static and dynamic translinear circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are...
TRANSCRIPT
![Page 1: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/1.jpg)
Static and Dynamic Translinear Circuits
Bradley A. Minch
Mixed Analog-Digital VLSI Circuits and Systems LabFranklin W. Olin College of Engineering
Needham, MA 02492–1200
May 20, 2010
![Page 2: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/2.jpg)
1
Translinear Circuits: What’s in a Name?
In 1975, Barrie Gilbert coined the term translinear to describe a class of circuits whoselarge-signal behavior hinges both on the precise exponential I/V relationship of the bipolartransistor and on the intimate thermal contact and close matching of monolithically integrateddevices.
The word translinear refers to the exponential I/V characteristic of the bipolar transistor—itstransconductance is linear in its collector current:
IC = IseVBE/UT =⇒ gm =
∂IC
∂VB= Ise
VBE/UT︸ ︷︷ ︸IC
· 1UT
=IC
UT.
Gilbert also meant the word translinear to refer to circuit analysis and design principles thatbridge the gap between the familiar territory of linear circuits and the uncharted domain ofnonlinear circuits.
![Page 3: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/3.jpg)
1
Translinear Circuits: What’s in a Name?
In 1975, Barrie Gilbert coined the term translinear to describe a class of circuits whoselarge-signal behavior hinges both on the precise exponential I/V relationship of the bipolartransistor and on the intimate thermal contact and close matching of monolithically integrateddevices.
The word translinear refers to the exponential I/V characteristic of the bipolar transistor—itstransconductance is linear in its collector current:
IC = IseVBE/UT =⇒ gm =
∂IC
∂VB= Ise
VBE/UT︸ ︷︷ ︸IC
· 1UT
=IC
UT.
Gilbert also meant the word translinear to refer to circuit analysis and design principles thatbridge the gap between the familiar territory of linear circuits and the uncharted domain ofnonlinear circuits.
![Page 4: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/4.jpg)
1
Translinear Circuits: What’s in a Name?
In 1975, Barrie Gilbert coined the term translinear to describe a class of circuits whoselarge-signal behavior hinges both on the precise exponential I/V relationship of the bipolartransistor and on the intimate thermal contact and close matching of monolithically integrateddevices.
The word translinear refers to the exponential I/V characteristic of the bipolar transistor—itstransconductance is linear in its collector current:
IC = IseVBE/UT =⇒ gm =
∂IC
∂VB= Ise
VBE/UT︸ ︷︷ ︸IC
· 1UT
=IC
UT.
Gilbert also meant the word translinear to refer to circuit analysis and design principles thatbridge the gap between the familiar territory of linear circuits and the uncharted domain ofnonlinear circuits.
![Page 5: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/5.jpg)
2
Gummel Plot of a Forward-Active Bipolar Transistor
���� ���� ��� ���� ��� ���� �������
����
����
��
���
����
��
���
����
��
�
��
���
��� ���
����������
��������
������
���������
����
��
�
���
��
�
![Page 6: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/6.jpg)
3
Translinearity of the Forward-Active Bipolar Transistor
�� ���� �� ��� ���� �� �������
����
���
����
��
���
����
�
� ��
�
��
���
�����
��
���������
�
���
������
���
����
gm =∂IC
∂VB=
IC
UT
δIC ≈ gmδVB
![Page 7: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/7.jpg)
4
The Translinear Principle
Consider a closed loop of base-emitter junctions of four closely matched npn bipolar transistorsbiased in the forward-active region and operating at the same temperature. Kirchhoff’s voltagelaw (KVL) implies that
V1 + V2 = V3 + V4
UT logI1
Is+ UT log
I2
Is= UT log
I3
Is+ UT log
I4
Is
logI1I2
I2s
= logI3I4
I2s
I1I2︸︷︷︸CCW
= I3I4︸︷︷︸CW
.
��
�� ��
��
��
�� ��
��
� !�"#$�%% %�%�&
�!'�&%(� !�"#$�%% %�%�&
This result is a particular case of Gilbert’s translinear principle (TLP): The product of theclockwise currents is equal to the product of the counterclockwise currents.
![Page 8: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/8.jpg)
5
Why Translinear Circuits?
![Page 9: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/9.jpg)
5
Why Translinear Circuits?
• Translinear circuits are universal. Conjecture: In principle, we can realize any systemwhose description we can write down as a nonlinear ODE with time as its independentvariable as a dynamic translinear circuit.
• Translinear circuits are synthesizable via highly structured methods. They should be veryamenable to the development both of CAD tools and of reconfigurable FPAA architecturesfor rapid prototyping and deployment of translinear analog signal processing systems.
• Translinear circuits are fundamentally large-signal circuits. Linear dynamic translinearcircuits are linear because of device nonlinearities rather than in spite of them.
