Philips Semiconductors

Product data sheet

NE1617A

Temperature monitor for microprocessor systems

2004 Oct 05

9

FUNCTIONAL DESCRIPTION

The NE1617A contains an integrating A-to-D converter, an analog

multiplexer, a status register, digital data registers, SMBus interface,

associated control logic and a local temperature sensor or channel.

The remote diode-type sensor or channel should be connected to

the D+ and D– pins properly.

Temperature measurements or conversions are either automatically

and periodically activated when the device is in free-running mode

(both STBY pin = HIGH, and the configuration register BIT6 = LOW)

or generated by one-shot command. The free-running period is

selected by changing the programmable data of the conversion rate

register as described later. For each conversion, the multiplexer

switches current sources through the remote and local temperature

sensors over a period of time, about 60 ms, and the voltages across

the diode-type sensors are sensed and converted into the

temperature data by the A-to-D converter. The resulting temperature

data is then stored in the temperature registers, in 8-bit, two’s

complement word format and automatically compared with the limits

which have been programmed in the temperature limit registers.

Results of the comparison are reflected accordingly by the flags

stored in the status register, an out-of-limit condition will set the

ALERT output pin to its LOW state. Because both channels are

automatically measured for each conversion, the results are

updated for both channels at the end of every successful

conversion.

Remote diode selection

The method of the temperature measurement is based on the

change of the diode VBE at two different operating current levels

given by:

DVBE + KT

q < LN(N)

where:

∆VBE = change in base emitter voltage drop at two current levels

K: Boltzman’s constant

T: absolute temperature in

° Kelvin

q: charge on the electron

N: ratio of the two currents

LN: natural logarithm

The NE1617A forces two well-controlled current sources of about

10

µA and 100 µA and measures the remote diode VBE. The

sensed voltage between two pins D+ and D– is limited between

0.25 V and 0.95 V. The external diode must be selected to meet this

voltage range at these two current levels. The diode-connected PNP

transistor provided on the microprocessor is typically used, or the

discrete diode-connected transistor 2N3904 or 2N3906 is

recommended as an alternative.

Even though the NE1617A integrating A-to-D converter has a good

noise performance, using the average of 10 measurement cycles,

high frequency noise filtering between D+ and D– should be

considered. An external capacitor of 2200 pF typical (but not higher

than 3300 pF) connected between D+ and D– is recommended.

Capacitance higher than 3300 pF will introduce measurement error

due to the rise time of the switched current source.

No calibration is required

As mentioned in the ‘Remote diode selection’ section, the NE1617A

uses two well-controlled current sources of 10:1 ratio to measure the

forward voltage of the diode (VBE). This technique eliminates the

dependent variable) and results in the forward voltage being

proportional to absolute temperature.

Address logic

The address pins of the NE1617A can be forced into one of three

the NE1617A samples and latches the address pins at the starting

of every conversion, it is suggested that those address pins should

be hard-wired to the logic applied, so that the logic is consistently

existed at the address pins. During the address sensing period, the

device forces a current at each address pin and compares the

voltage developed across the external connection with the

predefined threshold voltage in order to define the logic level. If an

external resistor is used for the connection of the address, then its

value should be less than 2 k

Ω to prevent the error in logic detection

from happening. Resistors of 1 k

Ω are recommended.