CN107451330A - A kind of acceleration Transient method for being used for Random telegraph noise in circuit - Google Patents
A kind of acceleration Transient method for being used for Random telegraph noise in circuit Download PDFInfo
- Publication number
- CN107451330A CN107451330A CN201710505382.7A CN201710505382A CN107451330A CN 107451330 A CN107451330 A CN 107451330A CN 201710505382 A CN201710505382 A CN 201710505382A CN 107451330 A CN107451330 A CN 107451330A
- Authority
- CN
- China
- Prior art keywords
- rtn
- circuit
- class
- time
- amplitude
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/10—Noise analysis or noise optimisation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Tests Of Electronic Circuits (AREA)
- Design And Manufacture Of Integrated Circuits (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses the acceleration Transient method of RTN in circuit a kind of, belong to microelectronic component and circuit reliability field.This method can solve the limitation sex chromosome mosaicism that traditional RTN emulation modes can not realize that RTN is emulated in the range of full-time, by classifying to RTN, Monte Carlo calculations are carried out to the RTN of respective classes in different time ranges, while the synergistic effect that RTN influences on circuit in the range of different time is considered by introducing amplitude card.It is of the invention that RTN is not computed repeatedly compared with the calculating process of traditional RTN emulation modes, do not increase the Monte Carlo calculations number to RTN, realize full time domain scale RTN emulation.
Description
Technical field
The invention belongs to microelectronic component and circuit reliability field, is related to Random telegraph noise in circuit and accelerates transient state
Emulation mode.
Background technology
In MOS device, Random telegraph noise (random telegraph noise, i.e. RTN) refers in gate oxide
The random capture of trap or release channel carrier cause the drain terminal electric current of device with the average period of a determination in high state
Fluctuated between low state.With being gradually reduced for MOS device yardstick, trap number is fewer and fewer in gate dielectric layer, and
Amplitude (the Δ V of Random telegraph noise caused by single trapTHOr Δ Id/Id) increasing, it is small that these make it that RTN becomes
Obvious dynamic fluctuation source in size MOS device.RTN random jump row is likely to result in several forthright failures of circuit, therefore,
, it is necessary to which RTN influence is meticulously considered in the Reliablility simulation of circuit is assessed.
Description to RTN characteristics mainly has three aspects：One is RTN amplitudes (Δ VTHOr Δ Id/Id), two other
It is launch time constant (τ respectivelye) and capture time constant (τc).Wherein amplitude and time constant all have obvious gate voltage
(VG) dependence.For different RTN, the amplitude between them has differences with time constant.Root is it is reported that specific
Under voltage, RTN amplitudes are obeyed normal state (Lognormal) under exponential distribution or logarithm and are distributed；Capture or launch time constant
Obey uniformly (Loguniform) distribution under logarithm., it is necessary to consider RTN width in the RTN simulation processes for carrying out actual circuit
The statistics sexual behaviour of degree and time constant.
What Fig. 1 was represented is the Monte-carlo Simulation Method of traditional single RTN random jumps behavior, specific mainly to pass through
Equally distributed random number is with occupying probability (Δ t/ τ between 0 to 1c) or release probability (Δ t/ τc) comparison come to RTN's
State is judged.According to current method, for very big (such as the capture time constant, τ of time constantc) RTN, once jump
The simulation of change behavior needs to carry out general τcThe Monte Carlo calculations that/Δ is t times, this is very time-consuming.Therefore, traditional
In RTN simulation processes, simulation time is limited to, is typically only capable to carry out the emulation of transient state for the RTN in particular time range；Separately
On the one hand, wider array of time range is considered when primary condition is set, but is constrained to the Transient time, the time is normal
The bigger RTN of number is that any saltus step behavior does not occur, and is also equivalent to not account for.Above to RTN Simulated hand
Section can be used for studying influences of the RTN to circuit, but if necessary to assess RTN exactly to circuit closer to actual conditions
Influence, it is necessary to consider influences of the RTN to circuit of full time domain scale.Therefore, RTN acceleration Transient means be there is an urgent need to
's.
The content of the invention
It is an object of the invention to propose the acceleration Transient method of RTN in circuit a kind of, full time domain scale is realized
RTN emulation.