• Translinear circuits are tunable electronically over a wide dynamic range of parameters(e.g., gains, corner frequencies, quality factors).
• Translinear circuits are robust. Carefully designed translinear circuits are temperatureinsensitive and do not depend on device or technology parameters.
![Page 10: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/10.jpg)
5
Why Translinear Circuits?
• Translinear circuits are universal. Conjecture: In principle, we can realize any systemwhose description we can write down as a nonlinear ODE with time as its independentvariable as a dynamic translinear circuit.
• Translinear circuits are synthesizable via highly structured methods. They should be veryamenable to the development both of CAD tools and of reconfigurable FPAA architecturesfor rapid prototyping and deployment of translinear analog signal processing systems.
• Translinear circuits are fundamentally large-signal circuits. Linear dynamic translinearcircuits are linear because of device nonlinearities rather than in spite of them.
• Translinear circuits are tunable electronically over a wide dynamic range of parameters(e.g., gains, corner frequencies, quality factors).
• Translinear circuits are robust. Carefully designed translinear circuits are temperatureinsensitive and do not depend on device or technology parameters.
![Page 11: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/11.jpg)
5
Why Translinear Circuits?
• Translinear circuits are universal. Conjecture: In principle, we can realize any systemwhose description we can write down as a nonlinear ODE with time as its independentvariable as a dynamic translinear circuit.
• Translinear circuits are synthesizable via highly structured methods. They should be veryamenable to the development both of CAD tools and of reconfigurable FPAA architecturesfor rapid prototyping and deployment of translinear analog signal processing systems.
• Translinear circuits are fundamentally large-signal circuits. Linear dynamic translinearcircuits are linear because of device nonlinearities rather than in spite of them.
• Translinear circuits are tunable electronically over a wide dynamic range of parameters(e.g., gains, corner frequencies, quality factors).
• Translinear circuits are robust. Carefully designed translinear circuits are temperatureinsensitive and do not depend on device or technology parameters.
![Page 12: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/12.jpg)
5
Why Translinear Circuits?
• Translinear circuits are universal. Conjecture: In principle, we can realize any systemwhose description we can write down as a nonlinear ODE with time as its independentvariable as a dynamic translinear circuit.
• Translinear circuits are synthesizable via highly structured methods. They should be veryamenable to the development both of CAD tools and of reconfigurable FPAA architecturesfor rapid prototyping and deployment of translinear analog signal processing systems.
• Translinear circuits are fundamentally large-signal circuits. Linear dynamic translinearcircuits are linear because of device nonlinearities rather than in spite of them.
• Translinear circuits are tunable electronically over a wide dynamic range of parameters(e.g., gains, corner frequencies, quality factors).
• Translinear circuits are robust. Carefully designed translinear circuits are temperatureinsensitive and do not depend on device or technology parameters.
![Page 13: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/13.jpg)
5
Why Translinear Circuits?
• Translinear circuits are universal. Conjecture: In principle, we can realize any systemwhose description we can write down as a nonlinear ODE with time as its independentvariable as a dynamic translinear circuit.
• Translinear circuits are synthesizable via highly structured methods. They should be veryamenable to the development both of CAD tools and of reconfigurable FPAA architecturesfor rapid prototyping and deployment of translinear analog signal processing systems.
• Translinear circuits are fundamentally large-signal circuits. Linear dynamic translinearcircuits are linear because of device nonlinearities rather than in spite of them.
• Translinear circuits are tunable electronically over a wide dynamic range of parameters(e.g., gains, corner frequencies, quality factors).
• Translinear circuits are robust. Carefully designed translinear circuits are temperatureinsensitive and do not depend on device or technology parameters.
![Page 14: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/14.jpg)
6
Simple EKV Model of the Saturated nMOS Transistor
We model the saturation current of an nMOS transistor by
Isat = SIs log2(1 + e(κ(VG−VT0)−VS)/2UT
)
≈
⎧⎪⎨⎪⎩
SIse(κ(VG−VT0)−VS)/UT, κ (VG − VT0) − VS < 0
SIs
4U2T
(κ (VG − VT0) − VS)2 , κ (VG − VT0) − VS > 0,
where
UT =kT
q, S =
W
L, Is =
2μCoxU2T
κ, and κ =
Cox
Cox + Cdep.
Weak inversion operation corresponds to Isat � SIs, moderate inversion operation cor-responds to Isat ≈ SIs, and strong inversion operation to Isat � SIs. Note that SIs isapproximately twice the saturation current at threshold.
