Background technology
Along with improving constantly of electronic equipment travelling speed, the operating voltage of its inside chip reduces gradually, and working current raises gradually, and the change of voltage and current can cause the transient voltage fluctuation, thereby may cause chip operation to be made mistakes.In order to address this problem, modal method is to use electric capacity.Correct use electric capacity must have sufficient understanding to the characteristic of electric capacity.
Electric capacity characteristic in actual applications is: presents capacitive at low frequency, presents perception at high frequency, and at capacitive and perceptual regional intersection generation resonance.Because this characteristic of electric capacity, electric capacity company can provide the equivalent RLC series connection model of electric capacity according to test data usually, promptly carries out the capacitor model test modeling.The equivalent RLC series connection model of electric capacity as shown in Figure 1.
Among Fig. 1, the equivalent RLC series connection model of electric capacity is composed in series by equivalence appearance value C, equivalent series resistance (ESR) and the equivalent series inductance (ESL) of electric capacity.The impedance curve of electric capacity as shown in Figure 2.
At present, the modeling method of capacitor equivalent model is: at first, get a bit in low frequency capacitive zone, and calculate the equivalence appearance value C of electric capacity according to formula Z=1/jwC, wherein, Z is impedance, j is the prefix of imaginary number, and w is an angular frequency, and w equals frequency and multiply by 2 and multiply by circular constant again.Then,, calculate the ESR of electric capacity,, calculate ESL according to formula f=1/ (2 * PI * sqrt (C * ESL)) at the capacitor resonance point according to formula Z=ESR at the capacitor resonance point; Wherein, PI is that circular constant, L are the equivalent series inductance of electric capacity, and C is the equivalence appearance value of electric capacity, and f is a frequency, and sqrt (*) is a square root.At last, the equivalent RLC series connection model of setting up electric capacity according to equivalence appearance value C, equivalent series resistance ESR and the equivalent series inductance ESL of the above-mentioned electric capacity that obtains.
For ceramic condenser, because its physical arrangement, the equivalent series resistance R of ceramic condenser increases with the rising of frequency, and equivalent series inductance L reduces with the rising of frequency.And equivalent series resistance and equivalent series inductance are constant in the above-mentioned modeling method, do not change with frequency change, so in high-frequency range, above-mentioned modeling method can't correctly reflect the impedance operator of ceramic condenser.
For tantalum electric capacity, because its physical arrangement, the impedance curve of tantalum electric capacity in low-frequency range can simply use an electric capacity to represent.And in low-frequency range, be approximately an ideal capacitance value in the above-mentioned modeling method, so in low-frequency range, above-mentioned modeling method can not correctly reflect the impedance variation characteristic of tantalum electric capacity.
The physical arrangement of ceramic condenser as shown in Figure 3.Among Fig. 3, the both positive and negative polarity plane of ceramic condenser is interlaced.At the physical arrangement of ceramic condenser, the capacitor equivalent RLC model of a kind of " transmission line " type has been proposed, as shown in Figure 4 at present.
For obtaining each equivalent parameters R, L and the C among Fig. 4, at first, need obtain the impedance curve of ceramic condenser by test, then, obtain each equivalent parameters according to the method for arithmetic according to the impedance curve of ceramic condenser.
Above-mentioned modeling method can not be used this modeler model only at ceramic condenser for other electric capacity different with the ceramic condenser physical arrangement, makes the scope of application of this modeling method very limited.In above-mentioned modeling method, each equivalent parameters is promptly to obtain by analytic method by simple arithmetic, the accuracy of equivalent parameters is affected, thereby make the impedance curve of modeler model and the practical impedance curve of ceramic condenser still have certain difference, can not well coincide with the practical impedance curve of ceramic condenser, make the ceramic condenser equivalent model can not reflect the impedance operator of ceramic condenser accurately, thereby make the accuracy of ceramic condenser equivalent model have shortcoming, finally cause simulation analysis result's low precision.
Summary of the invention
The objective of the invention is to, a kind of modeling method and device of capacitor equivalent model are provided, determine the equivalent parameters of capacitor equivalent model by choosing non-analytic method that equivalent parameters calculates the impedance of capacitor equivalent model and predetermined condition judging process, the impedance curve of capacitor equivalent module can be coincide with the practical impedance curve high precision of electric capacity, to realize improving the capacitor equivalent model accuracy, improve the simulation analysis purpose of precision as a result.
