SOT-23 at 1.5 watts

Why not use two in parallel. Maybe throw in some resistors.

--
 Thanks, 
    - Win

Is around 100 C the global minimum of dT/dI for that configuration?

Below it the temperature is climbing faster than the on resistance, above the on resistance is climbing faster than the temperature (till it melts)?

It should be OK with one. The LDO will have a 0.1 ohm resistor in the fet source lead, as a current measurement shunt, so I could increase that to shift some of the power dissipation if the fet were to get hot. But the FPGA boys (and girls) don't expect more than half an amp of core current, small FPGA, which is a mere 400 mW dissipation in the fet.

We have a big ZYNQ FPGA on the same board, and its core is powered by a 1V switcher. I don't want power supply noise on the smaller one, as it could modulate prop delays and make jitter, so it gets a private linear regulator, this homebrew LDO.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

maybe just a resistor sized to take most of the 0.8V at max current

Oi, that's my trick. It works well when you can lose some headroom.

NT

Your measurements are interesting

You have 50K/W whereas a standard minimum footprint SOT-23 has 500K/W (BC847)

It may depend on the leadframe from the die to the pins, so another manufacturer can show different figures

Cheers

Klaus

I agree it's probably fine at half a watt, but of course your IR camera is measuring the temperature in the top several microns of the package and not the actual die temperature, so that 105 C number has to be properly interpreted.

Cheers

Phil Hobbs

Even better, the 50K/W included the dissipation of the square of copperclad. Looks like the theta of the fet itself is roughly half that.

The key here is to really heatsink the collector lead. On a multilayer board, I could do better.

I think I'll Dremel one open and see what's inside.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

There seems to be a squarish blob of heat in the image, fairly uniform. Likely that is the chip outline. I'll carve one open.

Not bad for 7 cents.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

DSC00446.JPG

DSC00448.JPG

IR_0073.jpg

If you need the same sized pour for a SOT-23 as for a DPAK, how is the SOT-23 better? Pricing?

--
Tim Wescott 
Control systems, embedded software and circuit design 
I'm looking for work!  See my website if you're interested 
http://www.wescottdesign.com

Yes, and I'm using it other places on the same board, for relay drivers and such.

I can probably heat sink it through vias, to an inner layer pour, so it won't use much surface area.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

The data isn't super accurate, especially at lower powers. I probably didn't wait long enough for things to settle, either.

But it sure looks like I could dump a watt in that little part, half a watt easily.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

"You were so preoccupied with whether or not you could you didn't ask yourselves...whether you should..."

That makes sense.

That makes even more sense.

--
Tim Wescott 
Control systems, embedded software and circuit design 
I'm looking for work!  See my website if you're interested 
http://www.wescottdesign.com

Hey, why do all that hard thinking, when you can get your junk-food philosophy for free on YouTube?

At least it wasn't a music video.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

I knew I'd like the film "Jurassic Park" right about the time our heroes were on the test run of that automated park tour that according to the slimy lawyer is going to cost $2000 or $10000 a head and the exhibit isn't even around to be gawked at and they're just staring at trees.

It's hard to suspend one's disbelief for a film about bringing dinosaurs back from the dead, but that sort of sold it.

I am not sure. Your last link says the plastic package is already at

100C, meaning the inside is cooking way higher. That reminds me of a guy driving an Alfa Romeo Giulia. His normal range of engine operation was between 5000rpm and redline.

--
Regards, Joerg 

http://www.analogconsultants.com/

Put it in a cramped box & it'll be hotter still. Stick it in the sun in an overheating room etc. I know it's stating the obvious.

Long ago I squeezed way above rated output from some chips by heatsinking them. They were fine for 6 months, after which performance just deteriorated. What would it cost to use 2?

NT

With respect, the FET might not be intended to be driven like that.

However, having been a daily driver of a Giulia for 30 years, I need to say that those definitely *are*. And that guy has mad skills, I love the way he takes those chicanes. I had a hard-on the whole way through.

Clifford Heath. Traction is optional, velocity not.