![Page 15: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/15.jpg)
7
Saturation Current of an nMOS Transistor
� ��� � ��� � ��� ��)��
����
����
��
���
����
��
���
� �*
&��
��
���
�����������������
������+��
��*&���� !,��������)-��� ����
��*&��������)-��� ���
����������
��*&����
��������)-���
� �)
���
��*&
![Page 16: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/16.jpg)
8
Translinearity of the Saturated nMOS Transistor
���� ���� �� ��� ���� �� �����*&��
��
����
��
���
����
��
���
� ��
�
���
�����
����
����*&
�������������+��
����������
����
�������*&
��-�-��*&���� ����
�������*&
�)
���
��*&����*&
���)
gm =∂Isat
∂VG=
κIsat
UT, Isat � SIs
δIsat ≈ gmδVG
![Page 17: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/17.jpg)
9
Weak-Inversion MOS Translinear Principle
When designing weak-inversion MOS translinear circuits, if we restrict ourselves to usingtranslinear loops that alternate between clockwise and counterclockwise elements, we obtainGilbert’s original TLP, with no dependence on the body effect (i.e., κ).
stacked loop alternating loop
��
�� ��
��
���� ����
TLP: I1Iκ2 = Iκ
3 I4 TLP: I1I3 = I2I4
This restriction does not limit the class of systems that we can implement. However, designsbased on alternating loops generally consume more current but operate on a lower powersupply voltage than do designs based on stacked loops.
![Page 18: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/18.jpg)
10
Static Translinear Circuits: Squaring/Reciprocal
TLP =⇒ I2x = IyIz =⇒ Iz =
I2x
Iy
� ��
���� �
�� ��
![Page 19: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/19.jpg)
10
Static Translinear Circuits: Squaring/Reciprocal
�� ��� ���� �� ������
��
���
����
��
���
��
��
� �
��
�
����
�
���
�
� ��
��
�� ��
Iz =I2x
Iy
![Page 20: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/20.jpg)
10
Static Translinear Circuits: Squaring/Reciprocal
�� ��� ���� �� ������
��
���
����
��
���
��
��
� �
���
�����
����
� ��
��
�� ��
Iz =I2x
Iy
![Page 21: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/21.jpg)
11
Static Translinear Circuit Synthesis: Pythagorator
Synthesize a two-dimensional vector-magnitude circuit implementing
r =√
x2 + y2, where x > 0 and y > 0.
We represent each signal as a ratio of a signal current to the unit current:
x ≡ Ix
I1, y ≡ Iy
I1, and r ≡ Ir
I1.
We substitute these into the original equation and rearrange to obtain
Ir
I1=
√(Ix
I1
)2
+(
Iy
I1
)2
=⇒ I2r = I2
x + I2y =⇒ Ir =
I2x
Ir︸︷︷︸Ir1
+I2y
Ir︸︷︷︸Ir2
![Page 22: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/22.jpg)
11
Static Translinear Circuit Synthesis: Pythagorator
Synthesize a two-dimensional vector-magnitude circuit implementing
r =√
x2 + y2, where x > 0 and y > 0.
We represent each signal as a ratio of a signal current to the unit current:
x ≡ Ix
I1, y ≡ Iy
I1, and r ≡ Ir
I1.
We substitute these into the original equation and rearrange to obtain
Ir
I1=
√(Ix
I1
)2
+(
Iy
I1
)2
=⇒ I2r = I2
x + I2y =⇒ Ir =
I2x
Ir︸︷︷︸Ir1
+I2y
Ir︸︷︷︸Ir2
![Page 23: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/23.jpg)
11
Static Translinear Circuit Synthesis: Pythagorator
Synthesize a two-dimensional vector-magnitude circuit implementing
r =√
x2 + y2, where x > 0 and y > 0.
We represent each signal as a ratio of a signal current to the unit current:
x ≡ Ix
I1, y ≡ Iy
I1, and r ≡ Ir
I1.
We substitute these into the original equation and rearrange to obtain
Ir
I1=
√(Ix
I1
)2
+(
Iy
I1
)2
=⇒ I2r = I2
x + I2y =⇒ Ir =
I2x
Ir︸︷︷︸Ir1
+I2y
Ir︸︷︷︸Ir2
![Page 24: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/24.jpg)
11
Static Translinear Circuit Synthesis: Pythagorator
Synthesize a two-dimensional vector-magnitude circuit implementing
r =√
x2 + y2, where x > 0 and y > 0.
We represent each signal as a ratio of a signal current to the unit current:
x ≡ Ix
I1, y ≡ Iy
I1, and r ≡ Ir
I1.
We substitute these into the original equation and rearrange to obtain
Ir
I1=
√(Ix
I1
)2
+(
Iy
I1
)2
=⇒ I2r = I2
x + I2y =⇒ Ir =
I2x
Ir︸︷︷︸Ir1
+I2y
Ir︸︷︷︸Ir2
![Page 25: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/25.jpg)
11
Static Translinear Circuit Synthesis: Pythagorator
Synthesize a two-dimensional vector-magnitude circuit implementing
r =√
x2 + y2, where x > 0 and y > 0.
