A truly sweet, small, new SMD power FET package

Using in a new design: Vishay's PowerPAK 1212-8. Tiny, tough, easy-to-cool.

At 3.6 x 3.6mm footprint, it's a fraction the size of a high-power SO-8, but sports similar-class die-carrying capacity, Rds(on), theta(ja), etc.

theta(j->solder_point) ~= 2.4K/W, as compared to 1.8K/W for the much-larger power SO-8 variants with big metal pads, and 16 K/W for a BCP53/BCX53 (random choice) SOT-89/SOT-223.

Package footprint theta(j->s.p.)

-------- ------------ -------------- SOT-223 7.3 x 6.7 mm 16 K/W SOT-89 4.6 x 4.25mm 16 K/W PP1212-8 3.6 x 3.6 mm 2.4 K/W PP SO-8 6.2 x 5.2 mm 1.8 K/W

AN822 starting at page 8 of this datasheet[1] has a good discussion about this package AND empirical SMD heatsinking considerations in general, with spreading and graphs.

[1]
formatting link

The PowerPAK 1212-8's thermal performance is excellent. For this design, it was a game-changer. SOT-89, R.I.P.

Cheers, James Arthur

Reply to
dagmargoodboat
Loading thread data ...

That is a nice package. Source inductance will be very low, so fast switching should work nicely.

The appnote thermal notes are skimpy and in my opinion suspect. Or maybe dead wrong.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

They measured theta(j-a) for various pads with real parts on real boards (Fig. 5-6). Which results raised your flags?

Cheers, James

Reply to
dagmargoodboat

That just sounds so funny. Funny funny and funny haha.

Reply to
DecadentLinuxUserNumeroUno

There was no mention of vias, and the 2x2 board was likely the dominant thermal resistance to ambient. Without vias, the part is dumping heat directly into the epoxy-glass under its footprint.

They are basically concluding that this package can only dissipate a couple of watts. A sot-89 can do that.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

I'm designing one of those packages into a part. They're not the only one -- there are other pad-compatible packages out there, with different brand names but with dimensions that are similar enough to plop in place.

I suspect that getting a good solid solder joint is going to be critical

-- any voids underneath that honkin' big (compared to the package) drain pad is going to significantly increase the thermal resistance, and generally unmake your day. X-ray inspection may be mandatory, or at least some sort of load testing in production to make sure they don't get too hot.

--
Tim Wescott 
Wescott Design Services 
 Click to see the full signature
Reply to
Tim Wescott

I suspect that the PC board will dominate theta, and minor voids in the solder (which shouldn't happen anyhow) will be down in the noise.

Mounting this on a board and getting tens of watts of heat out should be interesting.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

You're right, they don't mention vias. AIUI they're demonstrating the performance of this package with a given top-side copper pour, with three solid copper planes beneath (but with no metallic connection) acting as spreaders.

Heat-sinking offers much higher power capability.

They're right though, as far as their presentation goes--if your heat sink is a small copper pour, dissipation is limited by that copper-pour to ~2W, and the package doesn't make a lot of difference.

I thought it was interesting that even a much larger copper pour didn't help much, because even that is spreading-resistance limited.

My gadget uses top and bottom pours to these parts tied with vias, and inner planes to spread the heat.

Another higher-power portion pipes heat out the bottom by direct contact with a massive anodized Al plate.

Cheers, James

Reply to
dagmargoodboat

I don't have any worry about getting a solid solder joint from part-to- board. Reflow and paste should handle that; molten solder wets and flows wonderfully. Old-hat D-PAKs are much larger.

Cheers, James Arthur

Reply to
dagmargoodboat

We were looking at a 240 amp laser driver (the project may be dead) that could have used these parts. It would be a polyphase switcher and we could use this package for both fets and schottky diodes. One possibility is to transfer heat through vias to the bottom of the board and dump it to a cold plate somehow. Gap-pads maybe. Spreading thermal resistance would be critical.

