Variable capacitance testing method based on CBCM
By using the clock signal and its delay control of the ports of the device to be tested in the CBCM test method, the parasitic capacitance is solved, and the problem of strict signal timing and large number of interfaces is achieved, and a higher accuracy variable capacitance measurement is achieved.
Patent Information
- Application Number
- CN202411098119.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-11-19
AI Technical Summary
The existing CBCM testing methods have strict signal timing, large number of test interfaces, and complex testing conditions, resulting in poor accuracy when measuring variable capacitance.
Use a clock signal and its delay clock cycle signal to control the two ports of the device to be tested. By measuring branches and reference branches, the parasitic capacitance of the device to be tested is blocked, the capacitance of the device to be tested is calculated, the test steps are simplified and the accuracy is improved. .
It significantly improves the accuracy of measuring variable capacitance, simplifies testing steps, reduces testing complexity, and improves the commercial value and application prospects of testing.
Smart Images

Figure CN118980864A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of semiconductor device testing, in particular to a variable capacitance testing method based on CBCM. Background Art
[0002] Traditional impedance parameter analysis and testing technology cannot accurately test tiny parasitic capacitance. Academics and industry tend to adopt the CBCM test method. The traditional CBCM circuit is composed of two sets of pseudo inverters of the same size and with independent gates. The capacitance is tested by charging and discharging the two sets of inverters at the same time, but it also has obvious shortcomings. First, due to the existence of process fluctuations, it is difficult to accurately realize the pseudo inverter structure that is completely symmetrical on the left and right sides, and an additional de-embedding structure is required, which increases the design difficulty of the test structure and also increases the layout area of the entire test circuit. On this basis, the academic community proposed a type of charge-based capacitance test circuit CIEF-CBCM without charge injection effect. It is composed of only one set of pseudo inverters. Compared with the traditional CBCM circuit, CIEF only needs one set of pseudo inverters, does not require de-embedding structure, eliminates mismatch effect, and two-step test completely avoids the influence of errors caused by charge injection on parasitic capacitance extraction; but it also has obvious shortcomings, that is, it requires 3 synchronous input clock signals at the same time, the signal timing is strict, and the accuracy is poor when measuring variable capacitance.
[0003] The CBCM test method in the prior art has problems such as strict signal timing, large number of test interfaces, complex test conditions, etc., and poor accuracy when measuring variable capacitance. Summary of the invention
[0004] The purpose of the present invention is to provide a variable capacitance test method based on CBCM, which is characterized by using a clock signal and its delayed clock cycle signal to control the voltage of one of the two ports of the device under test and thus control the test circuit of the device charge and discharge state, and by shielding the device under test to eliminate the influence of the circuit parasitic capacitance and then calculate the capacitance of the device under test at the voltage. The method is provided with a measurement branch and a reference branch, the two branches are connected to one of the ports of the device under test through a transmission gate, and the transmission gate is controlled to open and close by a clock signal and its delayed clock signal, thereby controlling the port voltage of the device under test, and the charging current is measured by the measurement branch controlling the port. It solves the problems of strict signal timing, large number of test interfaces, and complex test conditions. The de-embedding method for the parasitic capacitance in the measurement circuit is adopted, which greatly improves the accuracy of the variable capacitance measured by the measurement method, the method is simple, the use effect is good, and it has good application prospects and commercial value.
[0005] The specific technical solution for achieving the purpose of the present invention is:
[0006] A variable capacitance test method based on CBCM, characterized by:
[0007] A test circuit having a clock cycle signal and a delayed clock cycle signal connected to two transmission gates to control a port of the device under test connected to a measurement branch and a reference branch is set; the capacitance of the device under test under the port voltage is calculated by using the test circuit and shielding the device under test to eliminate the influence of the circuit parasitic capacitance, and the test specifically includes the following steps:
[0008] Step 1: First, raise a port of the device under test connected to the transmission gate to a voltage of Va through a measurement branch, and then lower it to a voltage of Vb, charge and discharge the device under test, and measure the current on the measurement branch and the reference branch during the charging process;
[0009] Step 2: Add a waveform at the other end of the device under test that does not overlap with the clock signal and the delayed clock signal, whose high and low levels are respectively equal to the voltage Va and the voltage Vb, and whose high-level pulse width completely covers the low-level pulse width of the clock waveform, to shield the device under test, and measure the charging current;
[0010] Step 3: Make the voltage difference between the two ports 0V and measure the current on the reference branch; change the voltage Va, repeat steps 1 and 2 to obtain the capacitance of the device under test at different voltages Va, and connect them to form a complete variable capacitance CV curve of the device under test.