We represent each signal as a ratio of a signal current to the unit current:
x ≡ Ix
I1, y ≡ Iy
I1, and r ≡ Ir
I1.
We substitute these into the original equation and rearrange to obtain
Ir
I1=
√(Ix
I1
)2
+(
Iy
I1
)2
=⇒ I2r = I2
x + I2y =⇒ Ir =
I2x
Ir︸︷︷︸Ir1
+I2y
Ir︸︷︷︸Ir2
![Page 26: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/26.jpg)
11
Static Translinear Circuit Synthesis: Pythagorator
Synthesize a two-dimensional vector-magnitude circuit implementing
r =√
x2 + y2, where x > 0 and y > 0.
We represent each signal as a ratio of a signal current to the unit current:
x ≡ Ix
I1, y ≡ Iy
I1, and r ≡ Ir
I1.
We substitute these into the original equation and rearrange to obtain
Ir
I1=
√(Ix
I1
)2
+(
Iy
I1
)2
=⇒ I2r = I2
x + I2y =⇒ Ir =
I2x
Ir︸︷︷︸Ir1
+I2y
Ir︸︷︷︸Ir2
![Page 27: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/27.jpg)
12
Static Translinear Circuit Synthesis: Pythagorator
TLP: Ir1Ir = I2x
Ir2Ir = I2y
KCL: Ir = Ir1+Ir2
![Page 28: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/28.jpg)
12
Static Translinear Circuit Synthesis: Pythagorator
TLP: Ir1Ir = I2x
Ir2Ir = I2y
KCL: Ir = Ir1+Ir2
��
��
��� ��
��
![Page 29: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/29.jpg)
12
Static Translinear Circuit Synthesis: Pythagorator
TLP: Ir1Ir = I2x
Ir2Ir = I2y
KCL: Ir = Ir1+Ir2
��
��
��� ��
��
��
�� ��
�� ���
![Page 30: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/30.jpg)
12
Static Translinear Circuit Synthesis: Pythagorator
TLP: Ir1Ir = I2x
Ir2Ir = I2y
KCL: Ir = Ir1+Ir2
��
��
��� ��
��
��
�� ���
![Page 31: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/31.jpg)
12
Static Translinear Circuit Synthesis: Pythagorator
TLP: Ir1Ir = I2x
Ir2Ir = I2y
KCL: Ir = Ir1+Ir2
����� ��
�� �� ��
� ����
![Page 32: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/32.jpg)
12
Static Translinear Circuit Synthesis: Pythagorator
TLP: Ir1Ir = I2x
Ir2Ir = I2y
KCL: Ir = Ir1+Ir2
� � � �
�������
�� �� ��
� ����
������ ������ ������ ������
![Page 33: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/33.jpg)
13
Dynamic Translinear Circuit Synthesis: Output Structures
noninverting inverting
��
��
��
��
���
��
��
��
����
���
��
��
In = Iτe(Vn−V0)/UT In = Iτe
κ(V0−Vn)/UT
∂In
∂Vn=
In
UT
∂In
∂Vn= −κIn
UT
![Page 34: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/34.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
Synthesize a first-order low-pass filter described by
τdy
dt+ y = x, where x > 0.
We represent each signal as a ratio of a signal current to the unit current:
x ≡ Ix
I1and y ≡ Iy
I1.
Substituting these into the ODE, we obtain
τd
dt
(Iy
I1
)+
Iy
I1=
Ix
I1=⇒ τ
dIy
dt+ Iy = Ix.
![Page 35: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/35.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
Synthesize a first-order low-pass filter described by
τdy
dt+ y = x, where x > 0.
We represent each signal as a ratio of a signal current to the unit current:
x ≡ Ix
I1and y ≡ Iy
I1.
Substituting these into the ODE, we obtain
τd
dt
(Iy
I1
)+
Iy
I1=
Ix
I1=⇒ τ
dIy
dt+ Iy = Ix.
![Page 36: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/36.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
Synthesize a first-order low-pass filter described by
τdy
dt+ y = x, where x > 0.
We represent each signal as a ratio of a signal current to the unit current:
x ≡ Ix
I1and y ≡ Iy
I1.
Substituting these into the ODE, we obtain
τd
dt
(Iy
I1
)+
Iy
I1=
Ix
I1=⇒ τ
dIy
dt+ Iy = Ix.
![Page 37: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/37.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
Synthesize a first-order low-pass filter described by
τdy
dt+ y = x, where x > 0.
We represent each signal as a ratio of a signal current to the unit current:
x ≡ Ix
I1and y ≡ Iy
I1.
Substituting these into the ODE, we obtain
τd
dt
(Iy
I1
)+
Iy
I1=
Ix
I1=⇒ τ
dIy
dt+ Iy = Ix.
