ConnectionResetError when Controlling an Oxford Triton Fridge via QCoDeS Interface

[johnnyzhaophysics]

I encountered a ConnectionResetError while controlling an Oxford Instruments Triton fridge using the QCoDeS interface. The traceback for the error is attached as below.

connectionerrormessage.docx

I have tried increasing the timeout before setting the field, adding a delay between commands but the issue persists.

Question:
Has anyone experienced similar issues when controlling the Triton magnet power supply via Qcodes?
Are there known fixes or best practices to prevent the connection from being reset?
Is there a recommended way to handle this in QCoDeS (e.g., reconnecting if the connection is lost)?

Any insights would be greatly appreciated! Thanks in advance.

[johnnyzhaophysics]

Here is my measurement code:

Measurement Code:

from qcodes.instrument_drivers.oxford.triton import Triton as qcodes_Triton
triton = qcodes_Triton(name='Oxford_T1', address='172.16.39.21', port=33576, timeout=100000, persistent=False)

magnet_field_x = Parameter('magnet_field_x', unit='T', label='$B_x$', set_cmd=triton.Bx, get_cmd=triton.Bx)
magnet_field_y = Parameter('magnet_field_y', unit='T', label='$B_y$', set_cmd=triton.By, get_cmd=triton.By)
magnet_field_z = Parameter('magnet_field_z', unit='T', label='$B_z$', set_cmd=triton.Bz, get_cmd=triton.Bz)

check_and_add_component(magnet_field_x)
check_and_add_component(magnet_field_y)
check_and_add_component(magnet_field_z)

def set_polar_r(r_val):
    pi = np.pi
    theta = 0
    phi = 90  # 90 degrees is π/2 radians
    
    # Convert angles from degrees to radians
    phi_rad = phi * pi / 180
    theta_rad = theta * pi / 180
    
    # Calculate the x, y, z components of the magnetic field
    x_target = r_val * np.sin(phi_rad) * np.cos(theta_rad)
    y_target = r_val * np.sin(phi_rad) * np.sin(theta_rad)
    z_target = r_val * np.cos(phi_rad)
    
    # Set the magnet field components
    magnet_field_x(x_target)
    print(f'set x_target to {x_target} T')
    x_current = triton.Bx()
    x_togo = np.abs(x_target - x_current)
    x_time = x_togo * 60 / 0.05
    sleep(x_time + 30)
    
    magnet_field_y(y_target)
    print(f'set y_target to {y_target} T')
    y_current = triton.By()
    y_togo = np.abs(y_target - y_current)
    y_time = y_togo * 60 / 0.05
    sleep(y_time + 30)
    
    magnet_field_z(z_target)
    print(f'set z_target to {z_target} T')
    z_current = triton.Bz()
    z_togo = np.abs(z_target - z_current)
    z_time = z_togo * 60 / 0.05
    sleep(z_time + 30)

magnet_field_r = Parameter('R', unit='T', label='R', set_cmd=set_polar_r, get_cmd=None)
triton.magnet_sweeprate(0.05)

[jenshnielsen]

johnnyzhaophysics I have not used the Triton driver for a long time. As I remember it there was a number of issues with the Triton interface provided and we usually used the Mercury driver directly.

I would probably debug first if the commands that do not require coms from the Triton to the magnet works well and if the magnet can be correctly connected from the Triton gui. Other than that perhaps check if something has changed in the Triton software since the qcodes driver was last updated