I am considering doing an in-depth ananysis into the heat transfer of our beloved PMD's.:rolleyes: (I have to do a project for school and can't think of anything better) I would like to find out the optimum surface area and convection rates for best heat dissipation.--Maby for future production or general knowledge if one is readily available--What I need to know is a maximum heat generation rate for this darn thing (PMD). If anyone has this or any additional information/ideas, it would be greatly appreciated. Thanks for any information.

BTG

How in depth are you going to take it? To really get into it, you'll need some backround or someone with background in thermal dynamics, chemistry (specificly specific heat of a certain few materials), and a little electrical engineering.

Hi there, engineer here, can I be of service?

so another finned heat sink? should use an alloy other then aluminum.

beryllium copper maybe, with a fin pattern.

somone here knows what the max heat is of the PMD is. that will be necessary.

Thanks for the quick replies. I am still unsure how far we are suppost to go with these projects. The prof hasn't exactly told us all of the guidelines, just to be thinking of things. The project is for my Heat Transfer and Fluid Mechanics lab.

Ok, I have finished Heat Transfer (aced it), and I am now in Advanced fluids, which sucks by the way, all matix calcualtions. FLuids will not be that involved in a problem such as this, all that is need to be known for the fluid is the density, kinematic viscosity at the under hood temp and the velocity to find the Reynolds number.

I am guessing that you will just want to look at the fin section. You will be dealing with extended surfaces obviously.

The under the hood temp will have to be known as well as a fluid velocity of the cooling medium, beeing air. Also, the conductivity of the material, the area of contact with the PMD, and the dimensions of the heat sink including the fin height, width, length, thickness, and there could also be a tapper in the fin making it like a triangle.

The best way though to start the analysis is to keep it simple and make assumptions to remove all of the extra variables.

Hi there, engineer here, can I be of service?

so another finned heat sink? should use an alloy other then aluminum.

beryllium copper maybe, with a fin pattern.

somone here knows what the max heat is of the PMD is. that will be necessary.

I'll second beryllium copper. The high performance heat sinks for PC CPU's are commonly made of copper alloys; they usually replace the factory aluminum ones.

Ok, I have finished Heat Transfer (aced it), and I am now in Advanced fluids, which sucks by the way, all matix calcualtions. FLuids will not be that involved in a problem such as this, all that is need to be known for the fluid is the density, kinematic viscosity at the under hood temp and the velocity to find the Reynolds number.

I am guessing that you will just want to look at the fin section. You will be dealing with extended surfaces obviously.

The under the hood temp will have to be known as well as a fluid velocity of the cooling medium, beeing air. Also, the conductivity of the material, the area of contact with the PMD, and the dimensions of the heat sink including the fin height, width, length, thickness, and there could also be a tapper in the fin making it like a triangle.

The best way though to start the analysis is to keep it simple and make assumptions to remove all of the extra variables.

What I was hoping to do is treat the PMD as an isolated item, ie: remotely mounted. This (I thought) would allow me to analyze it better/easier. From this and hopefully knowing the maximum heat generation I could then find an adhensive that would allow this heat to be transfered through it fast enough to get everything to the fin where stuff actually takes place. The area should just be that of one side of the PMD. (Again trying to simplify) Air density shouldn't be much of a problem.

What I am curious to find is if there is an optimal flow rate of the convection fluid (air) for this type of application. Was thinking along the lines that a faster moving fluid would effectively reduce contact time and therefore become a henderance.

From all of this information I would hope that all of the fin properties could then be optimized, fin height, material, etc...

High velocity is a must, you need to generate turbulent flow to remove the fluid boundary layer which inhibits heat transfer, so the more velocity, the better.

Also, heat conductive jelly is used on my PMD and heat sink that I use. Also, the heat sink is about twice the size of my PMD.

Thanks Chicago. Like I said, this is not a guarenteed project but one which I think is plausable. Do you know if a bad PMD will still generate heat? If it could, I might be able to aquire one for free from a generous person here at the place and do trials in the wind tunnel.

I think it does. Not sure, but pretty sure it does.

I do know that someone here did a small sutdy of the average PMD temperatures.

This is actaully a really simple project. I will dig out my book tomorrow and look up the correct formulas and correlations.

simple = good

I'm not quite sure what all you mean by formulas and correlations though. If it is just fin theory, I should have all of that in my text book. Although any extra help is appreciated in pointing me in the right direction.

Anybody here read Turbine Doc's "How hot does it get" thread....................

Anybody here read Turbine Doc's "How hot does it get" thread....................

that is what is needed, awesome!

Just be sure you reference it - the work was done here, first......................

Well, GM and Stanadyne did it first, then an engineering firm in Sweden, then .......

