Science desk
< November 1	<< Oct \| November \| Dec >>	November 3 >

Welcome to the Wikipedia Science Reference Desk Archives
The page you are currently viewing is a transcluded archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

November 2

Fluid with thermal conductivity over 10 W/m.K

Normally my actual cooling system works with water or something having a similar viscosity as water.

However the device to be cooled is so hot for the given surface (100 W/cm²) that the heat can t be transfered efficiently using water, even using a fast compressor. I know about cooling oils as fluid as water like Fluorinet but it turns out their thermal conductivity is worst than water.

I also checked using Aluminium substrated Peltier, since the surface is small, but if I understand correctly, Peltier modules don t works if the cold side temperature is above the host side.

Also, the coolent needs not to react with copper which excludes mercurial. It s so much heat that the external side in copper needs not to exceeds 50 Celcius degrees in order for the junction not to exceed 120 degrees which means this isn t hot enough to use sodium as a coolent. 2A01:CB05:8411:3C00:B908:1E54:561F:17A9 (talk) 09:39, 2 November 2020 (UTC)[reply]

Useful references are Water cooling and Thermoelectric effect (Peltier effect). Can the device hot surface area be increased by a finned Heat sink? 84.209.119.241 (talk) 10:18, 2 November 2020 (UTC)[reply]

Yes it that way that the temperature on the cooling side should not exceed 50 Celcius degrees.

A Peltier device (better link than above) can move heat from cold to hot or from hot to cold (just like a regular heat pump), depending on the direction of current. However, using a Peltier to carry heat in the "easy" direction is a tremendous waste of money in any normal circumstance, where a heat exchanger setup (i.e. use a moving fluid to carry heat away) would work better and cheaper.

Yes but I heard they are inefficient if the cold side temperature is above the cold side. The best bet would be to find a compressor no taller than 5 cm high. 37.166.221.172 (talk) 16:42, 8 November 2020 (UTC)[reply]

Your specification of 100 W/cm² seems so ridiculously high to me that I have to ask if that is not a typo. For reference that is the order of magnitude of the peak heat flux received on the walls of a gasoline internal combustion engine during the engine cycle, when the flame "impinges on" (~licks) the wall (source, slide 15). The average heat flux on the walls of an ICE is more like 2-10 W/cm², and the thermal management of this heat already requires a really large chunk of metal to spread out the heat and a complex cooling circuit. Outside a laboratory setting (high-power lasers anyone?), I do not really see how you generate that kind of heat flux without a torch, even less so how to move it away.

100 W/cm2 is not a ridiculously high heat flux. Moderns CPUs may consume 200-300 W and this heat is dissipated from the square area of merely 1-2 cm2. Ruslik_Zero 20:01, 2 November 2020 (UTC)[reply]

If the heat flux value is not a typo, I smell an XY problem, so please give more details about the problem you are trying to solve. Tigraan^{Click here to contact me} 13:23, 2 November 2020 (UTC)[reply]

It's probably important to know not just the heat flux, but the timeframe over which it needs to be cooled, the size of the device (which also tells us the actual amount of heat that needs to be transferred), and the temperature-range over which you can operate. This could help resolve some of Tigraan's XY-problem concerns. A short burst in a small area with some heating allowable followed by relaxation time is quite different than a steady-state production on a large surface that must be held at a constant low temp. DMacks (talk) 23:16, 2 November 2020 (UTC).[reply]

Die temperature should never go above 50 Celcius degrees so that the jonction doesn t exceed 120 Celcius degrees internally. Timeframe is during a month uniterrupted.

The problem is the manufacturer water cooling system isn t designed for working at room temperatures according to their manual (more exactly max is 5 Celcius degrees at 900 mètres off altitude). It s not even for overclocking at that point but to get things to work normally. 37.166.221.172 (talk) 16:42, 8 November 2020 (UTC)[reply]

OK, so you have a recirculation system set up, but it asks for 5°C water input.

Option 1, probably the safest, is to keep the setup as intended and match the required input.

