Even heating across a 4" x 4" PCB Heater

...snip...

Whoa! If you can get 30 W/mK out of solid alumina, wouldn't that work as a perfect heatspreader? Put the heatspreader between the heater and the constant temp unit and that should provide the heat uniformly as well as equalize the temperature of the unit needing the constant temp.

I guess we don't know exactly what it is that needs the constant temperature. If it can be mated well to a solid block of alumina that sounds great. But then I've never heard of an alumina heat sink for computers so I expect there is something wrong with this idea.

Rick

I guess that explains why we don't see solid alumina heat sinks...

Rick

Oops(tm). 400 W/mK. No excuse except sloppiness and forgetting to do a sanity check.

Actually, you do. The ceramic substrate used in hybrids and power amplifier blocks are aluminum oxide or aluminum nitride. Thick film resistors and conductors are silk screened and fired onto the substrate. Power devices are attached to the substrate with various methods (reflow, glue, etc). The substrate carries away the heat to a bonded aluminum heat sink. Heat conduction is so good, that rework with a soldering iron is difficult. Here's the not so wonderful results of my attempts to fix a metal migration problem on a 50 watt VHF power amp:

So does that mean I'm wrong twice? lol

Is Aluminum Nitride non-conductive? I know stainless is conductive. How do they use that as a substrate?

Rick

That is why it was mainly only HV insulative focus medium (LV stuff got thin "pads") Used mostly as a "thin" "transfer interface" for HV because it had a much higher punch through voltage than thin pads do, and had enough thermal capability for the task across a thin cross sectional thickness.

A REALLY good one is simple "hard anodization" on Aluminum. It is electrically non-conductive, and quite good at passing thermal energy though its matrix. There are three standardized levels of it, and the thickest one can even do "HV" up to a certain voltage, but was never trusted for that due to the occasional pin-hole, and that won't do for mission critical designs.

I'll bet those new skillet coatings would be pretty cool (or hot) interfaces. Maybe not.

Are the traces etched or painted on ??

Silk screened onto the substrate. Painting offers no control over the thickness of the traces. Silk screened resistors and traces are always the same thickness as the silk screen (for the same viscosity).

The layer cake is something like this (starting at the bottom):

Copper or aluminum heat spreader or sink. Alumina substrate Conductive traces. Bake in oven. Silk screen resistors over conductive traces. Bake in oven. Solder paste for devices with copper bottoms. Glue or reflow active devices. May be open chips or SMT devices. Wire bonds from chips to pads. Junction coat over active devices. Test, check for hot spots, cover, and ship.

It's really ancient technology. But if you want to get the heat out, there's nothing better. It's also a great idea if you want a constant temperature across the substrate for temperature tracking. I was working for a company in the late 1960's that made CTCSS tone encoders and decoders for mobile radios. We needed 50 ppm across the mil spec temperature range. Because all the resistors are made from the identical material with the same processes, they track each other. However, that doesn't work if there are hot spots on the PCB, so it was done on a ceramic substrate. You can do the same thing with thin film, but it costs more.

I'm not keeping score. (I've been wrong more times than I care to admit or recall).

It's very non-conductive. I know stainless is conductive.

As TheGlimmerMan mentioned, it's the same as your ceramic coated kitchen cooking skillet. The stainless is coated with some type of ceramic (matched for tracking the coefficient of thermal expansion for the stainless substrate). The circuit is then silk screened or printed on top of the insulating ceramic. Stainless steel isn't the best heat sink, but it's cheaper than copper and more temperature stable than aluminum.

This type of construction has been around for quite a long time. I ran into it on some Type 400 KTU cards in the 1960's. If you look carefully at the pictures at: you'll notice that the brown colored edges of the "PCB" are rather rounded. That's the ceramic coating over the steel core. Sharp edges on brittle ceramics tend to break off.

gn of film heaters? I am working on (with?) one that is heating to a modes t 37C but needs to be plus or minus half a degree Centigrade across the 4" x 4" surface. My contention is that the trace (and therefore heat density) needs to be higher at the edges than the center.

e analysis of everything the heater is contact with is in order.

iting ;-)

Thanks! I remember working with Minco in the late 80s on a "profiled" heat er. I was sure they had a nice summary of the design process somewhere.

EdV

gn of film heaters? I am working on (with?) one that is heating to a modes t 37C but needs to be plus or minus half a degree Centigrade across the 4" x 4" surface. My contention is that the trace (and therefore heat density) needs to be higher at the edges than the center.

e analysis of everything the heater is contact with is in order.

iting ;-)

Thanks much to all who have weighed in my topic. I think I have the info I need to move forward with a nice array of possible solutions.

EdV

of film heaters? I am working on (with?) one that is heating to a modest 37C but needs to be plus or minus half a degree Centigrade across the 4" x 4" surface. My contention is that the trace (and therefore heat density) needs to be higher at the edges than the center.

analysis of everything the heater is contact with is in order.

writing ;-)

I was sure they had a nice summary of the design process somewhere.

Still not going to cut it unless you can put a really sharp bound on the gradients due to external forcing.

Cheers

Phil Hobbs