Well, yeah, the first time it was done here, it was done here, first.:confused:

Cool project, BTG ... you're going to have to determine a few things first:

a) thermal flux of the PMD - how fast the PMD generates heat (basic calorimetry curve stuff). It's not just how MUCH heat is generated, it's how much heat per time unit. Obviously, that curve affects your choice of conductance and absorption materials, and influences the selection of heat-transfer media and surface configuration for dissipating it.

b) Heat absorption of the PMD itself during usage. All the heat doesn't go out through the heat sink on the bottom ... some is absorbed into the body and radiated from the (non-contact) surface. This should be a curve up to the max rate of radiation. If it continued to climb past that point, the damn things would internally fry no matter WHAT heatsink you used. The question is, does that surface radiation become surface accumulation when the PMD is mounted in the hot engine valley?

If you can determine the static flux rate, you can determine the mass/surface area ratio for heat-sink materials in different configurations.

What a cool (pun intended) idea for a thermo lab! You get to work on something personally valuable (like THAT ever happens...). Nice!

I have access to FLIR heat imaging cameras if you'd like any pictures taken; sure could provide for some neat visuals.

Remember, here - even on the hottest day, IP internal fuel temperature seldom gets above 140deg, maybe to 150deg is the highest I've ever seen it at lower fuel levels.

Part of that temp is the thermal input from the PMD - that's why your owner's manual suggests that you never repeatedly run with low fuel levels.

Other times it's 120-130deg - the problem is the sudden spike in engine bay temps to over 300deg when you kill the engine.

And, that spike occurs whether the FSD is mounted on top of the intake manifold, or on the firewall, or the fender - most areas in the engine bay, because of the exhaust manifolds and turbo, with avg EGT of ~650deg.

Avg engine bay temps running down the hi-way are ~150deg in some specific areas - other areas can be much hotter.

Also don't forget to factor moving or forced air as an aid to convection.

Last run I made to Jax. Fl, MT2500 scanner showed 147F fuel temp 3/4 full tank. So if fuel is the as designed coolant best it could ever be is 140+, assuming 100% heat transfer from driver to returned fuel, then mount the IP to a hot hunk of iron, (the engine block) which will reduce heat rejection rate further, then underhood heat 250F + at times, some more transfer inefficiency (what was GM thinking)

more background from FAQ http://dieselplace.com/forum/showthread.php?t=39436

also some thoughts from my tests & trials which led me to believe out of engine bay is best http://dieselplace.com/forum/showthread.php?t=10778

I believe the circuits ability to create it's own heat via transistors switching off/on, and inability to shed it sufficiently especially if anything is out of whack (poor grounds, loose transistor captive nuts, low fuel flow to IP, loose mounting, etc.) is what led Remar-Q to it's redesign of the whole circuit independent of heat rejection ability of existing driver design.

It hasn't been a bed of roses for them, early on it had it's own issues, I don't know of their current status for reliability, study ongoing at the Page, looks like from that reporting there has been an improvement on the RemarQ, but only to the point of a 5 yr warranty, if truly best solution I would think it would match competiton, or they have some doubts it could last as long.

I'm all for a review, technology changes all the time so maybe a better way to skin the cat, or "save the driver", kinda like a Green Peace chant "save the whales". "Save the driver" rallys, yeah that would be cause to take up, will be a challenge though to "study" the problem and come up with much better solutions than what are already on the table, also $$$.

The data I gathererd was a low $$$ way to get it, and only one scenario (in that realm not a quantitative science experiment or data gathering exercise), but is what I'd consider to be one of worst case; crawling at low speed, with low airflow on a hot summer day, can't get much worse than that IMO, driver solution I run now has done the same test multiple times and is still going, I can't argue with success to date almost 3 years without a hiccup.

Thanks for all of the info. Yes citing material is a must. If this project gets off the ground and actually happens, I'll have to find a way to attach the report--or parts of it--in here somewhere. It would be one really long post otherwise. (Lab reports are usually 20 pages or more)-:t

I still have my thermocouple/meter setup let me know what info I can provide, pdf files & xcel sheets are now capable of being posted to the site

High velocity is a must, you need to generate turbulent flow to remove the fluid boundary layer which inhibits heat transfer, so the more velocity, the better.

Also, heat conductive jelly is used on my PMD and heat sink that I use. Also, the heat sink is about twice the size of my PMD.

What about using a polymer compound instead of copper beryllium? Since the theory is to remove heat quickly aren't some of these more "transparent" ?

The idea would be to use the best material for the application. There are plenty of options but it would be unnecessary to use a more expensive material that can move/dissipate the heat at a higher rate than the PMD can generate it. Albeit this would be an unlikely senerio. Other considerations would be corrosion resistance, brittleness, machinability, etc for an all around best material. If this thing were to come together nicely, I could make a Pro-E drawing and throw it into a CNC machine and give it a real test trial this summer.

The people in my lab group decided to do a project on lump sum system analysis instead.:(