Ver 0.2 Preliminary

Dec 11, 2001

TEL: 886-3-5788833

http://www.gmt.com.tw

6

G771

Global Mixed-mode Technology Inc.

A/D Conversion Sequence

If a Start command is written (or generated automati-

cally in the free-running auto-convert mode), both two

channels are converted, and the results of both meas-

urements are available after the end of conversion. A

BUSY status bit in the status byte shows that the de-

vice is actually performing a new conversion; however,

even if the ADC is busy, the results of the previous

conversion are always available.

Remote-Diode Selection

Temperature

accuracy

depends

on

having

a

good-quality, diode-connected small-signal transistor.

Accuracy has been experimentally verified for all of

the devices listed in Table 1. The G771 can also di-

rectly measure the die temperature of CPUs and other

integrated

circuits

having

on-board

temperature-

sensing diodes. The transistor must be a small-signal

type with a relatively high forward voltage; otherwise,

the A/D input voltage range can be violated. The for-

ward voltage must be greater than 0.25V at 10µA;

check to ensure this is true at the highest expected

temperature. The forward voltage must be less than

0.95V at 200A; check to ensure this is true at the low-

est expected temperature. Large power transistors

don't work at all. Also, ensure that the base resistance

is less than 100

Ω. Tight specifications for forward

current gain (+50 to +150, for example) indicate that

the manufacturer has good process controls and that

the devices have consistent VBE characteristics.

Thermal Mass and Self-Heating

Thermal mass can seriously degrade the G771's ef-

fective accuracy. The thermal time constant of the

SSOP-16 package is about 140sec in still air. For the

G771 junction temperature to settle to within +1°C

after a sudden +100°C change requires about five

time constants or 12 minutes. The use of smaller

packages for remote sensors, such as SOT23s, im-

proves the situation. Take care to account for thermal

gradients between the heat source and the sen-

sor ,and ensure that stray air current across the sen-

sor package do not interfere with measurement accu-

racy.

Table 1. Remote-Sensor Transistor Manufacturers

MANUFACTURER

MODEL NUMBER

Philips

PMBS 3904

Motorola (USA)

MMBT3904

National Semiconductor (USA)

MMBT3904

Note:Transistors must be diode-connected (base short

-ed to collector).

ADC Noise Filtering

The ADC is an integrating type with inherently good

noise rejection, especially of low-frequency signals

such as 60Hz/120Hz power-supply hum. Micro-power

operation places constraints on high-frequency noise

rejection; therefore, careful PC board layout and

proper external noise filtering are required for high-

accuracy remote measurements in electrically noisy

environments.

High-frequency EMI is best filtered at DXP and DXN

with an external 2200pF capacitor. This value can be

increased to about 3300pF(max), including cable ca-

pacitance. Higher capacitance than 3300pF introduces

errors due to the rise time of the switched current

source.

Nearly all noise sources tested cause the ADC meas-

urements to be higher than the actual temperature,

typically by +1°C to 10°C, depending on the frequency

and amplitude (see Typical Operating Characteristics).

PC Board Layout

Place the G771 as close as practical to the remote

diode. In a noisy environment, such as a computer

motherboard, this distance can be 4 in. to 8 in. (typical)

or more as long as the worst noise sources (such as

CRTs, clock generators, memory buses, and ISA/PCI

buses) are avoided.

Do not route the DXP-DXN lines next to the deflection

coils of a CRT. Also, do not route the traces across a

fast memory bus, which can easily introduce +30°C

error, even with good filtering, Otherwise, most noise

sources are fairly benign.

Route the DXP and DXN traces in parallel and in close

proximity to each other, away from any high-voltage

traces such as +12VDC. Leakage currents from PC

board contamination must be dealt with carefully,

since a 20M

Ω leakage path from DXP to ground

causes about +1°C error.

Route the 2 pairs of DXP1-DXN and DXP2-DXN

traces independently (Figure 2a). Connect the com-

mon DXN as close as possible to the DXN pin on IC

(Figure 2a).

Connect guard traces to GND on either side of the

DXP-DXN traces (Figure 2b). With guard traces in place,

routing near high-voltage traces is no longer an issue.

Route through as few vias and crossunders as possi-

ble to minimize copper/solder thermocouple effects.

When introducing a thermocouple, make sure that

both the DXP and the DXN paths have matching

thermocouples. In general, PC board- induced ther-

mocouples are not a serious problem, A copper-solder

thermocouple exhibits 3µV/°C, and it takes about

200µV of voltage error at DXP-DXN to cause a +1°C

measurement error. So, most parasitic thermocouple

errors are swamped out.

Use wide traces. Narrow ones are more inductive and

tend to pick up radiated noise. The 10 mil widths and

spacing recommended on Figure 2 aren't absolutely

necessary (as they offer only a minor improvement in

leakage and noise), but try to use them where practical.