This page describes how to build simple optics rigs for microfluidics applications based on professional-grade caged optics sytems and scientific cameras.
This is a DropletKitchen project.
If not stated otherwise, the designs and documentation in this repository is Copyright (c) 2016 Martin Fischlechner and DropletKitchen Contributors, and made available under a Creative Commons Attribution 4.0 International (CC BY 4.0) License.
This project shows one way to package caged optics into systems with a small footprint and highly functional design. The example here is a simple microscope-setup for microfluidics experimentation. Instead of using an optics table, or an all-metal optics breadboard, a combination of laser-cut 5mm acrylic sheets with M6 nuts and bolts are used to assemble rigid and customized optical instruments. The system uses a simple, magnetic kinematic design for the x,y,z-stage and a high-power LED with cheap plastic focusing optics for illumination. The optics is based on an infinity-corrected objective, a lens and a global-shutter CCD or CMOS-camera to obtain videos or images. The output of an infinity-corrected objective (they usually have an ‘infinity’ symbol engraved) is a parallel beam of light. It therefore needs a lens to create an image. The advantage is that between the objective and the lens many optic components can be placed; the design can therefore also be used as a starting point for more complex setups, e.g. for measuring fluorescence. To get an idea which optics components are available and how much they cost, have a look at the websites of optics companies, for example Thorlabs (https://www.thorlabs.com/navigation.cfm).
Image of the rig at several stages of the build. A] bottom plate with the optics and an LED-driver enclosed in a little box. B] plate with LED and magnets to hold the stage. This part is used to immobilise the optics by clamping. C] XYZ-stage: Z and tilt is adjusted with three optics screws. X-Y movement is controlled with an inexpensive xy-stage. D] the fully assembled rig.
The model (./files/RigCageOpticsOpenSCAD.scad) was first developed in openSCAD (http://www.openscad.org/) and sliced for lasercutting. Later modifications were done in a .dxf-file with Draftsight (http://www.3ds.com/products-services/draftsight-cad-software/free-download/). Starting with a 3D-model is convenient to play with overall dimensions and shapes, but the design is quicker to modify in 2D-CAD at later stages (./files/RigCageOpticsPartsForCutting.dxf). The standard fixtures for optics on optics tables are M6, spaced 25mm apart. To have a flexible way of mounting (additional) parts, holes of 6mm in diameter with a spacing of 12.5mm (center to center) are additionally embedded into the design where possible. This makes setups adaptable and extendable.
Left: 3D-model in openSCAD. Right: Parts for laser cutting.
| Item | number | bought from | part number | approx cost (£/euro) |
|---|---|---|---|---|
| Structural components | ||||
| M6 Hex Screws (different lengths, eg l=16,20,25,30mm for flexibility) | many | RS or other hardware shop | NA | |
| M6 Nuts, Washers, ‘shake-proof washers’ | many | RS or other hardware shop | NA | |
| M6 Threaded Rods | 4x (~100mm) | hardware shop | NA | |
| M6 Threaded Rods | 3x (~150mm) | hardware shop | NA | |
| M3 Screws, Nuts and Washers (different lengths) | many | RS or other hardware shop | NA | |
| 3mm Acrylic Sheet (410x280mm) | 1 pack of 5 | www.techsoft.co.uk | TAAC3-R3-5-COL | 18 |
| 5mm Acrylic Sheet (410x280mm) | 1 pack of 5 | www.techsoft.co.uk | TAAC3-R5-5-COL | 25 |
| Double-Sided Tape (3M, 468MP or 467MP), optional but nice to have | 1 | RS | 842-1089 | 50 |
| XYZ-stage | ||||
| Fine Adjustment Screw with Knob, 1/4”-80, 2.00” Long | 3 | Thorlabs | FAS200 | 17 |
| Locking Phosphor-Bronze Bushing with Nut, 1/4”-80, L=0.50” | 3 | Thorlabs | N80L6P | 21 |
| XY-stage from eBay (search ‘microscope stage’) | 1 | eBay | max 20 | |
| Eclipse 20mm Threaded Hole Neodymium Ring Magnet | 1 | RS | 792-4565 | 7 |
| Eclipse 20mm Neodymium Disc Magnet (3 each; we use 5 for assembly) | 2 | RS | 695-0184 | 22 |
| Short Steel Rods (magnetisable) 20-40mm long, 5-6mm in diameter | 2 | NA | NA | NA |
| Light | ||||
| RECOM POWER RCD-24-0.70/W/X3 LED Driver Adjustable | 1 | Farnell | 1793223 | 13 |
| OPULENT REBEL-STAR-ES-NW200 High Brightness LED | 1 | Farnell | 2110405 | 4 |
| Polymer Optics 141/180, Lens Assembly | 1 | RS | 665-6573 | 3 |
| Stripboard, 10kOhm Potentiometer | eg Farnell, RS or other | NA | ||
| Slide Switch, Male/Female 0.1” PCB Connectors | eg Farnell, RS or other | NA | ||
| Sugru for isolation of wire connections… | 1 | cpc Farnell | MK00041 | 6 |
| Standard Epoxy Glue | NA | NA | NA | |
| Optics | ||||
| RMS10X - 10X Olympus Plan Achromat Objective, 0.25 NA, 10.6 mm WD | 1 | Thorlabs | RMS10X | 285 |
| RMS to SM1 adapter | 1 | Thorlabs | SM1A3TS | 16 |
| Cage Cube-Mounted Penta Prism | 1 | Thorlabs | CCM1-PS932/M | 110 |
| LENS: F=100mm 1”diameter (cheaper) OR Achromatic Doublet f=50 mm, Ø1” | 1 | Thorlabs | AC254-050-A-ML | 70 |
