Quote:
Originally Posted by rtchinn 
-----
In my opinion, (and you are certainly entitled to yours), that defines a way for failure. Three dots, 1/4 the size of a grain of rice?... are you serious?
-----
In my opinion, (and you are certainly entitled to yours), that defines a way for failure. Three dots, 1/4 the size of a grain of rice?... are you serious?
Yes, I am absolutely serious.
Based on the small surface area in this specific application ..any more than that amount would be sheer overkill. These mating surfaces are profoundly rectangular. If they were perfectly square mating surfaces with similar surface area, then I would recommend a single dot dead center no larger than half a grain of rice.
Quote:
With over 30 years of Electronics Industry work, over 25 of those as a design engineer, and I've never heard or read any thermal engineering source recommend such tactic.
With over 30 years of Electronics Industry work, over 25 of those as a design engineer, and I've never heard or read any thermal engineering source recommend such tactic.
I don't doubt your engineering prowess, and am not here to scuffle, but rather to inform. I'm sure you have heard of the center dot/small pea method which is most commonly used for mainframe and critical server CPU installations? This is ideal for the larger surface area perfectly square shaped PC class CPU heat spreaders. We are dealing with surfaces that are much smaller here and much less thermal compound suffices.
Quote:
YES... I'd agree that many people get themselves into trouble by following the "apply liberally" method of long ago, but likewise I believe, from my experience, your prescribed method goes to the opposite extreme.
YES... I'd agree that many people get themselves into trouble by following the "apply liberally" method of long ago, but likewise I believe, from my experience, your prescribed method goes to the opposite extreme.
My method is precisely perfect for this application. Additionally, spreading tests have shown that multiple dots spread evenly and do not introduce air gaps.
Quote:
Additionally, if both surfaces are not cleaned, ANYTHING left on the (in this case) DMD heat spreader, will prevent the thermal compound from spreading out to cover the entire mating surface.. and cause VERY quick failure.
Additionally, if both surfaces are not cleaned, ANYTHING left on the (in this case) DMD heat spreader, will prevent the thermal compound from spreading out to cover the entire mating surface.. and cause VERY quick failure.
I agree that it would be ideal to clean both surfaces, but not convenient nor essential in this particular case. When the DMD heat sink is removed, the thermal pad will be cleanly extracted and completely affixed to the base of the heat sink.
The DMD's heat spreader is recessed deep within a padded enclosure making cleaning impractical. One would likely introduce contamination to the heat spreader by probing the cavity in a cleaning attempt. Again, the pad will separate cleanly from the DMD heat spreader leaving behind no residue.
Quote:
My modifications to your instructions would be:
Use a "credit card" (or similar) to scrape the old material from both surfaces, THEN clean them with the prescribed Alcohol method, but use a Q-tip moistened in alcohol to clean the DMD surface IF there is any residue or pieces of the old thermal pad left. Both surfaces must be 100% CLEAN when done.
WAIT for both surfaces to completely dry, then apply a VERY thin layer of thermal compound to ONE surface (the heatsink), spreading it with the same (but CLEANED) credit card. The proper amount will be when there are streaks of the heatsink showing through the compound, but not large areas. Be CERTAIN that there are no "bumps" in the layer of compound, as this would indicate dried 'globs' of the base material, and would prevent the material from spreading out when reattached.
The goal is to apply enough across the entire mating surface that the pressure applied by the heatsink spring clip will cause the compound to "ooze" and spread out to fill-in the gaps - once the heat comes on.
On big problem with "dots" is that you have minimal thermal conduction to begin with, when the set is first operated. Then you are expecting these 3 "piles" of compound to spread out over the entire surface, by itself... and I can guarantee you that this method will leave areas of mating surface without thermally conductive compound... which means less heat transfer, and more prone to failure. Some "silver arctic" type compounds are very thick, and resist spreading out without significant pressure.
My modifications to your instructions would be:
Use a "credit card" (or similar) to scrape the old material from both surfaces, THEN clean them with the prescribed Alcohol method, but use a Q-tip moistened in alcohol to clean the DMD surface IF there is any residue or pieces of the old thermal pad left. Both surfaces must be 100% CLEAN when done.
WAIT for both surfaces to completely dry, then apply a VERY thin layer of thermal compound to ONE surface (the heatsink), spreading it with the same (but CLEANED) credit card. The proper amount will be when there are streaks of the heatsink showing through the compound, but not large areas. Be CERTAIN that there are no "bumps" in the layer of compound, as this would indicate dried 'globs' of the base material, and would prevent the material from spreading out when reattached.
The goal is to apply enough across the entire mating surface that the pressure applied by the heatsink spring clip will cause the compound to "ooze" and spread out to fill-in the gaps - once the heat comes on.
On big problem with "dots" is that you have minimal thermal conduction to begin with, when the set is first operated. Then you are expecting these 3 "piles" of compound to spread out over the entire surface, by itself... and I can guarantee you that this method will leave areas of mating surface without thermally conductive compound... which means less heat transfer, and more prone to failure. Some "silver arctic" type compounds are very thick, and resist spreading out without significant pressure.
Q-Tips should never be used to clean thermally critical surfaces, potential stray cotton fibers would be extremely detrimental to conduction.
The credit card spread method is foolish as it is known to incorporate micro air pockets within the compound no matter how careful you are.
Arctic Silver 5 (which I have recommended) has excellent initial spreading performance. The amount I specified is enough to fill all surface imperfections, cover the entire surface area and ensure that no dead air space will exist.