![Page 38: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/38.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
To implement the time derivative, we introduce a log-compressed voltage state variable, Vy.Using the chain rule, we can express the preceding equation as
τ∂Iy
∂Vy· dVy
dt+ Iy = Ix =⇒ τ
(− κ
UTIy
)dVy
dt+ Iy = Ix
=⇒ −κτ
UT· dVy
dt+ 1 =
Ix
Iy=⇒ − κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVy
dt︸ ︷︷ ︸Ic
+1 =Ix
Iy
=⇒ −Ic
Iτ+ 1 =
Ix
Iy=⇒ Iτ − Ic =
IτIx
Iy︸︷︷︸Ip
.
![Page 39: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/39.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
To implement the time derivative, we introduce a log-compressed voltage state variable, Vy.Using the chain rule, we can express the preceding equation as
τ∂Iy
∂Vy· dVy
dt+ Iy = Ix =⇒ τ
(− κ
UTIy
)dVy
dt+ Iy = Ix
=⇒ −κτ
UT· dVy
dt+ 1 =
Ix
Iy=⇒ − κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVy
dt︸ ︷︷ ︸Ic
+1 =Ix
Iy
=⇒ −Ic
Iτ+ 1 =
Ix
Iy=⇒ Iτ − Ic =
IτIx
Iy︸︷︷︸Ip
.
![Page 40: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/40.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
To implement the time derivative, we introduce a log-compressed voltage state variable, Vy.Using the chain rule, we can express the preceding equation as
τ∂Iy
∂Vy· dVy
dt+ Iy = Ix =⇒ τ
(− κ
UTIy
)dVy
dt+ Iy = Ix
=⇒ −κτ
UT· dVy
dt+ 1 =
Ix
Iy=⇒ − κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVy
dt︸ ︷︷ ︸Ic
+1 =Ix
Iy
=⇒ −Ic
Iτ+ 1 =
Ix
Iy=⇒ Iτ − Ic =
IτIx
Iy︸︷︷︸Ip
.
![Page 41: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/41.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
To implement the time derivative, we introduce a log-compressed voltage state variable, Vy.Using the chain rule, we can express the preceding equation as
τ∂Iy
∂Vy· dVy
dt+ Iy = Ix =⇒ τ
(− κ
UTIy
)dVy
dt+ Iy = Ix
=⇒ −κτ
UT· dVy
dt+ 1 =
Ix
Iy=⇒ − κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVy
dt︸ ︷︷ ︸Ic
+1 =Ix
Iy
=⇒ −Ic
Iτ+ 1 =
Ix
Iy=⇒ Iτ − Ic =
IτIx
Iy︸︷︷︸Ip
.
![Page 42: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/42.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
To implement the time derivative, we introduce a log-compressed voltage state variable, Vy.Using the chain rule, we can express the preceding equation as
τ∂Iy
∂Vy· dVy
dt+ Iy = Ix =⇒ τ
(− κ
UTIy
)dVy
dt+ Iy = Ix
=⇒ −κτ
UT· dVy
dt+ 1 =
Ix
Iy=⇒ − κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVy
dt︸ ︷︷ ︸Ic
+1 =Ix
Iy
=⇒ −Ic
Iτ+ 1 =
Ix
Iy=⇒ Iτ − Ic =
IτIx
Iy︸︷︷︸Ip
.
![Page 43: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/43.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
To implement the time derivative, we introduce a log-compressed voltage state variable, Vy.Using the chain rule, we can express the preceding equation as
τ∂Iy
∂Vy· dVy
dt+ Iy = Ix =⇒ τ
(− κ
UTIy
)dVy
dt+ Iy = Ix
=⇒ −κτ
UT· dVy
dt+ 1 =
Ix
Iy=⇒ − κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVy
dt︸ ︷︷ ︸Ic
+1 =Ix
Iy
=⇒ −Ic
Iτ+ 1 =
Ix
Iy=⇒ Iτ − Ic =
IτIx
Iy︸︷︷︸Ip
.
![Page 44: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/44.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
To implement the time derivative, we introduce a log-compressed voltage state variable, Vy.Using the chain rule, we can express the preceding equation as
τ∂Iy
∂Vy· dVy
dt+ Iy = Ix =⇒ τ
(− κ
UTIy
)dVy
dt+ Iy = Ix
=⇒ −κτ
UT· dVy
dt+ 1 =
Ix
Iy=⇒ − κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVy
dt︸ ︷︷ ︸Ic
+1 =Ix
Iy
=⇒ −Ic
Iτ+ 1 =
Ix
Iy=⇒ Iτ − Ic =
IτIx
Iy︸︷︷︸Ip
.
