Anybody have some experiences they would like to share regarding the design 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) ne eds to be higher at the edges than the center.
My other contention is that if you are going to design a heater a complete analysis of everything the heater is contact with is in order.
My contentions remain contentious if not in actual contempt as of this writ ing ;-)
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.
;-)
You're mostly right, but heat flow is complex and peverse. Air flow/convection is worse. 0.5C over that area will be difficult.
I assume that you are using the heater to heat some object, not just heat itself.
You could do zone heating (multiple sensors and heaters) or bond the heater to an aluminum block to reduce lateral thermal resistance.
formatting link
formatting link
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
gn of film heaters? I am working on (with?) one that is heating to a mod est 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 ;-)
Hmm, OK I would guess that for a thin plate (pcb?) the heat loss would not have much of an edge effect. So I'd try a uniform heater first. A heat spreader might help... can you put 2 oz. copper on one side? (Sorry, is this a pcb?) But we really need more info on the heat loss... is the plate in air, vacuum, is there some differential heat load from one side to the other?
ign of film heaters? I am working on (with?) one that is heating to a mode st 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.
te analysis of everything the heater is contact with is in order.
riting ;-)
Perverse is a nice way of putting it ;-)
The heater is part of a PCB that has other functions that require not havin g electrically conductive material in certain regions because there are pro ximity sensors for things moving in the z with respect to the x y of the he ater.
The heater sits below a 0.12" thick plastic plate which acts as a heat spre ader but is really there to provide material for slots that engage rails on the "thing" that gets heated. The "thing" is surrounded on three sides by more plastic and a PCB on top that does "other stuff.
We are getting +or- 1.3C in the regions of interest. Would prefer to use a single heater. I say FEA (Finite Element Analysis) by hand or in SW is th e next (should have been first) step)
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 ;-)
That's a pretty tight tolerance for a single zone. What are you going to do if somebody puts the thing right next to an A/C register? All the environmental forcing will be superimposed on your gradient.
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics
160 North State Road #203
Briarcliff Manor NY 10510
hobbs at electrooptical dot net
http://electrooptical.net
The "no electrically conductive" rule kind of boxes you into a corner.
I agree that you need uneven heating to get even temperature -- to me, your uneven temperature with even heating shouts that out.
Unless your board is going to sit in still air at a constant temperature, and you can wait a good long time for things to settle to equilibrium, I don't think a shaped heater is going to do the job -- I think you need some sort of a zone heater.
Maybe go shopping for a plastic material that's more thermally conductive than what you've got? I don't think you'll find joy, but maybe. Depending on your proximity sensor technology, maybe something that's metal-filled, but still not conductive (if such things exist -- I'm not sure if it would).
I assume a 4 x 4 x 0.12 plate of BeO would be out of the question, and might not work for you anyway.
Can you put a metal heat spreader on the back side of your plastic block?
End the end, I think you're going to have to accept either failure or zone heating.
--
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com
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 ;-)
electrically conductive material in certain regions because there are proximity sensors for things moving in the z with respect to the x y of the heater.
but is really there to provide material for slots that engage rails on the "thing" that gets heated. The "thing" is surrounded on three sides by more plastic and a PCB on top that does "other stuff.
single heater. I say FEA (Finite Element Analysis) by hand or in SW is the next (should have been first) step)
Plastic is a terrible heat spreader. Get some metal in there.
If there's an air gap between "thing" and the spreader, and between thing and the PCB above, the temperature at any point on thing will be some sort of weighted average of the bits above and below. Does thing dissipate any power? Does the PCB above dissipate much power? Does it have hot spots?
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
design 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 t he 4" x 4" surface. My contention is that the trace (and therefore heat den sity) needs to be higher at the edges than the center.
plete analysis of everything the heater is contact with is in order.
s writing ;-)
ving electrically conductive material in certain regions because there are proximity sensors for things moving in the z with respect to the x y of the heater.
preader but is really there to provide material for slots that engage rails on the "thing" that gets heated. The "thing" is surrounded on three sid es by more plastic and a PCB on top that does "other stuff.
use a single heater. I say FEA (Finite Element Analysis) by hand or in S W is the next (should have been first) step)
.highlandtechnology.com jlarkin at highlandtechnology dot com
Hey, I was just remembering these ads by (panasonic?) selling pyrolytic graphite sheets. That would make some nice thermal spreaders... something like 3x copper along the plane of the sheets (IIRC). I guess it might be too electrically conductive for the OP. (It'd be fun to play with some.)
In a past life, I was designing radios next to a group of engineers working on "modular products", all of which were potted. At one point, there was a problem getting uniform temperature distribution across the PCB, which I vaguely recall was about the same size. Filling the potting shell with epoxy was the standard method, which did not work, due to bubbles, voids, and poor thermal characteristics. Fluorinert worked, but nobody wanted to deal with the high cost and leakage problem.
The solution turned out to be thermally conductive alumina sand. The potting shell was filled with the sand through a large hole in the bottom PCB, shaken to distribute the sane, and sealed with epoxy. The sand kept the temperature even across the devices on the PCB. An added benefit was that the board could be easily de-potted, sand removed, and repaired.
