Well, I made my first attempt at a PMD cooler. I must say that the results were not as I had hoped. Casting seems to be more of an art than a science. Here are some pictures of what the "rough" casting looks like. In addition I posted some design ideas for future castings. Questions, comments, and concerns are alway welcome.
http://www.dieselplace.com/forum/uploads/tahoe2dr/2004-10-07_150116_casting1.JPG

The pattern for the standoffs colapsed during the pour... I have a new design apporach using a ceramic refractory.

http://www.dieselplace.com/forum/uploads/tahoe2dr/2004-10-07_150133_casting2.JPG

The area between the fins was filled by the pressure of the metal... I need to machine it out.

http://www.dieselplace.com/forum/uploads/tahoe2dr/2004-10-07_150154_heatsink.jpg

http://www.dieselplace.com/forum/uploads/tahoe2dr/2004-10-07_150203_heatsink2.jpg

http://www.dieselplace.com/forum/uploads/tahoe2dr/2004-10-07_150722_heatsinkwire.jpgEdited by: tahoe2dr

That's not too bad, you can probably resurface the area for the driver module by hand.


Actually, I like it alot. It's good for a first attmpt. Very creative. Edited by: quantum mechanic

Thanks QM.
I was hoping to see more replies... everyone has their own unique idea.
I test fitted the cooler last night and found it to be way overkill.
The PMD is only about 1/4 the size of the cooler.
So, the next one is going to be about 1/3 the size.
Also my standoffs were not properly spaced relative to the trans cooler bracket.
I think it is the trans cooler bracket (the one on the passenger's side).
The thickness, including fins, is also to large to allow it to fit behind the grille without modification.
I will go after some high density foam this week and make trial no. 2.
For a final finish I thinking of bead blasting and spraying it with just a falt black paint.
This should pretty well camoflage it in relation to the radiator.
QM, let me know if you would like to be a tester for one of the advanced prototypes.

without something on it or beside it to judge it's size I was thinking it looked bulky. I'm not studied in the heat transfer area so I don't really know if the fin or overall size help with the heat transfer issue. I would think that the number on fins would make a difference but may not be enough difference to worry about. I've seen alot cruder and less refined work outstanding right QM. http://www.dieselplace.com/forum/smileys/hihi.gif Edited by: bowtie

Hey t2,


I'd be happy to test it, but don't paint it black!!!

paint will definitely hinder heat transfer ability, I peaked on a radiator shop for painting my rad silver, they should know better. Die grinder, Sand blaster and wire wheel is what I suggest to clean it up, leaves a nice finish too.

in my opinion, id say your prototype is, as you say, bulky. your best bet with designing a heat sink is alot of surface area.you also need to force convection. i dont know any of the numbers on heat output. but id definately put a cooling fan on one side of it. you dont need the mass that is present in your design. the mass will help intially, but as time passes, it will not help. keep the size down and consider a cooling fan. i know you plan on the the motion of the truck to cool the pmd, but in my experience, it is when you stop that the temperatures rise and the electronics malfunction. my cooler is a small peice of finned aluminum and i mounted a computer cpu fan on the back side of it to force air flow over the fins. no problems since and its a black truck and im in florida. no problems all summer

Actually the fan on the engine provides enough cooling that the only heat the driver see's in front of the radiator is after shutdown and it's out of the soak!

Actually, QM is right.
With the engine running there is an ample amount of air being drawn over the cooler by the engine attached fan.
I had not really considered the air caused by movement to be a good source of cooling.
Yes it might be alright at about 30 mph, but any lower and you would have a problem with lack of airflow.
Yeah, it is way too bulky, but that is why it is a prototype.
According to the calculations the new design has over 220 square inches of surface area.
The thermal conductivity of aluminum 6061-T6 is 155.80 W/m*°C.
The thermal conductivity of water is 0.6030 W/m*°C.
Aluminum can transfer heat faster than water.
So, the only limiting factor is the air... hydrogen is like .06 something...
But with over a square foot of surface area exposed... it shouldn't be detrimental.
The mass shouldn't hurt too much, but it does need to be smaller.
Aluminum can dissapate heat very rapidly, but is dependant upon the potential difference.
I will get another casting done and take pictures of it next to a ruler.
I still think it looks neat... although I am a firm believer of function before form.

What are you using for aluminum in your pour? What are you using for a pattern? The casting sand looks a little coarse. Do you have the mill to finish it out? Over all it looks pretty good. Great pictures and drawings. Good job! Might want to sandwich the PMD between two heat sinks. No need in leaving one side bare if your going to remote mount. Leave some area in your casting for bolt hole bosses and you could just drill all the way thru and use screws with nuts on the outside to eliminate having to tap the threads.