![Page 45: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/45.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
To implement the time derivative, we introduce a log-compressed voltage state variable, Vy.Using the chain rule, we can express the preceding equation as
τ∂Iy
∂Vy· dVy
dt+ Iy = Ix =⇒ τ
(− κ
UTIy
)dVy
dt+ Iy = Ix
=⇒ −κτ
UT· dVy
dt+ 1 =
Ix
Iy=⇒ − κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVy
dt︸ ︷︷ ︸Ic
+1 =Ix
Iy
=⇒ −Ic
Iτ+ 1 =
Ix
Iy=⇒ Iτ − Ic =
IτIx
Iy︸︷︷︸Ip
.
![Page 46: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/46.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
TLP: IpIy = IxIτ KCL: Ic + Ip = Iτ
![Page 47: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/47.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
TLP: IpIy = IxIτ KCL: Ic + Ip = Iτ
���
��
���
��
![Page 48: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/48.jpg)
14
Dynamic Translinear Circuit Synthesis: First-Order LPF
TLP: IpIy = IxIτ KCL: Ic + Ip = Iτ
��
�
��
��
��
�
���
��
��
�
![Page 49: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/49.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
Synthesize an RMS-to-DC converter described by
x = w2, τdy
dt+ y = x, and z =
√y.
We can eliminate x and y from the system description by substituting
x = w2, y = z2, anddy
dt= 2z
dz
dt
into the linear ODE describing the low-pass filter, obtaining a first-order nonlinear ODEdescribing the system given by
2τzdz
dt+ z2 = w2.
![Page 50: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/50.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
Synthesize an RMS-to-DC converter described by
x = w2, τdy
dt+ y = x, and z =
√y.
We can eliminate x and y from the system description by substituting
x = w2, y = z2, anddy
dt= 2z
dz
dt
into the linear ODE describing the low-pass filter, obtaining a first-order nonlinear ODEdescribing the system given by
2τzdz
dt+ z2 = w2.
![Page 51: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/51.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
w+ ≡ Iw+
I1= 1
2
(1 + eκ(Vw−V0)/UT
)w− ≡ Iw−
I1= 1
2
(1 + e−κ(Vw−V0)/UT
)w ≡ Iw
I1= w+ − w−
w = sinhκ (Vw − V0)
UT
w′ ≡ Iw′
I1= w++w−−1
w′ = coshκ (Vw − V0)
UT
w2 = (w′)2 − 1
��
��.
��
�/
��.
���
�/
�/
��0
��
��.
���
��
![Page 52: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/52.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
The input signal, w, can be positive or negative. To remedy this situation, we adopt a sinhrepresentation for w and and define an associated signal, w′, as just described. Substitutingw2 = (w′)2 − 1 into the nonlinear ODE, we obtain
2τzdz
dt+ z2 = (w′)2 − 1.
We represent each signal as a ratio of a signal current to the unit current:
w′ ≡ Iw′
I1and z ≡ Iz
I1.
Substituting these into the nonlinear ODE, we obtain
2τIz
I1· d
dt
(Iz
I1
)+
(Iz
I1
)2
=(
Iw′
I1
)2
− 1 =⇒ 2τIzdIz
dt+ I2
z = I2w′ − I2
1 .
![Page 53: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/53.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
The input signal, w, can be positive or negative. To remedy this situation, we adopt a sinhrepresentation for w and and define an associated signal, w′, as just described. Substitutingw2 = (w′)2 − 1 into the nonlinear ODE, we obtain
2τzdz
dt+ z2 = (w′)2 − 1.
We represent each signal as a ratio of a signal current to the unit current:
w′ ≡ Iw′
I1and z ≡ Iz
I1.
Substituting these into the nonlinear ODE, we obtain
2τIz
I1· d
dt
(Iz
I1
)+
(Iz
I1
)2
=(
Iw′
I1
)2
− 1 =⇒ 2τIzdIz
dt+ I2
z = I2w′ − I2
1 .
![Page 54: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/54.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
The input signal, w, can be positive or negative. To remedy this situation, we adopt a sinhrepresentation for w and and define an associated signal, w′, as just described. Substitutingw2 = (w′)2 − 1 into the nonlinear ODE, we obtain
2τzdz
dt+ z2 = (w′)2 − 1.
We represent each signal as a ratio of a signal current to the unit current:
w′ ≡ Iw′
I1and z ≡ Iz
I1.
Substituting these into the nonlinear ODE, we obtain
2τIz
I1· d
dt
(Iz
I1
)+
(Iz
I1
)2
=(
Iw′
I1
)2
− 1 =⇒ 2τIzdIz
dt+ I2
z = I2w′ − I2
1 .
![Page 55: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/55.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
The input signal, w, can be positive or negative. To remedy this situation, we adopt a sinhrepresentation for w and and define an associated signal, w′, as just described. Substitutingw2 = (w′)2 − 1 into the nonlinear ODE, we obtain
2τzdz
dt+ z2 = (w′)2 − 1.