[0011] Furthermore, the clock signal has a delay of t1 clock, the low voltage of the clock signal is 0V, and the high voltage is V DD , the clock period is T.
[0012] Further, the step 1 specifically includes:
[0013] During charging and discharging, the low level of the clock signal causes the voltage of a port of the device under test connected to the measurement branch to rise to a predetermined voltage value Va, and the low level of the delayed clock signal causes the voltage of the port of the device under test to drop to a predetermined voltage value Vb, thereby obtaining the current I1 of the measurement branch and the current I REF , the charge Q1 of the parasitic capacitance in the device under test and the test circuit is calculated by the following formula (a):
[0014]
[0015] Among them, C DUT , C para are the capacitance of the device under test and the parasitic capacitance in the test circuit, and f is the clock signal frequency.
[0016] Further, the step 2 specifically includes:
[0017] Shielding the device under test means keeping the predetermined voltages Va and Vb in step 1 unchanged, applying a pulse signal to the other end of the device under test that does not overlap with the clock signal and the delayed clock signal and whose high and low levels are equal to Va and Vb respectively, and the high and low level pulse widths of the pulse signal completely cover the low level pulse widths of the clock signal and the delayed clock signal respectively, and the current I2 and the charging charge Q2 of the measurement branch are calculated by the following formula (b):
[0018]
[0019] Further, the step 3 specifically includes:
[0020] The voltage at the port of the device under test that is not connected to the measurement branch and the reference branch affects the parasitic capacitance value in the test circuit. To avoid repeated calculation of some parasitic capacitances, the voltage difference between the two ports of the device under test is set to 0V, and the current on the reference branch is measured and recorded as I REF0 ; Change the voltage Va, the current I measured by the reference branch in step 1 REF Subtract I REF0 And add it with I2 to get the final de-embedding current. The capacitance of the device under test is calculated by the following formula (c):
[0021]
[0022] The capacitance of the device under test calculated at different voltages Va is connected to form a variable capacitance CV curve.
[0023] Compared with the prior art, the present invention has the advantage of high measurement accuracy, and better solves the problems of strict signal timing, large number of test interfaces, and complex test conditions. A de-embedding method for parasitic capacitance in the measurement circuit is adopted to further improve the accuracy of the measurement method for measuring variable capacitance. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 A test circuit diagram of the present invention;
[0025] Figure 2 This is a schematic diagram of a clock signal generation module;
[0026] Figure 3 It is the signal timing diagram;
[0027] Figure 4 It is a comparison curve between the variable capacitance test method based on CBCM and the impedance parameter analysis test;
[0028] Figure 5 This is a fluctuation curve diagram of the measured capacitance value from golden. DETAILED DESCRIPTION
[0029] The present invention is described in detail below in conjunction with the accompanying drawings and embodiments.
[0030] The present invention uses a test circuit having a clock cycle signal and a delayed clock cycle signal respectively connected to two transmission gates to control a port of a device under test connected to a measurement branch and a reference branch. The device under test is shielded to eliminate the influence of circuit parasitic capacitance, and the capacitance of the device under test under port voltage is calculated. The test specifically includes the following steps:
[0031] Step 1: First, raise a port of the device under test connected to the transmission gate to a voltage of Va through a measurement branch, and then lower it to a voltage of Vb, charge and discharge the device under test, and measure the current on the measurement branch and the reference branch during the charging process;
[0032] Step 2: Add a waveform at the other end of the device under test that does not overlap with the clock signal and the delayed clock signal, whose high and low levels are equal to Va and Vb respectively, and whose high-level pulse width completely covers the low-level pulse width of the clock waveform to shield the device under test, and measure the charging current;
[0033] Step 3: Make the voltage difference between the two ports 0V and measure the current on the reference branch. Change the Va voltage, repeat steps 1 and 2 to obtain the capacitance of the device under test at different Va voltages, and connect them to form a complete variable capacitance CV curve of the device under test.
[0034] Example
[0035] See also Figure 1 , the test circuit in this embodiment is composed of 8 transmission gates, 3 inverters and 1 NOR gate, and the device to be tested is in the dotted box. Among them, the output of the NOR gate and the output through the inverter are respectively connected to the gates of the transmission gates NMOS and PMOS that control the transmission of the clock signals CLK_P1, CLK_P2, the VA1 end of the measurement branch and the VA2 end of the reference branch. The clock signal CLK_P1 and the clock signal through the inverter are respectively connected to the gates of the transmission gates NMOS and PMOS that control the transmission voltages of the VA1 end and the VA2 end; the clock signal CLK_P2 and the clock signal through the inverter are respectively connected to the gates of the transmission gates NMOS and PMOS that control the transmission voltages of the VA2 end and the VB end. The voltages of VA1, VA2 and VB are connected to the gate of the device to be tested after passing through the transmission gates; V DUT Connect the source and drain of the device under test, and short-circuit the source and drain.
