I recently made a batch of [∞] OpenRAMAN instruments and had the unpleasant surprise of discovering that most of the parts I received from Materialise for the liquid cuvette system were of bad quality and unusable in their state!
A picture of the bottom of the cuvette is given in Figure 1. You can spot that the bottom pocket hole is too thick and not very clean, making it impossible to insert the cylindrical lens. A similar problem occurs with the achromatic doublet hole which is also too tight to insert the lens properly. On the 5 parts I received, all of them were unusable though they did not all show the same degree of poor execution as in Figure 1. Some rework was then necessary to save the parts, which is the topic of this post.
I must confess I was not a big fan of my initial design and I’m probably the one to blame in the first place. Sliding the cylindrical lens was a pretty bad solution, although not much could be done about the loose tolerance in 3D printing of Materialise to fit the achromatic doublet. The biggest issue with the tolerance budget is that, if the part come too loose, you don’t get proper centering which is essential in optics as shown in our previous post [»] here. But applying too strict tolerances means you’ll have to rework the part yourself which can become even worse in terms of centering if you are not equipped properly (and time consuming, too).
In the new concept, the two lenses are mounted in a barrel which is inserted into a toleranced hole machined in the cuvette part. This design is shown in Figure 2 and the files for the barrel mount can be downloaded [∞] here for reproduction under our classical CERN open-hardware license. Although the drawing mentions brass, I 3D-printed mine using my Bambulab X1C printer which has excellent reproducibility. I then adapted the machine hole size to the barrel actual dimensions measured down to 0.001” using a precision caliper. This resulted in a very high-toleranced fit with no noticeable gap between the parts.
A cross-section of the barrel mount with the lenses is given in Figure 3. Lenses are held in place using DP190 thick epoxy glue and are centered on mechanical tolerance. It’s again the problem of achieving proper tolerances, except that it’s much easier to achieve accurate dimensions on a quality 3D printer for small cylindrical parts like this one. Also, note that the barrel is held in place using a DIN914 set-screw. The V-groove was placed in the barrel such that the set-screw would press the barrel down into the cuvette to ensure that the design distance between the lenses and the cuvette is respected.
While the design of Figure 2 and Figure 3 might seem complex at first sight, it’s actually state of the art practice in optomechanical engineering that I learned while working on precision optical instruments. It also has the advantage of keeping the complexity inside a more controllable part (the barrel) that you can eventually swap/remove more easily for cleaning or replace during maintenance operations.
The downside of this new design is that there is a rotational degree of freedom (optical roll axis) that you need to get right to achieve proper image quality. The system is not sensitive enough to require an alignment key (e.g. a DIN7 pin) which would have implied more complex rework of the cuvette holder part. Here, I choose to implement two small holes in the front of the barrel such that you could rotate the barrel using a spanner wrench until it aligns vertically.
This leaves the question of the rework of the cuvette holder part. For this job, I used my [∞] CNC STEP High-Z T 480 CNC milling machine. It’s relatively expensive to purchase as a hobbyist although it seems to be a popular choice among local FABLABs due to its low price and high accuracy. I bought mine first-day when starting my company because I felt it was a necessary tool for my startup.
After having firmly fixed the part, the initial achromat lens hole center was found using a scanning probe as shown in Figure 4. Getting the center position is a relatively critical task to achieve good lens centration and being off by a few tenths of millimeters could set your image quality and easiness of alignment back later on. A new hole was then machined into the part as shown in Figure 5 and Figure 6 with a 10 mm end-mill that was long enough to go to the bottom of the new hole. To achieve the best fit possible, I measured the actual barrel and endmill diameters using a precision caliper and adjusted the CNC configuration accordingly. The resulting fit was perfect with the barrel holding into the part with just enough friction to prevent it from dropping but still loose enough to remove it with a gentle tap in the palm of your end. I recommend testing the fit before you remove the part from the CNC, as shown in Figure 7, just in case you have to adapt the hole dimension. I initially tested the system on a scrap part to tune the tolerance to have the best fit possible. Last but not least, you may have to deburr the inside grooves of the original part to get the cylindrical lens fit properly with no obstruction (see Figure 8). Note that the NC program to machine the hole accounts for a perfect 27 mm hole diameter but use tool diameter compensation so that you can keep the program and just fine tune the tool diameter in your machine settings to achieve the proper results.
I hope this post has been useful for you :) don’t hesitate to share your comments on the [∞] community board to let me know!
I would also like to give a big thanks to Young, Sebastian, Alex, Stephen, Lilith, James, Jesse, Jon, Cory, Karel, Sivaraman, Samy, David, Themulticaster, Michael, Shaun, Tayyab, Kausban, Kirk, Marcel, Onur, Dennis, Benjamin, M, Sunanda and Natan who have supported this post through [∞] Patreon. I also take the occasion to invite you to donate through Patreon, even as little as $1. I cannot stress it more, you can really help me to post more content and make more experiments!
[⇈] Top of Page