Biases and Pulses#

Qibosoq supports the execution of modulated fast pulses, called here just pulses, and continuous DC biases.

Biases#

By bias we mean a DC current that gets turned on before the pulse sequence execution and gets turn off just at the end. Note that some care may be needed to fire continuous un-modulated pulses since the board itself may have baluns at the outputs.

Biases are controlled via the qibosoq.components.base.Qubit object:

from qibosoq.components.base import Qubit

qubit = Qubit(
    bias = 0.1,
    dac = 3
)

The bias parameter is a value relative to the maximum output voltage of the used dac so it’s defined in the range [-1, 1].

Pulses#

Differently from the biases, the pulses have a shape, a duration and are modulated. In Qibosoq there is a object qibosoq.components.pulses.Pulse:

from qibosoq.components.pulses import Pulse

pulse = Pulse(
    frequency = 4000,    # float in MHz
    amplitude = 0.5,     # float in [-1, 1]
    relative_phase = 0,  # int in degrees
    start_delay = 0,     # float in us
    duration = 1,        # float in us
    name = "id",         # str
    type = "drive",      # str in {"readout", "drive"}
    dac = 1,             # int
    adc = None,          # optional int
)

  • The start_delay is the difference in start time between this pulse and the one before it. We can consider the execution of a pulse always divided in two moments:

    • wait start_delay if it’s not zero

    • fire the pulse

  • The name parameter can be chosen to be whatever, but should be a unique identifier of the pulse.

  • The adc parameter is not needed for drive pulse, but it is for readout pulses. Every readout pulse is composed of a pulse fired through the dac and acquired by the adc, so both are required.

While the qibosoq.components.pulses.Pulse object can be used in the client, it cannot be used in execution time. In fact, the server expect a Pulse with a shape. The shape objects inherits from qibosoq.components.pulses.Pulse and share the same parameter + eventually others (… here are the qibosoq.components.pulses.Pulse parameters):

from qibosoq.components.pulses import Rectangular, Gaussian, Drag, Arbitrary, FlatTop

pulse = Rectangular(...)

pulse = Gaussian(
    ...
    rel_sigma = 5,  # float, sigma values as a fraction of the duration
)

pulse = Drag(
    ...
    rel_sigma = 5,  # float, sigma values as a fraction of the duration
    beta = 10,      # float (drag beta parameter)
)

pulse = FlatTop(
    ...
    rel_sigma = 5,  # float, sigma values as a fraction of the duration
)

pulse = Arbitrary(
    ...
    i_values = [...],  # list of floats
    q_values = [...],      # list of floats
)

Measurements#

From version 0.1.1, Qibosoq supports the execution of measurements decoupled from pulses. This allows to perform acquisition from an ADC without the need of genereting a pulse from a DAC.

from qibosoq.components.pulses import Measurement

meas = Measurement(
     type = "readout",
     frequency = 6400,    # float in MHz
     start_delay = 0,     # float in us
     duration = 1,        # float in us
     dac = 1,             # int
     adc = 0,          # int
)

The frequency parameter is the one of the signal we want to acquire. This is required by the downspampling scheme, to perform ADC acquisitions.

The DAC number, that could seem useless, can be used to ensure frequency matching between the ADC and a DAC. Namely, DACs and ADCs have different memory registers and we usually want to esure that have the same frequencies stored so that relative phases do not change during the experiment.