We represent each signal as a ratio of a signal current to the unit current:
w′ ≡ Iw′
I1and z ≡ Iz
I1.
Substituting these into the nonlinear ODE, we obtain
2τIz
I1· d
dt
(Iz
I1
)+
(Iz
I1
)2
=(
Iw′
I1
)2
− 1 =⇒ 2τIzdIz
dt+ I2
z = I2w′ − I2
1 .
![Page 56: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/56.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
To implement the time derivative, we introduce a log-compressed voltage state variable, Vz.Using the chain rule, we can express the preceding equation as
2τIz∂Iz
∂Vz· dVz
dt+ I2
z = I2w′ − I2
1 =⇒ 2τIz
(− κ
UTIz
)dVz
dt+ I2
z = I2w′ − I2
1
=⇒ −2wκ
UT· dVz
dt+ 1 =
I2w′
I2z
− I21
I2z
=⇒ − 2κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVz
dt︸ ︷︷ ︸Ic
+1 =I2w′
I2z
− I21
I2z
=⇒ −Ic
Iτ+ 1 =
I2w′
I2z
− I21
I2z
=⇒ Iτ − Ic =IτI
2w′
I2z︸ ︷︷ ︸
Ip
− IτI21
I2z︸︷︷︸
Iq
.
![Page 57: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/57.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
To implement the time derivative, we introduce a log-compressed voltage state variable, Vz.Using the chain rule, we can express the preceding equation as
2τIz∂Iz
∂Vz· dVz
dt+ I2
z = I2w′ − I2
1 =⇒ 2τIz
(− κ
UTIz
)dVz
dt+ I2
z = I2w′ − I2
1
=⇒ −2wκ
UT· dVz
dt+ 1 =
I2w′
I2z
− I21
I2z
=⇒ − 2κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVz
dt︸ ︷︷ ︸Ic
+1 =I2w′
I2z
− I21
I2z
=⇒ −Ic
Iτ+ 1 =
I2w′
I2z
− I21
I2z
=⇒ Iτ − Ic =IτI
2w′
I2z︸ ︷︷ ︸
Ip
− IτI21
I2z︸︷︷︸
Iq
.
![Page 58: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/58.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
To implement the time derivative, we introduce a log-compressed voltage state variable, Vz.Using the chain rule, we can express the preceding equation as
2τIz∂Iz
∂Vz· dVz
dt+ I2
z = I2w′ − I2
1 =⇒ 2τIz
(− κ
UTIz
)dVz
dt+ I2
z = I2w′ − I2
1
=⇒ −2wκ
UT· dVz
dt+ 1 =
I2w′
I2z
− I21
I2z
=⇒ − 2κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVz
dt︸ ︷︷ ︸Ic
+1 =I2w′
I2z
− I21
I2z
=⇒ −Ic
Iτ+ 1 =
I2w′
I2z
− I21
I2z
=⇒ Iτ − Ic =IτI
2w′
I2z︸ ︷︷ ︸
Ip
− IτI21
I2z︸︷︷︸
Iq
.
![Page 59: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/59.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
To implement the time derivative, we introduce a log-compressed voltage state variable, Vz.Using the chain rule, we can express the preceding equation as
2τIz∂Iz
∂Vz· dVz
dt+ I2
z = I2w′ − I2
1 =⇒ 2τIz
(− κ
UTIz
)dVz
dt+ I2
z = I2w′ − I2
1
=⇒ −2wκ
UT· dVz
dt+ 1 =
I2w′
I2z
− I21
I2z
=⇒ − 2κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVz
dt︸ ︷︷ ︸Ic
+1 =I2w′
I2z
− I21
I2z
=⇒ −Ic
Iτ+ 1 =
I2w′
I2z
− I21
I2z
=⇒ Iτ − Ic =IτI
2w′
I2z︸ ︷︷ ︸
Ip
− IτI21
I2z︸︷︷︸
Iq
.
![Page 60: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/60.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
To implement the time derivative, we introduce a log-compressed voltage state variable, Vz.Using the chain rule, we can express the preceding equation as
2τIz∂Iz
∂Vz· dVz
dt+ I2
z = I2w′ − I2
1 =⇒ 2τIz
(− κ
UTIz
)dVz
dt+ I2
z = I2w′ − I2
1
=⇒ −2wκ
UT· dVz
dt+ 1 =
I2w′
I2z
− I21
I2z
=⇒ − 2κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVz
dt︸ ︷︷ ︸Ic
+1 =I2w′
I2z
− I21
I2z
=⇒ −Ic
Iτ+ 1 =
I2w′
I2z
− I21
I2z
=⇒ Iτ − Ic =IτI
2w′
I2z︸ ︷︷ ︸
Ip
− IτI21
I2z︸︷︷︸
Iq
.
