacturers recommend at over 1k switching. Hardly any available PWM controllers will be able to handle the high current draw of a TEC and most that can are designed for 12volts. So what most people do is build a PWM booster which allows a low powered PWM controller to control any voltage and current. A PWM booster is simply a DC Solid State Rely (SSR) which is controlled by the PWM controller. I personally have had great success using a M-cube Tbalancer as the controller.

A PWM controller is not required but recommended you can pretty much plug the SSR into your motherboard with the aid of some 6 volt Zener diodes. You can then use motherboards software to adjust the response curve. Depending on how the SSR is wired in conjunction with your motherboard, this can become more of a thermal switch than a PWM controller but it still works and is VERY CHEAP.

It is recommend that the frequency of the PWM be of 1khz or more to ensure high efficiency. If the frequency is low like a thermal switch then when the TEC is in an off state the heat can (and will) flow backwards across the TEC. When the TEC is turned back on the TEC will have to move the heat it’s already moved back again. This can of course be a vicious circle leading to reduced efficiency.
Attachment:
bigNG-c.jpg


---

Q. I want to TEC cool my PC BUT i DONT want to go below ambient to avoid condensation.

A.
If this is you then TEC's ARE NOT for you . The whole point of using TEC's is so you CAN go below ambient. If you want to stay above ambient then your possible gains over a normal W/C cooled system is very minimal and not worth the money and time involved in TECing. If you want to stay above dew point then TEC’s may be worth investigating as long as you have an all important PWM controller.

---

Q. Do i have to insulate my Board and Block

A.
Yes
A TEC will cool your CPU far below dew point (IF no controller is in place to prevent this) especially when the CPU is at idle. So it will generate condensation. Water and electricity isn't a good mix. Also if the block and CPU isn't covered the TEC is then trying to cool your CPU and air in your case, which reduces it's efficiency.

---

Q. Which is cooler, a single stage TEC or a single stage Phase Change ?

A.
Comparing a single stage TEC v A single stage Phase then the Phase change will result in lower temps. Phase change has a greater Delta that's not dependent on it's input voltage. So excluding size or electricity used for the moment then a Phase change system should always result in lower temps. However if you are looking to keep your load at or around ambient then TECing will be far more practical.
Attachment:
dsc1032.jpg


---

Q. I’ve heard that TEC’s are very inefficient is this true. And what about phase change ?

A.
To begin with the Term “Efficient” is NOT the correct term to use when comparing the amount of electricity used to what It’s able to move. As this can exceed 100% by a long way. When talking about moving heat we use the term C.O.P (Coefficient of Performance)

The C.O.P number is ascertained by taking the Qc (Which may NOT be QMax) and dividing it by the input power. Higher the number equals higher the “Efficiency”

Now there is NO blanket rule to whether or not a TEC is “Efficient” (high COP). This is the case because the COP of a TEC changes depending on the input power relative to it’s max. As you increase the input power the COP is reduced (Less efficient).

If we compare the COP of common phase change systems to a TEC system including pumps and fans. The COP can be very similar PROVIDING you have a relatively low input power relative to the max of the TEC. However to achieve a TEC system with such high COP you’ll need to use many under voted TEC’s. Making the unit far to large to direct die cool your PC. You’ll have to make a water chillier for that kind of COP. A Single TEC direct die cooling a PC will require a high input power relative to it’s max giving it a comparably MUCH lower COP than phase change.

For a 400Qmax TEC running at 100% of Umax it would have a COP of about 0.46 (422Qc/902watts)
For a 400Qmax TEC running at 50% of Umax it would have a COP of about 1.4 (315Q/225watts)
Attachment:
Random.jpg


NOTE: As you decrease the input power relative to the Max of the TEC, the maximum achievable delta is also decreased. So if you have a phase change and TEC system running at the SAME COP the phase change system will always result in lower temps because the phase change unit will have a far greater delta.

---

Q. I have a cooler with a TEC that draws 75watts of power so it’s Qmax or Pmax is 75watts right ?

A.
You CANNOT make this deduction since you have no idea what the input power is relative to its un- known max. Theoretically the TEC could have a Qmax or Pmax greater or smaller than 75watts.

