Chapter 5 CHIP LAYOUT DESCRIPTION AND EXPERIMENTAL
5.3.1 Experimental Results of Type D
At first, we use voltage meter to measure the actual resistor values after fabrication, and the measurement result is shown in Table-16.
Table-16 Comparison of the design target and measured data of resistors of Type D circuit
Type D (n-diff. pair)
Element name Design target Measured data Tolerance RXN + RAN + R1N N+ diff 76.0k 71.6 k - 5.8 % RXN + R2AN N+ diff 174.0k 163.3k - 6.1 % RYN +R2BN N+ diff 174.0k 163.2k - 6.2 % R3N N+ diff 111.0k 104.5k - 5.9 % RS1N + RS2N PP+ diff 1100k 1082k - 1.6 %
Substitute the measured resistor value into the bandgap circuit as shown in Fig. 5-2 and re-simulate the circuit again. Next compare the simulated result with the measured data.
MS1 37.7k nodRBn
m=15
590k pod 10/20
m=11 RS1n 492k pod
17.5/15
5.6k nod Ryn MD2
MA02
RAn37.7k nod
20/20 R3n
104.5k nod
R1n28.3k nod
PART (I) Vref (Reference Voltage) vs. Vdd Measured Result of Type D
0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 3 . 0 3 . 5 4 . 0 4 . 5
0 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0 6 0 0 7 0 0 8 0 0 9 0 0
n-Vref (mV)
V d d ( V )
I C N o . 1 I C N o . 2 I C N o . 3 I C N o . 4 I C N o . 5 I C N o . 6 n - V r e f v s . V d d
Fig. 5-3 Measured n-Vref vs. Vdd of Type D
0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 3 . 0 3 . 5 4 . 0 4 . 5 4 0 0
5 0 0 6 0 0 7 0 0 8 0 0 9 0 0
M e a .
n-Vref (mV)
V d d ( V )
A v e r a g e o f m e a s u r e d n - V r e f S i m u l a t i o n a t T T
A v e r a g e v a l u e o f n - V r e f v s . V d d
S i m .
Fig. 5-4 Comparison of n-Vref vs. Vdd of Type D between simulation and measurement
Note : mean value = 829.5 mV STD value = 12.8mV
PART (II) Temperature Compensation Curve (1) Simulation Result of Type D at TT
Fig. 5-5 Simulated TC curve of Type D under typical condition axis X : temperature (℃) ; axis Y : n-Vref (mV)
At Vdd =1.3V from – 40 to 120℃
TCF(eff) ＝ 828mV
1 (
) 40 ( 120
7 . 827 7 . 829
−
−
− mV ) ＝14.9ppm/℃
(2) Measured Result of Type D at Vdd = 3.0V, 2.0V, 1.3V
- 4 0 - 2 0 0 2 0 4 0 6 0 8 0 1 0 0 1 2 0
8 1 8 8 2 0 8 2 2 8 2 4 8 2 6 8 2 8 8 3 0 8 3 2
V d d = 2 . 0 V 3 . 0 V
n-Vref (mV)
T e m p e r a t u r e ( C )
n - V r e f @ V d d = 3 . 0 V n - V r e f @ V d d = 2 . 0 V n - V r e f @ V d d = 1 . 3 V A v e r a g e T C C u r v e o f n - V r e f
V d d = 1 . 3 V
Fig. 5-6 Measured TC curve of Type D
At Vdd =1.3V，from – 40 to 120℃
1 826.5−822
(3) Experimental Results discussion
1. The TCF(eff) by Simulation is 15.0ppm/℃. The measured data of TCF(eff) is 34.1ppm/℃. It is acceptable.
2. It is at the General Case that we get TCF(eff) =34.1ppm/℃. Right now, we want to observe the extreme case, Vdd = 1.1V.
At Vdd = 1.1V, we get the TCF(eff) =73.6 ppm/℃, as shown Fig. 5-7.
- 4 0 - 2 0 0 2 0 4 0 6 0 8 0 1 0 0 1 2 0
8 1 8 8 2 0 8 2 2 8 2 4 8 2 6 8 2 8 8 3 0
n - V r e f @ V d d = 1 . 1 v A v e r a g e T C C u r v e o f n - V r e f
n-Vref (mV)
T e m p e r a t u r e ( C )
Fig. 5-7 Measured TC curve of Type D at the worst case
3. Next, we want to discuss an interesting phenomenon：
Why do the measurement data of TC curve look like Temp
Vref
,
rather than Temp
Vref
, which is presented by the simulation.
We think this phenomenon is related with the TSMC 0.35μm device characteristic.
According to theory, two temperature-compensated currents create bandgap output reference voltage. One is the PTAT current (coming from the thermal voltage ΦT)；the other is negative temperature coefficient current (coming from the VBE voltage ).
In general, the ΦT vs. temperature curve is like T ；and the VBE
vs. temperature curve is like T
VB E
When ΦT combine with the VBE, the whole TC curve was dominated by ΦT at low temperature and by VBE at high temperature. So, we can get the curve as shown below.
T
＋
TVBE
＝
TempVref
Right now, by the measurement result, we make an assumption that：for
TSMC 0.35μm device, the ΦT vs. temp. curve is still like T, but the
VBE vs. temp. curve is like T
VB E
.
