heat transfer question

Unfortunately not really... see below.

Ok, it's Friday (and work is slow), so I mathematically modeled it. Here's the numbers I used:

kglass=1.14 W/M-K kalum=167 W/M-K kair=.03 W/M-K k_oil=.1 W/M-K

tglass=.125/12/3.27 meters talum=.5/12/3.27 meters tair=.25/12/3.27 meters

Pr (si_oil)=10 v_oil_wall=2"/sec dynamic_viscosity_oil=.02 kg/m-s density_oil=920kg/m^3 Re=415 Nu_oil=22.7 = 3.657+.19*(Re*10*(7"/10"))^.8/(1+.117*(Re*10*(7"/10"))^.468) h_oil=12.7 W/M^2-K h_air=6.3 W/M^2-K

And here are the results:

For no internal fins on top alum plate (.25" air gap): Qtop=25.4 W Qsides=224.6 W T_alum=186 C T_glass=274 C

(Wow!)

That's REALLY HOT. In regards to my earlier calculations, I realized I wasn't consistent with the units, and those calculations are off...

For internal fins in the liquid:

Qtop=48.5 W Qsides=201.6 W T_alum=333 C T_glass=248 C

So, adding internal fins to the top will decrease the temperature of the sides SOMEWHAT, but will increase the temperature of the top... it's a trade off. This also makes sense cause the area of the top is only 18% the area of the sides.

On the other hand, increasing the surface area (12" dia by 15" tall) yields the following numbers:

Qtop=33.8 W Qsides=216.4 W Ttop=97 C Tsides=118 C

...which is MUCH cooler. This equates to a 260% increase in surface area, which is very effective.

Feel free to point out any errors I might have made.

Dave

BTW, this model doesn't include radiation. I would expect it to drop the temperatures a few dozen degrees for the sides (e=.95) and some for the top (e=.25 approx, unless it's anodized). I'll see if I can include the radiation component later...

Dave

///

Quite so, and to be even blunter, if you have no RF power design exposure, designing on the basis of holding insulation margins is fraught with unexpected consequences.

Still, as it happens, common-sense often carries a prudent engineer over enemy territory successfully.

Brian W

////

There's only one error in your modeling that is crucial: if you have good thermal conductivity into the head, this heat flow is easily removed convectively: either passively with exterior fins, with blown air, or with heat pipes.

Or to put it most succinctly:

for a thermally connected head, its temperature may be reduced to a small increment above ambient (or the temperature of the chilled forced air if used) as desired.

Brian W

It apparently does not model the effect of increasing heat flow out of the head by one of several available means

Brian W

Forced convection and/or external fins wasn't part of the original scenario that was presented, so I wouldn't call it an "error". Forced convection and/or external fins could be included in the model to see what happens... unfortunately I'm in Phoenix now, and don't have the program needed with me.

Dave

Correct. I wasn't aiming to model those scenarios at the time... just the original scenario, a scenario with internal fins in the oil, and a scenario with a larger container.

I don't have Matlab with me right now... but feel free to model it for me. :-)

Dave

Tsk, tsk: Dave, you were responding to my note [below], where you denied that a blown or finned head would provide as much cooling as wanted. Better if you just say something like: "Oops my mistake" ...as I found out the hard way, some years ago....

Brian W

You clearly said "dips fingers in the oil". I modeled this scenario. If you're not sure about that, then go back and read the post in question.

However, when you said I did not model the effect of increasing heat flow through several available means, I assumed you meant "either passively with exterior fins, with blown air, or with heat pipes", as you had stated in an earlier post. And, as I said above, I did NOT model these scenarios.

Don't confuse internal fins (that dip in the oil) with external fins (on the top).

On a side note, if by some misunderstanding you mean the increase of heat flux from the top due to increasing surface temperatures, then I did take that into account.

Dave

Holy gamoly,

I didn't expect all the effort on this but I certainly do appreciate it! Having been out of town for the last few days, I will need to go through the details for a more complete understanding on what you guys have done. Excellent!!.

clay

David Harper wrote:

Ed,

There are certainly some concerns here, high voltage being one of them. This is a piece of audio equipment. The high voltages in question are isolated in the system enclosure. Something we have significant experience with. The difference this time is that the entire electronics portion (not including power tranformers) are enclosed in a sealed chamber filled with dielectric fluid. Hence the heat transfer topic.

Your concerns regarding the high voltage are definitely a consideration however.

clay

Ed Ruf wrote:

Thank you for all the commentary & advice on this topic. It has definitely added to my understanding. Anyone interested in seeing what this product looks like can send me a direct email, and I will send a JPEG to you. Or alternately I think I can provide a web link.

thanks again for all you have helped.

clay

Brian Whatcott wrote:

I can add a substantial amount of surface area to the top by integrating heat sink fins. Upwards of 800% more surface area (within reason). More than that if necessary. I "may" also be able integrate a low speed fan in the top as long as it is not cosmetically or audibly invasive. the potential customers are absolute purists, so a fan may be "right out" however.

I cannot increase the diameter of the chamber at all, but, time permitting, I can increase the height. Increasing the wall thickness of the glass is not possible with this specific glass diameter, but using a larger diameter cylinder in the future this will be possible.

Another factor touched on here. or possibly a large wrench. We have the option of replacing the glass cylinder with thicker .250" wall acrylic. If we do so, based on Davids numbers previously and comment on thicker insulative walls, will this significantly cause the top aluminum plate to be the primary heat conductor/convector? Or is the acrylic going to get so hot that it just expands substantially and/or cracks/melts etc.....

Then, if the top is extracting a majority of the heat, is it capable of extracting "enough" to keep the inside surface of the acrylic below its failure temp. the softening temp of acrylic is way down at 112 C.

clay

Dear clay:

I wouldn't recommend the acrylic. Yes it is an insulator, so internal temperatures will go up. And if you don't modify the top cap, the outer surface of the acrylic will get about as hot as the quartz does now.

Increase the area. Or better still, add a fan per Brian's suggestions.

David A. Smith

Adding fins to the top will help some, but from a control volume standpoint, the heat flux that goes through the top only occurs through 18% of the control volume's surface area. Imagine trying to cool a hot glass of water only from the top. It's not as effective as cooling it from the sides (if possible).

Is it possible to use aluminum on the sides with vertical fins? That's probably out of the question, but that would help.

Dave

David,

I sent an email directly to you, apparently you did not recieve it?

It is preferrable to put a huge heat sink head on the top, and I have tremendous latitude in the thickness and surface area of the top, but I understand your analogy. We can add verticle aluminum fins to the sides, by bonding with a conductive adhesive, and may have to resort the that at this juncture. we have also considered making the upper chamber half aluminum , and the bottom glass. But would prefer not to have to go that route. I'll send you a picture, and it will make more sense why.

Can you provide the formula you used in modeling? and/or are you available on an hourly basis for consulting on this? We have a small but reasonable budget for doing so. Respond directly for contact info.

clay

David Harper wrote:

On a side note, with radiation the numbers for the first scenario become:

T_alum=119 C (down 67 deg) T_glass=128 C (down 146 deg)

That's more significant than I would have anticipated.

Dave