For achieving the above object, the invention provides a kind of model building device of capacitor equivalent model, comprising:
Memory module: the equivalent parameters span of memory capacitance equivalent model;
Equivalent parameters is chosen module: be connected with described memory module, and choose equivalent parameters to be determined from the equivalent parameters span of memory module storage, and export the described equivalent parameters of choosing to be determined;
Impedance acquisition module: choose module with described equivalent parameters and link to each other, receive the described equivalent parameters of choosing to be determined, and calculate the impedance of capacitor equivalent model, and export the impedance of described calculating according to this equivalent parameters to be determined;
Equivalent parameters determination module: link to each other with described impedance acquisition module, receive the impedance of described calculating, when the error of the condensance of the impedance of determining described reception and storage satisfies predetermined condition, with the equivalent parameters to be determined of the impedance correspondence of this calculating equivalent parameters as described capacitor equivalent model, and output.
The present invention also provides a kind of modeling method of capacitor equivalent model, comprises step:
A, equivalent parameters are chosen module and choose equivalent parameters to be determined from the equivalent parameters spans of the capacitor equivalent model of memory module storage, and export the described equivalent parameters of choosing to be determined;
B, impedance acquisition module receive the described equivalent parameters of choosing to be determined, and calculate the impedance of capacitor equivalent model according to the described equivalent parameters of choosing to be determined, and export the impedance of described calculating;
C, equivalent parameters determination module receive the impedance of described calculating, and when the error of the condensance of the impedance of determining described calculating and storage satisfies predetermined condition, with the equivalent parameters to be determined of the impedance correspondence of described calculating equivalent parameters as described capacitor equivalent model, and output.
The method of choosing equivalent parameters to be determined among the described step a comprises: the equivalent parameters that picked at random is to be determined.
The method of choosing equivalent parameters to be determined among the described step a specifically comprises: according to Monte Carlo random algorithm picked at random equivalent parameters to be determined from the equivalent parameters span of capacitor equivalent model.
The method that the error of determining the condensance of the impedance of described calculating and storage among the described step c satisfies predetermined condition specifically comprises: the error between the condensance of c1, the impedance of determining described calculating and storage is less than the predetermined convergence value.
The method of described step c also comprises:
When c2, the equivalent parameters determination module error between the condensance of the corresponding impedance of calculating of N group equivalent parameters of determining picked at random and storage all is not less than the predetermined convergence value, according to one group of least error correspondence equivalent parameters adjustment equivalent parameters span to be determined.
The method of determining error among described step c or c1 or the c2 specifically comprises:
Determine the impedance of described calculating and the error between the condensance according to least square method.
The condensance of the storage among described step c or c1 or the c2 obtains by the network analyzer test.
Described capacitor equivalent model is: ceramic condenser equivalent model or tantalum capacitor equivalent model.
Described capacitor equivalent model is: during the tantalum capacitor equivalent model, the equivalent model of described tantalum electric capacity comprises: a plurality of series resonant circuits by the equivalence appearance value and the equivalent series inductance of electric capacity are formed are connected in parallel by equivalent series resistance between series resonant circuit.
The equivalent parameters of described capacitor equivalent model comprises: the equivalence appearance value of electric capacity, equivalent series resistance and equivalent series inductance;
When described capacitor equivalent model was the ceramic condenser equivalent model, the equivalent parameters of described capacitor equivalent model comprised: the equivalent inductance of the pad via hole of the equivalent self-inductance of the equivalence appearance value of electric capacity, the equivalent series resistance of electric capacity, electric capacity, the equivalence of electric capacity welding inductance and electric capacity;
When described capacitor equivalent model was the tantalum capacitor equivalent model, the equivalent parameters of described capacitor equivalent model comprised: the equivalent appearance value of each of electric capacity, each equivalent series resistance and each equivalent series inductance.