Based on this kind of assumption, when ΦT combine with the VBE, the whole TC curve was dominated by VBE at low temperature and by ΦTat high temperature. So, we can get the curve as shown below.
T
＋
TVBE
＝
TempVref
ΦT
ΦT
ΦT
ΦT
Next we want to verify the assumption by measure the temperature coefficient of VBE and ΦT respectively.
(i) Verify the assumption for the temperature coefficient of ΦT
We verify the temperature coefficient of ΦT by measuring the ∆VBE . Note : ΔVBE = VBE1 － VBE2 = ΦT ㏑(m)
-20 0 20 40 60 80 100 120
65 70 75 80 85 90 95 100 105 110
ΔVBE vs. T @ Vdd=1.8v ΔVBE Temperature Coefficient Curve
ΔV BE (mV)
Temperature (C)
Fig. 5-8 ∆VBE Temperature Coefficient Curve
For -10 ~ 80 ℃ ：TC = (
) 10 ( 80
72 3 . 92
−
−
− mV ) = 0.225 Volt /℃
For 100 ~ 115 ℃ ：TC = (
) 100 ( 115
9 . 97 9 . 103
−
− mV ) = 0.4 Volt /℃
The temp coefficient of low temp ＜ The temp coefficient of high temp
So, the ΦT vs. temp. curve is like T , not as expect T
ΦT
ΦT
(ii) Verify the assumption by measuring the temperature coefficient of VBE.
-20 0 20 40 60 80 100 120
460 480 500 520 540 560 580 600 620 640 660 680 700 720 740
VBE vs. T @ Vdd=1.8v VBE Temperature Coefficient Curve
V BE (mV)
Temperature( C)
Fig. 5-9 VBE Temperature Coefficient Curve
For –10 ~ 80 ℃ ：TC = (
) 10 ( 80
720 552
−
−
− mV ) = - 1.866 Volt /℃
For 100 ~ 115 ℃ ：TC = (
) 100 ( 115
513 483
−
− mV ) = - 2 Volt /℃
The temperature coefficient for low temp. and high temp. are almost the same.
So, the VBE vs. temp. curve is like T
VBE
, not as expect T
VBE
Although the temperature coefficients of ΦT and VBE do not meet our expectation, however, it can prove that the whole TC curve is affected by the temperature coefficient of thermal voltage (ΦT) and result in the case that the curve face upward.
V Vref
ΦT
Put the measured data from Fig. 5-8 and Fig. 5-9 into (5.1), we can get the similar measured TC curve as shown in Fig. 5-10
) (
) 1
( )
(
2 1
3 2
1 2 1
3 1
3
B B
EB os
A A BE
REF R R
V R R V
R R
V R R V R
+ + +
+ Δ
= (5.1)
-20 0 20 40 60 80 100 120
750 760 770 780 790 800
p-Vref @ Vdd=1.8v No.5 p-Vref Temperature Compensation Curve
p-Vref (mV)
Temperature (C)
Fig. 5-10 Measured TC curve of Type C - No.5 p-Vref
PART (III) Transient Response (1) Simulation Result of Type D at TT
Fig. 5-11 Simulated transient response of Type D under typical condition upper axis X : time (sec) ; axis Y : Vdd (V)
lower axis X : time (sec) ; axis Y : n-Vref (mV) (2) Measured Result of Type D at AC mode
Fig. 5-12 Measured transient response of Type D at AC mode upper axis X : time (sec) ; axis Y : Vdd (V)
lower axis X : time (sec) ; axis Y : n-Vref (mV) Description：
1. The input signal (marked by the yellow curve) is shown on the DC mode but the output reference voltage (marked by the red curve) is shown on the AC mode.
2. The input voltage varies between 1.550V and 2.130V. The rise time ≒ 24.5us (as shown on the green circle), and the peak to peak voltage of
PART (IV) PSRR (Power Supply Rejection Ratio)
(1) Simulation Result of Type D under typical condition and Vdd =1.3V
Fig. 5-13 Simulated PSRR of Type D under typical condition axis X : frequency (Hz) ; axis Y : PSRR of n-Vref (dB)
(1) Measured Result of Type D by using Oscilloscope at Vdd = 1.5V, 2.0V
1 k 1 0 k 1 0 0 k
- 3 5 - 3 0 - 2 5 - 2 0 - 1 5 - 1 0 - 5 0 5
V d d = 1 . 5 V
PSRR (dB)
F r e q . ( H z )
V d d = 1 . 5 V V d d = 2 . 0 V P S R R ( n - V r e f ) f o r V d d = 1 . 5 V , 2 . 0 V
V d d = 2 . 0 V
Fig. 5-14 Measured PSRR of Type D
(2) Comparison：Simulation versus Measurement as shown in Table-17.
Table-17 Summary table of PSRR of Type D at Vdd = 1.5V
PSRR 1KHz 5KHz 10KHz 50KHz 100KHz Simulation -52dB -50dB -45dB -30dB -23dB Measurement -25dB -16dB -10dB -0.68dB -0.42dB
(4) Experimental Results discussion
The PSRR performance is not as good as we expect. We guess the reason of the poor PSRR performance may be related with the parasitic resistor and parasitic capacitance.