I don't recall the exact type of sand that was selected. I vaguely recall that aluminum oxide grinding or sand blasting sand was tested. The trick was to find the correct grit. Too fine (dust), and it wouldn't flow into the potting module. Too large (rocks) and it would have too few points of contact to provide decent thermal conductivity. The air gaps between grains are insulators. A solid block of alumina has a thermal conductivity of about 30 W/mK, but it will be much lower for sand.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
Also, you might also consider using PCB material with an aluminum backing as used in various microwave PCB's. Rogers is the leading supplier: For example: RT-Duriod is expensive, tricky to work with, probably overkill, but thermally superior PCB material.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
Seems like a bad idea: a few years of vibration later and it's vibratory-tumbled its way out of existence!
I was hoping your story finished with "then soaked the sand with epoxy". Maybe you tried, and it didn't work? Low viscosity epoxy (usually with little to no filler of its own) does exist; vacuum would probably be required to fully impregnate it. But vacuum probably would've come in handy for the original problem (with potting alone), too.
Tim
--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms
Have you tried aluminum core or heavy copper PCBs?
Aluminum core kind of sucks in some aspects, but it's my understanding that manufacturers can make them with vias and multiple layers. Just maybe not cheaply?
Tim
-- Deep Friar: a very philosophical monk. Website:
formatting link
My other contention is that if you are going to design a heater a complete analysis of everything the heater is contact with is in order.
My contentions remain contentious if not in actual contempt as of this writing ;-)
In the deep past I've mixed Al oxide with epoxy, (400 grit maybe?) but mostly just to thicken it up. (We dripped it over cheap connectors to make them leak tight... (10^-5 torr level))
Well, it was good enough that the military customer signed off on the idea. The only thing they complained about was the weight. I didn't get a look at that stuff that was used under the microscope, but I guess that the grains were not particularly sharp (to improve contact area). I don't recall if it was expected to survive much vibration.
Nope. After a thorough shaking and topping off with more grit, the blob of epoxy was poured in through a big hole in the bottom. I believe the theory was that if the grit was packed in hard enough and had no place to move around, it would not abrade the components.
I didn't try because it wasn't my project. I was one bench over belching enough RF to ruin everyone's measurements in the building. I recall seeing some of the engineers attaching a microphone to the case and listening for the sound of shifting sand.
The company had a vacuum degasification chamber for epoxy potting. That was standard as voids are a great way for components to fall apart in a high-g missile environment. However, this module probably didn't go into a missile, probably due to the weight.
The company did some work finding a potting epoxy with zero coefficient of thermal expansion. Any movement of the potting epoxy would rip the leads out of the component packages and out of the PCB.
This is roughly what the module looked like: The above module has a black bottom cover and no epoxy potting compound. With potting, there would be a large hole in the middle of the black bottom cover, which would be filled with epoxy and cured in a vacuum chamber.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
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.
My other contention is that if you are going to design a heater a complete analysis of everything the heater is contact with is in order.
My contentions remain contentious if not in actual contempt as of this writing ;-)
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.
My other contention is that if you are going to design a heater a complete analysis of everything the heater is contact with is in order.
My contentions remain contentious if not in actual contempt as of this writing ;-)
Nowadays there are potting compounds with better thermal conductivity.
--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------
You do not need to pot it. A thermal pad and proper case thickness will homogenize and carry off the heat as best it can. Potting only reduces serviceability and also usually changes some of the circuit operational characteristics.
I did some digging for typical thermally conductive epoxy potting compound. For comparison, solid aluminum oxide is 30 W/mK and copper is 4.0 W/mK. MG Epoxy 832TC 0.682 W/mK Epoxy Set EC-1015 broken link Epoxy Set EC-1006M-4 0.634 W/mK TimTronics 8006M-4 0.63 W/mK etc. Methinks 0.6 W/mK should work.
I noted that thermally conductive silicon potting compounds are about
1.2-1.5 W/mK. That might be better if nobody cares about repairs. That's the nice thing about sand. You can break the seal, pour out the sand, do the fix, put back the sand, shake well, and seal it again.
I don't know what thermal conductivity was achieved using just aluminum oxide sand. I suspect is was better than an epoxy plus aluminum oxide dust mix. The bulk of such a mix would be epoxy, which by itself is a rather miserable 0.006 W/mK (for JB Weld).
The original plan was to get 0.5C across a 16 sq-in area. That's not easy. I suspect it can't be done except perhaps with liquid coolant. For example, water is 0.6 W/mK. Getting a uniform head distribution will probably be impossible unless the entire assembly is inside an insulated box to prevent radiation losses. A silicon pad is great for distributing the head on the circuit side of a PCB but not so great on the component side. Since all the heat is generated on the component side, putting a heat spreader on the opposite side is not going to be very useful.
Coat the assembly with some manner of conformal coating. That should take care of any resistive changes. I don't know how to get rid of sensitivity to changes in capacitance other than a redesign.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
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.