Description by technique scheme as can be known, the present invention chooses equivalent parameters by employing and calculates the impedance of capacitor equivalent model and the non-analytic method of predetermined condition judging process, determine the equivalent parameters of capacitor equivalent model, avoided adopting simple arithmetic method and the ceramic condenser equivalent model that causes can not accurately reflect the phenomenon of the impedance operator of actual capacitance at the test impedance curve, make the accuracy height of equivalent parameters of the present invention, thereby the capacitor equivalent model that the present invention is set up can accurately reflect the impedance operator of actual capacitance at different frequency range; By convergency value being set, adjusting the equivalent parameters span, the impedance of the capacitor equivalent model that calculates and the error and the convergency value of actual capacitance impedance curve are compared, make capacitor equivalent model of the present invention can satisfy various simulation analysis results' accuracy requirement; The present invention does not limit the concrete form of capacitor equivalent model, as being ceramic condenser equivalent model, tantalum capacitor equivalent model etc., makes the scope of application of modeling method of the present invention and device extensive; The present invention can adopt Monte Carlo random algorithm etc., and the present invention does not limit the concrete form of random algorithm; Thereby realized raising capacitor equivalent model accuracy by technical scheme provided by the invention, improved the simulation analysis purpose of precision as a result.
Embodiment
The modeling method of capacitor equivalent model of the present invention and the core of device all are: at first, equivalent parameters is chosen module and choose equivalent parameters from the equivalent parameters span of the capacitor equivalent model of memory module storage, and export the equivalent parameters that this is chosen, then, the impedance acquisition module receives this equivalent parameters of choosing, and the impedance of calculating capacitor equivalent model according to this equivalent parameters of choosing, and export the impedance of this calculating, the equivalent parameters determination module receives the impedance of calculating, and when the error of impedance of determining to calculate and condensance satisfies predetermined condition, with the equivalent parameters of the impedance correspondence calculated equivalent parameters as capacitor equivalent model, and output.
Based on core concept of the present invention technical scheme provided by the invention is further described below.
In this embodiment, the model building device of capacitor equivalent model provided by the invention as shown in Figure 7.
Among Fig. 7, the model building device of capacitor equivalent model comprises: memory module, equivalent parameters are chosen module, impedance acquisition module and equivalent parameters determination module.
Memory module is mainly used in the equivalent parameters span of memory capacitance equivalent model, and the stored parameters span can be made amendment by the equivalent parameters determination module in the memory module.
Equivalent parameters is chosen module and is connected with memory module, be mainly used in picked at random equivalent parameters from the equivalent parameters span of memory module storage, and with the equivalent parameters output of choosing, the equivalent parameters of output can transfer to the impedance acquisition module, and the method for picked at random can adopt Monte Carlo random algorithm etc.Description among specific as follows embodiment.
The impedance acquisition module is chosen module with equivalent parameters and is connected, be mainly used in and receive the impedance that equivalent parameters is chosen module output, and the impedance of calculating capacitor equivalent model according to the equivalent parameters of its reception, and with the impedance output that calculates, the impedance of output can transfer to the equivalent parameters determination module.
The equivalent parameters determination module is connected with the impedance acquisition module, be mainly used in when the error of the impedance of determining the transmission of impedance acquisition module and condensance satisfies predetermined condition, with the equivalent parameters of this impedance correspondence equivalent parameters as capacitor equivalent model, and output.Predetermined condition can for the error of the impedance of calculating and condensance less than the predetermined convergence value etc., when the equivalent parameters determination module all is not less than the predetermined convergence value in the error of the computing impedance of many groups parameter value correspondence of determining picked at random and condensance, can adjust stored parameters span in the memory module according to one group of parameter value of error minimum, the method of adjusting the parameter value scope can the employing scope be dwindled ten times etc., the description among specific as follows embodiment.
Below in conjunction with accompanying drawing method and apparatus of the present invention is described in detail.
The present invention at first need be provided with the span of capacitor equivalent model equivalent parameters in memory module, capacitor equivalent model among the present invention can for as shown in Figure 1 general capacitor equivalent model, can be for the ceramic condenser equivalent model, also can be the tantalum capacitor equivalent model, the present invention does not limit the concrete manifestation form of electric capacity, be the present invention applicable to various capacitor equivalent models, capacitor equivalent model can be provided with fully according to actual needs.
For the capacitor equivalent model shown in the accompanying drawing 1, the span of equivalent parameters comprises: the span of the span of C, the span of ESR and ESL; For the ceramic condenser equivalent model shown in the accompanying drawing 4, the span of equivalent parameters comprises: the span of the span of C, the span of R, L_h (the equivalent self-inductance of electric capacity), the span of L_bottom (the equivalence welding inductance of electric capacity) and the span of L_padvia (equivalent inductance of the pad via hole of electric capacity); For the tantalum capacitor equivalent model, the span of equivalent parameters is relevant with the quantity of C, L, R, and tantalum capacitor equivalent model provided by the invention as shown in Figure 5.