There are commercially-available water chillers, usually based on a heat pump design (example). Most of those have cooling powers in the 500-2000W range so you should be fine if your generated heat is in the low hundreds. The problem is finding one that is not too expensive; the lab ones cost multiple thousand dollars but that's the price to pay for a really fine control of temperature (which you probably do not care much about). You might want to look around other applications to see if their providers have something less pricey (but still with a sufficient cooling capacity). Cold-water-fish tanks? Water dispensers?
You could make your own water cooler. Peltier cooling is less efficient than heat pump designs (in dollars per joule moved), but it is the option I would take because I am more comfortable fiddling with electronics than HVAC fluids. I found 200W-of-cooling Peltiers for about $50 apiece (example), the power supply and wiring stuff should be about the same cost or less. It probably needs to be tested first to see if natural convection is enough for the Peltiers to work, otherwise add some fancy finned heat sink on top of it. That option is therefore probably around $200 of material, but it requires a couple days of work.

Option 2 is to fiddle with the cooling fluid. The bad news is, if we restrict ourselves to conventional cooling, water is unbeatable in terms of heat capacity and conductivity »(well liquid sodium is "conventional" but it barely a worthy option for nuclear plants, so safety is a clear dealbreaker here). The good news is we can cheat by using phase-change fluids which have a very large effective heat capacity, using the Heat pump and refrigeration cycle. (Incredibly, we have an article about that but there is no reference in that section.) The bad news is that HVAC fluids are usually some nasty chemicals so it might damage the installed circulation system, and a 1h test run might not be enough to see what happens after weeks or months of use. I would guess the whole thing can be somewhat pricey, too.

Option 3 is to ditch the existing system and cool by Peltier. From an efficiency point of view, it cuts out the middleman (middefluid?) but it has its own issues. First, you will have to find a Peltier that moves enough heat, is small enough, and works at your temperature. Second, you still need something to evacuate the excess heat at the back of the Peltier - this probably means a heat sink (as 84.209.119.241 suggested) with conventional water cooling over a larger surface, but can you fit that next to the CPU? (Note that the Peltier backside temperature will necessarily be fairly hot, indeed as hot as the CPU would have been without the Peltier, so choose the materials accordingly.) Tigraan^{Click here to contact me} 10:49, 9 November 2020 (UTC)[reply]

I was talking about room temperature for 5 degrees. I didn t found a water chiller no taller than 8cm and and no larger than 5cm.

A Peltier is a good thing since the hot side can go above cpu max temperature and exchanging heat with a higher temperature from air is more efficient. But I heard it s useless if the cpu side is hotter than the hot side.

I d go with sodium if possible as it s a very dry environment but the problem is it s not liquid below target temperature. 2A01:CB05:8411:3C00:8D3E:5F93:ADEE:EDEA (talk) 20:40, 14 November 2020 (UTC)[reply]

What does Germany have the fewest of of all the countries in the world?

op is a long banned user. See WP:LTA/VXFC. --Jayron 32 18:51, 2 November 2020 (UTC)[reply]

The following discussion has been closed. Please do not modify it.

This question was posed on Ronan and Harriet's breakfast show on Magic radio this morning (answer at end of next sentence). Ronan suggested "pigs", but the answer is in fact...children. Of a population of eighty million 10.5 million are children, just 12%. Is this just a felicitous coincidence or is there a scientific reason for it? 81.139.213.56 (talk) 14:23, 2 November 2020 (UTC)[reply]

Not sure what you mean by coincidence, but the key factors are likely to be a sustained low birth rate combined with a high life expectancy. Assuming Germany being the lowest is accurate, it's not like it's an extreme outlier. I'm not sure what definition of "children" was used (but it's probably under 14 [1] [if you have problems try [2]]), but others are not that far off e.g. Italy [3] [if you have problems try [4]]. (Note it's closer to 12.6% than 12% for Germany if you're using 14 years old as your cut off.) Even Scotland is about 16% Demography of Scotland so not that much higher. Nil Einne (talk) 14:54, 2 November 2020 (UTC)[reply]

I fixed your links to be clickable. DMacks (talk) 15:13, 2 November 2020 (UTC)[reply]