![Page 61: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/61.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
To implement the time derivative, we introduce a log-compressed voltage state variable, Vz.Using the chain rule, we can express the preceding equation as
2τIz∂Iz
∂Vz· dVz
dt+ I2
z = I2w′ − I2
1 =⇒ 2τIz
(− κ
UTIz
)dVz
dt+ I2
z = I2w′ − I2
1
=⇒ −2wκ
UT· dVz
dt+ 1 =
I2w′
I2z
− I21
I2z
=⇒ − 2κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVz
dt︸ ︷︷ ︸Ic
+1 =I2w′
I2z
− I21
I2z
=⇒ −Ic
Iτ+ 1 =
I2w′
I2z
− I21
I2z
=⇒ Iτ − Ic =IτI
2w′
I2z︸ ︷︷ ︸
Ip
− IτI21
I2z︸︷︷︸
Iq
.
![Page 62: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/62.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
To implement the time derivative, we introduce a log-compressed voltage state variable, Vz.Using the chain rule, we can express the preceding equation as
2τIz∂Iz
∂Vz· dVz
dt+ I2
z = I2w′ − I2
1 =⇒ 2τIz
(− κ
UTIz
)dVz
dt+ I2
z = I2w′ − I2
1
=⇒ −2wκ
UT· dVz
dt+ 1 =
I2w′
I2z
− I21
I2z
=⇒ − 2κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVz
dt︸ ︷︷ ︸Ic
+1 =I2w′
I2z
− I21
I2z
=⇒ −Ic
Iτ+ 1 =
I2w′
I2z
− I21
I2z
=⇒ Iτ − Ic =IτI
2w′
I2z︸ ︷︷ ︸
Ip
− IτI21
I2z︸︷︷︸
Iq
.
![Page 63: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/63.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
To implement the time derivative, we introduce a log-compressed voltage state variable, Vz.Using the chain rule, we can express the preceding equation as
2τIz∂Iz
∂Vz· dVz
dt+ I2
z = I2w′ − I2
1 =⇒ 2τIz
(− κ
UTIz
)dVz
dt+ I2
z = I2w′ − I2
1
=⇒ −2wκ
UT· dVz
dt+ 1 =
I2w′
I2z
− I21
I2z
=⇒ − 2κτ
CUT︸ ︷︷ ︸1/Iτ
·CdVz
dt︸ ︷︷ ︸Ic
+1 =I2w′
I2z
− I21
I2z
=⇒ −Ic
Iτ+ 1 =
I2w′
I2z
− I21
I2z
=⇒ Iτ − Ic =IτI
2w′
I2z︸ ︷︷ ︸
Ip
− IτI21
I2z︸︷︷︸
Iq
.
![Page 64: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/64.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
TLP: IpI2z = IτI
2w′
IqI2z = IτI
21
KCL: Ic+Ip = Iτ +Iq
![Page 65: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/65.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
TLP: IpI2z = IτI
2w′
IqI2z = IτI
21
KCL: Ic+Ip = Iτ +Iq
��0 ��0�� �� ����
�� ��
![Page 66: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/66.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
TLP: IpI2z = IτI
2w′
IqI2z = IτI
21
KCL: Ic+Ip = Iτ +Iq
��0 ��0�� �� ����
�� ��
�� ���� �� ����
�� ��
![Page 67: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/67.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
TLP: IpI2z = IτI
2w′
IqI2z = IτI
21
KCL: Ic+Ip = Iτ +Iq
��0 ��0�� �� ����
�� ��
�� ����
��
![Page 68: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/68.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
TLP: IpI2z = IτI
2w′
IqI2z = IτI
21
KCL: Ic+Ip = Iτ +Iq
�� ��
�� ��0 ��0
��
��
�� �� ����
![Page 69: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/69.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
TLP: IpI2z = IτI
2w′
IqI2z = IτI
21
KCL: Ic+Ip = Iτ +Iq
�� ��
��
�� ��0
��0
��0
��0
��
��
�� ��
��
�� �� ��
�� ��
��
��
��
��
��
![Page 70: Static and Dynamic Translinear Circuitsmadvlsi.olin.edu/bminch/talks/100520_cmoset.pdfcircuits are linear because of device nonlinearities rather than in spite of them. • Translinear](https://reader035.vdocuments.mx/reader035/viewer/2022081521/5ed316939344585c674ec00f/html5/thumbnails/70.jpg)
15
Dynamic Translinear Circuit Synthesis: RMS-DC Converter
TLP: IpI2z = IτI
2w′
IqI2z = IτI
21
KCL: Ic+Ip = Iτ +Iq
�� ��
��
�� ��0
��0
��0
��0
��
��
�� ��
��
�� �� �� ��
�� ��
��
��
��
��
��