Technical scheme is as shown in figure 3, specifically include：
First, input RTN statistics parameters, including in device RTN number distribution, and RTN amplitude and time
The distribution of constant, each RTN have corresponding characteristic parameter；
Second step, classification n=1,2,3 ..., m-1, m are carried out to RTN according to RTN capture times constant, wherein, it is sorted
For journey according to uniformly being carried out under logarithm, n values are smaller to represent that time constant is smaller；
3rd step, tran emulation is carried out to the RTN of different classifications respectively, from low to high, the RTN of each classification is emulated classification
Time span is consistent, and Monte Carlo calculations are carried out to the RTN of respective classes in different time ranges.Emulating the n-th class
During RTN, initial time point is t0(t0Size it is consistent with the n-th class RTN time constant minimum border), when different
Punctum will be synchronously plus the degradation values of the corresponding moment points of amplitude card (n-1).It should be noted that imitated in every subtransient
During true, the probability that occupies of first moment point should be calculated with release probability according to the time actually undergone, i.e. Pc
=t0/τc, Pe=t0/τe.Wherein, in addition to considering the n-th class RTN, moment point is emulated at first to be needed to consider 1~n-1
Never the RTN of release behavior never occurs after the RTN of behavior is occupied in class RTN and is occupied.Wink is completed in the n-th class RTN
, it is necessary to be updated to RTN amplitude card, as RTN amplitude card after the calculating of state time span
(n)。
4th step, if n<M, continue lower a kind of RTN calculating according to three step process, if n=m, stop
Calculate.
Finally, according to result above, Δ t to 10 can rapidly be realizedmInfluences of the RTN to circuit in time range is entered
Row emulation.Due to Δ t to 10mCovering RTN time constant scope can be realized completely, it can be said that above method can be real
The now RTN of full time range emulation.
It is similar with above-mentioned DC situations flow for AC situations.Difference is：
1) in the first step, RTN be time constant and amplitude VGDependency need is accounted for, it is necessary to add
VGRelated amplitude and the model of time constant.
2) in second step, when classification to RTN in the case of AC, with high level (such as VDD) capture time
Classified on the basis of constant, sorting technique is similar with DC.
3) during the 3rd step carries out Transient to RTN, first of Transient is being carried out to every a kind of RTN
When moment point calculating occupies probability with release probability, calculation is as follows：Pc=DF*t0/τc_H, Pe=(1-DF) * t0/τe_L.Its
In, DF is the dutycycle of AC signals, transient waveform can be calculated by emulator, τc_HIt is the capture time of high level
Constant, τe_LIt is low level launch time constant.
Method provided by the invention, which can solve traditional RTN emulation modes, can not realize RTN emulation in the range of full-time
Sex chromosome mosaicism is limited, this method carries out covering spy in different time ranges by classifying to RTN to the RTN of respective classes
Carlow calculates, while considers the superposition effect that RTN influences on circuit in the range of different time by introducing amplitude card
Should；And this method is not computed repeatedly to RTN, not increased compared with the calculating process of traditional RTN emulation modes
Add the Monte Carlo calculations number to RTN.Therefore, the invention provides one to carry out acceleration wink to full time range RTN
The method of state emulation, the assessment to influences of the RTN to circuit have great importance.
Brief description of the drawings
Fig. 1 is the traditional analog flow of RTN random behaviors；
Fig. 2 is the restricted schematic diagram of traditional simulation method of RTN in circuit；
Fig. 3 is full time range RTN emulation mode flow charts of the invention.
Embodiment
By way of example and in conjunction with the accompanying drawings, the implementation method of the present invention is described in detail：
The RTN emulation modes of full time range are following in circuit (it is described in detail by taking DC situations as an example, wherein, for
Emulation per a kind of RTN, the present invention need to consider such before RTN caused by each device performance degenerate case, be this
RTN amplitude card (n) are introduced, represent that device performance degeneration synthesis result is in transient state caused by the 1 to the n-th class RTN
Changed with time in simulation process)：
1) the statistical distribution setting to RTN parameters all in all devices is realized first.The process can pass through RTN's
(log-uniform is distributed, and wherein distributed area is [10 for amplitude distribution (exponential distribution) and time constant distribution-a,10a]) feelings
Condition is operated by generating corresponding random number.
2) RTN time constant is classified.Wherein, assorting process is carried out according to the magnitude of capture time constant, such as
Fruit magnitude step-length is b, then can divide 2a/b classes RTN.
3) state to RTN and RTN amplitude card are initialized.Wherein, initial all RTN states can
To be arranged to 0.In initial RTN amplitude card, the degradation values of device are 0 corresponding to all moment points.
4) tran emulation is carried out to the RTN of the n-th class.In order to consider the synergistic effect between RTN, carrying out the n-th class RTN's
, it is necessary to produce an amplitude card (amount of degradation of each device caused by record RTN in real time) during Transient.Imitative
, it is necessary to by the amplitude card obtained in the (n-1)th class RTN simulation processes superpositions up when true n-th class RTN.Entering
First moment point of the row per a kind of RTN emulation, occupy probability and discharge being calculated as follows for probability：Pc=t0/τc,Pe
=t0/τe, wherein, t0Size it is consistent with respective class RTN time constant minimum border.When first moment point calculates, need
Consider the never RTN of generation release behavior after the RTN of behavior is never occupied in 1~n-1 classes RTN and is occupied.
