Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: Tony Viola who wrote (62543)	8/16/1998 12:58:00 AM
From: Paul Engel	Respond to of 186894

Tony - Re: "Do you happen to know how (hopefully somehow) Intel does prediction of temperature rise in chips, modules and systems?" These can be measured directly in a crude way, if appropriate care and precautions are taken, or more elegantly. 1. The Crude Way A thermocouple or RTD (Resistance_Temperature_Device) can be placed on the chip surface - or BACKSIDE in the case of C4 flip chip bonding - and the wires passed through a tiny hole drilled into the package, plastic case/thermal plate and heat sink/fan assembly. 2. The Elegant Way A more clever way is to use a calibrated forward biased p-n junction on the chip surface - the forward voltage drop is temperature sensitive -> about -2.4 millivolts/degree Centigrade. This can be calibrated by immersing the chip in a constant temperature bath, with no power to the device except the small CONSTANT CURRENT SOURCE (about 1 or 2 microamperes) through the sensing p-n junction, and measuring the forward voltage drop across the junction, Vpn. Increasing the bath temperature will provide a curve of Vpn vs temperature. This generates a calibration plot whereby electrical measurement of the Vpn (voltage drop under slight positive bias with a constant current source) will then yield the junction temperature. With a Pentium II running under "typical system conditions", and a special wiring to supply the pn junction test devices with a constant current, along with wires for measuring the pn voltage drop, the "temperature" of the chip can be measured. The thermal characteristics of the chip can then be characterized as a function of Power to the chip, frequency of the chip (which affects power), air flow over the package, etc. The special pn junctions can be designed in and can be wire bonded out - for engineering purposes. Thus, they can be electrically isolated from the active circuits on the chip, eliminating leakage currents which would otherwise disturb the sensitive measurements. Paul