Skip to main content

Build SPICE circuits and simulation environments using Python

Project description

Welcome to CircuitBrew

CircuitBrew provides a framework for building SPICE-based circuits and generating test vectors for arbitrarily complex transistor-based circuits directly in Python.

Some notable features of CircuitBrew over a schematic capture tool or a Domain Specific Language based circuit design tool:

  1. Built-in test vector generation: With most other tools, I end up needing to build a behavioral model in a high-level language like Python to generate test vectors. CircuitBrew provides the hooks to build the model inline with the circuit description, and uses a built-in event simulator to generate expected vectors.

  2. Simplicity: Schematic capture for complex digital circuit design can be tedious. And learning/re-learning a circuit design DSL creates a lot of overhead, especially when you want to jump in quickly and prototype a custom circuit. Here, subckts are just Python classes, and ports are just init args; all the power of Python is available to assemble complex digital circuits at the transistor level.

  3. Ease of maintenance: Extending a circuit design DSL, as SPICE models evolve and become more complex, becomes incredibly time-consuming, especially when the original tool authors have moved on from the code base. Everything here is built on standard Python features, and should be easily extensible.

The examples below show SPICE files emitted for the open-source SkyWater 130nm PDK, to avoid revealing any proprietary data, but I generally use this in my own work for advanced 7nm and smaller-node foundries.

Installation

    $ pip install circuitbrew

Usage

Please see the usage guide.

Disclaimer

The software is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

Simple example

Here's a simple inverter implemented directly in Python:

Inverter in Python

class Inv(Module):
    inp = InputPort()
    out = OutputPort()

    p = SupplyPort()

    def build(self):
        vdd, gnd = self.p.vdd, self.p.gnd
        self.pup = Pfet(g=self.inp, d=vdd, s=self.out, b=vdd)
        self.ndn = Nfet(g=self.inp, d=self.out, s=gnd, b=gnd)
        self.finalize()

class Main(Module):

    def build(self):
        self.inv = Inv('myinv')
        self.finalize()

SPICE for a simple inverter

CircuitBrew directly emits SPICE from this:

.subckt Main 
xmyinv inp_0 out_1 p_2 p_3 Inv
.ends

.subckt Inv inp out p.gnd p.vdd
xmn0 out inp p.gnd p.gnd sky130_fd_pr__nfet_01v8_lvt w=1.0 l=0.5
xmp0 p.vdd inp out p.vdd sky130_fd_pr__pfet_01v8_lvt w=1.0 l=0.5
.ends

xmain Main

More complex example with test vector generation

Here's a more complex example that generates a simulation harness with test vector generation and digital checking of outputs. We generate random values that get fed into the inverter, and verify them on the output in the SPICE sim. CircuitBrew contains custom Verilog-A modules that can handle driving and checking values digitally inside an analog hspice simulation.

Python for simulation/verification of an Inverter

class Inv(Module):
    inp = InputPort()
    out = OutputPort()

    p = SupplyPort()

    def build(self):
        vdd, gnd = self.p.vdd, self.p.gnd
        self.pup = Pfet(g=self.inp, d=vdd, s=self.out, b=vdd)
        self.ndn = Nfet(g=self.inp, d=self.out, s=gnd, b=gnd)
        self.finalize()

    async def sim(self):
        while True:
            val = await self.inp.recv()
            await self.out.send(1-val)
    
class Main(Module):

    def build(self):
        self.supply = Supply('vdd', self.sim_setup['voltage'], )
        p = self.supply.p
        vdd, gnd = p.vdd, p.gnd
        self.inv = Inv('myinv', p=p)

        self.clk_gen = VerilogClock('clk', freq=750e3, enable=vdd)
        src_clk = self.clk_gen.clk

        self.clk_buc = VerilogClock('clk2', freq=750e3, offset='100p', enable=vdd)
        sample_clk = self.clk_buc.clk

        self.src = VerilogSrc('src', [randint(0,1) for i in range(10)], 
                            d=self.inv.inp, clk=src_clk, _reset=vdd)

        self.bucket = VerilogBucket(name='buc', clk=sample_clk, _reset=vdd, d=self.inv.out)
        self.finalize()

Generated SPICE with test bench

.subckt Main 
xbuc sample_clk p.vdd d_0 VerilogBucket_0
xclk2 p.vdd sample_clk VerilogClock freq=750000.0 offset=100p
xclk p.vdd src_clk VerilogClock freq=750000.0 offset=0
xmyinv inp_1 d_0 gnd p.vdd Inv
xsrc src_clk p.vdd inp_1 VerilogSrc_0
xvdd gnd p.vdd Supply
.ends

.subckt Supply p.gnd p.vdd
Vvdd_vss p.gnd 0 0.0
Vvdd_vdd p.vdd 0 1.8
.ends

.subckt Inv inp out p.gnd p.vdd
xmn0 out inp p.gnd p.gnd sky130_fd_pr__nfet_01v8_lvt w=1.0 l=0.5
xmp0 p.vdd inp out p.vdd sky130_fd_pr__pfet_01v8_lvt w=1.0 l=0.5
.ends

.hdl hspice_clk.va
.hdl template_0_hspice_src.va
.hdl template_0_hspice_bucket.va

xmain Main

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

circuitbrew-0.1.9.tar.gz (4.2 MB view details)

Uploaded Source

Built Distribution

circuitbrew-0.1.9-py3-none-any.whl (65.4 kB view details)

Uploaded Python 3

File details

Details for the file circuitbrew-0.1.9.tar.gz.

File metadata

  • Download URL: circuitbrew-0.1.9.tar.gz
  • Upload date:
  • Size: 4.2 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: python-httpx/0.24.0

File hashes

Hashes for circuitbrew-0.1.9.tar.gz
Algorithm Hash digest
SHA256 6eb5ed7d1eb1ef52ff8ddc2d5850068f8411ce5fc7c011f6ba73b7591161127d
MD5 619430e25e290cd419ee97202ce6737f
BLAKE2b-256 3fdacad5ad8e9b07bfd183add5b0756d1f04376a5a92337c5dc6f727a30cb89b

See more details on using hashes here.

File details

Details for the file circuitbrew-0.1.9-py3-none-any.whl.

File metadata

File hashes

Hashes for circuitbrew-0.1.9-py3-none-any.whl
Algorithm Hash digest
SHA256 1447342045c1833c25514884a5bb0aaea4f2e7f9582cfdeed6c9e5fae7dbc54c
MD5 6f8edc50d507f62df38935543a9b0090
BLAKE2b-256 729703f47671d88ef452ec157aa1c0b3791fd8a5915fc599e2f7b5e8f418e38d

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page