5) above-mentioned Transient process is judged.If n=2a/b, then emulation terminates, if n<2a/b, that
Continue the simulation flow of the 3rd step of repetition.
6) result that full time range RTN influences on circuit is obtained.In summary simulation result, full time range distribution
The result of influences of the RTN to circuit can obtain.
Embodiment described above is not intended to limit the present invention, any those skilled in the art, is not departing from this hair
In bright spirit and scope, various change and retouching can be done, therefore protection scope of the present invention regards right institute circle
It is fixed.
Claims (8)
1. a kind of acceleration Transient method for being used for Random telegraph noise in circuit, step include：
1) input RTN statistics parameters, including in device RTN number distribution, and RTN amplitude and time constant divide
Cloth, each RTN have corresponding characteristic parameter；
2) classification n=1,2,3 ..., m-1, m are carried out to RTN according to RTN capture times constant, wherein, assorting process is according to right
Several lower uniformly progress, n values are smaller to represent that time constant is smaller；
3) carry out tran emulation to the RTN of different classifications respectively, classification from low to high, the RTN simulation time length of each classification
Unanimously, Monte Carlo calculations are carried out to the RTN of respective classes in different time ranges, in the process for the RTN for emulating the n-th class
In, will be synchronously plus the degradation values of the corresponding moment points of amplitude card (n-1) in point at different moments；
If 4) n<M, continue lower a kind of RTN calculating according to three step process, if n=m, stop calculating, realize
The emulation of influences of the RTN to circuit.
2. it is used for the acceleration Transient method of Random telegraph noise in circuit described in claim 1, it is characterised in that step 3)
During the RTN of the n-th class is emulated, initial time point is t0, t0Size and the n-th class RTN time constant minimum border one
Cause.
3. it is used for the acceleration Transient method of Random telegraph noise in circuit described in claim 1, it is characterised in that step 3)
During each Transient, first emulation moment point occupy probability with discharge probability should according to actually undergo when
Between calculated, wherein, in addition to considering the n-th class RTN, moment point is emulated at first to be needed to consider 1~n-1 classes RTN
In never occupy the RTN of behavior and the RTN of release behavior never occur after occupying.
4. it is used for the acceleration Transient method of Random telegraph noise in circuit described in claim 1, it is characterised in that step 3)
After the calculating that the n-th class RTN completes transient time length, RTN amplitude card are updated, as RTN
amplitude card(n)。
5. a kind of acceleration Transient method for being used for Random telegraph noise in circuit, step include：
1) RTN statistics parameters, including the distribution of RTN number, the distribution of RTN amplitude and time constant, V in device are inputtedG
Related amplitude and the model of time constant, each RTN have corresponding characteristic parameter；
2) classification n=1,2,3 ..., m-1, m are carried out to RTN on the basis of the capture time constant of high level, wherein, it is sorted
For journey according to uniformly being carried out under logarithm, n values are smaller to represent that time constant is smaller；
3) carry out Meng Takaluo calculating to the RTN of respective classes in different time ranges, classification from low to high, each classification
RTN Transients calculate time span it is consistent, and produce an amplitude card, when emulating the n-th class RTN, together
Step adds the degradation values that moment point is corresponded in amplitude card (n-1)；
If 4) n<M, continue lower a kind of RTN calculating according to three step process, if n=m, stop calculating, realize
The emulation of influences of the RTN to circuit.
6. it is used for the acceleration Transient method of Random telegraph noise in circuit described in claim 5, it is characterised in that step 3)
During the RTN of the n-th class is emulated, initial time point is t0, t0Size and the n-th class RTN time constant minimum border one
Cause.
7. it is used for the acceleration Transient method of Random telegraph noise in circuit described in claim 5, it is characterised in that step 3)
In when first moment point that Transient is carried out to every a kind of RTN calculates and occupies probability with release probability, calculation is such as
Under：Pc=DF*t0/τc_H, Pe=(1-DF) * t0/τe_L, wherein, DF is the dutycycle of AC signals, by emulator to transient wave
Shape is calculated, and initial time point is t0, τc_HIt is the capture time constant of high level, τe_LIt is that low level launch time is normal
Number.
