English
Language : 

MIC2776_05 Datasheet, PDF (6/8 Pages) Micrel Semiconductor – Micro-Power Low Voltage Supervisor
MIC2776
Application Information
Programming the Voltage Threshold
Referring to the “Typical Application Circuit”, the voltage
threshold is calculated as follows:
VTH
=
VREF
×
(R1+ R2)
R2
where VREF = 0.300V
In order to provide the additional criteria needed to solve
for the resistor values, the resistors can be selected such
that the two resistors have a given total value, that is, R1
+ R2 = RTOTAL. Imposing this condition on the resistor val-
ues provides two equations that can be solved for the two
unknown resistor values. A value such as 1MΩ for RTOTAL
is a reasonable choice since it keeps quiescent current to a
generally acceptable level while not causing any measurable
errors due to input bias currents. The larger the resistors, the
larger the potential errors due to input bias current (IIN). The
maximum recommended value of RTOTAL is 3MΩ.
Applying this criteria and rearranging the VTH expression to
solve for the resistor values gives:
( )( ) R2 = RTOTAL VREF
VTH
R1 = RTOTAL – R2
Application Example
Figure 1 below illustrates a hypothetical MIC2776 application
in which the MIC2776 is used to monitor the core supply of a
high-performance CPU or DSP. The core supply, VCORE, in
this example is 1.0V ±5%. The main power rail and I/O volt-
age, VI/O, is 2.5V ±5%. As shown in Figure 1, the MIC2776
is powered by VI/O. The minimum value of VI/O is 2.5V –5%
= 2.375V; the maximum is 2.5V +5% = 2.625V. This is well
within the MIC2776’s power supply range of 1.5V to 5.5V.
Resistors R1 and R2 must be selected to correspond to the
VCORE supply of 1.0V. The goal is to insure that the core supply
voltage is adequate to insure proper operation, i.e., VCORE
≥ (1.0V –5%) = 0.950V. Because there is always a small
degree of uncertainty due to the accuracy of the resistors,
variations in the devices’ voltage reference, etc., the threshold
will be set slightly below this value. The potential variation in
the MIC2776’s voltage reference is specified as ±1.5%. The
resistors chosen will have their own tolerance specification.
This example will assume the use of 1% accurate resistors.
The potential worst-case error contribution due to input bias
current can be calculated once the resistor values are chosen.
If the guidelines above regarding the maximum total value of
R1+R2 are followed, this error contribution will be very small
thanks to the MIC2776’s very low input bias current.
Micrel, Inc
To summarize, the various potential error sources are:
• Variation in VREF: specified at ±1.5%
• Resistor tolerance:
chosen by designer (typically ≤ ±1%)
• Input bias current, IIN:
calculated once resistor values are known, typically
very small
Taking the various potential error sources into account, the
threshold voltage will be set slightly below the minimum VCORE
specification of 0.950V so that when the actual threshold
voltage is at its maximum, it will not intrude into the normal
operating range of VCORE. The target threshold voltage will
be set as follows:
Given that the total tolerance on VTH is [VREF tolerance] +
[resistor tolerance]
= ±1.5% + ±1% = ±2.5%,
and VTH(max) = VCORE(min),
then VCORE(min) = VTH + 2.5% VTH = 1.025 VTH,
therefore, solving for VTH results in
VTH
=
VCORE(min)
1.025
=
0.950
1.025
= 0.9268V
Solving for R1 and R2 using this value for VTH and the equa-
tions above yields:
R1 = 676.3kΩ ≈ 673kΩ
R2 = 323.7kΩ ≈ 324kΩ
The resulting circuit is shown in Figure 1.
Input Bias Current Effects
Now that the resistor values are known, it is possible to cal-
culate the maximum potential error due to input bias current,
IIN. As shown in the “Electrical Characteristics” table, the
maximum value of IIN is 10nA. (Note that the typical value
is a much smaller 5pA!) The magnitude of the offset caused
by IIN is given by:
VERROR = IIN(max) × (R1||R2) =
VERROR = ±1 × 10-8A × 2.189 ×105Ω =
VERROR = ±2.189 × 10-3V =
VERROR = ±2.189mV
The typical error is about three orders of magnitude lower
than this - close to one microvolt! Generally, the error
due to input bias can be discounted. If it is to be taken
into account, simply adjust the target threshold voltage
downward by this amount and recalculate R1 and R2. The
resulting value will be very close to optimum. If accuracy
is more important than the quiescent current in the
resistors, simply reduce the value of RTOTAL to minimize
offset errors.
MIC2776
6
November 2005