| 1” Lens tubes (1 adjustable + other (adjust Lens-Camera-distance)) | depends on Lens | look at Thorlabs | ~50 | |
| Adapter with External C-Mount Threads and Internal SM1 Threads | 1 | Thorlabs | SM1A9 | 14 |
| Global Shutter Camera: eg Firefly MV USB2.0 Camera (Point Grey) | 1 | Point Grey | FMVU-03MTM-CS | ~200 |
One has an overwhelming choice of optics components. Think of which magnification you need for the objective. To focus the image onto the camera we use a F=50mm double achromat lens, but a simple F=100mm lens
works equally well (but needs more space). As the rig uses standard 1” optics, any camera can be mounted with appropriate adapters. We have chosen the FireflyMV because it uses USB 2.0, has a global shutter and can be triggered, its image binned, etc. (Firefly MV Technical Reference Manual, https://www.ptgrey.com/support/downloads/10116). It runs well on Ubuntu Linux (and Win) and even on Raspberry Pi. Such ‘inexpensive’ scientific cameras will never give you high-speed video (max ~ 400fps with the FireflyMV when binning to a very small image), but bright illumination and minimising the exposure time yields sharp images of fast events, which is good enough for most microfluidics applications. One can enhance the camera greatly by strobing the LED light source (see here for an example: https://github.com/DropletKitchen/strobe). If you need high-speed video, you will have to use a true ‘fast camera’ with on-board RAM, which start around 6k GBP (eg a Miro Phantom).
For assembly you have to first attach the lens to the camera such that it is focused to ‘infinity’. Get the camera to run on a computer, connect the camera and lens with a tube (the distance between lens camera-chip should equal the focal length of the lens; you will need an adjustable SM1 lens tube in your assembly to get it right). Then focus the image on a distant object (e.g. a tree/house outside somewhere, or a feature in a far-away corner of the lab). Fix the distance with a locking ring. Now connect this assembly with the 30 mm cage containing a Penta Prism or a 45 degree mirror, screw in your objective (with the right adapter) and you are done.
Assembling the rig itself is pretty straightforward; have a look at the image above. Make sure that your components fit - if you change the design, just measure components with a caliper and use exactly those dimensions for holes etc. Laser cut components usually result in a tight fit. The optics assembly is held in place by clamping it between the bottom plate and the upper part using M6 threaded rod posts. Use washers, or better ‘shake-proof washers’ where possible.
The x,y,z-stage is mounted on three optics screws with spherical ends in a kinematic fashion - this defines the position of the stage by resting on 6 points of contact. This is ideally 3 points on one, 2 on the next, and one on the third. Here strong magnets are used to fix the stage on its support - a ring magnet on the first side (a small cheat but works well), 2 stacked disk magnets that magnetize two small steel rods on top, and a stack of three disk magnets on the third point. This assembly alone allows you not only to control z distance (focus), but you can tilt and therefore have some movement in x,y. For convenience, adding an inexpensive x,y stage is recommended. They all differ, so make sure you adapt the files accordingly.
For the lightsource, a star-shaped white high power LED is used which is controlled by an adjustable LED driver (Recom RCD-24-0.70, http://www.recom-power.com/pdf/Lightline/RCD-24.pdf). An M3 screw acts as a heat sink and also holds the LED in place. A potentiometer is used to control brightness. Along with an on/off switch, the components are soldered onto a stripboard and then packaged in an acrylic enclosure. It does not have to bear large loads, so the enclosure is based on cut acrylic sheets with double-sided tape added to one side before cutting. This allows the structural components to be bonded together. Inexpensive plastic lenses are used to focus the LED light onto the chip. For assembly, just use a dab of epoxy glue. Do not use ‘Superglue’, as the cyanoacrylate vapours cloud the optics components over time. Change the distance between LED and the stage such that the image on the camera is overexposed when the LED is fully on and the camera is set to the smallest exposure time possible. This allows you to get the best possible images of fast-moving objects with your setup. Below is the circuit diagram for wiring if you use the Recom LED-driver. If you want an on/off switch, just wire it to the voltage or ground wire. Almost any power supply will do (>=1A, 5-36V).
Circuit diagram for the Recom LED driver.
It is relatively easy to adjust designs to other components. Use a caliper to measure parts and change design files accordingly.
Rig with a Miro EX2 fast camera and an objective turret. The file for laser cut parts can be found here (./files/MiroEX2RigLasercutParts.dxf).