0603 self heating

Do you have a thermal imager? ...Jim Thompson

-- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

Hey simple physics! So it's mostly Al2O3(?) and heated from the top? about 0.3mm thick?

Al2O3 has a conductivity of 30W/m*K with a thickness of 0.3 mm I get a thermal resistance of ~55 K/W (assuming no mistakes?) (from top to bottom) And everything has a heat capacity of about 3J/cm^3 so I get a heat capacity of ~ 1.6E-4 J and a TC of 55*1.6E-4 ~10ms

You'll have to check my math though :^)

George H.

oops J/K.

Den onsdag den 28. maj 2014 20.33.07 UTC+2 skrev John Larkin:

http://www.cetti.ro/v2/download/materiale_bibliografice/Thermal%20Management-SMD-Vishay.pdf

-Lasse

If you don't get reliable answers, you could buy one of these for $2, solder it down, and test it yourself.

The interesting part might be figuring out how to put a constant power into the RTD. Intuition (and a quick derivative calculation) tells me that putting it in series with a resistor of equal value midway through the expected temperature swing and feeding it from a fixed voltage should be close.

and the power dissipation will be constant within a few tenths of a percent. If you're guessing 300K/W, then 1/3W => about 13.2V across the pair.

The temperature can be calculated directly from the RTD voltage drop (or use a Wheatstone bridge so the ADC doesn't have to work so hard, but you probably have some very good ADCs kicking around), using

Or just put it on a power supply and measure the voltage required to get the resistance at some reasonable temperature rise, but that doesn't give you the nice multiple (?) time-constant curve.

--sp

Den onsdag den 28. maj 2014 23.24.19 UTC+2 skrev Spehro Pefhany:

ms in series

it's close, ~336 to ~338mW

Version 4 SHEET 1 2536 680 WIRE 80 16 -32 16 WIRE 352 48 304 48 WIRE 384 48 352 48 WIRE -32 64 -32 16 WIRE 80 112 80 96 WIRE 176 112 80 112 WIRE 208 112 176 112 WIRE 80 128 80 112 WIRE -32 224 -32 144 WIRE 80 224 80 208 FLAG 80 224 0 FLAG -32 224 0 FLAG 384 128 0 FLAG 304 128 0 FLAG 352 48 trtd FLAG 176 112 vrtd SYMBOL res 64 112 R0 SYMATTR InstName R1 SYMATTR Value R= if((V(trtd)-0.5), rzero*(1+3.9083e-3*V(trtd)-0.5775e-7* V(trtd)*V(trtd)), rzero*(1+3.9083e-3*V(trtd)-0.5775e-7*V(trtd)*V(trtd) -4.1

That's nice. I've only got limited experience in this sort of thing, but there is also a surface contact resistance that seems to be missing. Maybe between the copper and FR4?

George H.

I've done that with a 1206 RTD. In free air, with tiny wires, tau is around 10 seconds. Soldered to lots of copper, it's more like 0.5 seconds.

A 1206 mounted on a PCB, above a layer 2 ground plane, with normal pads and skinny traces, runs about 80 K/W.

The Vishay paper that Lasse cited shows 350 K/W for an 0603, but it looks like the PCB had no internal copper planes.

I'm probably safe assuming 300 K/W for an 0603 over a ground plane.

The issue is self-heating error of a Susumu thinfilm resistor with about 4 milliwatts dissipated. Temp rise will be roughly 1 deg C, not much effect on resistance.

Den torsdag den 29. maj 2014 05.00.28 UTC+2 skrev John Larkin:

hms in series

nd 10

nd

PG

like

on

there's another one here:

-Lasse

The poor-man's thermal imager is a test board, a selection of temperature-indicating lacquer/crayons, a power supply, a clock/stopwatch, and a videocamera.

It'd be a good exercise for a lab full of students, assign three or four different heat levels to see how the thermal timing varies and fit a lot of experimental results to the model.