As beefy as that looks ever considered plumbing water to it maybe put a copper tubing in mold with several U bends in it before pouring in Aluminum, put a little 12v circ pump on it with a timer to keep it running for a bit after shutdown.


As far as size relation to FSD size to heat sink; the Heath kit I have, the FSD is centered on a 6 x 12 plate, superconductive alloy is what draws away heat from FSD.Edited by: Turbine Doc

I used a recycled 6061 Al. The pattern was a "rough" (very rough) design cut from high density foam. The sand is a dry silica base. The sand actually has a very fine texture. What you are seeing in the pictures is the impression left by the reough shaping of the foam. I used a band saw to get some idea of what I wanted. I work in a machine shop so I wanted something that I could machine. I would like to surface grind the PMD contact area, drill and tap both the PMD mounting holes and the standoffs.


I actually considered the copper tubing, but just as a means to more evenly distribute the heat load. As far as fans, circulation pumps, and other gizmos... I want to stay as far away as possible. The design objective here is to produce something that is easily installed, asthetically pleasing, and maintanence free. Well, as maintanence free as an FSD can get. The 6x12 plate you refer to, is it mounted outside the engnie bay? I would think that might be a little over kill... but obviously someone ran the necessary tests and found that size to be optimal. Do you know what the superconductive alloy is?I was seriously considering a piece of copper in the mold cavity to serve as a ready-to-go surface for mounting the PMD. If I were to mill dovetails in it or something then it would have a good mechanical bond. I don't think the heat transfer would be all that great though. The other option was to use a lead alloy to solder the two materials together. However, all of this gets away from the intial design criterion of being relatively simple. I do like the water cooled idea though. I might have to run with that one. Although I am sure that would be overkill. Good brainstorming though.

actually if you look closely you can see where the bandsaw blade cut the foam.

The 6x12 plate you refer to, is it mounted outside the engine bay? I would think that might be a little over kill... but obviously someone ran the necessary tests and found that size to be optimal. Do you know what the superconductive alloy is?


Yes mounted out of the engine bay, mounted here can't get much more airflow than this http://www.dieselplace.com/forum/forum_posts.asp?TID=16035&a (http://www.dieselplace.com/forum/forum_posts.asp?TID=16035&PN=3) mp;a mp;a mp;PN=3 the faster I go (officer I'm not speeding, if I go slow the truck will shut down when my FSD overheatshttp://www.dieselplace.com/forum/smileys/hihi.gif) the more flow I get, and I've not taken away any cooling or blocked cooling from anything else.


Don't know what the alloy is, that is proprietary info, but I'd bet its real close to what they use in those thaw plates you see in late nite infomercials, piece of aluminum looking plate that "thaws" a frozen steak in about 10 minutes. My guess is the 6x12 plate size is to pull away heat when there is no or minimal air flow, from my tests this is critical, as the other "coolers" out do a good job so long as they have air flow across their fins.


They work better out of the engine bay as when engine is shut down and no air flowing under hood no way for heat to be exchanged if under hood temp is greater than the "cooler".Edited by: Turbine Doc

I posted these pics before but they didn't turn out the way I wanted, I'm going to try again. My cooler originally had fins on both sides so I had to machine them off which gave me an added benefit of creating a bare aluminum surface to mount the FSD to. I mounted it to the trans cooler brackets. I like that spot because even with the truck not moving the fan draws air past it. I drilled the holes to allow some air to pass through as opposed to all the air being forced around it. I idled my truck with the air conditioning on for about a half hour in 75 degree weather and when I touched the FSD and cooler you could barely tell that they were a little bit warmer than the ambient air temp. I didn't measure the actual temp but Id say well under 100 degrees. Next time I'll measure the temp, I have one of those cool infrared temp measuring thingies, Just point and shoot.http://www.dieselplace.com/forum/uploads/Bobt250/2004-10-12_064708_IMG_0125.JPGnot http://www.dieselplace.com/forum/uploads/Bobt250/2004-10-12_064738_IMG_0126.JPGEdited by: Bobt250

So TD the Heath kit plate has no fins. It's just a 6X12 inch slat made from any unknown metal? Interesting, I didn't know that till now. It makes that picture behind your front plate make my sense to now.

Hmmm, the Heath kit has no fins... that would lead me to believe that the FSD does not generate as much heat as some claim. I know that aluminum dissipates heat very rapidly, but it does not radiate heat all that well. To make that a little more clear... if you had a piece of steel and a piece of aluminum, and heated them to the same temperature, the steel would "feel" warmer without actually touching it. The aluminum on the other hand would require you to actually touch it, or be very near to the surface to get and accurate representation of the temp. With that in mind, know that when heat is induced into a piece of aluminum it releases this heat energy in a more controlled fashion. This is true to about 500*F anyway. I would think that the lack of cooling fins would hinder its ability to dissipate the heat load. also the surface area is only 72sq.in. x 2 - the area occupied by the PMD. My design has over 200 square inches on the "cooling fin surface". Which is really making me think about thinner section widths, smaller overall size, and shorter fins. I would assume if I can achieve a casting with minimal porosity and a fairly fine grained surface, that 140 sq. in. would surfice. I will definately crunch some numbers on this and let you know what I find. As of right now, all work on the cooler has ceased until I find the source of my stalling problems. Should have a solution soon though.

On a side note,
Does anyone have a caliper and a loose PMD that could post dimensions? I am interested in WxLxH and corner radius, as well as the dimensions and offset of the connector. Thanks

t2d, your design is very nice, but to cast it like on the drawing, you'll need a form that is in two halfes, upper side and lower side.

Such forms are usually made of steel (which may need to be hardened after the form has been cut out), eventually electrically heated and/or water cooled inside. A great deal of experience and some testing is required for making the metal spread quickly and evenly enough in the form during the casting. Provisions must be made to get the metal out of the form once it has cooled down.

I've once asked a collegue of mine --- he actually manufactures casting forms and such --- how finned coolers are made. He told me that they are neither cast, nor cut out of a block, but drawn (like wire).


> Casting seems to be more of an art than a science.

That is very true!

Well, it isn't as much fun as casting, but buying a sizeable finned cooler and starting from there is easier ... But I'll see if I can ask my collegue for some suggestions on casting tomorrow.

A self-cast PMD cooler would be a very cool thing :)

T2 yes I have an old one why don't I trace it and you will have a exact template for sizing; PM me your fax # and I can send it. Me & GMCTD going to do some study on it maybe fix it next time I get over his way, it does get quite hot; go to my how hot does it get post I have some data taken while in operation of the transistor cavity with a thermocouple touching the transistor.


Yes a flat plate, unknown metal to you or I; not Bill, he says it's last one you will ever need, I have not heard differently from anyone else using one before I gave on using a finned one.

Yeah, you could always use an "extruded" piece for your heatsink, but where is the fun in that? I also wanted something with standoffs to mount it. As far as the mold being two pieces, yes if it was a permanent or reusable mold. Since I am not in this for mass production, lost foam casting works just fine. All I have to do is hape the piece I want out of foam and bury it in silica sand. When the metal is poured over the foam the foam burns away leaving you with a metal impression of what the foam looked like. It's a really neat, and smokey process.

OK, I'll look out for advice on casting, especially casting fine structures like fins, what to use for the mold, what kind of aluminuim might be best and the percentage the metal will shrink when cooling down. I'm pretty sure that there's a whealth of knowledge among my collegues, since they're dealing with mass production of casted aluminium parts that have to comply to quality standards for decades now. Something interesting will show up :)


For the standoffs, my impression from your pictures and from the casted parts I happened to see at work is, they are very likely weak spots because they are somewhat 'strangulated' from the rest of the material. After casting, they'll cool down at a different rate than the rest of the material. Also, my impression from the patterns that show on your pics is that you poured the hot aluminuim into the mold too slowly. The metal looks like what I'd describe as 'uneven in itself': The metal poured in first cools down somewhat before the rest of it is poured in, and thus you get kind of 'layers' within the block.

Well, I'll ask about that, too.

Try using forged pistons for you melt. The pistons have a mixure of alloys that make excellant donor material. Stay away from cast pistons or reusing old melted aluminum to get best results. Auto transmissions also have quality alloys in them. 6061 aluminum is good stuff but for casting you need the extra alloys that get lost in melting aluminum. Pistons seem to have more than enough alloys in them.


Your ""lost mold" foam castings look pretty good. I know you put a lot of effort into it. Make sure your stand offs are bolted to something cool or they might end up putting heat into your cooler instead of out.


The PMD can generate heat in excess of 200 degrees within minutes if the truck is worked hard.

any updates on this cast cooler?

I've been asking around and ended up with no actual knowledge on sand casting, but with some tables describing various kinds of aluminium alloys. Those tables can help to find an alloy that is especially suited to the purpose.

I'll just need to find some time to study the tables before I can come up with suggestions. Sorry that this isn't much, much less than I expected, but it's all I could get.