---

TEC's in Parallel or Series ( stacking, Cascade )
(The physical placement of the TEC's NOT their electrical wiring configuration)

ALL MATH IS FOR DEMONSTRATION PURPOSES ONLY

Parallel TEC's is when you have two or more TECs that are placed NEXT to each other

TEC's in Parallel are used to increases the Q but keep the deltas the same

If one TEC has a Qmax of 400watts and a Delta of 62c then two in parallel will result in a combined total of their Qmax so 400 + 400 = 800watts of Q . Their delta will stay the same of 62c. The delta is not added together.
PIC 1
Attachment:
pic1.jpg


Series TEC's is when you mount one TEC on top of the other. So you have your load, then the cold side of TEC 1(lower TEC). Then you mount the cold side of TEC 2 (upper TEC) to the hot side of TEC 1. You then cool your load and both TEC's by cooling the hot side of TEC 2 with your water block.
Attachment:
threeStage.png


TEC's in series is used to increase the Delta But the Q is significantly reduced below that of a single TEC.

The delta is increased in a series config because you have one TEC cooling the other.

If your hot side is 20c and you have only ONE TEC then in a best case scenario your cold side will be 62c below it or -42. Now it doesn't matter if you have one or a thousand TEC's in PARALLEL the cold sides will still be -42c

In an overly simplistic example, if you had 2 TECs, one cooling the other then the upper TEC would cool the lower TEC's HOT side to -42c . Then the lower TEC would cool your load from its hot side temp of -42 to a cold side temp of -104c

Now the reality is you don't get a delta of 124 (62+62), if you were to try the above for 2 reasons

1 As you decrease the hot side of the lower TEC it's maximum Delta and Q is reduced.
PIC2
Attachment:
pic2.jpg


2 The max delta (dTmax) is achieved when NO load is applied to the cold side of a TEC . In a Series config the upper TEC is ALWAYS under load because it's having to move the heat ( electricty ) from the lower TEC which is ATLEATS the electrical wattage applied to the lower TEC. (explained more below ) This load then reduces the maximum achievable delta and a Q


The BIG disadvantage of stacking TEC's is that the Q is reduced in the process

With ONE TEC or TEC's in PARALLEL all of their rated Q is available to move heat. If a TEC has a Qmax of 360 then this TEC is able to move 360 watts of heat from one side to the other to a Delta of 0. (like it should)

However this is not the case in a Series config. This is because the upper TEC has to not only move the load applied to the bottom TEC but ALSO any electricity that the LOWER TEC uses.

An example of how important this factor is follows. If we used 2 400Qmax TECs and ran them at the Umax (full power) the lower TEC would be able to move 422watts to a delta od 0while consuming 902watts of electricity to heat. Now the combined total is 1324watts. This load then gets passed to our upper TEC to be moved. Now since the upper TEC can ONLY move 422 watts we have a VERY large problem.. actually the above is impossible.

SO to get it to work you have to reduce the voltage applied to the lower TEC so it's not consuming 902 watts of electricity . If we reduced the lower TEC's input voltage to say 30% of Umax to increase its COP it would now be only consuming 81watts (to a delta of 0) . However by reducing the input voltage to the lower TEC you have also decreased it's delta and Q .. so the lower TEC can now only move a maximum of 214 watts of Q. Unfortunately 214watts of Q isn't enough to cooler a modern over clocked CPU do a delta you want.


Stacking can work with larger loads, but only with a combined parallel series stack. You use the Series stack to increase the Delta. You then have multiple stacks in parallel to increase the Q. Such a cooler would require a lot of TEC's and space so Direct die cooling a CPU is out of the question leaving a water to water chiller the only option. Such a chiller has never been built to my knowledge

Now for some practical examples of Single, Parallel and Series Configs . I will presume we are able to keep the hottest side of the TEC's at 27c (though this is possibly unrealistic)


load of 250 watts ( OC'd i7)

One 400watts of Q TEC running at 100% of Umax would = a cold side of about 2c. It would be using about 800 watts of electricity

One 400watts of Q TEC running at 50% of Umax would = a cold side of about 18c. It would be using about 210 watts of electricity
Pic3
Attachment:
pic3.jpg


Two PARALLEL 400watts of Q TECs running at 100% of Umax would = a cold side of about -16c. It would be using about 1500 watts of electricity