Tantalum capacitor equivalent model of the present invention comprises: a plurality of series resonant circuits that a plurality of equivalent appearance values and equivalent series inductance are formed, be connected in parallel by equivalent series resistance between series resonant circuit, equivalent model as the tantalum electric capacity among Fig. 5 comprises: 5 equivalent appearance value C1 to C5,5 equivalent series inductance L1 to L5 and 5 equivalent series resistance R1 to R5, each equivalent appearance value can be identical, also can be different, each equivalent series inductance can be identical, also can be different, each equivalent series resistance can be identical, also can be different; Among Fig. 5, C1, L1 form series resonant circuit, and then connect with R1, and note is Circuit1; C2 and L2 form series resonant circuit, and be in parallel with Circuit1 again, connects with R2 then, and note is Circuit2; C3 and L3 form series resonant circuit, and be in parallel with Circuit2 again, connects with R3 then, and note is Circuit3; C4 and L4 form series resonant circuit, and be in parallel with Circuit3 again, connects with R4 then, and note is Circuit4; At last, C5 and L5 form series resonant circuit, and be in parallel with Circuit4 again, connects with R5 then, and note is Circuit5; Above-mentioned Circuit5 is a tantalum capacitor equivalent model circuit.
For the tantalum capacitor equivalent model shown in the accompanying drawing 5, the span of the equivalent parameters of storing in the memory module comprises: the span of the span of L1, the span of C2, R2, the span of L2 ... 15 of the spans of the span of C5, the span of R5, L5 etc., certainly, if C1 is identical with the equivalence appearance value of C2, then the span of C1 and C2 can be identical, and other spans also have similar situation.
After the span of the capacitor equivalent model equivalent parameters in being provided with memory module, equivalent parameters of the present invention is chosen module and is chosen the method for equivalent parameters can be picked at random equivalent parameters from span from the span of memory module storage, certainly, can choose equivalent parameters from span according to predetermined rule, the present invention does not limit the method for choosing equivalent parameters from span yet.After choosing equivalent parameters, calculate the impedance of capacitor equivalent model according to the equivalent parameters of choosing.The method of picked at random equivalent parameters includes but not limited to Monte Carlo (Monte Carlo) random algorithm, and the present invention does not limit the concrete form that equivalent parameters is chosen module picked at random equivalent parameters.
Choose the impedance of the next equivalent parameters calculating capacitor equivalent model of module transmission according to equivalent parameters at the impedance acquisition module after, the impedance of calculating is transferred to the equivalent parameters determination module, the equivalent parameters determination module need compare impedance and the condensance that calculates, if the impedance of calculating and the error of condensance are less than the predetermined convergence value, then the equivalent parameters of this computing impedance correspondence can be defined as the equivalent parameters of capacitor equivalent model, if the impedance of calculating and the error of condensance are not less than the predetermined convergence value, then need to utilize once more said process to carry out the modeling method of capacitor equivalent model of the present invention.In this process predetermined condition for the error of the impedance of calculating and condensance less than the predetermined convergence value, and, the computing method of the impedance of calculating and the error of condensance can be for least square method etc., can certainly adopt other method, and the present invention does not limit the computing method of error.Condensance can be the measurement acquisition by electric capacity in this process, as the condensance by using network analyzer and test fixture testing capacitor to obtain.
The present invention is choosing many group equivalent parameterss in the parameter value scope, as choose and organize R more, L, the C parameter value, when if the error of the computing impedance of choosing of respectively organizing the parameter value correspondence and condensance all is not less than the predetermined convergence value, then can adjust the parameter area of capacitor equivalent model according to the computing impedance corresponding parameters value of error minimum, then, adopt the foregoing description process to carry out the modeling method of capacitor equivalent model of the present invention according to adjusted parameter area, in this process predetermined condition still for the error of the impedance of calculating and condensance less than the predetermined convergence value, equally, the computing method of the impedance of calculating and the error of condensance can be for least square method etc., and the present invention does not limit the computing method of error.