[0036] During the measurement process, the input voltages of VA1 and VA2 are kept equal, and the voltage of VB is less than the voltages of VA1 and VA2. The low level of the input clock signal at CLK_P1 corresponds to a voltage at one port of the device under test equal to VA1 (VA2); after a certain delay, the same clock signal is input at CLK_P2 to make the voltage at this port of the device under test equal to VB, and the charging voltage range of the device under test is the voltage difference between VA1 and VB. Since the transmission gate controlled by CLK_P1 is turned off when the device under test is discharging, the average current value measured at VA1 will only include the charging current of the device under test. The voltages at VA1 (VA2) and VB remain unchanged. The second measurement only needs to be performed once. A pulse signal with no overlap with the CLK_P1 and CLK_P2 signals and high and low levels equal to VA1 (VA2) and VB is applied to both ends of the device under test. The high-level pulse width of the VDUT waveform completely covers the low-level pulse width of the CLK_P1 waveform to ensure that when the transmission gate controlled by CLK_P1 is turned on, the voltages at both ends of the device under test are completely consistent. At this time, V DUT The terminal signal will completely shield the device under test. The current at the reference branch VA2 terminal under the port voltage of each device under test is measured, and the current obtained at the VA1 terminal is corrected to obtain the capacitance of the device under different port voltages.
[0037] See also Figure 2 From the figure, it can be seen that the input CLKIN directly outputs CLK_P1 after the buffer with enhanced driving capability, while the clock signal CLK_P2 with a certain phase difference only needs to pass CLKIN through an additional series inverter chain (2N+1) for a delay of t1 and then pass through the buffer for output.
[0038] See also Figure 1-2 , the measurement starts, the signals closel and rowel are set to 1, and the transmission gate for transmitting the clock signal and the voltage of one port of the device under test is opened.
[0039] Step 1: Measure the sum of the measurement branch charging currents of the DUT and parasitic capacitance and the current of the reference branch as follows:
[0040] 1) Assume that the input voltages of VA1 and VA2 are Va, and the input voltage of VB is Vb (Va>Vb).
[0041] 2) The CLK_P1 signal is low level first, and the two transmission gates in the measurement branch and the reference branch controlled by it are turned on at the same time. The voltage of one port of the device under test connected to the measurement branch and the reference branch will rise to Va. When the input of the CLK_P2 terminal is low level, the voltage of the port of the device under test will drop to Vb. In one cycle, the current I1 charged by the device under test and the parasitic capacitor at the VA1 terminal and the reference current I at the VA2 terminal areREF , the charge Q1 at the VA1 terminal is calculated by the following formula (a):
[0042]
[0043] Among them, C DUT , C para are the capacitance of the device under test and other parasitic capacitances respectively, and f is the clock signal frequency.
[0044] Step 2: Keep the voltages provided by VA1, VA2 and VB unchanged, add a waveform at the other end of the DUT that does not overlap with the clock signal and the delayed clock signal, with the high and low levels equal to Va and Vb respectively, and the high-level pulse width completely covering the low-level pulse width of the clock waveform to shield the DUT, and measure the charging current to obtain the capacitance of the DUT under the voltage at the VA1 terminal. Follow the steps below:
[0045] 1) Keep Va and Vb unchanged in step 1, apply a pulse signal to the device under test that does not overlap with the clock signal and the delayed clock signal and whose high and low levels are equal to Va and Vb respectively, and the high and low level pulse widths of the pulse signal completely cover the low level pulse widths of the clock signal and the delayed clock signal respectively. Step 2 only needs to be performed once, and the current I2 and the charging charge Q2 of the measurement branch are calculated by the following formula (b):
[0046]
[0047] Step 3: Make the voltage difference between the two ports 0V and measure the current on the reference branch. Change the Va voltage, repeat steps 1 and 2 to obtain the capacitance of the device under test at different Va voltages, and connect them to form a complete variable capacitance CV curve of the device under test. Follow the steps below:
[0048] 1) The voltage at the port of the device under test that is not connected to the measurement branch and the reference branch will affect the parasitic capacitance value in the test circuit. To avoid repeated calculation of some parasitic capacitances, the voltage difference between the two ports of the device under test is set to 0V, and the current on the reference branch is measured and recorded as I REF0 ; Change the voltage Va, the current I measured by the reference branch in step 1 REF Subtract I REF0 And add it to I2 to get the final de-embedding current. The capacitance of the device under test is calculated by the following formula (c):
[0049]
[0050] The capacitance of the device under test calculated at different Va voltages is connected to form a variable capacitance curve.