8. it is used for the acceleration Transient method of Random telegraph noise in circuit described in claim 5, it is characterised in that step 3)
After the calculating that the n-th class RTN completes transient time length, RTN amplitude card are updated, as RTN
amplitude card(n)。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710505382.7A CN107451330B (en) | 2017-06-28 | 2017-06-28 | Accelerated transient simulation method for random telegraph noise in circuit |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710505382.7A CN107451330B (en) | 2017-06-28 | 2017-06-28 | Accelerated transient simulation method for random telegraph noise in circuit |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107451330A true CN107451330A (en) | 2017-12-08 |
| CN107451330B CN107451330B (en) | 2020-07-28 |
Family
ID=60487202
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710505382.7A Active CN107451330B (en) | 2017-06-28 | 2017-06-28 | Accelerated transient simulation method for random telegraph noise in circuit |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107451330B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6122704A (en) * | 1989-05-15 | 2000-09-19 | Dallas Semiconductor Corp. | Integrated circuit for identifying an item via a serial port |
| CN1804847A (en) * | 2004-11-08 | 2006-07-19 | 国际商业机器公司 | Method and system for converting globally clock-gated circuits to locally clock-gated circuits |
| US20110112784A1 (en) * | 2009-11-09 | 2011-05-12 | Nikitin Alexei V | Method and apparatus for adaptive real-time signal conditioning and analysis |
| CN102157556A (en) * | 2011-01-27 | 2011-08-17 | 北京大学 | Oxidizing-dephlegmation-based silicon-based wrap gate transistor with buried-channel structure and preparation method thereof |
| US20120065920A1 (en) * | 2010-09-10 | 2012-03-15 | Renesas Electronics Corporation | Evaluation method, evaluation apparatus, and simulation method of semiconductor device |
-
2017
- 2017-06-28 CN CN201710505382.7A patent/CN107451330B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6122704A (en) * | 1989-05-15 | 2000-09-19 | Dallas Semiconductor Corp. | Integrated circuit for identifying an item via a serial port |
| CN1804847A (en) * | 2004-11-08 | 2006-07-19 | 国际商业机器公司 | Method and system for converting globally clock-gated circuits to locally clock-gated circuits |
| US20110112784A1 (en) * | 2009-11-09 | 2011-05-12 | Nikitin Alexei V | Method and apparatus for adaptive real-time signal conditioning and analysis |
| US20120065920A1 (en) * | 2010-09-10 | 2012-03-15 | Renesas Electronics Corporation | Evaluation method, evaluation apparatus, and simulation method of semiconductor device |
| CN102157556A (en) * | 2011-01-27 | 2011-08-17 | 北京大学 | Oxidizing-dephlegmation-based silicon-based wrap gate transistor with buried-channel structure and preparation method thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107451330B (en) | 2020-07-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Toledano-Luque et al. | Response of a single trap to AC negative bias temperature stress | |
| Brandwajn et al. | Severity indices for contingency screening in dynamic security assessment | |
| CN106971066A (en) | Method based on Neural Network Models To Prediction geomagnetic storm | |
| Lakshmi et al. | Optimal placement of distributed generation using firefly algorithm | |
| US7865795B2 (en) | Methods and apparatuses for generating a random sequence of commands for a semiconductor device | |
| CN102799112B (en) | Method for determining aircraft lightning stroke attachment point position | |
| Cheni et al. | Reducing latency in a converted spiking video segmentation network | |
| Trondoli et al. | Configurable stochastic model for high impedance faults simulations in electrical distribution systems | |
| Wadı | Five different distributions and metaheuristics to model wind speed distribution | |
| CN107451330A (en) | A kind of acceleration Transient method for being used for Random telegraph noise in circuit | |
| Samusenko et al. | Computer simulation of ionic wind in the point-torus electrodes system | |
| EP1382975A1 (en) | Method of generating a test pattern for the simulation and/or test of the layout of an integrated circuit | |
| Fang et al. | Understanding the impact of transistor-level BTI variability | |
| Gómez-Castañeda et al. | Photovoltaic panel emulator in FPGA technology using ANFIS approach | |
| Zhang | Artificial neural network model-based design and fixed-point FPGA implementation of hénon map chaotic system for brain research | |
| CN107862379A (en) | Neutral net FPGA | |
| Cai et al. | An algorithm Q-PSO for community detection in complex networks | |
| Martínez et al. | LFSR generation for high test coverage and low hardware overhead | |
| EP1382976A1 (en) | Method of processing test patterns for an integrated circuit | |
| CN116056124A (en) | Wireless network simulator based on deep reinforcement learning | |
| Isa et al. | Development of artificial neural network and adaptive neuro-fuzzy inference system based techniques and algorithms for protection of transmission line | |
| CN108629093A (en) | A kind of dynamic voltage frequency adjusting transitional sequence design method of noise perception | |
| CN102298659A (en) | Method for simulating impulse current and device | |
| Abba et al. | An “artificial retina” processor for track reconstruction at the full LHC crossing rate | |
| Yuan-Liang et al. | Study on automatic test generation of digital circuits using particle swarm optimization |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |