Interesting data, thanks for posting it. Why do you think you might be
50% low? You have one extra grease interface not accounted for, but even so your measurements are within the range of published polycrystalline AlN conductivities, with Accuratus claiming 140 to 180 W/m.k, Azom claiming
60 to 177, and the abstract of the book "Thermal conductivity of polycrystalline aluminum nitride (AlN) ceramics" claiming 17 to 285 (compared to 319 for a pure single crystal). What did your supplier claim for conductivity?
I've contacted two companies on MY quest for larger print! One was a company that makes shampoo and conditioner. I use whatever is in the caddie, but my wife changed brands, in the shower I could not tell shampoo from conditioner, so I have just used hand soap for several years. When my wife came dripping wet out of the shower to ask me to read the label, something needed to be said. "hey nice (.) (.) " Er, I mean, Dear shampoo company, Make your print larger! She got $18 worth of coupons, for my letter. The second letter was to a company that makes a microwaveable, curried chicken and rice. I could not read the cooking directions. I've seen no action from either, buy I'm sure that's a bit of a process.
Does your Chinese supplier sell them retail, or was your batch a special? I have this high power density laser gizmo going, and could use a leg up with the thermal design.
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510
http://electrooptical.net
http://hobbs-eo.com
The plausibilty isn't impressive. The thermal resistance of the AlN pad is being claimed to be the difference between 1.3K/W without the pad - which i ncludes one greased copper to aluminium block interface, and 1.7K/W with th e pad in place which includes two greased interfaces.
The implication is that two greased interfaces in series each have a lower thermal resistance than the single one with which you started off.
This can be understood if neither the end of your copper rod nor the surfac e of your aluminium plate is as flat as the carefully ground upper and lowe r faces of your AlN insulator, so both layers of grease are - on average - half as thick as your "reference" layer.
I think I got a lower thermal resistance contact with a layer of carbon fib re cloth back in 1993 (but the area involved was probably larger).
Your customers will have to be just as gullible as you are to take this all that seriously.
If you want to measure the temperature drop across the insulator itself, yo u need sensors glued to either face of the insulator, and gapped from the c opper rod and the aluminium plated by little wells drilled in both surfaces .
The setup is pretty crude, and I expected bigger numbers. I did this to be sure that the stuff is really AlN; I've bought so-called AlN that isn't.
Neither claimed anything; they just say "aluminum nitride."
Most people spec numbers like 150 to 170. The machinable stuff has some boron nitride mixed in, so is usually spec'd around 100.
I got 100 of the Chinese TO-220 parts custom made for $2.20 each. I've seen quantity pricing below $1. Domestic pricing is often in the $25 range.
I have a circuit that needs a high ratio of thermal conductivity to capacitance. BeO is a little better but has the ROHS problem. Solid diamond is best.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
It is not! That's a rubber band! It's pretty ugly.
I sent them dims and they made the parts for me in a week. Really nice work.
formatting link
I could send you some of the TO-220s if they would help, they are 60 mils thick. I also have some bigger square slabs of thinner stuff. Thinner is better unless capacitance matters.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
I feel your pain. I can't read much of anything in the shower. I'm ok in my shower because I don't need to read, but anywhere else I can't tell what is what. They label shower products to sell from store shelves, so the br and name and flowery adjectives are printed very large and the "shampoo" or "conditioner" are very tiny. They often use color to distinguish them, bu t you have to memorize that. I'm tired of memorizing crap! Then I think a bout my 91 1/2 year old friend who is legally blind (is there such a thing as illegally blind?)... I got nothing to complain about, but I don't need to let the shampoo makers know that.
In the kitchen I have the same problems. Even with glasses it is hard to r ead the directions on some products, but they don't actually make it to my kitchen. The ones that tick me off are the ones that write the ingredients or nutritional data far too small or in very indistinct colors to be read without a magnifying glass (I'm talking about YOU Haagen Dazs and yogurt ma kers). Worse are the shelf labels with the unit price so tiny I have to ge t within a foot to read. I'm not bending down to get that close to the bot tom shelf, so I peel the label off and bring it up to my eyes instead. Yea h, it usually doesn't get put back as well as it was attached, but not much I can do about that. Bending down to get that label used to be a lot wors e until I got the hip fixed.
No, the world is designed by people with good vision and the rest of us hav e to cope as best we can.
I always wear contact lenses, and the doc says my vision is corrected to 20/25. one eye is a little flaky. I was on the computer wearing readers, so I could easily read the label on the shampoo for my wife. I just went and got two bottles out of the shower, a 6.8oz and a
15.9oz. The 6.8oz bottle has 0.075" tall letters, and the 15.9oz has
0.085" tall letters. I can read them both without my glasses, but this is after I found out where to read, knew what it said and I have good lighting. The odd thing, both bottles are conditioner, and there is no shampoo in the shower. I would think she could tell the difference between shampoo and conditioner by the feel. Btw, the Curried chicken with rice, has 0.054" tall letters and they are gray on white. Mikek
For product bottles in the shower I've found colored electrical tape work s well. Red for shampoo blue for conditioner and so on. Magic markers work too. I 'll use them for marking the tops of pill bottles.
The silicone grease, on not-optically-flat surfaces, probably adds theta. Maybe I need a better rubber band, or (seriously) a better mechanical setup. I did calculate thermal conductivity based on the contact area of the 1/4" copper post. There's a bit of fringing that I ignored.
In the real product, I'll have machined surfaces (not ground flat) and silicone grease, so this is a sort of reasonable model of my real situation.
I don't really need to know the thermal conductivity of the AlN; I need to know if my board will work.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
But you might need to know if it was worth grinding - or lapping - some of the mating surfaces involved.
formatting link
Machinists can make essentially optically flat metal surfaces (as in Johan sson blocks, which can be wrung together). If you'd bothered to have that d one to the end of your copper rod and the top of the lump of aluminium you were using as your heat sink, you might have a better idea of what was goin g on, and how you might be able optimise the product you shipped (perhaps b y spending more money).
I didn't mean it was good looking, just that those are pretty cool insulators.
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510
http://electrooptical.net
https://hobbs-eo.com
It's a pity that John Larkin couldn't characterise them more precisely.
It's nice that they will probably do the job he wants them to, but the two layers of heat-sink grease that he has built into his measurements make the results less clear than they might be.
Properly lapped mating surfaces might do even better.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.