[0051] See also Figure 3From the figure, we can see the timing diagram of CLK_P1, CLK_P2, the input signal waveform of the device under test, and the current I_VA1 and I_DUT of VA1 and the device under test during the measurement process.
[0052] See also Figure 4 The figure shows the comparison of the test results of the same device under test under the variable capacitance test method based on CBCM and the impedance parameter analysis test. The root mean square error RMS of the variable capacitance test method based on CBCM and the impedance parameter analysis test is less than 1%, which fully proves the effectiveness and accuracy of the measurement method.
[0053] See also Figure 5 The figure shows the fluctuation range of gate capacitance values from the golden value when the gate port voltage is 0V for PMOS devices with W / L of 0.25 / 0.04, 0.5 / 0.04 and 1.0 / 0.04 in 9 dies. The fluctuation range of these three sets of data is within ±1.2%. It can be seen that the variable capacitance test method based on CBCM of the present invention, under certain process manufacturing deviations, the capacitance values measured are all within the acceptable fluctuation range.
[0054] The above specific implementations are only for further explanation of the present invention, and are not intended to limit the present invention. Any equivalent implementations of the present invention should be included in the scope of the claims of the present invention.
Claims
1. A variable capacitance testing method based on CBCM, characterized in that: A test circuit having a clock cycle signal and a delayed clock cycle signal connected to two transmission gates to control a port of the device under test connected to the measurement branch and the reference branch is provided; The test circuit is adopted and the influence of the circuit parasitic capacitance is eliminated by shielding the device under test, and the capacitance of the device under test under the port voltage is calculated. The test specifically includes the following steps: Step 1: First, raise a port of the device under test connected to the transmission gate to a voltage of Va through a measurement branch, and then lower it to a voltage of Vb, charge and discharge the device under test, and measure the current on the measurement branch and the reference branch during the charging process; Step 2: Add a waveform at the other end of the device under test that does not overlap with the clock signal and the delayed clock signal, whose high and low levels are respectively equal to the voltage Va and the voltage Vb, and whose high-level pulse width completely covers the low-level pulse width of the clock waveform, to shield the device under test, and measure the charging current; Step 3: Make the voltage difference between the two ports 0V and measure the current on the reference branch; change the voltage Va, repeat steps 1 and 2 to obtain the capacitance of the device under test at different voltages Va, and connect them to form a complete variable capacitance CV curve of the device under test.
2. The variable capacitance testing method based on CBCM according to claim 1, characterized in that: The clock signal has a delay of t1 clock, the low voltage of the clock signal is 0V, and the high voltage is V DD , the clock period is T.
3. The variable capacitance testing method based on CBCM according to claim 1, characterized in that: The step 1 specifically includes: During charging and discharging, the low level of the clock signal causes the voltage of a port of the device under test connected to the measurement branch to rise to a predetermined voltage value Va, and the low level of the delayed clock signal causes the voltage of the port of the device under test to drop to a predetermined voltage value Vb, thereby obtaining the current I1 of the measurement branch and the current I REF , the charge Q1 of the parasitic capacitance in the device under test and the test circuit is calculated by the following formula (a): Among them, C DUT , C para are the capacitance of the device under test and the parasitic capacitance in the test circuit, and f is the clock signal frequency.
4. The variable capacitance testing method based on CBCM according to claim 1, characterized in that: The step 2 specifically includes: Shielding the device under test means keeping the predetermined voltages Va and Vb in step 1 unchanged, applying a pulse signal to the other end of the device under test that does not overlap with the clock signal and the delayed clock signal and whose high and low levels are equal to Va and Vb respectively, and the high and low level pulse widths of the pulse signal completely cover the low level pulse widths of the clock signal and the delayed clock signal respectively, and the current I2 and the charging charge Q2 of the measurement branch are calculated by the following formula (b):
5. The variable capacitance testing method based on CBCM according to claim 1, characterized in that: The step 3 specifically includes: The voltage at the port of the device under test that is not connected to the measurement branch and the reference branch affects the parasitic capacitance value in the test circuit. To avoid repeated calculation of some parasitic capacitances, the voltage difference between the two ports of the device under test is set to 0V, and the current on the reference branch is measured and recorded as I REF0 ; Change the voltage Va, the current I measured by the reference branch in step 1 REF Subtract I REF0 And add it to I2 to get the final de-embedding current. The capacitance of the device under test is calculated by the following formula (c): The capacitance of the device under test calculated at different voltages Va is connected to form a variable capacitance CV curve.
Citation Information
Cited By
Semiconductor dynamic capacitance test system
CN119395392A