Equivalent parameters of the present invention is chosen module when choosing many group equivalent parameterss in the parameter value scope, also can be directly with one group of parameter value of the computing impedance correspondence of error minimum equivalent parameters as capacitor equivalent model, when the parameter group of choosing is abundant, also can guarantee the accuracy of capacitor equivalent model equivalent parameters in the parameter value scope.Predetermined condition is the impedance of calculating and the error minimum of condensance in this process.
Be example with Monte Carlo random algorithm, least square method below, in conjunction with the accompanying drawings 6, the modeling method of 7 pairs of capacitor equivalent models of the present invention of accompanying drawing is described.
Among Fig. 6, in step 600, the times N of choosing parameter value in the Monte Carlo algorithm is set, N is that constant and N are the positive integer more than or equal to 1, is 100 as N.Be set to the initial value of the inferior count value of N note, it is 0 that i promptly is set.The equivalent parameters determination module reads condensance Z_meas from the file of the condensance data that store test, the condensance Z_meas that stores in the storage file can obtain by network analyzer and test fixture test.
To step 610, equivalent parameters is chosen module and is chosen parameter value for each stochastic parameter of capacitor equivalent model by Monte Carlo random algorithm from the parameter value scope of the capacitor equivalent model of memory module storage, and transfer to the impedance acquisition module, simultaneously, the count value inferior for the N note increases by 1, i.e. i=i+1; Then, the impedance acquisition module is given capacitor equivalent model with the parameter value of choosing, and calculates the impedance Z _ model of capacitor equivalent model, and Z_model is transferred to the equivalent parameters determination module.
To step 620, the equivalent parameters determination module calculate Z_model according to least square method and the Z_meas that from file, reads between error e rror,
To step 630, whether the equivalent parameters determination module judges error less than predetermined convergency value, if error is less than predetermined convergency value, to step 631, with the equivalent parameters of this Z_model corresponding parameters value as capacitor equivalent model, the modeling method of this capacitor equivalent model finishes.
In step 630, if error is not less than predetermined convergency value, to step 640, the equivalent parameters determination module is judged as the inferior count value i of N note and whether surpasses N, if i does not surpass N, the notice equivalent parameters is chosen module and is continued to choose the capacitor equivalent parameter, to step 610, the module of choosing equivalent parameters continues to choose parameter value for each stochastic parameter of capacitor equivalent model in the parameter value scope of the capacitor equivalent model stored by Monte Carlo random algorithm from memory module, and transfer to the impedance acquisition module, simultaneously, the count value inferior for the N note increases by 1, i.e. i=i+1; Then, the impedance acquisition module is given capacitor equivalent model with the parameter value of choosing, and calculates the impedance Z _ model of capacitor equivalent model, and Z_model is transferred to the equivalent parameters determination module to step 620.
In step 640, if the count value i inferior for the N note surpasses N, to step 650, the equivalent parameters determination module is chosen one group of parameter value of error minimum from above-mentioned each error that obtains, with this parameter value is the span that benchmark is adjusted the equivalent parameters of storing in the memory module, as span is dwindled ten/first-class, simultaneously, the notice equivalent parameters is chosen module and is continued to choose the capacitor equivalent parameter.Then, to step 610, the module of choosing equivalent parameters continues to choose parameter value for each stochastic parameter of capacitor equivalent model in the parameter value scope of the capacitor equivalent model stored by Monte Carlo random algorithm from memory module, and transfer to the impedance acquisition module, simultaneously, the count value inferior for the N note increases by 1, i.e. i=i+1; Then, the impedance acquisition module is given capacitor equivalent model with the parameter value of choosing, and calculates the impedance Z _ model of capacitor equivalent model, and Z_model is transferred to the equivalent parameters determination module to step 620.
Flow process by above-mentioned Fig. 6, the model building device of capacitor equivalent module can accurately obtain the equivalent parameters of capacitor equivalent model, thereby the capacitor equivalent model that the model building device of capacitor equivalent module of the present invention is set up can accurately reflect the impedance operator of actual capacitance at different frequency range.Predetermined convergence value of the present invention can be set according to the accuracy requirement of capacitor equivalent model.
Though described the present invention by embodiment, those of ordinary skills know, the present invention has many distortion and variation and do not break away from spirit of the present invention, and the claim of application documents of the present invention comprises these distortion and variation.