Thanks but they should be already for most readers I think. Not sure why you need the http links, but you may want to look at your browser or ISP. I prefer the protocol-relative URLs since for most readers, both https and http should work. If they're visiting from http for some odd reason (which is difficult nowadays since Wikimedia has disabled support), I'm assuming they may need http for the external links too. If they're visiting from https, https would be better for security reasons. You unfortunately seem to fall into the odd category of someone visiting from https but who can't handle https for either site. (Or maybe you just have a very old browser without support for protocol relative URLs?) I've left both versions for anyone else with the same problems, but for most, the protocol relative would be a better bet. (Since https works, if I wasn't using protocol relative, I would use https so they probably still wouldn't work!) Theoretically one day it's possible Wikipedia will be available via some other protocol but the websites may not be. But despite what Wikipedia:Protocol-relative URL says, protocol relative links are still not uncommon so I assume if that ever happens we will probably forcefully convert them. I checked that https works for me, and http should work since even sites forcing https still allow http as a redirect and checked again both http (as a redirect) and https work for me with both sites so I think it's probably a problem on your end. Nil Einne (talk) 15:40, 2 November 2020 (UTC)[reply]

Thanks for double-checking! Looks like there was a proxy here that broke the link if it was https even though other https links worked fine. Craaaazy. DMacks (talk) 18:08, 2 November 2020 (UTC)[reply]

World Bank (based on UN numbers for 2019) has Germany as only tenth-least.[5]. When using vaguely-defined terms, it's so easy to be surprised at trivia or play "gotcha!" no matter what answer one gives at first. DMacks (talk) 15:16, 2 November 2020 (UTC)[reply]

I found out we have List of countries by age structure which seems to be based on 2017 World Bank (or UN figures) which still puts Hong Kong and Japan with lower percentages. Putting aside that Hong Kong is part of China and so probably isn't considered a country, Japan clearly is. Besides definition issues, the figures are very close anyway, and at that level data quality also matters .While both are developed countries with I assume excellent government statistic departments, Germany does have relatively open borders. I think we can safely say Scotland or the UK as a whole are higher, to give one random example. But whether Germany is the lowest, well that will depend. Nil Einne (talk) 15:46, 2 November 2020 (UTC)[reply]

I noticed that list doesn't include Vatican City. I'd be shocked if they didn't have the lowest percentage of kids. Matt Deres (talk) 18:37, 2 November 2020 (UTC)[reply]

The OP is a sock of a banned user who just got off his most recent block. His refusal to be a grownup goes along with his bizarre notion that a low birthrate is "felicitous". ←Baseball Bugs ^{What's up, Doc?} carrots→ 18:07, 2 November 2020 (UTC)[reply]

No objection to nuking this section. DMacks (talk) 18:10, 2 November 2020 (UTC)[reply]

Masculine and feminine energy

Is it pop science that people are attracted to those who possess the opposite masculine/feminine energies to them as this blog suggests? Or is there actual scientific evidence for this? http://artoflovingaman.com/blog/2016/3/8/uzej4t4o86xwvobz8wwakzjxl0v79b-m38n3 — Preceding unsigned comment added by 90.194.52.17 (talk) 21:17, 2 November 2020 (UTC)[reply]

Not even pop science. Simply pure and utter garbage. HiLo48 (talk) 23:27, 2 November 2020 (UTC)[reply]

Maybe kind of a lame variation of Men Are from Mars, Women Are from Venus. ←Baseball Bugs ^{What's up, Doc?} carrots→ 11:04, 3 November 2020 (UTC)[reply]

The terms "masculine energy" and "feminine energy" have no defined meaning. Therefore the statement is meaningless. Rather than opposites attracting each other, there is considerable evidence that similarity is a positive factor. See further at Similarity attraction effect. --Lambiam 14:20, 3 November 2020 (UTC)[reply]

That makes it flexible enough to accommodate any interpretation. Kind of like astrology. ←Baseball Bugs ^{What's up, Doc?} carrots→ 15:59, 3 November 2020 (UTC)[reply]

I think the term we're looking for here is Psychobabble. ←Baseball Bugs ^{What's up, Doc?} carrots→ 13:55, 4 November 2020 (UTC)[reply]

The problem is the lack of precision in the terms "masculine energy" and "feminine energy". These terms may be literal but they also may be figurative. Any legitimacy involves a dollop of poetic license. Bus stop (talk) 03:34, 5 November 2020 (UTC)[reply]