Quote:
This is the precise reason most manufacturers have chosen the thermal pad method... it eliminates human error in applying the precise amount required. Thermal pads are NOT BAD... in general. Some are great, some are not so... but when properly engineered and applied, they do the job VERY well.
This is the precise reason most manufacturers have chosen the thermal pad method... it eliminates human error in applying the precise amount required. Thermal pads are NOT BAD... in general. Some are great, some are not so... but when properly engineered and applied, they do the job VERY well.
True, but the thermal pads chosen for use in these TVs are exceptionally poor.
Quote:
One problem with "thermal paste" is that MANY products have a liquid carrier that completely evaporates in VERY short order with the applied heat. What happens then is that the actual thermal transfer material becomes hard (or even brittle), and when the device heats up and thermal creep happens between the two surfaces, the dried up base material "bunches up", and no longer "goes with the flow"... it no longer allows for movement... even on the small scale we're talking about. This obviously results in thermal failure.
One problem with "thermal paste" is that MANY products have a liquid carrier that completely evaporates in VERY short order with the applied heat. What happens then is that the actual thermal transfer material becomes hard (or even brittle), and when the device heats up and thermal creep happens between the two surfaces, the dried up base material "bunches up", and no longer "goes with the flow"... it no longer allows for movement... even on the small scale we're talking about. This obviously results in thermal failure.
Yes, aka cheapo silicone/zinc oxide based concoctions. ...Hence my recommendation for AS5.
Quote:
This "creep" is CRITICALLY important when the heatsink is not directly fastened TO the device generating the heat. If there is no room for "creep", then it is much less of a problem. HOWEVER... in our application (Samsung DLP), the heatsink is NOT "attached" to the DMD... it's attached to the mechanical hardware around the DMD.
This "creep" is CRITICALLY important when the heatsink is not directly fastened TO the device generating the heat. If there is no room for "creep", then it is much less of a problem. HOWEVER... in our application (Samsung DLP), the heatsink is NOT "attached" to the DMD... it's attached to the mechanical hardware around the DMD.
Yes, I am already aware of potential binding issues due to the aging of poorly designed thermal compounds
...Hence my recommendation for AS5. AS5 is guaranteed to never creep nor harden.
Quote:
Well engineered Thermal Pads have high temperature (silicone) in them (which is expensive), which acts as a permanent lubricant, and does not dry out during the lifetime of the product.
I just pulled the heatsink from one of my used spares (yes, Samsung LED DLP), and the pad (light yellow colored) was still pliable, and able to 'slide' across the DMD surface easily... which tells me Samsung Engineers probably did their homework.
Well engineered Thermal Pads have high temperature (silicone) in them (which is expensive), which acts as a permanent lubricant, and does not dry out during the lifetime of the product.
I just pulled the heatsink from one of my used spares (yes, Samsung LED DLP), and the pad (light yellow colored) was still pliable, and able to 'slide' across the DMD surface easily... which tells me Samsung Engineers probably did their homework.
I grade the result of their thermodynamics homework a resounding F.
Quote:
I'm not saying it can't be improved upon, but I am saying that if it ain't broke, don't TRY to fix it! Without proper test conditions (measuring the actual DMD device operating temperature with BOTH types of thermal transfer material), you really do not 'know' that you have done it any good... or bad.
I'm not saying it can't be improved upon, but I am saying that if it ain't broke, don't TRY to fix it! Without proper test conditions (measuring the actual DMD device operating temperature with BOTH types of thermal transfer material), you really do not 'know' that you have done it any good... or bad.
...And this is where I think you will listen to my reason:
I have measured the net benefit of losing the pad.
Prior to initial HS removal, I detached the DMD cooling fan and powered up the TV....then tuned to a dead-signal, static NTSC channel. (TI datasheets specify that this will stress the DMD to the max and generate the most heat)
***After 2 minutes of on-time the DMD heat sink _was still room temperature (71F)_***
I then replaced the pad with AS5 exactly as I have recommended. Would you like to know what happened next?....
***After 2 minutes of on-time the heat sink was very warm to the touch (121F)_***
The results need no further explanation.
Quote:
BTW... stuck pixels can't be directly tied to a thermal problem. Let's face it... the DMD is a bit of a "magical mystery" anyway... it's an amazing technology, but it is still an electro-mechanical device... those million+ microscopic mirrors are moving billions of times, and SOME WILL FAIL... maybe not in my set, maybe not in yours... but some will fail.
BTW... stuck pixels can't be directly tied to a thermal problem. Let's face it... the DMD is a bit of a "magical mystery" anyway... it's an amazing technology, but it is still an electro-mechanical device... those million+ microscopic mirrors are moving billions of times, and SOME WILL FAIL... maybe not in my set, maybe not in yours... but some will fail.
Sorry Sir, you are misinformed and incorrect.
There is official documentation direct from TI that emphasizes the dire importance of efficient heat dissipation on the DMD panel. There is a profound DMD sensitivity to heat. Failures (aka stuck pixels) are common within 2000hrs of service if the guidelines are not followed.
I have easy access to all of the TI datasheets I could ever want as a member of their engineering discussion board.
Quote:
Sorry if I came across too harshly, but I'm passionate about engineering, and sometimes these kinds of issues get my ire up.
Sorry if I came across too harshly, but I'm passionate about engineering, and sometimes these kinds of issues get my ire up.
I am as well, and tend to get worked up over issues myself. It's fine..no hard feelings.
. The TV is about 3 years old. Anyone know what causes this? Picture data is still visible in this shadowed area, it's just darker than everything else. This anomaly occurs on all 3 HDMI inputs.