Two PARALLEL 400watts of Q TECs running at 50% of Umax would = a cold side of about -0c. It would be using about 360 watts of electricity
Pic4
Attachment:
pic4.jpg


Two SERIES 400watts of Q TECs running at 100% of Umax. This config would be unable to cool a load of 250 because the Upper TEC would be unable to cool the 1070watts (250 CPU load + Electricity used by the lower TEC 820) being applied to it because the max load is 422watts which is far less than the required 1070watts applied to it
Pic 5
Attachment:
pic5.jpg


Two SERIES 400watts of Q TECs running at 50% of Umax. This config would be unable to cool a load of 250 because the Upper TEC would be unable to cool the 460watts being applied to it, as Max load is 315watts
Pic6
Attachment:
pic6.jpg


Now i know there will be some saying but i dont have a OC'd i7 using 250watts i only have a 60watt CPU. Well it might be 60 watts stock but really unless your planning to over clock it there's no point in TEC cooling it let alone a series stack. So it will far exceeding the 60 watts. However i will proceed with our unrealistic 60 watts


load of 60 watts ( OC'd commodore 64)

One 400watts of Q TEC running at 100% of Umax would = a cold side of about -26c. It would be using about 724 watts of electricity

One 400watts of Q TEC running at 50% of Umax would = a cold side of about -10c. It would be using about 163 watts of electricity
Pic 7
Attachment:
pic7.jpg


Two PARALLEL 400watts of Q TECs running at 100% of Umax would = a cold side of about -30c. It would be using about 1420 watts of electricity

Two PARALLEL 400watts of Q TECs running at 50% of Umax would = a cold side of about -14c. It would be using about 312 watts of electricity
Pic8
Attachment:
pic8.jpg


Two SERIES 30watts of Q TECs running at 100% of Umax. This config would be unable to cool a load of 240 because the Upper TEC would be unable to cool the 784watts (60 CPU load + Electricity used by the lower TEC 724) being applied to it because the upper tec’s max load is only 315watt
Pic9
Attachment:
pic9.jpg


Two SERIES 360watts of Q TECs running at 50% of Umax Now this works just

The lower Level would be able to move the 60watt load to a delta of 33c while using 150 watts of electricity. With a max load of 282

The upper level would just be able to move the lower levels load of 210 (60 +150) to a delta of 15.29c while using 200 watts of electricity
So the combined delta is now -72c (33+15) and the load is now -21, while consuming a combined total of 350watts of electricity.
Pic10
Attachment:
pic10.jpg


Two SERIES 400watts of Q TECs. The lower TEC running at 50% of Umax and the upper TEC running at 100% of Umax
The lower Level would be able to move the 60watt load to a delta of 28c. While using 134watts of electricity with a max load of 245watt
The upper level would be able to move the lower levels load of 175 (60 +135) to a delta of 33c while using 790watts of electricity.
The total delta is now 28.3 + 33.36 = 61.9 so our load is now -34.66c while using a total of 925watts of electricity.
Pic11
Attachment:
pic11.jpg


If you are intending on making Series Parallel water chiller please ensure you have a heat spreader between the 2 levels IF you use different TEC's for the lower and upper levels. That includes TEC's with the same physical dimensions but have a different number of couples.++

---

Q. So how much Qmax do i need?

A.
This depends on many factors, but mainly the heat load your trying to cool and the temperature your trying to achieve. For a starting point you should have at lest twice ( two times for the yanks) the Qmax than the load your trying to cool. So if you have a CPU load of 200 watts you'll want to have at lest 400watts of Qmax . doing so will get your load to a proximately ambient. This is not core temps but cold plate temps or water temps (in a water to water chiller). For those of you who want cooler temps than that i would suggest you shoot for 4 times the amount of Q than you have in load ie 800Qmax.

---

Q.So how many radiators do i need?

A.
As a guide It has been suggested that you need a 240 (120.2) radiator for each TEC you have. Obviously this is dependant on the percentage of Umax (voltage) your applying to your TEC's. The 240 per TEC is based on 75% of Umax and is only a ruff guide.
(MindChill)

---

Q.What TEC should i use and how much power will it use?