Or use those surface-mount wingy heat sinks and blow a lot of air across the top of the board.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Yes, I was thinking that anyone who uses it will have to become a PCB thermal specialist. I'm thinking along the lines of using lots of vias to a back side plane that'll be heat-sunk.

Somewhere floating around on the web is a white paper on thermal conductivity of vias. I need to see if I can find it. I would assume that a big variable is how much plating there actually is on the inside of a via -- if the board is 1 ounce can you count on 1 ounce down the hole, or will it be less?

--
Tim Wescott 
Wescott Design Services 
 Click to see the full signature
Reply to
Tim Wescott

Via plating can be surprisingly thick. A 20 mil drill plated 1 oz is roughly 70 K/w from top to bottom on a 62 mil board, less to inner layers. Use lots of vias, under the part and outside on a layer 1 pour, to suck heat away.

Solder-filled vias are better, ballpark 2:1.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Den fredag den 5. februar 2016 kl. 23.50.35 UTC+1 skrev John Larkin:

these are kinda cute:

formatting link

-Lasse

Reply to
Lasse Langwadt Christensen

My understanding is that the plating is done all over the board including the top and bottom layers. So they start with say 1/2 oz copper and plate another 1/2 oz to make it 1 oz. The most the vias will have is 1/2 oz and I would expect less as the plating is not totally uniform in the via.

A larger number of smaller vias will conduct more heat until you reach minimum spacing issues. Anyone know what limits how closely packed vias can be as a function of drill size and plating thickness?

Plating vias closed can be useful, but will result in heavy deposits on the surface layers which mean you need larger space/trace sizes.

--

Rick
Reply to
rickman

Some big slabs of aluminum? (If it was a pulsed thing, there might be two time scales) George H.

Reply to
George Herold

I did a power opamp with thermal vias underneath. There wasn't that much power, 3W max, 1W typ. I just put in 20 holes, and it worked. (the worst was getting the cad program to not give errors everywhere... I think I gave up and had it made.)

And recently two big holes under a smaller thermal pad. But hand assembled and filled with solder. In theory I think the solder wins. (it's a little worse than copper, but much better than fiber glass.) I never tried to measure. (Thermal design (at least for me) is much more of a hack, you do some numbers, make something, it works, because you put in a 20C/ 50% fudge factor. or it doesn't and you find the mistake in your equation/ theory.)

Someone should make a thermal "spice" model. (surface contact resistance would be a variable>)

George H.

Reply to
George Herold

What's wrong with a regular Spice model? Resistance is resistance.

Solder is a much better conductor than fiberglass, but it is about 6 times worse than copper depending on exactly which solder. So you need

6 times more area of solder to get the same conductivity as copper. Silver solder can do the job almost as well as copper. Or even a silver based paste.

formatting link

They claim "a thermal conductivity of more than 200 W/mK" vs. 385 for copper. That's three or four times better than conventional solders.

--

Rick
Reply to
rickman

I wrote a spreadsheet that figures all that thermal (and electrical) c> Has anyone filled vias with solder to improve thermal and electrical

I wrote a spreadsheet I use for evaluating those trade offs. Here's a sample:

25uM hole plating, 1.6mm board thickness, lead-free solder

(K/W per via) Plated hole via solder- copper- diameter alone plugged plugged

----------- ----- ------- ------ .010" (.254) 190 K/W 139 K/W 57.3 K/W .025" (.635) 80.2 41 11.3 .039" (1.0) 51.6 20.5 4.8

For Cu plugs the sneaky way is to mount a dummy through-hole part with copper leads, e.g. a SIP, a .025" header row, or a 2W resistor. =============

Cheers, James Arthur

Reply to
dagmargoodboat

What type of solder did you assume? There can be a large difference.

--

Rick
Reply to
rickman

If you have space for them, copper pins/wires that extend out of the board some distance (ie, not just plugs) make a pin-type heat sink.

--
Cats, coffee, chocolate...vices to live by 
Please don't feed the trolls. Killfile and ignore them so they will go away.
Reply to
Ecnerwal

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.