That was me. You can find the pictures (here) and (here).
You may lower the thermal resistance ever so slightly at the expense of blowing any resistance reduction completely out of the water by warping the package when you mount it.
The package engineers at TI are usually pretty frugal with the materials. If they could shave off a few µg of copper by making the tab thinner, I'm sure they would.
If you have to resort to these sorts of tricks to make your design work, you're doing something fundamentally wrong.
The die is located under the thick plastic part, so you can cut quite a bit of the tab off before you start cutting into the die, bond wires, etc. As you say, though, it's not exactly a production flow. 🙂
Tom
Henry, do you have the insulated or non-insulated amp chips? There should be no problem with taking of a few microns of the exposed metal and plastic on the back of the non-insulated units. I would believe the same is true for the insulated version -- just be careful that you don't expose metal.
Thanks for the reply Brain. I've got the insulated lm3886tf. I was going to wrap a piece of 1000 grit wet/dry emery around a large parallel bar, that is flat w/in .00005 , and lightly rub the back of the lm3886 on the paper a few times.
However I put the edge of a machinists precision knife edge bar against the back and slid it around looking for any light that would show gaps and it appears very smooth and flat. I now think it doesn't need sanding. A very light coating of goop should do it.
However I put the edge of a machinists precision knife edge bar against the back and slid it around looking for any light that would show gaps and it appears very smooth and flat. I now think it doesn't need sanding. A very light coating of goop should do it.
I have ground/sanded the back of a few To220 and To264 devices. 600 wet or dry seems good enough for me.
It takes off the not quite flat solder coating and often reveals that the copper backplate is not flat, but has a step, or steps, from the extruding/rolling process.
I keep grinding until the step on that backplate is flat. Then a very small amount of thermal goop fills in any remaining scratches/imperfections at the interface.
It takes off the not quite flat solder coating and often reveals that the copper backplate is not flat, but has a step, or steps, from the extruding/rolling process.
I keep grinding until the step on that backplate is flat. Then a very small amount of thermal goop fills in any remaining scratches/imperfections at the interface.
I haven't managed to annoy mine quite that much yet. How close is sparkling clean and waxed kitchen to that point? Oh, I'd better report to the kitchen for breakfast, coffee, and other something non-boring directly.
Not entirely off topic for amplifiers, because it is required that a good one must bring it, most of the time, in most cases. Well, that IS what they have to do, because that's how a human can tell if they work right.
Please standby for the next message. Perhaps it will be more illustrative? 😀
Luckily mine is limited to making knives so sharp they can thinly slice ripe tomatoes without them exploding and really shiny shoes. God help me if I ever get into car detailing...
I think I need a new soldering station. My weller just died after only two years and very little use. I've seen some negative posts about weller so don't want to get another.
What are considered good quality units. I've read a lot of positive posts for Hakko and don't mind spending about $150. Any thoughts on it or other quality brands will be appreciated.
Thanks,
henry
What are considered good quality units. I've read a lot of positive posts for Hakko and don't mind spending about $150. Any thoughts on it or other quality brands will be appreciated.
Thanks,
henry
This is about goop between the lm3886tf and the heatsink. I noticed the lm3886 has four round recessed depressions about 3/32" in Dia. at the four corners on the heatsink side. I thought it would be good to ensure there was some grease in those spots to try to minimize air gaps. I put only a small amount of the goop on the mounting surface and slid the chip in all directions to minimize goop and maximize contact area.
After doing this with one chip and heatsink I pulled them apart and there was still a very light layer of the goop between the two surfaces. It was thin enough that the back of the chip looked almost bare. However there was an extremely thin layer of the goop.
Since I lapped the the area of the heatsink, where the chip will in contact, I question the need for the goop. If actual contact is better than goop, which is better than air, why use goop if the surfaces will be in intimate contact?
After doing this with one chip and heatsink I pulled them apart and there was still a very light layer of the goop between the two surfaces. It was thin enough that the back of the chip looked almost bare. However there was an extremely thin layer of the goop.
Since I lapped the the area of the heatsink, where the chip will in contact, I question the need for the goop. If actual contact is better than goop, which is better than air, why use goop if the surfaces will be in intimate contact?
I get 40 W into 8 Ω at low THD with the goop versus 32-35 W without the goop. It makes a difference.
Re. soldering iron: Love my Metcal MX-500. My Weller TCP from 1988 is still going strong after hours and hours of use. I'd recommend it. A better place for soldering iron research is one of the many threads dedicated to the topic in the tools section, though.
Tom
In case any of you are curious about the transient response and clipping response of the MOD86, I just posted a stack of measurements in the Vendors Thread.
Tom
Tom
Henry, the chip is entirely inside the big portion of the package. It is roughly 0.2" square, if I remember correctly, and centered left-to-right inside the package. The main path of heat is straight 'down' (from die through package to heat sink). If you looked at how much the heat spreads out, you'd see that it spreads at maybe a 45 degree angle. With good contact between the middle of the package to heat sink, the majority of the package contributes little to thermal conductivity.
"Intimate contact" is a relative term. This has all been thoroughly studied and analyzed. As has been mentioned, you need only the slightest amount of heat sink compound. It is better than nothing (air).
"Intimate contact" is a relative term. This has all been thoroughly studied and analyzed. As has been mentioned, you need only the slightest amount of heat sink compound. It is better than nothing (air).
I've used Weller temperature-controlled irons for years (like 40). I've not had issues. One of my older ones did die on my son recently but it was due to the bending/twisting of the cord where it came out of the handle of the iron. Not bad for one so old. I wouldn't consider them unreliable.
At work we have very expensive Metcal irons for tiny SMT work and their very expensive heater/tip assemblies fail regularly.
At work we have very expensive Metcal irons for tiny SMT work and their very expensive heater/tip assemblies fail regularly.
I agree, if you have both surfaces lapped close to 'perfectly' flat and very fine surface condition then there is no real need for HS grease compounds.
Theory as has been explained is that the very fine filler powder fills minor voids and scratch lines providing improved wrt air thermal coupling.
The challenge is ensuring that the HS grease film is so thin that it only serves to fill micro voids, but does not actually separate the two thermal surfaces.....close to mission impossible in practical usage.
For epoxy insulated thermal surfaces (transistors and flat pack ics), I was always surprised at the non flatness revealed after each stroke when lapping on 3000 W&D water fixed to a 2cm thick glass sheet.
Heatsinks too, especially anodised, exhibit similar roughness and I always found flatness defects.
When fitting one off supplied service parts I grew to not trust that damage had not occurred in the supply chain.....dropped on a hard floor, improperly packed etc..I have seen minor flatness damage with these causes.
So, if you can achieve precision flatness quickly and easily I say go for it.
An interesting comparison would be between highly polished surfaces and ground/sanded finish, without and with HS greases (Std Silicone/Zinc Oxide, Arctic Silver 5/silver, aluminum oxide, zinc oxide, and boron nitride in a polysynthetic oil base, etc)....if you have an IR thermometer you may/may not measure temp differences, but likely very minor.
You may also cause minor subjective change according to type of HS grease, and none may be better sounding.
Dan.
You must eliminate all air from the interface.
That is the sole purpose of the thermal goop.
