This shows you the differences between two versions of the page.
Next revision | Previous revision | ||
soft:cadence:monte_carlo:start [2009/07/15 11:26] dominik.przyborowski created |
soft:cadence:monte_carlo:start [2019/03/08 14:08] (current) |
||
---|---|---|---|
Line 1: | Line 1: | ||
====== Monte Carlo analysis ====== | ====== Monte Carlo analysis ====== | ||
+ | ===== AMS 3.70 ===== | ||
<code>HIT-KIT 3.70; spectra</code> \\ | <code>HIT-KIT 3.70; spectra</code> \\ | ||
<code>details: file:///opt/cds/designKits/ams_v3.70/www/hitkit/circuit_sim/montecarlo/mismatch.html</code> | <code>details: file:///opt/cds/designKits/ams_v3.70/www/hitkit/circuit_sim/montecarlo/mismatch.html</code> | ||
- | ===== Models setting ===== | + | ==== Models setting ==== |
To run MC simulations you must set Monte Carlo sections of elements models. \\ | To run MC simulations you must set Monte Carlo sections of elements models. \\ | ||
So run //virtuoso analog design environment// and click **setup -> model libraries**. Next change in libraries (cmos53, res, cap, etc) name of section, from *tm to *mc. For example cmostm -> cmosmc. | So run //virtuoso analog design environment// and click **setup -> model libraries**. Next change in libraries (cmos53, res, cap, etc) name of section, from *tm to *mc. For example cmostm -> cmosmc. | ||
- | ===== Run of simulation ===== | + | ==== Run of simulation ==== |
After change the models, you run the simulation: **tools -> monte carlo** \\ | After change the models, you run the simulation: **tools -> monte carlo** \\ | ||
Gauss distibution requiere over 30 samples. So **number of runs** should be higher than 30. To see all results, change ** save data between runs ** checkbox on true. | Gauss distibution requiere over 30 samples. So **number of runs** should be higher than 30. To see all results, change ** save data between runs ** checkbox on true. | ||
- | ===== Results ===== | + | ==== Results ==== |
Results as characteristic family: \\ | Results as characteristic family: \\ | ||
In //setup outputs// set //data type// on //wave//. \\ | In //setup outputs// set //data type// on //wave//. \\ | ||
Line 17: | Line 18: | ||
In //setup outputs// set //data type// on //scalar//. \\ | In //setup outputs// set //data type// on //scalar//. \\ | ||
+ | |||
+ | ===== AMS 4.00 ===== | ||
+ | <code>HIT-KIT 4.00; spectra</code> \\ | ||
+ | |||
+ | ==== Models setting ==== | ||
+ | To run MC simulations you must set Monte Carlo sections of elements models. \\ | ||
+ | **ICFB -> Hit-Kit Utilities -> Simulation Utilities -> Change Model Files -> Set MC models** | ||
+ | |||
+ | ==== Run of simulation ==== | ||
+ | Monte Carlo simulations are available in ADE XL. If cellview //adexl// doesn't exist: in schematic window click **Launch -> ADE XL** In dialog box choose **Create new view** and click OK (if adexl cellview exists - simple open it). \\ | ||
+ | In toolbar **Run** change //single run...// on // Monte Carlo Sampling//. In //Simulation options// (first button on right) set method (mismatch, process or all), specified number of iterations and finally mark //Save Data to Allow Family Plots//. | ||
+ | Finally run the simulations (green button on **run** toolbar). | ||
+ | |||
+ | \\ \\ | ||
+ | |||
+ | ** Remark ** All simulations type (dc, ac, etc) and outputs are set on **Data View** in **Tests** branch. | ||
+ | |||
+ | ===== Phase Margin, GBW and open loop gain ===== | ||
+ | Due to lack of corelations between instances on test bench, it is not possible to check above parameters in the standard AC simulations on replica circuits. \\ | ||
+ | The solution is to run MC on //stb// analysis and use following expressions to obtain histograms of DC open loop gain, phase margin and gain-bandwidth product: \\ | ||
+ | - DC Open loop gain: <code>dB20(value(getData("loopGain" ?result "stb") 0))</code> | ||
+ | - Phase Margin: <code>getData("/phaseMargin" ?result "stb_margin")</code> | ||
+ | - GainBandWidth: <code>getData("/phaseMarginFreq" ?result "stb_margin")</code> |