SETUP and CALIBRATION

Sampler

Assemble Z-drive, arm, DL needle
Build stand (with desired stage heights)
Loosely attach brackets w/ fasteners to stage bottom
Attach stage (near end of cantilever)
Attach Z-drive assembly
- linear guide should be perpendicular to stage surface
Set up controller and software(?), attach cables
Attach deck to stage
- Be aware that arm could be obstructed by taller vessels in certain orientations, preventing access to shorter vessels

HW: adjust hardware limits, alignment with Z axis
HW/DECK: set convenient zero/home corner in config
DECK: Determine transform to HW

Fraction Collector

Assemble slider, arm
subset of Sampler, w/ arm, stage attachment difference



In [ ]:

    
from acqpack import Motor, AsiController, Autosampler, FractionCollector
from acqpack import gui
from acqpack import utils as ut

import yaml
import numpy as np
import pandas as pd

1 HW: set imits; align XY with Z

physical adjustments

adjust hall effect stops to maximize range
make sure:
- all relevant points on deck are accessible (check with 384-well plate)
- the stage does not collide with Z-axis (reorientation of the stage/deck may be necessary)
- at max Z, needle is at or just below top of deck



In [ ]:

    
a = Autosampler(Motor('config/motor.yaml'),  AsiController('config/input.yaml'))
gui.stage_control(a)

2 HW/DECK: set convenient XY zero/home corner in config



In [ ]:

    
# in input.yaml, set desired init_dir (x_dir and y_dir)

3 DECK: Determine transform to hardware

This is one way to do this; it assumes things are generally rectilinear (90deg rotations)
Could also do regression between frame1 and frame2 (generalizable way?)

XY rotation / reflection must be re-determined any time:
- deck is rotated relative to stage
- zero/home corner is changed in config
offset must be re-determined any time:
- deck is rotated relative to stage
- zero/home corner is changed in config
- hall-effect sensors are moved
- z-drive and stage are shifted (in any direction) relative to one another
- needle is changed to a different one

a x' y' z' w'
x 1. 0. 0. 0.
y 0. 1. 0. 0.
z 0. 0. -1. .0
w 0. 0. 0. 1.



In [ ]:

    
a = Autosampler(Motor('config/motor.yaml'),  AsiController('config/input.yaml'))
gui.stage_control(a)



In [ ]:

    
# DECK is row; HW is col
# determine XY axes mapping (rotations / reflections)

# determine XYZ offset (i.e. when DECK is 0,0,0, what is HW ?)

# save affine in input_deck.txt

4 PLATES: Make frames in deck coordinates

This is one way to do this; it assumes things are generally rectilinear (90deg rotations)
Could also do regression between frame1 and frame2 (generalizable way?)

x,y,z of all well positions are in 'DECK coordinates', i.e. relative to fixed point on deck (ref point, coordinate origin)
this allows platemaps to be reused so long as:
- deck is the same
- position AND orientation of plate on deck is the same
could also use sequence of transforms (PLATE -> DECK -> HARDWARE), but this is simpler for now



In [ ]:

    
a = Autosampler(Motor('config/motor.yaml'),  AsiController('config/input.yaml'))
a.add_frame('deck', 'config/input_deck.txt')
gui.stage_control(a)



In [ ]:

    
# determine plate 'A01' offset relative to deck origin using above GUI
# i.e. deck coordinates of 'A01'
offset = (2.493, 12.749, 0)

# generate platemap with offset
num_rc = (8,12)   # ct, ct
space_rc = (9,9)  # mm, mm (can do negative if row or column reflected)
platemap = ut.generate_position_table(num_rc, space_rc, offset, True)

# save
filename = '96plate.txt'
platemap



In [ ]:

    
# determine plate 'A01' offset relative to deck origin using above GUI
# i.e. deck coordinates of 'A01'
offset = (8.192, -18.952, 37.700)

# generate platemap with offset
num_rc = (1,4)   # ct, ct
space_rc = (0,28.375)  # mm, mm (can do negative if row or column reflected)
platemap = ut.generate_position_table(num_rc, space_rc, offset, True)

# save
filename = '4scint.txt'
platemap



In [ ]:

    
a.add_frame('plate', 'config/input_deck.txt', 'config/96plate.txt')
a.add_frame('wash', 'config/input_deck.txt', 'config/4scint.txt')



In [ ]:

    
a.goto('plate','name', 'H10')



In [ ]:

    
a.where()

Runtime: option to check deck transform?

be sure to clear deck!!!!!!
go to XY (0,0)
drop Z manually to check (0,)
can:
- tweak physically to bring back into alignment
- update deck coordinate frame



In [ ]:

    
a.goto('deck', 'xyz', (0,0,0))

Generalized prodcedure to determine transform between two coordinate frames?

regression between set of points from frame1 and frame2

TODO

Determine transforms
TEST
- push version
needle

yaml np.array load/store format (better to print to save for readability? or use csv?)
- or use RS matrix + offset vector



In [ ]:

RUN!



In [2]:

    
from acqpack import Motor, AsiController, Autosampler, FractionCollector
from acqpack import gui
from acqpack import utils as ut

Autosampler



In [3]:

    
a = Autosampler(Motor('config/motor.yaml'),  AsiController('config/input.yaml'))
gui.stage_control(a)









    



(-0.000, 0.000, 0.000) hardware
(0.330, -28.698, 99.600) plate
(0.330, -28.698, 99.600) wash



In [33]:

    
a.add_frame('plate', 'config/input_deck.txt', 'config/96plate.txt')
a.add_frame('wash', 'config/input_deck.txt', 'config/4scint.txt')



In [60]:

    
a.goto('plate', 'name', 'A01', zh_travel=35)



In [34]:

    
a.goto('plate', 'xyz', (0,0,0))

FractionCollector



In [55]:

    
f = FractionCollector(AsiController('config/output.yaml'))
gui.stage_control(f)









    



(101.493, -75.749) hardware
(101.493, 75.749) plate



In [56]:

    
f.add_frame('plate', 'config/output_deck.txt', 'config/96plate.txt')



In [58]:

    
f.goto('plate', 'name', 'H12')



In [62]:

    
f.where('plate')









    Out[62]:





(101.49339999999999, 75.748900000000006)



In [54]:

    
f.exit()