A.
Here is some thermal modelling of different TEC's and water blocks. The modelling is quite advanced with proper pump and TEC curves applied to it.

http://thermo-electric-cooling.com/pic/Block%20Test.xlsx

Image

---

Q.I have noticed that the TEC’s use less power the longer I leave them running ?

A.
This occurrence is most notable in water chillers where there is a large volume of water on the cold side. When everything is off and both sides are at ambient when you turn everything on the TEC’s are not only having to move your CPU’s load but also the sorted energy in the loop which is reflected in the cold side being at ambient. So when you turn them on your TEC’s maybe moving 100watts of CPU load and 50watts of loop heat. Once the cold side loop has reached a point of equilibrium the TEC’s are removing 100watts from the CPU only. Now all of the above affects the electricity the TEC’s use because as the load is reduced the electricity is also reduced, thus when you turn your TEC’s on the electricity used is greater than when your CPU is at normal operating temperature.

---

Q.Where can i get a TEC calculator app to do all the math for me ?

A.
You can get the one i wrote from here
Attachment:
TECCAL.jpg


http://www.ultrasonic2.com

This app doesn't let you add your own TEC's to it because then people would add all these dodge E-bay TEC's in there that are rated at Pmax and or have 50c hot side temps. making the app useless. However i have included the common E-bay TEC's

---

Q.So i downloaded your calculator and entered the relevant data then went out and tried it but i didn't get the temps your program said i should why ?

A.
First off the calculator isn't perfect but is pretty much correct and there is a thered dealing with that issue.
There are many reasons why you dont get CPU temps = to the cold side indicated in my app in the example above that s 2.85c. A quick run down of excluded factors from the calculator are below

*Thermal reisstance of the hot sides water block and radiator
*Thermal resistance of the cold side
*The thermal resistance of the CPU it's self

Fist off the TEC's hot side is always significantly higher than ambient. because the Radiator has a thermal resistance and so does the the copper waterblock cooling the TEC's. so if you'r air is 20c the water maybe 25c and the TEC's hot side being 30c.

You then have to transfer the load to the cold side of the TEC ether by water for a chiller or via a copper cold plate. Both have significant resistance. If your TEC's cold side is 2.85 that may mean while transferring your heat load the temperature of the surface touching the cpu maybe 4c

When looking at applications like CoreTemp its IMPORTANT to understand that the reported temperatures are not the temperature of the integrated heat spreader ( the bit touching the cold plate or water block ) but inside the Silicone core it's self. Silicone is a poor conductor of heat so the the core will be MANY degrees higher than the temperature of the cold plate. 30c is normal at load.

So in our example with an ambient temp of 20c the core would read 34c at load for example.

---

Q.What's best a Direct Die cooler or a Water to Water chiller ?

A.
Direct Die setup is where the hot side of the TEC's is water cooled and the cold side touches the CPU directly via a cold plate.
Attachment:
62-1.jpg


A Water chiller has two SEPARATE loops one to cool the hot side of the TEC and the other takes the CPU's load to the cold side of the TEC.
Image

Both have advantages and disadvantage

Disadvantages
Direct Die: limited in size and therefore limited in Max QMAX

Water Chiller: When cooling below dew point not only does the cpu socket need to be insulated but also the tubing and the chiller it's self
Water Chiller: Two loops are required so more pumps and piping .... and therefore cost is greater to implement.
Water Chiller: Because the extra Pumps and so on, more space is required to implement a chiller.
Water Chiller: chillers are more complex than Direct die blocks and therefore normally more costly.

Advantages
Direct Die: Less space is required
Direct Die: Less complex and therefore cheaper
Direct Die: Less condensation issues

Water Chiller: Because space is unlimited Qmax is unlimited
Water Chiller: Chillers don't become obsolete when a new socket is released by intel.

After reading this you might be thinking why would anyone want a TEC chiller. If loads are relatively low a Direct Die block would be the right answer. However the only advantage of a chiller is a BIG one Unlimited Qmax, which means you can get the results you want and/or while using a minimum amount of electricity in the process.

---

Q. Where can i buy TEC Direct Die blocks and Water to Water chillers?

A.
I make some TEC Direct Die blocks and Water to Water chillers and you can get them here

http://thermo-electric-cooling.com

Image


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 14 posts ]  Go to page Previous  1, 2


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  

cron

Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
CoDFaction Style by Daniel St. Jules of Gamexe.net